Content-Type: multipart/mixed; boundary="===============8052356165775126479=="
MIME-Version: 1.0
Subject: PostgreSQL Weekly News - June 27, 2021
To: PostgreSQL Announce <pgsql-announce@lists.postgresql.org>
From: PWN via PostgreSQL Announce <announce-noreply@postgresql.org>
Reply-To: david@fetter.org
Date: Mon, 28 Jun 2021 11:45:22 +0000
Message-ID: <162488072281.697.15663151974959065115@wrigleys.postgresql.org>
Auto-Submitted: auto-generated
Archived-At: 
 <https://www.postgresql.org/message-id/162488072281.697.15663151974959065115%40wrigleys.postgresql.org>
Precedence: bulk

--===============8052356165775126479==
Content-Type: multipart/alternative;
 boundary="===============0330074949510022916=="
MIME-Version: 1.0

--===============0330074949510022916==
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable

# PostgreSQL Weekly News - June 27, 2021

PostgreSQL 14 Beta 2 released.
[Test!](https://www.postgresql.org/about/news/postgresql-14-beta-2-released=
-2249/)

The 12th PostgreSQL Conference Cuba (PostgresqlCUBA, @PgCuba) will take pla=
ce on
November 18-19, 2021, at the Hotel Habana Libre.  This event is part of the
TECNOGET conference, and we will host one track dedicated to PostgreSQL rel=
ated
talks. For more details, reach out to cu AT postgresql DOT org.

# PostgreSQL Product News

JDBC 42.2.22
[released](https://jdbc.postgresql.org/documentation/changelog.html#version=
_42.2.22)

pgpoolAdmin 4.2.0, the administration tool for Pgpool-II,
[released](https://www.pgpool.net/docs/pgpoolAdmin/NEWS).

pgCenter 0.9.0, a command-line admin tool for observing and troubleshooting
PostgreSQL, [released](https://github.com/lesovsky/pgcenter/releases/tag/v0=
.9.0)

# PostgreSQL Jobs for June

[https://archives.postgresql.org/pgsql-jobs/2021-06/](https://archives.post=
gresql.org/pgsql-jobs/2021-06/)

# PostgreSQL in the News

Planet PostgreSQL: [https://planet.postgresql.org/](https://planet.postgres=
ql.org/)

PostgreSQL Weekly News is brought to you this week by David Fetter

Submit news and announcements by Sunday at 3:00pm PST8PDT to david@fetter.o=
rg.

# Applied Patches

Tom Lane pushed:

- Work around portability issue with newer versions of mktime(). Recent gli=
bc
  versions have made mktime() fail if tm_isdst is inconsistent with the
  prevailing timezone; in particular it fails for tm_isdst =3D 1 when the z=
one is
  UTC.  (This seems wildly inconsistent with the POSIX-mandated treatment of
  "incorrect" values for the other fields of struct tm, so if you ask me it=
's a
  bug, but I bet they'll say it's intentional.)  This has been observed to =
cause
  cosmetic problems when pg_restore'ing an archive created in a different
  timezone.  To fix, do mktime() using the field values from the archive, a=
nd if
  that fails try again with tm_isdst =3D -1.  This will give a result that'=
s off
  by the UTC-offset difference from the original zone, but that was true be=
fore,
  too.  It's not terribly critical since we don't do anything with the resu=
lt
  except possibly print it.  (Someday we should flush this entire bit of lo=
gic
  and record a standard-format timestamp in the archive instead.  That's not
  okay for a back-patched bug fix, though.)  Also, guard our only other use=
 of
  mktime() by having initdb's build_time_t() set tm_isdst =3D -1 not 0.  Th=
is case
  could only have an issue in zones that are DST year-round; but I think so=
me do
  exist, or could in future.  Per report from Wells Oliver.  Back-patch to =
all
  supported versions, since any of them might need to run with a newer glib=
c.
  Discussion:
  [https://postgr.es/m/CAOC+FBWDhDHO7G-i1_n_hjRzCnUeFO+H-Czi1y10mFhRWpBrew@=
mail.gmail.com](https://postgr.es/m/CAOC+FBWDhDHO7G-i1_n_hjRzCnUeFO+H-Czi1y=
10mFhRWpBrew@mail.gmail.com)
  [https://git.postgresql.org/pg/commitdiff/f807e3410fdfc29ced6590c7c2afa76=
637e001ad](https://git.postgresql.org/pg/commitdiff/f807e3410fdfc29ced6590c=
7c2afa76637e001ad)

- Remove orphaned expected-result file. This should have been removed in
  43e084197, which removed the corresponding spec file.  Noted while fooling
  about with the isolationtester.
  [https://git.postgresql.org/pg/commitdiff/ffbe9dec13599fa786ea6567df1c6a3=
f3ee3c673](https://git.postgresql.org/pg/commitdiff/ffbe9dec13599fa786ea656=
7df1c6a3f3ee3c673)

- Update variant expected-result file. This should have been updated in
  d2d8a229b, but it was overlooked. According to 31a877f18 which added it, =
this
  file is meant to show the results you get under default_transaction_isola=
tion
  =3D serializable. We've largely lost track of that goal in other isolation
  tests, but as long as we've got this one, it should be right.  Noted while
  fooling about with the isolationtester.
  [https://git.postgresql.org/pg/commitdiff/0a1e80c5c4f094087257fc4284a87e0=
bc7bca591](https://git.postgresql.org/pg/commitdiff/0a1e80c5c4f094087257fc4=
284a87e0bc7bca591)

- Remove another orphan expected-result file. aborted-keyrevoke_2.out was
  apparently needed when it was added (in commit 0ac5ad513) to handle the c=
ase
  of serializable transaction mode. However, the output in serializable mode
  actually matches the regular aborted-keyrevoke.out file, and AFAICT has d=
one
  so for a long time. There's no need to keep dragging this variant along.
  [https://git.postgresql.org/pg/commitdiff/f6352a0d4e437ac8bc266f77df22d06=
4592056c9](https://git.postgresql.org/pg/commitdiff/f6352a0d4e437ac8bc266f7=
7df22d064592056c9)

- Update another variant expected-result file. This should have been update=
d in
  533e9c6b0, but it was overlooked. Given the lack of complaints, I won't b=
other
  back-patching.
  [https://git.postgresql.org/pg/commitdiff/d3c878499c9d639ff06e0664d06b8c7=
31e30c2fc](https://git.postgresql.org/pg/commitdiff/d3c878499c9d639ff06e066=
4d06b8c731e30c2fc)

- Improve SQLSTATE reporting in some replication-related code. I started out
  with the goal of reporting ERRCODE_CONNECTION_FAILURE when walrcv_connect=
()
  fails, but as I looked around I realized that whoever wrote this code was=
 of
  the opinion that errcodes are purely optional.  That's not my understandi=
ng of
  our project policy.  Hence, make sure that an errcode is provided in each
  ereport that (a) is ERROR or higher level and (b) isn't arguably an inter=
nal
  logic error. Also fix some very dubious existing errcode assignments.  Wh=
ile
  this is not per policy, it's also largely cosmetic, since few of these ca=
ses
  could get reported to applications.  So I don't feel a need to back-patch.
  Discussion:
  [https://postgr.es/m/2189704.1623512522@sss.pgh.pa.us](https://postgr.es/=
m/2189704.1623512522@sss.pgh.pa.us)
  [https://git.postgresql.org/pg/commitdiff/6b787d9e32005867ee3660d1ea20f44=
7810a403d](https://git.postgresql.org/pg/commitdiff/6b787d9e32005867ee3660d=
1ea20f447810a403d)

- Fix plancache refcount leak after error in ExecuteQuery. When stuffing a =
plan
  from the plancache into a Portal, one is not supposed to risk throwing an
  error between GetCachedPlan and PortalDefineQuery; if that happens, the p=
lan
  refcount incremented by GetCachedPlan will be leaked.  I managed to break=
 this
  rule while refactoring code in 9dbf2b7d7.  There is no visible consequence
  other than some memory leakage, and since nobody is very likely to trigge=
r the
  relevant error conditions many times in a row, it's not surprising we hav=
en't
  noticed.  Nonetheless, it's a bug, so rearrange the order of operations to
  remove the hazard.  Noted on the way to looking for a better fix for bug
  #17053. This mistake is pretty old, so back-patch to all supported branch=
es.
  [https://git.postgresql.org/pg/commitdiff/131ea3e908d3c97a2fe1ab25cce5046=
dd5cb905f](https://git.postgresql.org/pg/commitdiff/131ea3e908d3c97a2fe1ab2=
5cce5046dd5cb905f)

- Centralize the logic for protective copying of utility statements. In the
  "simple Query" code path, it's fine for parse analysis or execution of a
  utility statement to scribble on the statement's node tree, since that'll=
 just
  be thrown away afterwards.  However it's not fine if the node tree is in =
the
  plan cache, as then it'd be corrupted for subsequent executions.  Up to n=
ow
  we've dealt with that by having individual utility-statement functions ap=
ply
  copyObject() if they were going to modify the tree.  But that's prone to
  errors of omission.  Bug #17053 from Charles Samborski shows that CREATE/=
ALTER
  DOMAIN didn't get this memo, and can crash if executed repeatedly from pl=
an
  cache.  In the back branches, we'll just apply a narrow band-aid for that=
, but
  in HEAD it seems prudent to have a more principled fix that will close of=
f the
  possibility of other similar bugs in future. Hence, let's hoist the
  responsibility for doing copyObject up into ProcessUtility from its child=
ren,
  thus ensuring that it happens for all utility statement types.  Also, mod=
ify
  ProcessUtility's API so that its callers can tell it whether a copy step =
is
  necessary.  It turns out that in all cases, the immediate caller knows wh=
ether
  the node tree is transient, so this doesn't involve a huge amount of code
  thrashing.  In this way, while we lose a little bit in the execute-from-c=
ache
  code path due to sometimes copying node trees that wouldn't be mutated an=
yway,
  we gain something in the simple-Query code path by not copying throwaway =
node
  trees.  Statements that are complex enough to be expensive to copy are al=
most
  certainly ones that would have to be copied anyway, so the loss in the ca=
che
  code path shouldn't be much.  (Note that this whole problem applies only =
to
  utility statements. Optimizable statements don't have the issue because we
  long ago made the executor treat Plan trees as read-only.  Perhaps someda=
y we
  will make utility statement execution act likewise, but I'm not holding my
  breath.)  Discussion:
  [https://postgr.es/m/931771.1623893989@sss.pgh.pa.us](https://postgr.es/m=
/931771.1623893989@sss.pgh.pa.us)
  Discussion:
  [https://postgr.es/m/17053-3ca3f501bbc212b4@postgresql.org](https://postg=
r.es/m/17053-3ca3f501bbc212b4@postgresql.org)
  [https://git.postgresql.org/pg/commitdiff/7c337b6b527b7052e6a751f966d5734=
c56f668b5](https://git.postgresql.org/pg/commitdiff/7c337b6b527b7052e6a751f=
966d5734c56f668b5)

- Improve version reporting in pgbench. Commit 547f04e73 caused pgbench to =
start
  printing its version number, which seems like a fine idea, but it needs a=
 bit
  more work: * Print the server version number too, when different. * Print=
 the
  PG_VERSION string, not some reconstructed approximation.  This patch copi=
es
  psql's well-tested code for the same purpose.  Discussion:
  [https://postgr.es/m/1226654.1624036821@sss.pgh.pa.us](https://postgr.es/=
m/1226654.1624036821@sss.pgh.pa.us)
  [https://git.postgresql.org/pg/commitdiff/84bee9610965331d5110971d8de390a=
5bbe2effc](https://git.postgresql.org/pg/commitdiff/84bee9610965331d5110971=
d8de390a5bbe2effc)

- Fix misbehavior of DROP OWNED BY with duplicate polroles entries. Ordinar=
ily,
  a pg_policy.polroles array wouldn't list the same role more than once; but
  CREATE POLICY does not prevent that.  If we perform DROP OWNED BY on a ro=
le
  that is listed more than once, RemoveRoleFromObjectPolicy either suffered=
 an
  assertion failure or encountered a tuple-updated-by-self error.  Rewrite =
it to
  cope correctly with duplicate entries, and add a CommandCounterIncrement =
call
  to prevent the other problem.  Per discussion, there's other cleanup that
  ought to happen here, but this seems like the minimum essential fix.  Per=
 bug
  #17062 from Alexander Lakhin.  It's been broken all along, so back-patch =
to
  all supported branches.  Discussion:
  [https://postgr.es/m/17062-11f471ae3199ca23@postgresql.org](https://postg=
r.es/m/17062-11f471ae3199ca23@postgresql.org)
  [https://git.postgresql.org/pg/commitdiff/d21fca084356946664bfce19d66d2df=
2bb873cbd](https://git.postgresql.org/pg/commitdiff/d21fca084356946664bfce1=
9d66d2df2bb873cbd)

- Provide feature-test macros for libpq features added in v14. We had a req=
uest
  to provide a way to test at compile time for the availability of the new
  pipeline features.  More generally, it seems like a good idea to provide =
a way
  to test via #ifdef for all new libpq API features.  People have been usin=
g the
  version from pg_config.h for that; but that's more likely to represent the
  server version than the libpq version, in the increasingly-common scenario
  where they're different.  It's safer if libpq-fe.h itself is the source of
  truth about what features it offers.  Hence, establish a policy that star=
ting
  in v14 we'll add a suitable feature-is-present macro to libpq-fe.h when w=
e add
  new API there. (There doesn't seem to be much point in applying this poli=
cy
  retroactively, but it's not too late for v14.)  Tom Lane and Alvaro Herre=
ra,
  per suggestion from Boris Kolpackov.  Discussion:
  [https://postgr.es/m/boris.20210617102439@codesynthesis.com](https://post=
gr.es/m/boris.20210617102439@codesynthesis.com)
  [https://git.postgresql.org/pg/commitdiff/6991e774e0304f5ef488cf1ae4fa795=
78b6ae3d5](https://git.postgresql.org/pg/commitdiff/6991e774e0304f5ef488cf1=
ae4fa79578b6ae3d5)

- Stabilize test case added by commit f61db909d. Buildfarm members ayu and =
tern
  have sometimes shown a different plan than expected for this query.  I'd =
been
  unable to reproduce that before today, but I finally realized what is
  happening. If there is a concurrent open transaction (probably an autovac=
uum
  run in the buildfarm, but this can also be arranged manually), then the i=
ndex
  entries for the rows removed by the DELETE a few lines up are not killed
  promptly, causing a change in the planner's estimate of the extremal valu=
e of
  ft2.c1, which moves the rowcount estimate for "c1 > 1100" by enough to ch=
ange
  the join plan from nestloop to hash.  To fix, change the query condition =
to
  "c1 > 1000", causing the hash plan to be preferred whether or not a concu=
rrent
  open transaction exists.  Since this UPDATE is tailored to be a no-op, no=
thing
  else changes.  Report:
  [https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=3Dayu&dt=3D2021-=
06-09%2022%3A45%3A48](https://buildfarm.postgresql.org/cgi-bin/show_log.pl?=
nm=3Dayu&dt=3D2021-06-09%2022%3A45%3A48)
  Report:
  [https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=3Dayu&dt=3D2021-=
06-13%2022%3A38%3A18](https://buildfarm.postgresql.org/cgi-bin/show_log.pl?=
nm=3Dayu&dt=3D2021-06-13%2022%3A38%3A18)
  Report:
  [https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=3Dtern&dt=3D2021=
-06-20%2004%3A55%3A36](https://buildfarm.postgresql.org/cgi-bin/show_log.pl=
?nm=3Dtern&dt=3D2021-06-20%2004%3A55%3A36)
  [https://git.postgresql.org/pg/commitdiff/5843659d091bfb6f2c60e010ea1fd00=
e55ee6ada](https://git.postgresql.org/pg/commitdiff/5843659d091bfb6f2c60e01=
0ea1fd00e55ee6ada)

