agora inbox for pgsql-hackers@postgresql.orghelp / color / mirror / Atom feed
[PATCH 2/2] Fix libxml leaks in contrib/xml2 xpath_table 249+ messages / 2 participants [nested] [flat]
* [PATCH 2/2] Fix libxml leaks in contrib/xml2 xpath_table @ 2026-05-25 21:11 Andrey Chernyy <andrey.cherny@tantorlabs.com> 0 siblings, 0 replies; 249+ messages in thread From: Andrey Chernyy @ 2026-05-25 21:11 UTC (permalink / raw) xpath_table() did not release libxml objects allocated while evaluating XPath expressions. xmlXPathCompiledEval() returns an xmlXPathObjectPtr that must be freed, and string results copied into the tuple input array must be freed after BuildTupleFromCStrings() has consumed them. Track the current libxml objects across the existing PG_TRY block so they are also released on error. --- contrib/xml2/xpath.c | 47 +++++++++++++++++++++++++++++++++++++++----- 1 file changed, 42 insertions(+), 5 deletions(-) diff --git a/contrib/xml2/xpath.c b/contrib/xml2/xpath.c index 94819961787..ac140a640e0 100644 --- a/contrib/xml2/xpath.c +++ b/contrib/xml2/xpath.c @@ -643,6 +643,10 @@ xpath_table(PG_FUNCTION_ARGS) StringInfoData query_buf; PgXmlErrorContext *xmlerrcxt; volatile xmlDocPtr doctree = NULL; + xmlXPathContextPtr volatile ctxt = NULL; + xmlXPathObjectPtr volatile res = NULL; + xmlXPathCompExprPtr volatile comppath = NULL; + xmlChar *volatile resstr = NULL; InitMaterializedSRF(fcinfo, MAT_SRF_USE_EXPECTED_DESC); @@ -662,7 +666,7 @@ xpath_table(PG_FUNCTION_ARGS) attinmeta = TupleDescGetAttInMetadata(rsinfo->setDesc); - values = (char **) palloc(rsinfo->setDesc->natts * sizeof(char *)); + values = (char **) palloc0(rsinfo->setDesc->natts * sizeof(char *)); xpaths = (xmlChar **) palloc(rsinfo->setDesc->natts * sizeof(xmlChar *)); /* @@ -732,10 +736,6 @@ xpath_table(PG_FUNCTION_ARGS) { char *pkey; char *xmldoc; - xmlXPathContextPtr ctxt; - xmlXPathObjectPtr res; - xmlChar *resstr; - xmlXPathCompExprPtr comppath; HeapTuple ret_tuple; /* Extract the row data as C Strings */ @@ -780,6 +780,11 @@ xpath_table(PG_FUNCTION_ARGS) had_values = false; for (j = 0; j < numpaths; j++) { + ctxt = NULL; + res = NULL; + comppath = NULL; + resstr = NULL; + ctxt = xmlXPathNewContext(doctree); if (ctxt == NULL || pg_xml_error_occurred(xmlerrcxt)) xml_ereport(xmlerrcxt, @@ -798,6 +803,7 @@ xpath_table(PG_FUNCTION_ARGS) /* Now evaluate the path expression. */ res = xmlXPathCompiledEval(comppath, ctxt); xmlXPathFreeCompExpr(comppath); + comppath = NULL; if (res != NULL) { @@ -842,8 +848,16 @@ xpath_table(PG_FUNCTION_ARGS) * result tuple. */ values[j + 1] = (char *) resstr; + resstr = NULL; + } + + if (res != NULL) + { + xmlXPathFreeObject(res); + res = NULL; } xmlXPathFreeContext(ctxt); + ctxt = NULL; } /* Now add the tuple to the output, if there is one. */ @@ -854,6 +868,16 @@ xpath_table(PG_FUNCTION_ARGS) heap_freetuple(ret_tuple); } + /* BuildTupleFromCStrings() has copied the values. */ + for (j = 1; j < rsinfo->setDesc->natts; j++) + { + if (values[j] != NULL) + { + xmlFree((xmlChar *) values[j]); + values[j] = NULL; + } + } + rownr++; } while (had_values); } @@ -870,6 +894,19 @@ xpath_table(PG_FUNCTION_ARGS) } PG_CATCH(); { + if (resstr != NULL) + xmlFree(resstr); + for (j = 1; j < rsinfo->setDesc->natts; j++) + { + if (values[j] != NULL) + xmlFree((xmlChar *) values[j]); + } + if (res != NULL) + xmlXPathFreeObject(res); + if (comppath != NULL) + xmlXPathFreeCompExpr(comppath); + if (ctxt != NULL) + xmlXPathFreeContext(ctxt); if (doctree != NULL) xmlFreeDoc(doctree); -- 2.54.0 --MP_/T7YUrG7W2jPQOBsjEYwqj9B-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