- Restore the portal-level snapshot for simple expressions, too. Commits
  84f5c2908 et al missed the need to cover plpgsql's "simple expression" co=
de
  path.  If the first thing we execute after a COMMIT/ROLLBACK is one of th=
ose,
  rather than a full-fledged SPI command, we must explicitly do
  EnsurePortalSnapshotExists() to make sure we have an outer snapshot.  Note
  that it wouldn't be good enough to just push a snapshot for the duration =
of
  the expression execution: what comes back might be toasted, so we'd better
  have a snapshot protecting it.  The test case demonstrating this fact che=
ats a
  bit by marking a SQL function immutable even though it fetches from a tab=
le.
  That's nothing that users haven't been seen to do, though.  Per report fr=
om
  Jim Nasby.  Back-patch to v11, like the previous fix.  Discussion:
  [https://postgr.es/m/378885e4-f85f-fc28-6c91-c4d1c080bf26@amazon.com](htt=
ps://postgr.es/m/378885e4-f85f-fc28-6c91-c4d1c080bf26@amazon.com)
  [https://git.postgresql.org/pg/commitdiff/d102aafb6259a6a412803d4b1d8c4f0=
0aa17f67e](https://git.postgresql.org/pg/commitdiff/d102aafb6259a6a412803d4=
b1d8c4f00aa17f67e)

- Use annotations to reduce instability of isolation-test results. We've lo=
ng
  contended with isolation test results that aren't entirely stable.  Some =
test
  scripts insert long delays to try to force stable results, which is not
  terribly desirable; but other erratic failure modes remain, causing
  unrepeatable buildfarm failures.  I've spent a fair amount of time trying=
 to
  solve this by improving the server-side support code, without much succes=
s:
  that way is fundamentally unable to cope with diffs that stem from chance
  ordering of arrival of messages from different server processes.  We can
  improve matters on the client side, however, by annotating the test scrip=
ts
  themselves to show the desired reporting order of events that might occur=
 in
  different orders.  This patch adds three types of annotations to deal wit=
h (a)
  test steps that might or might not complete their waits before the
  isolationtester can see them waiting; (b) test steps in different sessions
  that can legitimately complete in either order; and (c) NOTIFY messages t=
hat
  might arrive before or after the completion of a step in another session.=
  We
  might need more annotation types later, but this seems to be enough to de=
al
  with the instabilities we've seen in the buildfarm.  It also lets us get =
rid
  of all the long delays that were previously used, cutting more than a min=
ute
  off the runtime of the isolation tests.  Back-patch to all supported bran=
ches,
  because the buildfarm instabilities affect all the branches, and because =
it
  seems desirable to keep isolationtester's capabilities the same across all
  branches to simplify possible future back-patching of tests.  Discussion:
  [https://postgr.es/m/327948.1623725828@sss.pgh.pa.us](https://postgr.es/m=
/327948.1623725828@sss.pgh.pa.us)
  [https://git.postgresql.org/pg/commitdiff/741d7f1047fe52da7ced6fa9cea661c=
e9320c8d4](https://git.postgresql.org/pg/commitdiff/741d7f1047fe52da7ced6fa=
9cea661ce9320c8d4)

- Improve display of query results in isolation tests. Previously,
  isolationtester displayed SQL query results using some ad-hoc code that
  clearly hadn't had much effort expended on it. Field values longer than 14
  characters weren't separated from the next field, and usually caused
  misalignment of the columns too.  Also there was no visual separation of a
  query's result from subsequent isolationtester output.  This made test re=
sult
  files confusing and hard to read.  To improve matters, let's use libpq's
  PQprint() function.  Although that's long since unused by psql, it's still
  plenty good enough for the purpose here.  Like 741d7f104, back-patch to a=
ll
  supported branches, so that this isn't a stumbling block for back-patching
  isolation test changes.  Discussion:
  [https://postgr.es/m/582362.1623798221@sss.pgh.pa.us](https://postgr.es/m=
/582362.1623798221@sss.pgh.pa.us)
  [https://git.postgresql.org/pg/commitdiff/4a054069a36032a59afceb07f3b837f=
09ab1a2e9](https://git.postgresql.org/pg/commitdiff/4a054069a36032a59afceb0=
7f3b837f09ab1a2e9)

- Don't assume GSSAPI result strings are null-terminated. Our uses of
  gss_display_status() and gss_display_name() assumed that the gss_buffer_d=
esc
  strings returned by those functions are null-terminated.  It appears that=
 they
  generally are, given the lack of field complaints up to now.  However, the
  available documentation does not promise this, and some man pages for
  gss_display_status() show examples that rely on the gss_buffer_desc.length
  field instead of expecting null termination.  Also, we now have a report =
that
  on some implementations, clang's address sanitizer is of the opinion that=
 the
  byte after the specified length is undefined.  Hence, change the code to =
rely
  on the length field instead.  This might well be cosmetic rather than fix=
ing
  any real bug, but it's hard to be sure, so back-patch to all supported
  branches. While here, also back-patch the v12 changes that made pg_GSS_er=
ror
  deal honestly with multiple messages available from gss_display_status.  =
Per
  report from Sudheer H R.  Discussion:
  [https://postgr.es/m/5372B6D4-8276-42C0-B8FB-BD0918826FC3@tekenlight.com]=
(https://postgr.es/m/5372B6D4-8276-42C0-B8FB-BD0918826FC3@tekenlight.com)
  [https://git.postgresql.org/pg/commitdiff/126cdaf47af275f76b2f2ddb023bfdc=
6f018ae30](https://git.postgresql.org/pg/commitdiff/126cdaf47af275f76b2f2dd=
b023bfdc6f018ae30)

- Doc: fix confusion about LEAKPROOF in syntax summaries. The syntax summar=
ies
  for CREATE FUNCTION and allied commands made it look like LEAKPROOF is an
  alternative to IMMUTABLE/STABLE/VOLATILE, when of course it is an orthogo=
nal
  option.  Improve that.  Per gripe from aazamrafeeque0.  Thanks to David
  Johnston for suggestions.  Discussion:
  [https://postgr.es/m/162444349581.694.5818572718530259025@wrigleys.postgr=
esql.org](https://postgr.es/m/162444349581.694.5818572718530259025@wrigleys=
.postgresql.org)
  [https://git.postgresql.org/pg/commitdiff/2031e1668e5577e64cfed29da69a349=
03d5a5227](https://git.postgresql.org/pg/commitdiff/2031e1668e5577e64cfed29=
da69a34903d5a5227)

- Allow non-quoted identifiers as isolation test session/step names. For no
  obvious reason, isolationtester has always insisted that session and step
  names be written with double quotes.  This is fairly tedious and does lit=
tle
  for test readability, especially since the names that people actually cho=
ose
  almost always look like normal identifiers.  Hence, let's tweak the lexer=
 to
  allow SQL-like identifiers not only double-quoted strings.  (They're SQL-=
like,
  not exactly SQL, because I didn't add any case-folding logic.  Also there=
's no
  provision for U&"..." names, not that anyone's likely to care.)  There is=
 one
  incompatibility introduced by this change: if you write "foo""bar" with no
  space, that used to be taken as two identifiers, but now it's just one
  identifier with an embedded quote mark.  I converted all the
  src/test/isolation/ specfiles to remove unnecessary double quotes, but st=
opped
  there because my eyes were glazing over already.  Like 741d7f104, back-pa=
tch
  to all supported branches, so that this isn't a stumbling block for
  back-patching isolation test changes.  Discussion:
  [https://postgr.es/m/759113.1623861959@sss.pgh.pa.us](https://postgr.es/m=
/759113.1623861959@sss.pgh.pa.us)
  [https://git.postgresql.org/pg/commitdiff/a443c1b2d6a646cf90a8afc193c07ed=
12a2bf045](https://git.postgresql.org/pg/commitdiff/a443c1b2d6a646cf90a8afc=
193c07ed12a2bf045)

- Further stabilize postgres_fdw test. The queries involving ft1_nopw don't
  stably return the same row anymore.  I surmise that an autovacuum hitting=
 "S
  1"."T 1" right after the updates introduced by f61db909d/5843659d0 freed =
some
  space, changing where subsequent insertions get stored. It's only by good=
 luck
  that these results were stable before, though, since a LIMIT without ORDE=
R BY
  isn't well defined, and it's not like we've ever treated that table as
  append-only in this test script.  Since we only really care whether these
  commands succeed or not, just replace `"SELECT *"` with "SELECT 1".  Repo=
rt:
  [https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=3Dcrake&dt=3D202=
1-06-23%2019%3A52%3A08](https://buildfarm.postgresql.org/cgi-bin/show_log.p=
l?nm=3Dcrake&dt=3D2021-06-23%2019%3A52%3A08)
  [https://git.postgresql.org/pg/commitdiff/802177090992511c610804da54a4603=
d4f50c594](https://git.postgresql.org/pg/commitdiff/802177090992511c610804d=
a54a4603d4f50c594)

- Doc: remove commit f560209c6 from v14 release notes. Now that this has be=
en
  back-patched, it's no longer a new feature for v14.
  [https://git.postgresql.org/pg/commitdiff/8a80562d732c0da1ddcc9fb88dfb976=
f4b846577](https://git.postgresql.org/pg/commitdiff/8a80562d732c0da1ddcc9fb=
88dfb976f4b846577)

- Remove unnecessary failure cases in RemoveRoleFromObjectPolicy(). It's not
  really necessary for this function to open or lock the relation associated
  with the pg_policy entry it's modifying.  The error checks it's making on=
 the
  rel are if anything counterproductive (e.g., if we don't want to allow
  installation of policies on system catalogs, here is not the place to pre=
vent
  that).  In particular, it seems just wrong to insist on an ownership chec=
k.
  That has the net effect of forcing people to use superuser for DROP OWNED=
 BY,
  which surely is not an effect we want.  Also there is no point in rebuild=
ing
  the dependencies of the policy expressions, which aren't being changed.
  Lastly, locking the table also seems counterproductive; it's not helping =
to
  prevent race conditions, since we failed to re-read the pg_policy row aft=
er
  acquiring the lock.  That means that concurrent DDL would likely result in
  "tuple concurrently updated/deleted" errors; which is the same behavior t=
his
  code will produce, with less overhead.  Per discussion of bug #17062.
  Back-patch to all supported versions, as the failure cases this eliminates
  seem just as undesirable in 9.6 as in HEAD.  Discussion:
  [https://postgr.es/m/1573181.1624220108@sss.pgh.pa.us](https://postgr.es/=
m/1573181.1624220108@sss.pgh.pa.us)
  [https://git.postgresql.org/pg/commitdiff/5a0f1c8c0193f0dd7fba50c22d96781=
fa2414007](https://git.postgresql.org/pg/commitdiff/5a0f1c8c0193f0dd7fba50c=
22d96781fa2414007)

- Remove undesirable libpq dependency on stringinfo.c. Commit c0cb87fbb unw=
isely
  introduced a dependency on the StringInfo machinery in fe-connect.c.  We =
must
  not use that in libpq, because it will do a summary exit(1) if it hits OO=
M,
  and that is not appropriate behavior for a general-purpose library.  The =
goal
  of allowing arbitrary line lengths in service files doesn't seem like it's
  worth a lot of effort, so revert back to the previous method of using a
  stack-allocated buffer and failing on buffer overflow.  This isn't an exa=
ct
  revert though.  I kept that patch's refactoring to have a single exit pat=
h, as
  that seems cleaner than having each error path know what to do to clean u=
p.
  Also, I made the fixed-size buffer 1024 bytes not 256, just to push off t=
he
  need for an expandable buffer some more.  There is more to do here; in
  particular the lack of any mechanical check for this type of mistake now =
seems
  pretty hazardous.  But this fix gets us back to the level of robustness w=
e had
  in v13, anyway.  Discussion:
  [https://postgr.es/m/daeb22ec6ca8ef61e94d766a9b35fb03cabed38e.camel@vmwar=
e.com](https://postgr.es/m/daeb22ec6ca8ef61e94d766a9b35fb03cabed38e.camel@v=
mware.com)
  [https://git.postgresql.org/pg/commitdiff/8ec00dc5cd70e0e579e9fbf8661bc46=
f5ccd8078](https://git.postgresql.org/pg/commitdiff/8ec00dc5cd70e0e579e9fbf=
8661bc46f5ccd8078)

- Doc: update v14 release notes for revert of commit c0cb87fbb.
  [https://git.postgresql.org/pg/commitdiff/dcffc9ba8a1e0ab1b0a57e9b9d38e3d=
c9960f83f](https://git.postgresql.org/pg/commitdiff/dcffc9ba8a1e0ab1b0a57e9=
b9d38e3dc9960f83f)

- Remove memory leaks in isolationtester. specscanner.l leaked a kilobyte of
  memory per token of the spec file. Apparently somebody thought that the
  introductory code block would be executed once; but it's once per yylex()
  call.  A couple of functions in isolationtester.c leaked small amounts of
  memory due to not bothering to free one-time allocations.  Might as well
  improve these so that valgrind gives this program a clean bill of health.
  Also get rid of an ugly static variable.  Coverity complained about one o=
f the
  one-time leaks, which led me to try valgrind'ing isolationtester, which l=
ed to
  discovery of the larger leak.
  [https://git.postgresql.org/pg/commitdiff/642c0697c96b9c6ba5d194653a329f7=
820565f01](https://git.postgresql.org/pg/commitdiff/642c0697c96b9c6ba5d1946=
53a329f7820565f01)

Micha=C3=ABl Paquier pushed:

- Remove forced toast recompression in VACUUM FULL/CLUSTER. The extra checks
  added by the recompression of toast data introduced in bbe0a81 is proving=
 to
  have a performance impact on VACUUM or CLUSTER even if no recompression is
  done.  This is more noticeable with more toastable columns that contain
  non-NULL values.  Improvements could be done to make those extra checks l=
ess
  expensive, but that's not material for 14 at this stage, and we are not s=
ure
  either if the code path of VACUUM FULL/CLUSTER is adapted for this job.  =
Per
  discussion with several people, including Andres Freund, Robert Haas, =C3=
=81lvaro
  Herrera, Tom Lane and myself.  Discussion:
  [https://postgr.es/m/20210527003144.xxqppojoiwurc2iz@alap3.anarazel.de](h=
ttps://postgr.es/m/20210527003144.xxqppojoiwurc2iz@alap3.anarazel.de)
  [https://git.postgresql.org/pg/commitdiff/dbab0c07e5ba1f19a991da2d72972a8=
fe9a41bda](https://git.postgresql.org/pg/commitdiff/dbab0c07e5ba1f19a991da2=
d72972a8fe9a41bda)

- Improve handling of dropped objects in pg_event_trigger_ddl_commands(). An
  object found as dropped when digging into the list of objects returned by
  pg_event_trigger_ddl_commands() could cause a cache lookup error, as the =
calls
  grabbing for the object address and the type name would fail if the objec=
t was
  missing.  Those lookup errors could be seen with combinations of ALTER TA=
BLE
  sub-commands involving identity columns.  The lookup logic is changed in =
this
  code path to get a behavior similar to any other SQL-callable function by
  ignoring objects that are not found, taking advantage of 2a10fdc.  The
  back-branches are not changed, as they require this commit that is too
  invasive for stable branches.  While on it, add test cases to exercise ev=
ent
  triggers with identity columns, and stress more cases with the event
  ddl_command_end for relations.  Author: Sven Klemm, Aleksander Alekseev,
  Michael Paquier Discussion:
  [https://postgr.es/m/CAMCrgp2R1cEXU53iYKtW6yVEp2_yKUz+z=3D3-CTrYpPP+xryRt=
g@mail.gmail.com](https://postgr.es/m/CAMCrgp2R1cEXU53iYKtW6yVEp2_yKUz+z=3D=
3-CTrYpPP+xryRtg@mail.gmail.com)
  [https://git.postgresql.org/pg/commitdiff/2d689babe3cb50dcb29f6ed595a61d5=
6e518c0d8](https://git.postgresql.org/pg/commitdiff/2d689babe3cb50dcb29f6ed=
595a61d56e518c0d8)

- doc: Apply markup <productname> to OpenSSL more consistently. Author: Dan=
iel
  Gustafsson Discussion:
  [https://postgr.es/m/CE12DD5C-4BB3-4166-BC9A-39779568734C@yesql.se](https=
://postgr.es/m/CE12DD5C-4BB3-4166-BC9A-39779568734C@yesql.se)
  [https://git.postgresql.org/pg/commitdiff/f80979f659d39e238e95444e6752142=
799428078](https://git.postgresql.org/pg/commitdiff/f80979f659d39e238e95444=
e6752142799428078)