* [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners @ 2026-06-03 17:41 Andres Freund <andres@anarazel.de> 0 siblings, 0 replies; 249+ messages in thread From: Andres Freund @ 2026-06-03 17:41 UTC (permalink / raw) Without windows-vs is about 10min slower than the others tasks (windows-vs is currently 31m40s, windows-mingw is 20m45s), which makes for a frustrating experience. It seems worth "sacrificing" one of the 20 available concurrent jobs to avoid that. This reduces the time down to about 18m. ci-os-only: windows --- .github/workflows/pg-ci.yml | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pg-ci.yml b/.github/workflows/pg-ci.yml index 93b17af46df..83ee93ce5d6 100644 --- a/.github/workflows/pg-ci.yml +++ b/.github/workflows/pg-ci.yml @@ -723,8 +723,13 @@ jobs: # Job: Windows - Visual Studio + # + # If we were to execute tests in this job serially, this would be the + # slowest job by a good margin. To avoid that, use a matrix in combination + # with meson test's --slice SLICE/NUM_SLICES mechanism to split the tests + # across two runners. windows-vs: - name: Windows - Visual Studio + name: Windows - Visual Studio - Slice ${{ matrix.slice}}/${{ matrix.num_slices}} needs: [setup, sanity-check] if: | !cancelled() && @@ -732,6 +737,17 @@ jobs: needs.sanity-check.result != 'failure' runs-on: windows-2022 timeout-minutes: 60 + + # As described at the top of the task, split the tests across two runners + # for performance. The gains from additional concurrency diminish + # relatively quickly, due to each instance having to install dependencies + # and build postgres. + strategy: + fail-fast: false + matrix: + num_slices: [2] + slice: [1, 2] + env: # Avoid port conflicts between concurrent tap tests PG_TEST_USE_UNIX_SOCKETS: 1 @@ -885,6 +901,11 @@ jobs: - name: Test world env: + # As described at the top of the task, split the tests across two + # runners for performance. It's not the prettiest to implement this + # by prepending to MTEST_TARGET, but a more complicated solution + # doesn't seem worth it. + MTEST_TARGET: --slice ${{ matrix.slice}}/${{ matrix.num_slices}} ${{env.MTEST_TARGET}} ADDITIONAL_SETUP: | call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64 run: *meson_test_world_cmd -- 2.54.0.380.gc69baaf57b --lyfxwjjve3vodszg-- ^ permalink raw reply [nested|flat] 249+ messages in thread
end of thread, other threads:[~2026-06-03 17:41 UTC | newest] Thread overview: 249+ messages (download: mbox mbox.gz follow: Atom feed) -- links below jump to the message on this page -- 2026-05-25 21:11 [PATCH 2/2] Fix libxml leaks in contrib/xml2 xpath_table Andrey Chernyy <andrey.cherny@tantorlabs.com> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de> 2026-06-03 17:41 [PATCH v9a 22/22] ci: Slice tests on windows-vs across two runners Andres Freund <andres@anarazel.de>
This inbox is served by agora; see mirroring instructions for how to clone and mirror all data and code used for this inbox