- Fix pattern matching logic for logs in TAP tests of pgbench. The logic
  checking for the format of per-thread logs used grep() with directly "$re=
",
  which would cause the test to consider all the logs as a match without ca=
ring
  about their format at all.  Using "/$re/" makes grep() perform a regex te=
st,
  which is what we want here.  While on it, improve some of the tests to be=
 more
  picky with the patterns expected and add more comments to describe the te=
sts.
  Issue discovered while digging into a separate patch.  Author: Fabien Coe=
lho,
  Michael Paquier Discussion:
  [https://postgr.es/m/YNPsPAUoVDCpPOGk@paquier.xyz](https://postgr.es/m/YN=
PsPAUoVDCpPOGk@paquier.xyz)
  Backpatch-through: 11
  [https://git.postgresql.org/pg/commitdiff/c13585fe9e55813cf9feac67fe7b65d=
3a78fff92](https://git.postgresql.org/pg/commitdiff/c13585fe9e55813cf9feac6=
7fe7b65d3a78fff92)

- doc: Move remove_temp_files_after_crash to section for developer options.=
 The
  main goal of this option is to allow inspecting temporary files for debug=
ging
  purposes, so moving the parameter there is natural.  Oversight in cd91de0.
  Reported-by: Justin Pryzby Author: Euler Taveira Discussion:
  [https://postgr.es/m/20210612004347.GP16435@telsasoft.com](https://postgr=
.es/m/20210612004347.GP16435@telsasoft.com)
  [https://git.postgresql.org/pg/commitdiff/797b0fc0b078c7b4c46ef9f2d9e47aa=
2d98c6c63](https://git.postgresql.org/pg/commitdiff/797b0fc0b078c7b4c46ef9f=
2d9e47aa2d98c6c63)

- Add more debugging information with log checks in TAP tests of pgbench.
  fairywren is not happy with the pattern checks introduced by c13585f. I a=
m not
  sure if this outlines a bug in pgbench or if the regex patterns used in t=
he
  tests are too restrictive for this buildfarm member's environment.  This =
adds
  more debugging information to show the log entries that do not match with=
 the
  expected pattern, to help in finding out what's happening.  That seems li=
ke a
  good addition in the long-term anyway as that may not be the only issue in
  this area.  Discussion:
  [https://postgr.es/m/YNUad2HvgW+6eXyo@paquier.xyz](https://postgr.es/m/YN=
Uad2HvgW+6eXyo@paquier.xyz)
  [https://git.postgresql.org/pg/commitdiff/87b2124dfa0782a697ea7b90aff15a6=
f6117a720](https://git.postgresql.org/pg/commitdiff/87b2124dfa0782a697ea7b9=
0aff15a6f6117a720)

- doc: Add acronyms for MITM and SNI. This adds MITM and SNI as acronyms, a=
s the
  documentation already had them marked up with <acronym>.  While on it, ma=
ke
  sure to spell man-in-the-middle with dashes consistently, and add acronym=
s for
  those new terms where appropriate.  Author: Daniel Gustafsson Reviewed-by:
  Michael Paquier Discussion:
  [https://postgr.es/m/CE12DD5C-4BB3-4166-BC9A-39779568734C@yesql.se](https=
://postgr.es/m/CE12DD5C-4BB3-4166-BC9A-39779568734C@yesql.se)
  [https://git.postgresql.org/pg/commitdiff/15ff5401d1719aaf6c9a47e5abea517=
cc2bcbaf1](https://git.postgresql.org/pg/commitdiff/15ff5401d1719aaf6c9a47e=
5abea517cc2bcbaf1)

- Cleanup some code related to pgbench log checks in TAP tests. This fixes a
  couple of problems within the so-said code of this commit subject: - Repl=
ace
  the use of open() with slurp_file(), fixing an issue reported by buildfarm
  member fairywren whose perl installation keep around CRLF characters, cau=
sing
  the matching patterns for the logs to fail. - Remove the eval block, whic=
h is
  not really necessary.  This set of issues has come into light after fixin=
g a
  different issue with c13585fe, and this is wrong since this code has been
  introduced.  Reported-by: Andrew Dunstan, and buildfarm member fairywren
  Author: Michael Paquier Reviewed-by: Andrew Dunstan Discussion:
  [https://postgr.es/m/0f49303e-7784-b3ee-200b-cbf67be2eb9e@dunslane.net](h=
ttps://postgr.es/m/0f49303e-7784-b3ee-200b-cbf67be2eb9e@dunslane.net)
  Backpatch-through: 11
  [https://git.postgresql.org/pg/commitdiff/38ff135d9466c35b506b1049fedef73=
047907be0](https://git.postgresql.org/pg/commitdiff/38ff135d9466c35b506b104=
9fedef73047907be0)

- Remove some useless logs from the TAP tests of pgbench. 002_pgbench_no_se=
rver
  was printing some array pointers instead of the actual contents of those
  arrays for the expected outputs of stdout and stderr for a tested command.
  This does not add any new information that can help with debugging as the=
 test
  names allow to track failure locations, if any.  This commit simply remov=
es
  those logs as the rest of the printed information is redundant with
  command_checks_all().  Per discussion with Andrew Dunstan and =C3=81lvaro=
 Herrera.
  Discussion:
  [https://postgr.es/m/YNXNFaG7IgkzZanD@paquier.xyz](https://postgr.es/m/YN=
XNFaG7IgkzZanD@paquier.xyz)
  Backpatch-through: 11
  [https://git.postgresql.org/pg/commitdiff/704e1dbd9aa29a0b46c356f1803ad55=
cbdef2c20](https://git.postgresql.org/pg/commitdiff/704e1dbd9aa29a0b46c356f=
1803ad55cbdef2c20)

- Remove non-existing variable reference in MSVC's Solution.pm. The version
  string is grabbed from PACKAGE_VERSION in pg_config.h in the MSVC build s=
ince
  8f4fb4c6, but an error message referenced a variable that existed before =
that.
  This had no consequences except if one messes up enough with the version
  number of the build.  Author: Anton Voloshin Discussion:
  [https://postgr.es/m/af79ee1b-9962-b299-98e1-f90a289e19e6@postgrespro.ru]=
(https://postgr.es/m/af79ee1b-9962-b299-98e1-f90a289e19e6@postgrespro.ru)
  Backpatch-through: 13
  [https://git.postgresql.org/pg/commitdiff/d5a2c413fcdd187dc16c4fab16610af=
7d4849cc1](https://git.postgresql.org/pg/commitdiff/d5a2c413fcdd187dc16c4fa=
b16610af7d4849cc1)

Bruce Momjian pushed:

- doc:  add PG 14 relnote item about array function references. User-defined
  objects that reference some built-in array functions will need to be recr=
eated
  in PG 14.  Reported-by: Justin Pryzby  Discussion:
  [https://postgr.es/m/20210608225618.GR16435@telsasoft.com](https://postgr=
.es/m/20210608225618.GR16435@telsasoft.com)
  [https://git.postgresql.org/pg/commitdiff/25dfb5a831a1b8a83a8a68453b4bbe3=
8a5ef737e](https://git.postgresql.org/pg/commitdiff/25dfb5a831a1b8a83a8a684=
53b4bbe38a5ef737e)

- doc:  PG 14 relnote updates. Reported-by: Justin Pryzby  Discussion:
  [https://postgr.es/m/20210612034551.GU16435@telsasoft.com](https://postgr=
.es/m/20210612034551.GU16435@telsasoft.com)
  [https://git.postgresql.org/pg/commitdiff/a2559d4093725592a3fd5da8a4c7ac7=
c883115a0](https://git.postgresql.org/pg/commitdiff/a2559d4093725592a3fd5da=
8a4c7ac7c883115a0)

- doc:  PG 14 relnotes fixes. Items related to logical replication attribut=
ion
  and BRIN indexes  Reported-by: Tomas Vondra, John Naylor  Discussion:
  [https://postgr.es/m/0db66294-a668-2caa-2b5e-a8db60b30662@enterprisedb.co=
m,](https://postgr.es/m/0db66294-a668-2caa-2b5e-a8db60b30662@enterprisedb.c=
om,)
  CAFBsxsH21KnteYdk33F1oZu2O726NSD6_XBq51Tn0jytsA1AnA@mail.gmail.com
  [https://git.postgresql.org/pg/commitdiff/86b222b09071d3918c5c55b164d22b2=
236d3f872](https://git.postgresql.org/pg/commitdiff/86b222b09071d3918c5c55b=
164d22b2236d3f872)

- doc:  add mention of +4GB windows file handling in PG14 relnotes. Reporte=
d-by:
  Masahiko Sawada  Discussion:
  [https://postgr.es/m/CAD21AoCTHyouoGv-xt1qNjjvPbGMErLi0AJncByTvr66Nq7j8g@=
mail.gmail.com](https://postgr.es/m/CAD21AoCTHyouoGv-xt1qNjjvPbGMErLi0AJncB=
yTvr66Nq7j8g@mail.gmail.com)
  [https://git.postgresql.org/pg/commitdiff/90855908b751d40f67352fa0252e0fc=
daa7e317b](https://git.postgresql.org/pg/commitdiff/90855908b751d40f67352fa=
0252e0fcdaa7e317b)

- doc:  adjust PG 14 relnotes to be current.
  [https://git.postgresql.org/pg/commitdiff/69a58bfe4ab05567a8fab8bdce7f309=
5ed06b99c](https://git.postgresql.org/pg/commitdiff/69a58bfe4ab05567a8fab8b=
dce7f3095ed06b99c)

=C3=81lvaro Herrera pushed:

- Fix logic bug in 1632ea43682f. I overlooked that one condition was logica=
lly
  inverted.  The fix is a little bit more involved than simply negating the
  condition, to make the code easier to read.  Fix some outdated comments l=
eft
  by the same commit, while at it.  Author: Masahiko Sawada
  <sawada.mshk@gmail.com> Author: =C3=81lvaro Herrera <alvherre@alvh.no-ip.=
org>
  Reviewed-by: Amit Kapila <amit.kapila16@gmail.com> Discussion:
  [https://postgr.es/m/YMRlmB3/lZw8YBH+@paquier.xyz](https://postgr.es/m/YM=
RlmB3/lZw8YBH+@paquier.xyz)
  [https://git.postgresql.org/pg/commitdiff/33c509956704e1d918077b51ccd954f=
2e201a8f5](https://git.postgresql.org/pg/commitdiff/33c509956704e1d918077b5=
1ccd954f2e201a8f5)

- Add test case for obsoleting slot with active walsender. The code to sign=
al a
  running walsender when its reserved WAL size grows too large is completely
  uncovered before this commit; this adds coverage for that case.  This test
  involves sending SIGSTOP to walsender and walreceiver and running a check=
point
  while advancing WAL, then sending SIGCONT.  There's no precedent for this
  coding in Perl tests, and my reading of relevant manpages says it's likel=
y to
  fail on Windows.  Because of this, this test is always skipped on that
  platform.  Author: =C3=81lvaro Herrera <alvherre@alvh.no-ip.org> Discussi=
on:
  [https://postgr.es/m/202106102202.mjw4huiix7lo@alvherre.pgsql](https://po=
stgr.es/m/202106102202.mjw4huiix7lo@alvherre.pgsql)
  [https://git.postgresql.org/pg/commitdiff/09126984a2631db8dd6299f23954e0d=
ede69735f](https://git.postgresql.org/pg/commitdiff/09126984a2631db8dd6299f=
23954e0dede69735f)

- Revert "Add test case for obsoleting slot with active walsender". This re=
verts
  commit 09126984a263; the test case added there failed once in circumstanc=
es
  that remain mysterious.  It seems better to remove the test for now so th=
at
  14beta2 doesn't have random failures built in.
  [https://git.postgresql.org/pg/commitdiff/96795176810b979a2bf1f4563bdcb16=
1679d5b95](https://git.postgresql.org/pg/commitdiff/96795176810b979a2bf1f45=
63bdcb161679d5b95)

- Add test case for obsoleting slot with active walsender, take 2. The code=
 to
  signal a running walsender when its reserved WAL size grows too large is
  completely uncovered before this commit; this adds coverage for that case.
  This test involves sending SIGSTOP to walsender and walreceiver, then
  advancing enough WAL for a checkpoint to trigger, then sending SIGCONT.
  There's no precedent for STOP signalling in Perl tests, and my reading of
  relevant manpages says it's likely to fail on Windows.  Because of this, =
this
  test is always skipped on that platform.  This version fixes a couple of
  rarely hit race conditions in the previous attempt 09126984a263; most not=
ably,
  both LOG string searches are loops, not just the second one; we acquire t=
he
  start-of-log position before STOP-signalling; and reference the correct
  process name in the test description.  All per Tom Lane.  Author: =C3=81l=
varo
  Herrera <alvherre@alvh.no-ip.org> Discussion:
  [https://postgr.es/m/202106102202.mjw4huiix7lo@alvherre.pgsql](https://po=
stgr.es/m/202106102202.mjw4huiix7lo@alvherre.pgsql)
  [https://git.postgresql.org/pg/commitdiff/24043c27b46f873211177e3460ab09d=
c011802a5](https://git.postgresql.org/pg/commitdiff/24043c27b46f873211177e3=
460ab09dc011802a5)

Noah Misch pushed:

- Remove pg_wait_for_backend_termination(). It was unable to wait on a back=
end
  that had already left the procarray. Users tolerant of that limitation can
  poll pg_stat_activity.  Other users can employ the "timeout" argument of
  pg_terminate_backend().  Reviewed by Bharath Rupireddy.  Discussion:
  [https://postgr.es/m/20210605013236.GA208701@rfd.leadboat.com](https://po=
stgr.es/m/20210605013236.GA208701@rfd.leadboat.com)
  [https://git.postgresql.org/pg/commitdiff/5f1df62a459b51ab3bb625a0ee5e655=
429254257](https://git.postgresql.org/pg/commitdiff/5f1df62a459b51ab3bb625a=
0ee5e655429254257)

- Copy-edit text for the pg_terminate_backend() "timeout" parameter. Revert=
 the
  pg_description entry to its v13 form, since those messages usually remain
  shorter and don't discuss individual parameters.  No catversion bump, sin=
ce
  pg_description content does not impair backend compatibility or applicati=
on
  compatibility.  Justin Pryzby  Discussion:
  [https://postgr.es/m/20210612182743.GY16435@telsasoft.com](https://postgr=
.es/m/20210612182743.GY16435@telsasoft.com)
  [https://git.postgresql.org/pg/commitdiff/0aac73e6a2602696d23aa7a96862049=
65f9093dc](https://git.postgresql.org/pg/commitdiff/0aac73e6a2602696d23aa7a=
9686204965f9093dc)

- Finish rename of PQtraceSetFlags() to PQsetTraceFlags(). Jie Zhang
  Discussion:
  [https://postgr.es/m/TYWPR01MB767844835390EDD8DB276D75F90A9@TYWPR01MB7678=
.jpnprd01.prod.outlook.com](https://postgr.es/m/TYWPR01MB767844835390EDD8DB=
276D75F90A9@TYWPR01MB7678.jpnprd01.prod.outlook.com)
  [https://git.postgresql.org/pg/commitdiff/047a259e35b9dde2dad5fd0e5d5d784=
bb327b848](https://git.postgresql.org/pg/commitdiff/047a259e35b9dde2dad5fd0=
e5d5d784bb327b848)

Amit Kapila pushed:

- Fix decoding of speculative aborts. During decoding for speculative inser=
ts,
  we were relying for cleaning toast hash on confirmation records or next c=
hange
  records. But that could lead to multiple problems (a) memory leak if ther=
e is
  neither a confirmation record nor any other record after toast insertion =
for a
  speculative insert in the transaction, (b) error and assertion failures i=
f the
  next operation is not an insert/update on the same table.  The fix is to =
start
  queuing spec abort change and clean up toast hash and change record durin=
g its
  processing. Currently, we are queuing the spec aborts for both toast and =
main
  table even though we perform cleanup while processing the main table's sp=
ec
  abort record. Later, if we have a way to distinguish between the spec abo=
rt
  record of toast and the main table, we can avoid queuing the change for s=
pec
  aborts of toast tables.  Reported-by: Ashutosh Bapat Author: Dilip Kumar
  Reviewed-by: Amit Kapila Backpatch-through: 9.6, where it was introduced
  Discussion:
  [https://postgr.es/m/CAExHW5sPKF-Oovx_qZe4p5oM6Dvof7_P+XgsNAViug15Fm99jA@=
mail.gmail.com](https://postgr.es/m/CAExHW5sPKF-Oovx_qZe4p5oM6Dvof7_P+XgsNA=
Viug15Fm99jA@mail.gmail.com)
  [https://git.postgresql.org/pg/commitdiff/4daa140a2f50e9a160bc180c3997ee1=
3c7aabf9e](https://git.postgresql.org/pg/commitdiff/4daa140a2f50e9a160bc180=
c3997ee13c7aabf9e)

- Document a few caveats in synchronous logical replication. In a synchrono=
us
  logical setup, locking [user] catalog tables can cause deadlock. This is
  because logical decoding of transactions can lock catalog tables to access
  them so exclusively locking those in transactions can lead to deadlock. To
  avoid this users must refrain from having exclusive locks on catalog tabl=
es.
  Author: Takamichi Osumi Reviewed-by: Vignesh C, Amit Kapila Backpatch-thr=
ough:
  9.6 Discussion:
  [https://www.postgresql.org/message-id/20210222222847.tpnb6eg3yiykzpky%40=
alap3.anarazel.de](https://www.postgresql.org/message-id/20210222222847.tpn=
b6eg3yiykzpky%40alap3.anarazel.de)
  [https://git.postgresql.org/pg/commitdiff/3cb828dbe26087e7754f49f3cfe3ed0=
36d5af439](https://git.postgresql.org/pg/commitdiff/3cb828dbe26087e7754f49f=
3cfe3ed036d5af439)

- Handle no replica identity index case in RelationGetIdentityKeyBitmap. Co=
mmit
  e7eea52b2d has introduced a new function RelationGetIdentityKeyBitmap whi=
ch
  omits to handle the case where there is no replica identity index on a
  relation.  Author: Mark Dilger Reviewed-by: Takamichi Osumi, Amit Kapila
  Discussion:
  [https://www.postgresql.org/message-id/4C99A862-69C8-431F-960A-81B1151F1B=
89@enterprisedb.com](https://www.postgresql.org/message-id/4C99A862-69C8-43=
1F-960A-81B1151F1B89@enterprisedb.com)
  [https://git.postgresql.org/pg/commitdiff/2731ce1bd550d08f3fdd7bcb1497af4=
b95170976](https://git.postgresql.org/pg/commitdiff/2731ce1bd550d08f3fdd7bc=
b1497af4b95170976)

- Doc: Update caveats in synchronous logical replication. Reported-by: Simon
  Riggs Author: Takamichi Osumi Reviewed-by: Amit Kapila Backpatch-through:=
 9.6
  Discussion:
  [https://www.postgresql.org/message-id/20210222222847.tpnb6eg3yiykzpky@al=
ap3.anarazel.de](https://www.postgresql.org/message-id/20210222222847.tpnb6=
eg3yiykzpky@alap3.anarazel.de)
  [https://git.postgresql.org/pg/commitdiff/c66fb78ebb4f473bb4fd8773de3cda9=
339e14f81](https://git.postgresql.org/pg/commitdiff/c66fb78ebb4f473bb4fd877=
3de3cda9339e14f81)

- Doc: Update logical replication message formats. Commits 9de77b5453 and
  ac4645c015 missed to update the logical replication message formats secti=
on in
  the docs.  Author: Brar Piening Reviewed-by: Amit Kapila Discussion:
  [https://www.postgresql.org/message-id/cc70956c-e578-e54f-49e6-b5d68c8957=
6f@gmx.de](https://www.postgresql.org/message-id/cc70956c-e578-e54f-49e6-b5=
d68c89576f@gmx.de)
  [https://git.postgresql.org/pg/commitdiff/f08722cf83ab1fabee948a4e5754bf6=
236ad700b](https://git.postgresql.org/pg/commitdiff/f08722cf83ab1fabee948a4=
e5754bf6236ad700b)

- Doc: Fix minor formatting issue in logicaldecoding.sgml. Author: Guillaume
  Lelarge Discussion:
  [https://www.postgresql.org/message-id/CAECtzeXf3_oZoU6mgFCOy5+pDZ5n4XtH0=
Da4a5n_KacraVWiHQ@mail.gmail.com](https://www.postgresql.org/message-id/CAE=
CtzeXf3_oZoU6mgFCOy5+pDZ5n4XtH0Da4a5n_KacraVWiHQ@mail.gmail.com)
  [https://git.postgresql.org/pg/commitdiff/847c62ee76cbc237b3a204ca94b1b12=
811d698e3](https://git.postgresql.org/pg/commitdiff/847c62ee76cbc237b3a204c=
a94b1b12811d698e3)

Alexander Korotkov pushed:

- Support for unnest(multirange) and cast multirange as an array of ranges.=
 It
  has been spotted that multiranges lack of ability to decompose them into
  individual ranges.  Subscription and proper expanded object representation
  require substantial work, and it's too late for v14.  This commit provide=
s the
  implementation of unnest(multirange) and cast multirange as an array of
  ranges, which is quite trivial.  unnest(multirange) is defined as a
  polymorphic procedure.  The catalog description of the cast underlying
  procedure is duplicated for each multirange type because we don't have
  anyrangearray polymorphic type to use here.  Catversion is bumped.
  Reported-by: Jonathan S. Katz Discussion:
  [https://postgr.es/m/flat/60258efe-bd7e-4886-82e1-196e0cac5433%40postgres=
ql.org](https://postgr.es/m/flat/60258efe-bd7e-4886-82e1-196e0cac5433%40pos=
tgresql.org)
  Author: Alexander Korotkov Reviewed-by: Justin Pryzby, Jonathan S. Katz,
  Zhihong Yu
  [https://git.postgresql.org/pg/commitdiff/29854ee8d1ca4a46adb7e84deb17e6f=
b18e531cc](https://git.postgresql.org/pg/commitdiff/29854ee8d1ca4a46adb7e84=
deb17e6fb18e531cc)

- Add missing type name "multirange" in docs chapter title. Discussion:
  [https://postgr.es/m/CAFj8pRDioOxiJgmgw9TqQqZ3CxnJC4P5B2Oospf2eMgAjJuewA%=
40mail.gmail.com](https://postgr.es/m/CAFj8pRDioOxiJgmgw9TqQqZ3CxnJC4P5B2Oo=
spf2eMgAjJuewA%40mail.gmail.com)
  Author: Pavel Stehule, Alexander Korotkov Reviewed-by: Justin Pryzby, Tom=
 Lane
  [https://git.postgresql.org/pg/commitdiff/ad2da246c69dd5ec55764d4b6066f3b=
0c0d24ca2](https://git.postgresql.org/pg/commitdiff/ad2da246c69dd5ec55764d4=
b6066f3b0c0d24ca2)

- Revert 29854ee8d1 due to buildfarm failures. Reported-by: Tom Lane Discus=
sion:
  [https://postgr.es/m/CAPpHfdvcnw3x7jdV3r52p4%3D5S4WUxBCzcQKB3JukQHoicv1LS=
Q%40mail.gmail.com](https://postgr.es/m/CAPpHfdvcnw3x7jdV3r52p4%3D5S4WUxBCz=
cQKB3JukQHoicv1LSQ%40mail.gmail.com)
  [https://git.postgresql.org/pg/commitdiff/817bb0a7d1e02df4643d37304aed857=
4cf43f629](https://git.postgresql.org/pg/commitdiff/817bb0a7d1e02df4643d373=
04aed8574cf43f629)

Peter Geoghegan pushed:

- Remove unneeded field from VACUUM state. Bugfix commit 5fc89376 effective=
ly
  made the lock_waiter_detected field from vacuumlazy.c's global state stru=
ct
  into private state owned by lazy_truncate_heap().  Finish this off by
  replacing the struct field with a local variable.
  [https://git.postgresql.org/pg/commitdiff/958cfbcf2dd338e3179c2d8a35f48bd=
e020eba60](https://git.postgresql.org/pg/commitdiff/958cfbcf2dd338e3179c2d8=
a35f48bde020eba60)

- Support disabling index bypassing by VACUUM. Generalize the INDEX_CLEANUP
  VACUUM parameter (and the corresponding reloption): make it into a ternary
  style boolean parameter.  It now exposes a third option, "auto".  The "au=
to"
  option (which is now the default) enables the "bypass index vacuuming"
  optimization added by commit 1e55e7d1.  "VACUUM (INDEX_CLEANUP TRUE)" is
  redefined to once again make VACUUM simply do any required index vacuumin=
g,
  regardless of how few dead tuples are encountered during the first scan o=
f the
  target heap relation (unless there are exactly zero).  This gives users a=
 way
  of opting out of the "bypass index vacuuming" optimization, if for whatev=
er
  reason that proves necessary.  It is also expected to be used by PostgreS=
QL
  developers as a testing option from time to time.  "VACUUM (INDEX_CLEANUP
  FALSE)" does the same thing as it always has: it forcibly disables both i=
ndex
  vacuuming and index cleanup.  It's not expected to be used much in Postgr=
eSQL
  14.  The failsafe mechanism added by commit 1e55e7d1 addresses the same
  problem in a simpler way. INDEX_CLEANUP can now be thought of as a testin=
g and
  compatibility option.  Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By:
  Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Justin Pryzby
  <pryzby@telsasoft.com> Discussion:
  [https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXcrhqsv6OHQ@=
mail.gmail.com](https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8F=
cXcrhqsv6OHQ@mail.gmail.com)
  [https://git.postgresql.org/pg/commitdiff/3499df0dee8c4ea51d264a674df5b5e=
31991319a](https://git.postgresql.org/pg/commitdiff/3499df0dee8c4ea51d264a6=
74df5b5e31991319a)

- Remove overzealous VACUUM failsafe assertions. The failsafe can trigger w=
hen
  index processing is already disabled. This can happen when VACUUM's
  INDEX_CLEANUP parameter is "off" and the failsafe happens to trigger.  Re=
move
  assertions that assume that index processing is directly tied to the fail=
safe.
  Oversight in commit c242baa4, which made it possible for the failsafe to
  trigger in a two-pass strategy VACUUM that has yet to make its first call=
 to
  lazy_vacuum_all_indexes().
  [https://git.postgresql.org/pg/commitdiff/e8f201ab82be234b2f103234cf9f262=
f9afeaeba](https://git.postgresql.org/pg/commitdiff/e8f201ab82be234b2f10323=
4cf9f262f9afeaeba)

- Add list of ignorable pgindent commits for git-blame. Add a
  .git-blame-ignore-revs file with a list of pgindent, pgperlyidy, and
  reformat-dat-files commit hashes.  Postgres hackers that configure git to=
 use
  the ignore file will get git-blame output that avoids attributing line ch=
anges
  to the ignored indent commits.  This makes git-blame output much easier to
  work with in practice.  Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By:=
 Tom
  Lane <tgl@sss.pgh.pa.us> Discussion:
  [https://postgr.es/m/CAH2-Wz=3DcVh3GHTP6SdLU-Gnmt2zRdF8vZkcrFdSzXQ=3DWhbW=
m9Q@mail.gmail.com](https://postgr.es/m/CAH2-Wz=3DcVh3GHTP6SdLU-Gnmt2zRdF8v=
ZkcrFdSzXQ=3DWhbWm9Q@mail.gmail.com)
  [https://git.postgresql.org/pg/commitdiff/8e638845ff6bb255ad1dea15831089a=
234535391](https://git.postgresql.org/pg/commitdiff/8e638845ff6bb255ad1dea1=
5831089a234535391)

Andrew Dunstan pushed:

- Further refinement of stuck_on_old_timeline recovery test. TestLib::perl2=
host
  can take a file argument as well as a directory argument, so that code be=
comes
  substantially simpler. Also add comments on why we're using forward slash=
es,
  and why we're setting PERL_BADLANG=3D0.  Discussion:
  [https://postgr.es/m/e9947bcd-20ee-027c-f0fe-01f736b7e345@dunslane.net](h=
ttps://postgr.es/m/e9947bcd-20ee-027c-f0fe-01f736b7e345@dunslane.net)
  [https://git.postgresql.org/pg/commitdiff/54a5ed22016940d7ad5060ed62d2347=
3924756a1](https://git.postgresql.org/pg/commitdiff/54a5ed22016940d7ad5060e=
d62d23473924756a1)

- Don't set a fast default for anything but a plain table. The fast default=
 code
  added in Release 11 omitted to check that the table a fast default was be=
ing
  added to was a plain table. Thus one could be added to a foreign table, w=
hich
  predicably blows up. Here we perform that check.  In addition, on the back
  branches, since some of these might have escaped into the wild, if we
  encounter a missing value for an attribute of something other than a plain
  table we ignore it.  Fixes bug #17056  Backpatch to release 11,  Reviewed=
 by:
  Andres Freund, =C3=81lvaro Herrera and Tom Lane
  [https://git.postgresql.org/pg/commitdiff/0a4efdc7ebf2584257b166c87e82797=
eb92815b5](https://git.postgresql.org/pg/commitdiff/0a4efdc7ebf2584257b166c=
87e82797eb92815b5)

Heikki Linnakangas pushed:

- Fix outdated comment that talked about seek position of WAL file. Since c=
ommit
  c24dcd0cfd, we have been using pg_pread() to read the WAL file, which doe=
sn't
  change the seek position (unless we fall back to the implementation in
  src/port/pread.c). Update comment accordingly.  Backpatch-through: 12, wh=
ere
  we started to use pg_pread()
  [https://git.postgresql.org/pg/commitdiff/d0303bc8d2670d11c9df9220aa08a2c=
33529e100](https://git.postgresql.org/pg/commitdiff/d0303bc8d2670d11c9df922=
0aa08a2c33529e100)

- Tidy up GetMultiXactIdMembers()'s behavior on error. One of the error pat=
hs
  left `*members` uninitialized. That's not a live bug, because most caller=
s don't
  look at `*members` when the function returns -1, but let's be tidy. One c=
aller,
  in heap_lock_tuple(), does "if (members !=3D NULL) pfree(members)", but A=
FAICS
  it never passes an invalid 'multi' value so it should not reach that error
  case.  The callers are also a bit inconsistent in their expectations.
  heap_lock_tuple() pfrees the 'members' array if it's not-NULL, others pfr=
ee()
  it if "nmembers >=3D 0", and others if "nmembers > 0". That's not a live =
bug
  either, because the function should never return 0, but add an Assert for=
 that
  to make it more clear. I left the callers alone for now.  I also moved the
  line where we set `*nmembers`. It wasn't wrong before, but I like to do t=
hat
  right next to the 'return' statement, to make it clear that it's always s=
et on
  return.  Also remove one unreachable return statement after ereport(ERROR=
),
  for brevity and for consistency with the similar if-block right after it.
  Author: Greg Nancarrow with the additional changes by me Backpatch-throug=
h:
  9.6, all supported versions
  [https://git.postgresql.org/pg/commitdiff/d24c5658a80c8f5037e9e1948de311d=
3f3350f12](https://git.postgresql.org/pg/commitdiff/d24c5658a80c8f5037e9e19=
48de311d3f3350f12)

- Prevent race condition while reading relmapper file. Contrary to the comm=
ent
  here, POSIX does not guarantee atomicity of a read(), if another process =
calls
  write() concurrently. Or at least Linux does not. Add locking to
  load_relmap_file() to avoid the race condition.  Fixes bug #17064. Thanks=
 to
  Alexander Lakhin for the report and test case.  Backpatch-through: 9.6, a=
ll
  supported versions. Discussion:
  [https://www.postgresql.org/message-id/17064-bb0d7904ef72add3@postgresql.=
org](https://www.postgresql.org/message-id/17064-bb0d7904ef72add3@postgresq=
l.org)
  [https://git.postgresql.org/pg/commitdiff/b6d8d2073f228d9f7198f1f9826d8f6=
ab643c219](https://git.postgresql.org/pg/commitdiff/b6d8d2073f228d9f7198f1f=
9826d8f6ab643c219)

- Another fix to relmapper race condition. In previous commit, I missed that
  relmap_redo() was also not acquiring the RelationMappingLock. Thanks to T=
homas
  Munro for pointing that out.  Backpatch-through: 9.6, like previous commi=
t.
  Discussion:
  [https://www.postgresql.org/message-id/CA%2BhUKGLev%3DPpOSaL3WRZgOvgk217e=
t%2BbxeJcRr4eR-NttP1F6Q%40mail.gmail.com](https://www.postgresql.org/messag=
e-id/CA%2BhUKGLev%3DPpOSaL3WRZgOvgk217et%2BbxeJcRr4eR-NttP1F6Q%40mail.gmail=
.com)
  [https://git.postgresql.org/pg/commitdiff/9b8ed0f52bffceaccf6fa650ffe482e=
7da20fbf2](https://git.postgresql.org/pg/commitdiff/9b8ed0f52bffceaccf6fa65=
0ffe482e7da20fbf2)

Tom=C3=A1=C5=A1 Vondra pushed:

- Fix copying data into slots with FDW batching. Commit b676ac443b optimized
  handling of tuple slots with bulk inserts into foreign tables, so that the
  slots are initialized only once and reused for all batches. The data was
  however copied into the slots only after the initialization, inserting
  duplicate values when the slot gets reused. Fixed by moving the ExecCopyS=
lot
  outside the init branch.  The existing postgres_fdw tests failed to catch=
 this
  due to inserting data into foreign tables without unique indexes, and then
  checking only the number of inserted rows. This adds a new test with both=
 a
  unique index and a check of inserted values.  Reported-by: Alexander Pyha=
lov
  Discussion:
  [https://postgr.es/m/7a8cf8d56b3d18e5c0bccd6cd42d04ac%40postgrespro.ru](h=
ttps://postgr.es/m/7a8cf8d56b3d18e5c0bccd6cd42d04ac%40postgrespro.ru)
  [https://git.postgresql.org/pg/commitdiff/99cea49d6525e30bc3768e4ffb95377=
e7584dea1](https://git.postgresql.org/pg/commitdiff/99cea49d6525e30bc3768e4=
ffb95377e7584dea1)

Fujii Masao pushed:

- Make archiver process handle barrier events. Commit d75288fb27 made WAL
  archiver process an auxiliary process. An auxiliary process needs to hand=
le
  barrier events but the commit forgot to make archiver process do that.
  Reported-by: Thomas Munro Author: Fujii Masao Reviewed-by: Thomas Munro
  Discussion:
  [https://postgr.es/m/CA+hUKGLah2w1pWKHonZP_+EQw69=3Dq56AHYwCgEN8GDzsRG_Hg=
w@mail.gmail.com](https://postgr.es/m/CA+hUKGLah2w1pWKHonZP_+EQw69=3Dq56AHY=
wCgEN8GDzsRG_Hgw@mail.gmail.com)
  [https://git.postgresql.org/pg/commitdiff/981524d2e3aa3f28d364c472e34f538=
6be846811](https://git.postgresql.org/pg/commitdiff/981524d2e3aa3f28d364c47=
2e34f5386be846811)

Peter Eisentraut pushed:

- amcheck: Fix code comments. Code comments were claiming that verify_heapa=
m()
  was checking privileges on the relation it was operating on, but it didn't
  actually do that.  Perhaps earlier versions of the patch did that, but no=
w the
  access is regulated by privileges on the function.  Remove the wrong comm=
ents.
  [https://git.postgresql.org/pg/commitdiff/97b7134186490b36e01efc9d2feaf70=
38c666457](https://git.postgresql.org/pg/commitdiff/97b7134186490b36e01efc9=
d2feaf7038c666457)

- Translation updates. Source-Git-URL:
  git://git.postgresql.org/git/pgtranslation/messages.git Source-Git-Hash:
  70796ae860c444c764bb591c885f22cac1c168ec
  [https://git.postgresql.org/pg/commitdiff/a7bb0ce58f56ee8907c3f49c52d99f5=
02536c796](https://git.postgresql.org/pg/commitdiff/a7bb0ce58f56ee8907c3f49=
c52d99f502536c796)

- Remove redundant variable pageSize in gistinitpage. In gistinitpage, page=
Size
  variable looks redundant, instead just pass BLCKSZ. This will be consiste=
nt
  with its peers BloomInitPage, brin_page_init and SpGistInitPage.  Author:
  Bharath Rupireddy <bharath.rupireddy@enterprisedb.com> Discussion:
  [https://www.postgresql.org/message-id/flat/CALj2ACWj=3DV1k5591eeZK2sOg2F=
YuBUp6azFO8tMkBtGfXf8PMQ@mail.gmail.com](https://www.postgresql.org/message=
-id/flat/CALj2ACWj=3DV1k5591eeZK2sOg2FYuBUp6azFO8tMkBtGfXf8PMQ@mail.gmail.c=
om)
  [https://git.postgresql.org/pg/commitdiff/a60c4c5c1a99746485123ae93fbd3e5=
8c78e5d62](https://git.postgresql.org/pg/commitdiff/a60c4c5c1a99746485123ae=
93fbd3e58c78e5d62)

- doc: Change reloption data type spelling for consistency. Use "floating p=
oint"
  rather than "float4", like everywhere else in this context.  Author:
  Shinya11.Kato@nttdata.com Discussion:
  [https://www.postgresql.org/message-id/flat/TYAPR01MB28965989AF84B57FC351=
B97BC40F9@TYAPR01MB2896.jpnprd01.prod.outlook.com](https://www.postgresql.o=
rg/message-id/flat/TYAPR01MB28965989AF84B57FC351B97BC40F9@TYAPR01MB2896.jpn=
prd01.prod.outlook.com)
  [https://git.postgresql.org/pg/commitdiff/63e6d05bf34da58eef7cd1ed461b9c8=
315177a38](https://git.postgresql.org/pg/commitdiff/63e6d05bf34da58eef7cd1e=
d461b9c8315177a38)

- Fixes in ALTER SUBSCRIPTION DROP PUBLICATION code. ALTER SUBSCRIPTION DROP
  PUBLICATION does not actually support copy_data option, so remove it from=
 tab
  completion.  Also, reword the error message that is thrown when all the
  publications from a subscription are specified to be dropped.  Also, made=
 few
  doc and cosmetic adjustments.  Author: Vignesh C <vignesh21@gmail.com>
  Reviewed-by: Bharath Rupireddy <bharath.rupireddy@enterprisedb.com>
  Reviewed-by: Japin Li <japinli@hotmail.com> Discussion:
  [https://www.postgresql.org/message-id/flat/CALDaNm21RwsDzs4xj14ApteAF7au=
yyomHNnp+NEL-sH8m-jMvQ@mail.gmail.com](https://www.postgresql.org/message-i=
d/flat/CALDaNm21RwsDzs4xj14ApteAF7auyyomHNnp+NEL-sH8m-jMvQ@mail.gmail.com)
  [https://git.postgresql.org/pg/commitdiff/e59d428f34297cd0eb67e3b4e4b8c8b=
c58504921](https://git.postgresql.org/pg/commitdiff/e59d428f34297cd0eb67e3b=
4e4b8c8bc58504921)

- Put option listing back into alphabetical order.
  [https://git.postgresql.org/pg/commitdiff/3af10943ce21450e299b3915b9cad47=
cd90369e9](https://git.postgresql.org/pg/commitdiff/3af10943ce21450e299b391=
5b9cad47cd90369e9)

- Error message refactoring. Take some untranslatable things out of the mes=
sage
  and replace by format placeholders, to reduce translatable strings and re=
duce
  translation mistakes.
  [https://git.postgresql.org/pg/commitdiff/c302a6139072fed9410204fa9e751d9=
5930e05ff](https://git.postgresql.org/pg/commitdiff/c302a6139072fed9410204f=
a9e751d95930e05ff)

David Rowley pushed:

- Fix assert failure in expand_grouping_sets. linitial_node() fails in asse=
rt
  enabled builds if the given pointer is not of the specified type.  Here t=
he
  type is IntList.  The code thought it should be expecting List, but it was
  wrong.  In the existing tests which run this code the initial list elemen=
t is
  always NIL.  Since linitial_node() allows NULL, we didn't trigger any ass=
ert
  failures in the existing regression tests.  There is still some discussio=
n as
  to whether we need a few more tests in this area, but for now, since beta=
2 is
  looming, fix the bug first.  Bug: #17067 Discussion:
  [https://postgr.es/m/17067-665d50fa321f79e0@postgresql.org](https://postg=
r.es/m/17067-665d50fa321f79e0@postgresql.org)
  Reported-by: Yaoguang Chen
  [https://git.postgresql.org/pg/commitdiff/8d29d45d9b3cab95a866efbcdd9138b=
3d76741b3](https://git.postgresql.org/pg/commitdiff/8d29d45d9b3cab95a866efb=
cdd9138b3d76741b3)

Andres Freund pushed:

- Use correct horizon when vacuuming catalog relations. In dc7420c2c92 I
  (Andres) accidentally used RelationIsAccessibleInLogicalDecoding() as the=
 sole
  condition to use the non-shared catalog horizon in
  GetOldestNonRemovableTransactionId(). That is incorrect, as
  RelationIsAccessibleInLogicalDecoding() checks whether wal_level is logic=
al.
  The correct check, as done e.g. in GlobalVisTestFor(), is to check
  IsCatalogRelation() and RelationIsAccessibleInLogicalDecoding().  The obs=
erved
  misbehavior of this bug was that there could be an endless loop in
  lazy_scan_prune(), because the horizons used in heap_page_prune() and the
  individual tuple liveliness checks did not match. Likely there are other
  potential consequences as well.  A later commit will unify the determinat=
ion
  which horizon has to be used, and add additional assertions to make it ea=
sier
  to catch a bug like this.  Reported-By: Justin Pryzby <pryzby@telsasoft.c=
om>
  Diagnosed-By: Matthias van de Meent <boekewurm+postgres@gmail.com> Author:
  Matthias van de Meent <boekewurm+postgres@gmail.com> Discussion:
  [https://postgr.es/m/CAEze2Wg32Y9+WJfw=3DaofkRx1ZRFt_Ev6bNPc4PSaz7PjSFtZg=
Q@mail.gmail.com](https://postgr.es/m/CAEze2Wg32Y9+WJfw=3DaofkRx1ZRFt_Ev6bN=
Pc4PSaz7PjSFtZgQ@mail.gmail.com)
  [https://git.postgresql.org/pg/commitdiff/5a1e1d83022b976ebdec5cfa8f255c4=
278b75b8e](https://git.postgresql.org/pg/commitdiff/5a1e1d83022b976ebdec5cf=
a8f255c4278b75b8e)

Joe Conway pushed:

- Stamp 14beta2.
  [https://git.postgresql.org/pg/commitdiff/bafad2c5b261a1449bed60ebac9e791=
8ebcc42b4](https://git.postgresql.org/pg/commitdiff/bafad2c5b261a1449bed60e=
bac9e7918ebcc42b4)

Thomas Munro pushed:

- Prepare for forthcoming LLVM 13 API change. LLVM 13 (due out in September=
) has
  changed the semantics of LLVMOrcAbsoluteSymbols(), so we need to bump some
  reference counts to avoid a double-free that causes crashes and bad query
  results.  A proactive change seems necessary to avoid having a window of =
time
  where our respective latest releases would interact badly.  It's possible=
 that
  the situation could change before then, though.  Thanks to Fabien Coelho =
for
  monitoring bleeding edge LLVM and Andres Freund for tracking down the cha=
nge.
  Back-patch to 11, where the JIT code arrived.  Discussion:
  [https://postgr.es/m/CA%2BhUKGLEy8mgtN7BNp0ooFAjUedDTJj5dME7NxLU-m91b85si=
A%40mail.gmail.com](https://postgr.es/m/CA%2BhUKGLEy8mgtN7BNp0ooFAjUedDTJj5=
dME7NxLU-m91b85siA%40mail.gmail.com)
  [https://git.postgresql.org/pg/commitdiff/9b4e4cfe66ff133717c1b8ba3c2725d=
525c3e67c](https://git.postgresql.org/pg/commitdiff/9b4e4cfe66ff133717c1b8b=
a3c2725d525c3e67c)

# Pending Patches

Nitin Jadhav sent in another revision of a patch to make a way to track the
progress of the operations performed during the startup process.

Yuzuko Hosoya sent in a patch to control whether autoanalyze runs for parti=
tion
operations via new GUCs autovacuum_analyze_attach_partition,
autovacuum_analyze_detach_partition, and autovacuum_analyze_drop_partition.

Peter Eisentraut sent in a patch to add an index OID macro argument to
DECLARE_INDEX.  Instead of defining symbols such as AmOidIndexId explicitly,
include them as an argument of DECLARE_INDEX() and have genbki.pl generate =
the
way as the table OID symbols from the CATALOG() declaration

Filip Gospodinov sent in a patch to intended to fix a bug that manifested a=
s the
shipped pkg-config file is broken for statically linking libpq because
libpgcommon and libpgport are missing. The fix adds those two missing priva=
te
dependencies.

Heikki Linnakangas sent in another revision of a patch to split xlog.c into
xlog.c and xlogrecovery.c.  This moves the functions related to performing =
WAL
recovery into the new xlogrecovery.c source file, leaving xlog.c responsibl=
e for
maintaining the WAL buffers, coordinating the startup and switch from recov=
ery
to normal operations, and other miscellaneous stuff that have always been in
xlog.c

Greg Nancarrow sent in two revisions of a patch to remove the useless int64
range checks on BIGINT sequence MINVALUE/MAXVALUE values.

Peter Geoghegan sent in three revisions of a patch to add a list of ignorab=
le
pgindent commits for "git blame".

David Rowley sent in another revision of a patch to add a new hash table ty=
pe
which has stable pointers, and use same to speed up SMgr.

Peter Smith and Ajin Cherian traded patches to add an option to support pre=
pared
transactions to built-in logical replication.

Daniel Gustafsson and Micha=C3=ABl Paquier traded patches to document some =
SSL/TLS
related acronyms, and replace usage of SSL with SSL/TLS, the latter being m=
ore
accurate and up to date.

Simon Riggs sent in three revisions of a patch to add a documentation chapt=
er on
hash indexes.

Atsushi Torikoshi sent in another revision of a patch to add a function to =
log
the complete query string and its plan for the query currently running on t=
he
backend with the specified process ID.

Li Japin sent in a patch to make a minor code beautification in
RelationGetIdentityKeyBitmap.

Bertrand Drouvot sent in another revision of a patch to enable minimal logi=
cal
decoding on standbys.

Alexander Pyhalov sent in another revision of a patch to allow pushing CASE
expressions to foreign servers.

Peter Eisentraut sent in a patch to Make the Unicode makefile more
parallel-safe, Make UCS_to_most.pl process encodings in sorted order, remove
some whitespace in generated C output to get it in line with the coding sty=
le of
the rest of the project, simplify the code-generation code, and fix indenta=
tion
in generated output.

Maxim Orlov sent in another revision of a patch to fix parallel worker fail=
ed
assertion and coredump.

Vigneshwaran C sent in two more revisions of a patch to add schema level su=
pport
for PUBLICATIONs.

Mike Fiedler  sent in a patch to emit the namespace in post-copy output.

Emre Hasegeli sent in a patch to decouple operator classes from index access
methods.

Jacob Champion sent in two revisions of a patch to decouple the SASL framew=
ork
from the SCRAM code, as it may be useful separately.

Yugo Nagata sent in another revision of a patch to address some Pgbench err=
ors
by using the Variables structure for client variables.  This is most import=
ant
when it is used to reset client variables during the repeating of transacti=
ons
after serialization/deadlock failures.  Don't allocate Variable structs one=
 by
one. Instead, add a constant margin each time it overflows. Also included, =
a fix
for pgbench errors and serialization/deadlock retries.

Jacob Champion, Micha=C3=ABl Paquier, and Tom Lane traded patches to make i=
t possible
to use jsonapi with libpq.

Gurjeet Singh sent in a patch to add automatic notification for top transac=
tion
IDs.

Haiying Tang sent in another revision of a patch to support tab completion =
with
a query result for upper case inputs in psql.

Peter Eisentraut sent in a patch to make genbki error handling more useful =
by
counting the errors and error out at the end instead of just writing warnin=
gs
for invalid cross-catalog lookups.

Daniel Gustafsson sent in another revision of a patch to support NSS as a T=
LS
backend for libpq.

Andrey V. Lepikhov sent in a patch to inform genericcostestimate()'s index
selectivity value of the fact that in the case of an unique one-row btree
index, scan only one row can be returned.

Tom=C3=A1=C5=A1 Vondra sent in another revision of a patch to make it possi=
ble to
replicate sequences in logical decoding.

Ranier Vilela sent in a patch to fix an uninitialized copy_data var in
src/backend/commands/subscriptioncmds.c.

Peter Eisentraut sent in another revision of a patch to improve error messa=
ges
about mismatching relkind by making the primary error message shorter and s=
aying
more directly that what was attempted is not possible.

Peter Eisentraut sent in a patch to add tests for UNBOUNDED syntax ambiguit=
y.

Simon Riggs sent in a patch to make pgbench use COPY FREEZE.

Justin Pryzby sent in a patch to avoid double parentheses, and fix a commen=
t in
tablefunc.c to refer to the correct comment.

Craig Ringer sent in a patch to add some more PGDLLIMPORTs to expose some o=
f the
things heretofore unavailable on Windows.

Amit Kapila sent in a patch to allow streaming the changes after speculative
aborts.

Hayato Kuroda sent in a patch to make ECPG's new DECLARE STATEMENT work for
DEALLOCATE and DESCRIBE.

Bruce Momjian sent in another revision of a patch to implement cluster file
encryption.

Jie Zhang sent in a patch to bring the generated sample postgresql.conf into
line with what's actually the default for `shared_buffers`.

Richard Guo sent in a patch to use each rel as both outer and inner for anti
joins.

Tom Lane sent in a patch to prevent any calls of abort() or exit() from ins=
ide
libpq.

Alexander Korotkov sent in a patch to fix some small inconsistencies in the
catalog definition of multirange operators.

Julien Rouhaud sent in another revision of a patch to preserve param locati=
on
when generating a const node.

Andrey Borodin sent in another revision of a patch to reorganize the pglz
compression code.

Julien Rouhaud and Ranier Vilela traded patches to add a pg_get_query_def()=
 to
deparse and print a rewritten SQL statement.

Tom=C3=A1=C5=A1 Vondra sent in a PoC patch to do cardinality estimation usi=
ng runtime
sampling.
--===============0330074949510022916==
Content-Type: text/html; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable


<!doctype html>
<html>
  <head>
    <meta name=3D"viewport" content=3D"width=3Ddevice-width">
    <meta http-equiv=3D"Content-Type" content=3D"text/html; charset=3DUTF-8=
">
    <title>PostgreSQL Weekly News - June 27, 2021</title>
    <style>

    @media only screen and (max-width: 620px) {
      table[class=3Dbody] h1 {
        font-size: 28px !important;
        margin-bottom: 10px !important;
      }
      table[class=3Dbody] p,
            table[class=3Dbody] ul,
            table[class=3Dbody] ol,
            table[class=3Dbody] td,
            table[class=3Dbody] span,
            table[class=3Dbody] a {
        font-size: 16px !important;
      }
      table[class=3Dbody] .wrapper,
            table[class=3Dbody] .article {
        padding: 10px !important;
      }
      table[class=3Dbody] .content {
        padding: 0 !important;
      }
      table[class=3Dbody] .container {
        padding: 0 !important;
        width: 100% !important;
      }
      table[class=3Dbody] .main {
        border-left-width: 0 !important;
        border-radius: 0 !important;
        border-right-width: 0 !important;
      }
      table[class=3Dbody] .btn table {
        width: 100% !important;
      }
      table[class=3Dbody] .btn a {
        width: 100% !important;
      }
      table[class=3Dbody] .img-responsive {
        height: auto !important;
        max-width: 100% !important;
        width: auto !important;
      }
    }

    @media all {
      .ExternalClass {
        width: 100%;
      }
      .ExternalClass,
            .ExternalClass p,
            .ExternalClass span,
            .ExternalClass font,
            .ExternalClass td,
            .ExternalClass div {
        line-height: 100%;
      }
      .apple-link a {
        color: inherit !important;
        font-family: inherit !important;
        font-size: inherit !important;
        font-weight: inherit !important;
        line-height: inherit !important;
        text-decoration: none !important;
      }
      #MessageViewBody a {
        color: inherit;
        text-decoration: none;
        font-size: inherit;
        font-family: inherit;
        font-weight: inherit;
        line-height: inherit;
      }
      .btn-primary table td:hover {
        background-color: #34495e !important;
      }
      .btn-primary a:hover {
        background-color: #34495e !important;
        border-color: #34495e !important;
      }
    }
    </style>
  </head>
  <body class=3D"" style=3D"background-color: #f6f6f6; font-family: sans-se=
rif; -webkit-font-smoothing: antialiased; font-size: 14px; line-height: 1.4=
; margin: 0; padding: 0; -ms-text-size-adjust: 100%; -webkit-text-size-adju=
st: 100%;">
    <table border=3D"0" cellpadding=3D"0" cellspacing=3D"0" class=3D"body" =
style=3D"border-collapse: separate; mso-table-lspace: 0pt; mso-table-rspace=
: 0pt; width: 100%; background-color: #f6f6f6;">
      <tr>
        <td style=3D"font-family: sans-serif; font-size: 14px; vertical-ali=
gn: top;">&nbsp;</td>
        <td class=3D"container" style=3D"font-family: sans-serif; font-size=
: 14px; vertical-align: top; display: block; Margin: 0 auto; max-width: 580=
px; padding: 10px; width: 580px;">
          <div class=3D"content" style=3D"box-sizing: border-box; display: =
block; Margin: 0 auto; max-width: 580px; padding: 10px;">


            <span class=3D"preheader" style=3D"color: transparent; display:=
 none; height: 0; max-height: 0; max-width: 0; opacity: 0; overflow: hidden=
; mso-hide: all; visibility: hidden; width: 0;"></span>
            <table class=3D"main" style=3D"border-collapse: separate; mso-t=
able-lspace: 0pt; mso-table-rspace: 0pt; width: 100%; background: #ffffff; =
border-radius: 3px;">


              <tr>
                <td class=3D"wrapper" style=3D"font-family: sans-serif; fon=
t-size: 14px; vertical-align: top; box-sizing: border-box; padding: 20px;">
                  <table border=3D"0" cellpadding=3D"0" cellspacing=3D"0" s=
tyle=3D"border-collapse: separate; mso-table-lspace: 0pt; mso-table-rspace:=
 0pt; width: 100%;">
                    <tr>
                      <td style=3D"font-family: sans-serif; font-size: 14px=
; vertical-align: top;">

<div>
<h1 style=3D"color: #000; font-family: sans-serif; line-height: 1.4; margin=
: 0; margin-bottom: 30px; font-size: 25px; font-weight: 300; text-align: ce=
nter">PostgreSQL Weekly News - June 27, 2021</h1>
</div>
<h1 style=3D"color: #000; font-family: sans-serif; line-height: 1.4; margin=
: 0; margin-bottom: 30px; font-size: 25px; font-weight: 300; text-align: ce=
nter">PostgreSQL Weekly News - June 27, 2021</h1>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">PostgreSQL 14 Beta 2 released.
<a href=3D"https://www.postgresql.org/about/news/postgresql-14-beta-2-relea=
sed-2249/" style=3D"color: #3498db; text-decoration: underline">Test!</a></=
p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">The 12th PostgreSQL Conference Cuba (Postgr=
esqlCUBA, @PgCuba) will take place on
November 18-19, 2021, at the Hotel Habana Libre.  This event is part of the
TECNOGET conference, and we will host one track dedicated to PostgreSQL rel=
ated
talks. For more details, reach out to cu AT postgresql DOT org.</p>
<h1 style=3D"color: #000; font-family: sans-serif; line-height: 1.4; margin=
: 0; margin-bottom: 30px; font-size: 25px; font-weight: 300; text-align: ce=
nter">PostgreSQL Product News</h1>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">JDBC 42.2.22
<a href=3D"https://jdbc.postgresql.org/documentation/changelog.html#version=
_42.2.22" style=3D"color: #3498db; text-decoration: underline">released</a>=
</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">pgpoolAdmin 4.2.0, the administration tool =
for Pgpool-II,
<a href=3D"https://www.pgpool.net/docs/pgpoolAdmin/NEWS" style=3D"color: #3=
498db; text-decoration: underline">released</a>.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">pgCenter 0.9.0, a command-line admin tool f=
or observing and troubleshooting
PostgreSQL, <a href=3D"https://github.com/lesovsky/pgcenter/releases/tag/v0=
.9.0" style=3D"color: #3498db; text-decoration: underline">released</a></p>
<h1 style=3D"color: #000; font-family: sans-serif; line-height: 1.4; margin=
: 0; margin-bottom: 30px; font-size: 25px; font-weight: 300; text-align: ce=
nter">PostgreSQL Jobs for June</h1>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px"><a href=3D"https://archives.postgresql.org/=
pgsql-jobs/2021-06/" style=3D"color: #3498db; text-decoration: underline">h=
ttps://archives.postgresql.org/pgsql-jobs/2021-06/</a></p>
<h1 style=3D"color: #000; font-family: sans-serif; line-height: 1.4; margin=
: 0; margin-bottom: 30px; font-size: 25px; font-weight: 300; text-align: ce=
nter">PostgreSQL in the News</h1>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Planet PostgreSQL: <a href=3D"https://plane=
t.postgresql.org/" style=3D"color: #3498db; text-decoration: underline">htt=
ps://planet.postgresql.org/</a></p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">PostgreSQL Weekly News is brought to you th=
is week by David Fetter</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Submit news and announcements by Sunday at =
3:00pm PST8PDT to david@fetter.org.</p>
<h1 style=3D"color: #000; font-family: sans-serif; line-height: 1.4; margin=
: 0; margin-bottom: 30px; font-size: 25px; font-weight: 300; text-align: ce=
nter">Applied Patches</h1>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Tom Lane pushed:</p>
<ul style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal;=
 margin: 0; margin-bottom: 15px">
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Work around portability issue with newer ve=
rsions of mktime(). Recent glibc
  versions have made mktime() fail if tm_isdst is inconsistent with the
  prevailing timezone; in particular it fails for tm_isdst =3D 1 when the z=
one is
  UTC.  (This seems wildly inconsistent with the POSIX-mandated treatment of
  "incorrect" values for the other fields of struct tm, so if you ask me it=
's a
  bug, but I bet they'll say it's intentional.)  This has been observed to =
cause
  cosmetic problems when pg_restore'ing an archive created in a different
  timezone.  To fix, do mktime() using the field values from the archive, a=
nd if
  that fails try again with tm_isdst =3D -1.  This will give a result that'=
s off
  by the UTC-offset difference from the original zone, but that was true be=
fore,
  too.  It's not terribly critical since we don't do anything with the resu=
lt
  except possibly print it.  (Someday we should flush this entire bit of lo=
gic
  and record a standard-format timestamp in the archive instead.  That's not
  okay for a back-patched bug fix, though.)  Also, guard our only other use=
 of
  mktime() by having initdb's build_time_t() set tm_isdst =3D -1 not 0.  Th=
is case
  could only have an issue in zones that are DST year-round; but I think so=
me do
  exist, or could in future.  Per report from Wells Oliver.  Back-patch to =
all
  supported versions, since any of them might need to run with a newer glib=
c.
  Discussion:
  <a href=3D"https://postgr.es/m/CAOC+FBWDhDHO7G-i1_n_hjRzCnUeFO+H-Czi1y10m=
FhRWpBrew@mail.gmail.com" style=3D"color: #3498db; text-decoration: underli=
ne">https://postgr.es/m/CAOC+FBWDhDHO7G-i1_n_hjRzCnUeFO+H-Czi1y10mFhRWpBrew=
@mail.gmail.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/f807e3410fdfc29ced6590c=
7c2afa76637e001ad" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/f807e3410fdfc29ced6590c7c2afa76637e00=
1ad</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Remove orphaned expected-result file. This =
should have been removed in
  43e084197, which removed the corresponding spec file.  Noted while fooling
  about with the isolationtester.
  <a href=3D"https://git.postgresql.org/pg/commitdiff/ffbe9dec13599fa786ea6=
567df1c6a3f3ee3c673" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/ffbe9dec13599fa786ea6567df1c6a3f3ee=
3c673</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Update variant expected-result file. This s=
hould have been updated in
  d2d8a229b, but it was overlooked. According to 31a877f18 which added it, =
this
  file is meant to show the results you get under default_transaction_isola=
tion
  =3D serializable. We've largely lost track of that goal in other isolation
  tests, but as long as we've got this one, it should be right.  Noted while
  fooling about with the isolationtester.
  <a href=3D"https://git.postgresql.org/pg/commitdiff/0a1e80c5c4f094087257f=
c4284a87e0bc7bca591" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/0a1e80c5c4f094087257fc4284a87e0bc7b=
ca591</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Remove another orphan expected-result file.=
 aborted-keyrevoke_2.out was
  apparently needed when it was added (in commit 0ac5ad513) to handle the c=
ase
  of serializable transaction mode. However, the output in serializable mode
  actually matches the regular aborted-keyrevoke.out file, and AFAICT has d=
one
  so for a long time. There's no need to keep dragging this variant along.
  <a href=3D"https://git.postgresql.org/pg/commitdiff/f6352a0d4e437ac8bc266=
f77df22d064592056c9" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/f6352a0d4e437ac8bc266f77df22d064592=
056c9</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Update another variant expected-result file=
. This should have been updated in
  533e9c6b0, but it was overlooked. Given the lack of complaints, I won't b=
other
  back-patching.
  <a href=3D"https://git.postgresql.org/pg/commitdiff/d3c878499c9d639ff06e0=
664d06b8c731e30c2fc" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/d3c878499c9d639ff06e0664d06b8c731e3=
0c2fc</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Improve SQLSTATE reporting in some replicat=
ion-related code. I started out
  with the goal of reporting ERRCODE_CONNECTION_FAILURE when walrcv_connect=
()
  fails, but as I looked around I realized that whoever wrote this code was=
 of
  the opinion that errcodes are purely optional.  That's not my understandi=
ng of
  our project policy.  Hence, make sure that an errcode is provided in each
  ereport that (a) is ERROR or higher level and (b) isn't arguably an inter=
nal
  logic error. Also fix some very dubious existing errcode assignments.  Wh=
ile
  this is not per policy, it's also largely cosmetic, since few of these ca=
ses
  could get reported to applications.  So I don't feel a need to back-patch.
  Discussion:
  <a href=3D"https://postgr.es/m/2189704.1623512522@sss.pgh.pa.us" style=3D=
"color: #3498db; text-decoration: underline">https://postgr.es/m/2189704.16=
23512522@sss.pgh.pa.us</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/6b787d9e32005867ee3660d=
1ea20f447810a403d" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/6b787d9e32005867ee3660d1ea20f447810a4=
03d</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Fix plancache refcount leak after error in =
ExecuteQuery. When stuffing a plan
  from the plancache into a Portal, one is not supposed to risk throwing an
  error between GetCachedPlan and PortalDefineQuery; if that happens, the p=
lan
  refcount incremented by GetCachedPlan will be leaked.  I managed to break=
 this
  rule while refactoring code in 9dbf2b7d7.  There is no visible consequence
  other than some memory leakage, and since nobody is very likely to trigge=
r the
  relevant error conditions many times in a row, it's not surprising we hav=
en't
  noticed.  Nonetheless, it's a bug, so rearrange the order of operations to
  remove the hazard.  Noted on the way to looking for a better fix for bug
  #17053. This mistake is pretty old, so back-patch to all supported branch=
es.
  <a href=3D"https://git.postgresql.org/pg/commitdiff/131ea3e908d3c97a2fe1a=
b25cce5046dd5cb905f" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/131ea3e908d3c97a2fe1ab25cce5046dd5c=
b905f</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Centralize the logic for protective copying=
 of utility statements. In the
  "simple Query" code path, it's fine for parse analysis or execution of a
  utility statement to scribble on the statement's node tree, since that'll=
 just
  be thrown away afterwards.  However it's not fine if the node tree is in =
the
  plan cache, as then it'd be corrupted for subsequent executions.  Up to n=
ow
  we've dealt with that by having individual utility-statement functions ap=
ply
  copyObject() if they were going to modify the tree.  But that's prone to
  errors of omission.  Bug #17053 from Charles Samborski shows that CREATE/=
ALTER
  DOMAIN didn't get this memo, and can crash if executed repeatedly from pl=
an
  cache.  In the back branches, we'll just apply a narrow band-aid for that=
, but
  in HEAD it seems prudent to have a more principled fix that will close of=
f the
  possibility of other similar bugs in future. Hence, let's hoist the
  responsibility for doing copyObject up into ProcessUtility from its child=
ren,
  thus ensuring that it happens for all utility statement types.  Also, mod=
ify
  ProcessUtility's API so that its callers can tell it whether a copy step =
is
  necessary.  It turns out that in all cases, the immediate caller knows wh=
ether
  the node tree is transient, so this doesn't involve a huge amount of code
  thrashing.  In this way, while we lose a little bit in the execute-from-c=
ache
  code path due to sometimes copying node trees that wouldn't be mutated an=
yway,
  we gain something in the simple-Query code path by not copying throwaway =
node
  trees.  Statements that are complex enough to be expensive to copy are al=
most
  certainly ones that would have to be copied anyway, so the loss in the ca=
che
  code path shouldn't be much.  (Note that this whole problem applies only =
to
  utility statements. Optimizable statements don't have the issue because we
  long ago made the executor treat Plan trees as read-only.  Perhaps someda=
y we
  will make utility statement execution act likewise, but I'm not holding my
  breath.)  Discussion:
  <a href=3D"https://postgr.es/m/931771.1623893989@sss.pgh.pa.us" style=3D"=
color: #3498db; text-decoration: underline">https://postgr.es/m/931771.1623=
893989@sss.pgh.pa.us</a>
  Discussion:
  <a href=3D"https://postgr.es/m/17053-3ca3f501bbc212b4@postgresql.org" sty=
le=3D"color: #3498db; text-decoration: underline">https://postgr.es/m/17053=
-3ca3f501bbc212b4@postgresql.org</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/7c337b6b527b7052e6a751f=
966d5734c56f668b5" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/7c337b6b527b7052e6a751f966d5734c56f66=
8b5</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Improve version reporting in pgbench. Commi=
t 547f04e73 caused pgbench to start
  printing its version number, which seems like a fine idea, but it needs a=
 bit
  more work: * Print the server version number too, when different. * Print=
 the
  PG_VERSION string, not some reconstructed approximation.  This patch copi=
es
  psql's well-tested code for the same purpose.  Discussion:
  <a href=3D"https://postgr.es/m/1226654.1624036821@sss.pgh.pa.us" style=3D=
"color: #3498db; text-decoration: underline">https://postgr.es/m/1226654.16=
24036821@sss.pgh.pa.us</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/84bee9610965331d5110971=
d8de390a5bbe2effc" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/84bee9610965331d5110971d8de390a5bbe2e=
ffc</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Fix misbehavior of DROP OWNED BY with dupli=
cate polroles entries. Ordinarily,
  a pg_policy.polroles array wouldn't list the same role more than once; but
  CREATE POLICY does not prevent that.  If we perform DROP OWNED BY on a ro=
le
  that is listed more than once, RemoveRoleFromObjectPolicy either suffered=
 an
  assertion failure or encountered a tuple-updated-by-self error.  Rewrite =
it to
  cope correctly with duplicate entries, and add a CommandCounterIncrement =
call
  to prevent the other problem.  Per discussion, there's other cleanup that
  ought to happen here, but this seems like the minimum essential fix.  Per=
 bug
  #17062 from Alexander Lakhin.  It's been broken all along, so back-patch =
to
  all supported branches.  Discussion:
  <a href=3D"https://postgr.es/m/17062-11f471ae3199ca23@postgresql.org" sty=
le=3D"color: #3498db; text-decoration: underline">https://postgr.es/m/17062=
-11f471ae3199ca23@postgresql.org</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/d21fca084356946664bfce1=
9d66d2df2bb873cbd" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/d21fca084356946664bfce19d66d2df2bb873=
cbd</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Provide feature-test macros for libpq featu=
res added in v14. We had a request
  to provide a way to test at compile time for the availability of the new
  pipeline features.  More generally, it seems like a good idea to provide =
a way
  to test via #ifdef for all new libpq API features.  People have been usin=
g the
  version from pg_config.h for that; but that's more likely to represent the
  server version than the libpq version, in the increasingly-common scenario
  where they're different.  It's safer if libpq-fe.h itself is the source of
  truth about what features it offers.  Hence, establish a policy that star=
ting
  in v14 we'll add a suitable feature-is-present macro to libpq-fe.h when w=
e add
  new API there. (There doesn't seem to be much point in applying this poli=
cy
  retroactively, but it's not too late for v14.)  Tom Lane and Alvaro Herre=
ra,
  per suggestion from Boris Kolpackov.  Discussion:
  <a href=3D"https://postgr.es/m/boris.20210617102439@codesynthesis.com" st=
yle=3D"color: #3498db; text-decoration: underline">https://postgr.es/m/bori=
s.20210617102439@codesynthesis.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/6991e774e0304f5ef488cf1=
ae4fa79578b6ae3d5" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/6991e774e0304f5ef488cf1ae4fa79578b6ae=
3d5</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Stabilize test case added by commit f61db90=
9d. Buildfarm members ayu and tern
  have sometimes shown a different plan than expected for this query.  I'd =
been
  unable to reproduce that before today, but I finally realized what is
  happening. If there is a concurrent open transaction (probably an autovac=
uum
  run in the buildfarm, but this can also be arranged manually), then the i=
ndex
  entries for the rows removed by the DELETE a few lines up are not killed
  promptly, causing a change in the planner's estimate of the extremal valu=
e of
  ft2.c1, which moves the rowcount estimate for "c1 &gt; 1100" by enough to=
 change
  the join plan from nestloop to hash.  To fix, change the query condition =
to
  "c1 &gt; 1000", causing the hash plan to be preferred whether or not a co=
ncurrent
  open transaction exists.  Since this UPDATE is tailored to be a no-op, no=
thing
  else changes.  Report:
  <a href=3D"https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=3Dayu&=
amp;dt=3D2021-06-09%2022%3A45%3A48" style=3D"color: #3498db; text-decoratio=
n: underline">https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=3Dayu=
&amp;dt=3D2021-06-09%2022%3A45%3A48</a>
  Report:
  <a href=3D"https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=3Dayu&=
amp;dt=3D2021-06-13%2022%3A38%3A18" style=3D"color: #3498db; text-decoratio=
n: underline">https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=3Dayu=
&amp;dt=3D2021-06-13%2022%3A38%3A18</a>
  Report:
  <a href=3D"https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=3Dtern=
&amp;dt=3D2021-06-20%2004%3A55%3A36" style=3D"color: #3498db; text-decorati=
on: underline">https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=3Dte=
rn&amp;dt=3D2021-06-20%2004%3A55%3A36</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/5843659d091bfb6f2c60e01=
0ea1fd00e55ee6ada" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/5843659d091bfb6f2c60e010ea1fd00e55ee6=
ada</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Restore the portal-level snapshot for simpl=
e expressions, too. Commits
  84f5c2908 et al missed the need to cover plpgsql's "simple expression" co=
de
  path.  If the first thing we execute after a COMMIT/ROLLBACK is one of th=
ose,
  rather than a full-fledged SPI command, we must explicitly do
  EnsurePortalSnapshotExists() to make sure we have an outer snapshot.  Note
  that it wouldn't be good enough to just push a snapshot for the duration =
of
  the expression execution: what comes back might be toasted, so we'd better
  have a snapshot protecting it.  The test case demonstrating this fact che=
ats a
  bit by marking a SQL function immutable even though it fetches from a tab=
le.
  That's nothing that users haven't been seen to do, though.  Per report fr=
om
  Jim Nasby.  Back-patch to v11, like the previous fix.  Discussion:
  <a href=3D"https://postgr.es/m/378885e4-f85f-fc28-6c91-c4d1c080bf26@amazo=
n.com" style=3D"color: #3498db; text-decoration: underline">https://postgr.=
es/m/378885e4-f85f-fc28-6c91-c4d1c080bf26@amazon.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/d102aafb6259a6a412803d4=
b1d8c4f00aa17f67e" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/d102aafb6259a6a412803d4b1d8c4f00aa17f=
67e</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Use annotations to reduce instability of is=
olation-test results. We've long
  contended with isolation test results that aren't entirely stable.  Some =
test
  scripts insert long delays to try to force stable results, which is not
  terribly desirable; but other erratic failure modes remain, causing
  unrepeatable buildfarm failures.  I've spent a fair amount of time trying=
 to
  solve this by improving the server-side support code, without much succes=
s:
  that way is fundamentally unable to cope with diffs that stem from chance
  ordering of arrival of messages from different server processes.  We can
  improve matters on the client side, however, by annotating the test scrip=
ts
  themselves to show the desired reporting order of events that might occur=
 in
  different orders.  This patch adds three types of annotations to deal wit=
h (a)
  test steps that might or might not complete their waits before the
  isolationtester can see them waiting; (b) test steps in different sessions
  that can legitimately complete in either order; and (c) NOTIFY messages t=
hat
  might arrive before or after the completion of a step in another session.=
  We
  might need more annotation types later, but this seems to be enough to de=
al
  with the instabilities we've seen in the buildfarm.  It also lets us get =
rid
  of all the long delays that were previously used, cutting more than a min=
ute
  off the runtime of the isolation tests.  Back-patch to all supported bran=
ches,
  because the buildfarm instabilities affect all the branches, and because =
it
  seems desirable to keep isolationtester's capabilities the same across all
  branches to simplify possible future back-patching of tests.  Discussion:
  <a href=3D"https://postgr.es/m/327948.1623725828@sss.pgh.pa.us" style=3D"=
color: #3498db; text-decoration: underline">https://postgr.es/m/327948.1623=
725828@sss.pgh.pa.us</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/741d7f1047fe52da7ced6fa=
9cea661ce9320c8d4" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/741d7f1047fe52da7ced6fa9cea661ce9320c=
8d4</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Improve display of query results in isolati=
on tests. Previously,
  isolationtester displayed SQL query results using some ad-hoc code that
  clearly hadn't had much effort expended on it. Field values longer than 14
  characters weren't separated from the next field, and usually caused
  misalignment of the columns too.  Also there was no visual separation of a
  query's result from subsequent isolationtester output.  This made test re=
sult
  files confusing and hard to read.  To improve matters, let's use libpq's
  PQprint() function.  Although that's long since unused by psql, it's still
  plenty good enough for the purpose here.  Like 741d7f104, back-patch to a=
ll
  supported branches, so that this isn't a stumbling block for back-patching
  isolation test changes.  Discussion:
  <a href=3D"https://postgr.es/m/582362.1623798221@sss.pgh.pa.us" style=3D"=
color: #3498db; text-decoration: underline">https://postgr.es/m/582362.1623=
798221@sss.pgh.pa.us</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/4a054069a36032a59afceb0=
7f3b837f09ab1a2e9" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/4a054069a36032a59afceb07f3b837f09ab1a=
2e9</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Don't assume GSSAPI result strings are null=
-terminated. Our uses of
  gss_display_status() and gss_display_name() assumed that the gss_buffer_d=
esc
  strings returned by those functions are null-terminated.  It appears that=
 they
  generally are, given the lack of field complaints up to now.  However, the
  available documentation does not promise this, and some man pages for
  gss_display_status() show examples that rely on the gss_buffer_desc.length
  field instead of expecting null termination.  Also, we now have a report =
that
  on some implementations, clang's address sanitizer is of the opinion that=
 the
  byte after the specified length is undefined.  Hence, change the code to =
rely
  on the length field instead.  This might well be cosmetic rather than fix=
ing
  any real bug, but it's hard to be sure, so back-patch to all supported
  branches. While here, also back-patch the v12 changes that made pg_GSS_er=
ror
  deal honestly with multiple messages available from gss_display_status.  =
Per
  report from Sudheer H R.  Discussion:
  <a href=3D"https://postgr.es/m/5372B6D4-8276-42C0-B8FB-BD0918826FC3@teken=
light.com" style=3D"color: #3498db; text-decoration: underline">https://pos=
tgr.es/m/5372B6D4-8276-42C0-B8FB-BD0918826FC3@tekenlight.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/126cdaf47af275f76b2f2dd=
b023bfdc6f018ae30" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/126cdaf47af275f76b2f2ddb023bfdc6f018a=
e30</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Doc: fix confusion about LEAKPROOF in synta=
x summaries. The syntax summaries
  for CREATE FUNCTION and allied commands made it look like LEAKPROOF is an
  alternative to IMMUTABLE/STABLE/VOLATILE, when of course it is an orthogo=
nal
  option.  Improve that.  Per gripe from aazamrafeeque0.  Thanks to David
  Johnston for suggestions.  Discussion:
  <a href=3D"https://postgr.es/m/162444349581.694.5818572718530259025@wrigl=
eys.postgresql.org" style=3D"color: #3498db; text-decoration: underline">ht=
tps://postgr.es/m/162444349581.694.5818572718530259025@wrigleys.postgresql.=
org</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/2031e1668e5577e64cfed29=
da69a34903d5a5227" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/2031e1668e5577e64cfed29da69a34903d5a5=
227</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Allow non-quoted identifiers as isolation t=
est session/step names. For no
  obvious reason, isolationtester has always insisted that session and step
  names be written with double quotes.  This is fairly tedious and does lit=
tle
  for test readability, especially since the names that people actually cho=
ose
  almost always look like normal identifiers.  Hence, let's tweak the lexer=
 to
  allow SQL-like identifiers not only double-quoted strings.  (They're SQL-=
like,
  not exactly SQL, because I didn't add any case-folding logic.  Also there=
's no
  provision for U&amp;"..." names, not that anyone's likely to care.)  Ther=
e is one
  incompatibility introduced by this change: if you write "foo""bar" with no
  space, that used to be taken as two identifiers, but now it's just one
  identifier with an embedded quote mark.  I converted all the
  src/test/isolation/ specfiles to remove unnecessary double quotes, but st=
opped
  there because my eyes were glazing over already.  Like 741d7f104, back-pa=
tch
  to all supported branches, so that this isn't a stumbling block for
  back-patching isolation test changes.  Discussion:
  <a href=3D"https://postgr.es/m/759113.1623861959@sss.pgh.pa.us" style=3D"=
color: #3498db; text-decoration: underline">https://postgr.es/m/759113.1623=
861959@sss.pgh.pa.us</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/a443c1b2d6a646cf90a8afc=
193c07ed12a2bf045" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/a443c1b2d6a646cf90a8afc193c07ed12a2bf=
045</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Further stabilize postgres_fdw test. The qu=
eries involving ft1_nopw don't
  stably return the same row anymore.  I surmise that an autovacuum hitting=
 "S
  1"."T 1" right after the updates introduced by f61db909d/5843659d0 freed =
some
  space, changing where subsequent insertions get stored. It's only by good=
 luck
  that these results were stable before, though, since a LIMIT without ORDE=
R BY
  isn't well defined, and it's not like we've ever treated that table as
  append-only in this test script.  Since we only really care whether these
  commands succeed or not, just replace <code>"SELECT *"</code> with "SELEC=
T 1".  Report:
  <a href=3D"https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=3Dcrak=
e&amp;dt=3D2021-06-23%2019%3A52%3A08" style=3D"color: #3498db; text-decorat=
ion: underline">https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=3Dc=
rake&amp;dt=3D2021-06-23%2019%3A52%3A08</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/802177090992511c610804d=
a54a4603d4f50c594" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/802177090992511c610804da54a4603d4f50c=
594</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Doc: remove commit f560209c6 from v14 relea=
se notes. Now that this has been
  back-patched, it's no longer a new feature for v14.
  <a href=3D"https://git.postgresql.org/pg/commitdiff/8a80562d732c0da1ddcc9=
fb88dfb976f4b846577" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/8a80562d732c0da1ddcc9fb88dfb976f4b8=
46577</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Remove unnecessary failure cases in RemoveR=
oleFromObjectPolicy(). It's not
  really necessary for this function to open or lock the relation associated
  with the pg_policy entry it's modifying.  The error checks it's making on=
 the
  rel are if anything counterproductive (e.g., if we don't want to allow
  installation of policies on system catalogs, here is not the place to pre=
vent
  that).  In particular, it seems just wrong to insist on an ownership chec=
k.
  That has the net effect of forcing people to use superuser for DROP OWNED=
 BY,
  which surely is not an effect we want.  Also there is no point in rebuild=
ing
  the dependencies of the policy expressions, which aren't being changed.
  Lastly, locking the table also seems counterproductive; it's not helping =
to
  prevent race conditions, since we failed to re-read the pg_policy row aft=
er
  acquiring the lock.  That means that concurrent DDL would likely result in
  "tuple concurrently updated/deleted" errors; which is the same behavior t=
his
  code will produce, with less overhead.  Per discussion of bug #17062.
  Back-patch to all supported versions, as the failure cases this eliminates
  seem just as undesirable in 9.6 as in HEAD.  Discussion:
  <a href=3D"https://postgr.es/m/1573181.1624220108@sss.pgh.pa.us" style=3D=
"color: #3498db; text-decoration: underline">https://postgr.es/m/1573181.16=
24220108@sss.pgh.pa.us</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/5a0f1c8c0193f0dd7fba50c=
22d96781fa2414007" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/5a0f1c8c0193f0dd7fba50c22d96781fa2414=
007</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Remove undesirable libpq dependency on stri=
nginfo.c. Commit c0cb87fbb unwisely
  introduced a dependency on the StringInfo machinery in fe-connect.c.  We =
must
  not use that in libpq, because it will do a summary exit(1) if it hits OO=
M,
  and that is not appropriate behavior for a general-purpose library.  The =
goal
  of allowing arbitrary line lengths in service files doesn't seem like it's
  worth a lot of effort, so revert back to the previous method of using a
  stack-allocated buffer and failing on buffer overflow.  This isn't an exa=
ct
  revert though.  I kept that patch's refactoring to have a single exit pat=
h, as
  that seems cleaner than having each error path know what to do to clean u=
p.
  Also, I made the fixed-size buffer 1024 bytes not 256, just to push off t=
he
  need for an expandable buffer some more.  There is more to do here; in
  particular the lack of any mechanical check for this type of mistake now =
seems
  pretty hazardous.  But this fix gets us back to the level of robustness w=
e had
  in v13, anyway.  Discussion:
  <a href=3D"https://postgr.es/m/daeb22ec6ca8ef61e94d766a9b35fb03cabed38e.c=
amel@vmware.com" style=3D"color: #3498db; text-decoration: underline">https=
://postgr.es/m/daeb22ec6ca8ef61e94d766a9b35fb03cabed38e.camel@vmware.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/8ec00dc5cd70e0e579e9fbf=
8661bc46f5ccd8078" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/8ec00dc5cd70e0e579e9fbf8661bc46f5ccd8=
078</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Doc: update v14 release notes for revert of=
 commit c0cb87fbb.
  <a href=3D"https://git.postgresql.org/pg/commitdiff/dcffc9ba8a1e0ab1b0a57=
e9b9d38e3dc9960f83f" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/dcffc9ba8a1e0ab1b0a57e9b9d38e3dc996=
0f83f</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Remove memory leaks in isolationtester. spe=
cscanner.l leaked a kilobyte of
  memory per token of the spec file. Apparently somebody thought that the
  introductory code block would be executed once; but it's once per yylex()
  call.  A couple of functions in isolationtester.c leaked small amounts of
  memory due to not bothering to free one-time allocations.  Might as well
  improve these so that valgrind gives this program a clean bill of health.
  Also get rid of an ugly static variable.  Coverity complained about one o=
f the
  one-time leaks, which led me to try valgrind'ing isolationtester, which l=
ed to
  discovery of the larger leak.
  <a href=3D"https://git.postgresql.org/pg/commitdiff/642c0697c96b9c6ba5d19=
4653a329f7820565f01" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/642c0697c96b9c6ba5d194653a329f78205=
65f01</a></p>
</li>
</ul>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Micha=C3=ABl Paquier pushed:</p>
<ul style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal;=
 margin: 0; margin-bottom: 15px">
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Remove forced toast recompression in VACUUM=
 FULL/CLUSTER. The extra checks
  added by the recompression of toast data introduced in bbe0a81 is proving=
 to
  have a performance impact on VACUUM or CLUSTER even if no recompression is
  done.  This is more noticeable with more toastable columns that contain
  non-NULL values.  Improvements could be done to make those extra checks l=
ess
  expensive, but that's not material for 14 at this stage, and we are not s=
ure
  either if the code path of VACUUM FULL/CLUSTER is adapted for this job.  =
Per
  discussion with several people, including Andres Freund, Robert Haas, =C3=
=81lvaro
  Herrera, Tom Lane and myself.  Discussion:
  <a href=3D"https://postgr.es/m/20210527003144.xxqppojoiwurc2iz@alap3.anar=
azel.de" style=3D"color: #3498db; text-decoration: underline">https://postg=
r.es/m/20210527003144.xxqppojoiwurc2iz@alap3.anarazel.de</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/dbab0c07e5ba1f19a991da2=
d72972a8fe9a41bda" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/dbab0c07e5ba1f19a991da2d72972a8fe9a41=
bda</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Improve handling of dropped objects in pg_e=
vent_trigger_ddl_commands(). An
  object found as dropped when digging into the list of objects returned by
  pg_event_trigger_ddl_commands() could cause a cache lookup error, as the =
calls
  grabbing for the object address and the type name would fail if the objec=
t was
  missing.  Those lookup errors could be seen with combinations of ALTER TA=
BLE
  sub-commands involving identity columns.  The lookup logic is changed in =
this
  code path to get a behavior similar to any other SQL-callable function by
  ignoring objects that are not found, taking advantage of 2a10fdc.  The
  back-branches are not changed, as they require this commit that is too
  invasive for stable branches.  While on it, add test cases to exercise ev=
ent
  triggers with identity columns, and stress more cases with the event
  ddl_command_end for relations.  Author: Sven Klemm, Aleksander Alekseev,
  Michael Paquier Discussion:
  <a href=3D"https://postgr.es/m/CAMCrgp2R1cEXU53iYKtW6yVEp2_yKUz+z=3D3-CTr=
YpPP+xryRtg@mail.gmail.com" style=3D"color: #3498db; text-decoration: under=
line">https://postgr.es/m/CAMCrgp2R1cEXU53iYKtW6yVEp2_yKUz+z=3D3-CTrYpPP+xr=
yRtg@mail.gmail.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/2d689babe3cb50dcb29f6ed=
595a61d56e518c0d8" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/2d689babe3cb50dcb29f6ed595a61d56e518c=
0d8</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">doc: Apply markup &lt;productname&gt; to Op=
enSSL more consistently. Author: Daniel
  Gustafsson Discussion:
  <a href=3D"https://postgr.es/m/CE12DD5C-4BB3-4166-BC9A-39779568734C@yesql=
.se" style=3D"color: #3498db; text-decoration: underline">https://postgr.es=
/m/CE12DD5C-4BB3-4166-BC9A-39779568734C@yesql.se</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/f80979f659d39e238e95444=
e6752142799428078" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/f80979f659d39e238e95444e6752142799428=
078</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Fix pattern matching logic for logs in TAP =
tests of pgbench. The logic
  checking for the format of per-thread logs used grep() with directly "$re=
",
  which would cause the test to consider all the logs as a match without ca=
ring
  about their format at all.  Using "/$re/" makes grep() perform a regex te=
st,
  which is what we want here.  While on it, improve some of the tests to be=
 more
  picky with the patterns expected and add more comments to describe the te=
sts.
  Issue discovered while digging into a separate patch.  Author: Fabien Coe=
lho,
  Michael Paquier Discussion:
  <a href=3D"https://postgr.es/m/YNPsPAUoVDCpPOGk@paquier.xyz" style=3D"col=
or: #3498db; text-decoration: underline">https://postgr.es/m/YNPsPAUoVDCpPO=
Gk@paquier.xyz</a>
  Backpatch-through: 11
  <a href=3D"https://git.postgresql.org/pg/commitdiff/c13585fe9e55813cf9fea=
c67fe7b65d3a78fff92" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/c13585fe9e55813cf9feac67fe7b65d3a78=
fff92</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">doc: Move remove_temp_files_after_crash to =
section for developer options. The
  main goal of this option is to allow inspecting temporary files for debug=
ging
  purposes, so moving the parameter there is natural.  Oversight in cd91de0.
  Reported-by: Justin Pryzby Author: Euler Taveira Discussion:
  <a href=3D"https://postgr.es/m/20210612004347.GP16435@telsasoft.com" styl=
e=3D"color: #3498db; text-decoration: underline">https://postgr.es/m/202106=
12004347.GP16435@telsasoft.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/797b0fc0b078c7b4c46ef9f=
2d9e47aa2d98c6c63" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/797b0fc0b078c7b4c46ef9f2d9e47aa2d98c6=
c63</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Add more debugging information with log che=
cks in TAP tests of pgbench.
  fairywren is not happy with the pattern checks introduced by c13585f. I a=
m not
  sure if this outlines a bug in pgbench or if the regex patterns used in t=
he
  tests are too restrictive for this buildfarm member's environment.  This =
adds
  more debugging information to show the log entries that do not match with=
 the
  expected pattern, to help in finding out what's happening.  That seems li=
ke a
  good addition in the long-term anyway as that may not be the only issue in
  this area.  Discussion:
  <a href=3D"https://postgr.es/m/YNUad2HvgW+6eXyo@paquier.xyz" style=3D"col=
or: #3498db; text-decoration: underline">https://postgr.es/m/YNUad2HvgW+6eX=
yo@paquier.xyz</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/87b2124dfa0782a697ea7b9=
0aff15a6f6117a720" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/87b2124dfa0782a697ea7b90aff15a6f6117a=
720</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">doc: Add acronyms for MITM and SNI. This ad=
ds MITM and SNI as acronyms, as the
  documentation already had them marked up with &lt;acronym&gt;.  While on =
it, make
  sure to spell man-in-the-middle with dashes consistently, and add acronym=
s for
  those new terms where appropriate.  Author: Daniel Gustafsson Reviewed-by:
  Michael Paquier Discussion:
  <a href=3D"https://postgr.es/m/CE12DD5C-4BB3-4166-BC9A-39779568734C@yesql=
.se" style=3D"color: #3498db; text-decoration: underline">https://postgr.es=
/m/CE12DD5C-4BB3-4166-BC9A-39779568734C@yesql.se</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/15ff5401d1719aaf6c9a47e=
5abea517cc2bcbaf1" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/15ff5401d1719aaf6c9a47e5abea517cc2bcb=
af1</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Cleanup some code related to pgbench log ch=
ecks in TAP tests. This fixes a
  couple of problems within the so-said code of this commit subject: - Repl=
ace
  the use of open() with slurp_file(), fixing an issue reported by buildfarm
  member fairywren whose perl installation keep around CRLF characters, cau=
sing
  the matching patterns for the logs to fail. - Remove the eval block, whic=
h is
  not really necessary.  This set of issues has come into light after fixin=
g a
  different issue with c13585fe, and this is wrong since this code has been
  introduced.  Reported-by: Andrew Dunstan, and buildfarm member fairywren
  Author: Michael Paquier Reviewed-by: Andrew Dunstan Discussion:
  <a href=3D"https://postgr.es/m/0f49303e-7784-b3ee-200b-cbf67be2eb9e@dunsl=
ane.net" style=3D"color: #3498db; text-decoration: underline">https://postg=
r.es/m/0f49303e-7784-b3ee-200b-cbf67be2eb9e@dunslane.net</a>
  Backpatch-through: 11
  <a href=3D"https://git.postgresql.org/pg/commitdiff/38ff135d9466c35b506b1=
049fedef73047907be0" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/38ff135d9466c35b506b1049fedef730479=
07be0</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Remove some useless logs from the TAP tests=
 of pgbench. 002_pgbench_no_server
  was printing some array pointers instead of the actual contents of those
  arrays for the expected outputs of stdout and stderr for a tested command.
  This does not add any new information that can help with debugging as the=
 test
  names allow to track failure locations, if any.  This commit simply remov=
es
  those logs as the rest of the printed information is redundant with
  command_checks_all().  Per discussion with Andrew Dunstan and =C3=81lvaro=
 Herrera.
  Discussion:
  <a href=3D"https://postgr.es/m/YNXNFaG7IgkzZanD@paquier.xyz" style=3D"col=
or: #3498db; text-decoration: underline">https://postgr.es/m/YNXNFaG7IgkzZa=
nD@paquier.xyz</a>
  Backpatch-through: 11
  <a href=3D"https://git.postgresql.org/pg/commitdiff/704e1dbd9aa29a0b46c35=
6f1803ad55cbdef2c20" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/704e1dbd9aa29a0b46c356f1803ad55cbde=
f2c20</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Remove non-existing variable reference in M=
SVC's Solution.pm. The version
  string is grabbed from PACKAGE_VERSION in pg_config.h in the MSVC build s=
ince
  8f4fb4c6, but an error message referenced a variable that existed before =
that.
  This had no consequences except if one messes up enough with the version
  number of the build.  Author: Anton Voloshin Discussion:
  <a href=3D"https://postgr.es/m/af79ee1b-9962-b299-98e1-f90a289e19e6@postg=
respro.ru" style=3D"color: #3498db; text-decoration: underline">https://pos=
tgr.es/m/af79ee1b-9962-b299-98e1-f90a289e19e6@postgrespro.ru</a>
  Backpatch-through: 13
  <a href=3D"https://git.postgresql.org/pg/commitdiff/d5a2c413fcdd187dc16c4=
fab16610af7d4849cc1" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/d5a2c413fcdd187dc16c4fab16610af7d48=
49cc1</a></p>
</li>
</ul>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Bruce Momjian pushed:</p>
<ul style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal;=
 margin: 0; margin-bottom: 15px">
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">doc:  add PG 14 relnote item about array fu=
nction references. User-defined
  objects that reference some built-in array functions will need to be recr=
eated
  in PG 14.  Reported-by: Justin Pryzby  Discussion:
  <a href=3D"https://postgr.es/m/20210608225618.GR16435@telsasoft.com" styl=
e=3D"color: #3498db; text-decoration: underline">https://postgr.es/m/202106=
08225618.GR16435@telsasoft.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/25dfb5a831a1b8a83a8a684=
53b4bbe38a5ef737e" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/25dfb5a831a1b8a83a8a68453b4bbe38a5ef7=
37e</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">doc:  PG 14 relnote updates. Reported-by: J=
ustin Pryzby  Discussion:
  <a href=3D"https://postgr.es/m/20210612034551.GU16435@telsasoft.com" styl=
e=3D"color: #3498db; text-decoration: underline">https://postgr.es/m/202106=
12034551.GU16435@telsasoft.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/a2559d4093725592a3fd5da=
8a4c7ac7c883115a0" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/a2559d4093725592a3fd5da8a4c7ac7c88311=
5a0</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">doc:  PG 14 relnotes fixes. Items related t=
o logical replication attribution
  and BRIN indexes  Reported-by: Tomas Vondra, John Naylor  Discussion:
  <a href=3D"https://postgr.es/m/0db66294-a668-2caa-2b5e-a8db60b30662@enter=
prisedb.com," style=3D"color: #3498db; text-decoration: underline">https://=
postgr.es/m/0db66294-a668-2caa-2b5e-a8db60b30662@enterprisedb.com,</a>
  CAFBsxsH21KnteYdk33F1oZu2O726NSD6_XBq51Tn0jytsA1AnA@mail.gmail.com
  <a href=3D"https://git.postgresql.org/pg/commitdiff/86b222b09071d3918c5c5=
5b164d22b2236d3f872" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/86b222b09071d3918c5c55b164d22b2236d=
3f872</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">doc:  add mention of +4GB windows file hand=
ling in PG14 relnotes. Reported-by:
  Masahiko Sawada  Discussion:
  <a href=3D"https://postgr.es/m/CAD21AoCTHyouoGv-xt1qNjjvPbGMErLi0AJncByTv=
r66Nq7j8g@mail.gmail.com" style=3D"color: #3498db; text-decoration: underli=
ne">https://postgr.es/m/CAD21AoCTHyouoGv-xt1qNjjvPbGMErLi0AJncByTvr66Nq7j8g=
@mail.gmail.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/90855908b751d40f67352fa=
0252e0fcdaa7e317b" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/90855908b751d40f67352fa0252e0fcdaa7e3=
17b</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">doc:  adjust PG 14 relnotes to be current.
  <a href=3D"https://git.postgresql.org/pg/commitdiff/69a58bfe4ab05567a8fab=
8bdce7f3095ed06b99c" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/69a58bfe4ab05567a8fab8bdce7f3095ed0=
6b99c</a></p>
</li>
</ul>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">=C3=81lvaro Herrera pushed:</p>
<ul style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal;=
 margin: 0; margin-bottom: 15px">
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Fix logic bug in 1632ea43682f. I overlooked=
 that one condition was logically
  inverted.  The fix is a little bit more involved than simply negating the
  condition, to make the code easier to read.  Fix some outdated comments l=
eft
  by the same commit, while at it.  Author: Masahiko Sawada
  <a style=3D"color: #3498db; text-decoration: underline">&#115;&#97;&#119;=
&#97;&#100;&#97;&#46;&#109;&#115;&#104;&#107;&#64;&#103;&#109;&#97;&#105;&#=
108;&#46;&#99;&#111;&#109;</a> Author: =C3=81lvaro Herrera <a style=3D"colo=
r: #3498db; text-decoration: underline">&#97;&#108;&#118;&#104;&#101;&#114;=
&#114;&#101;&#64;&#97;&#108;&#118;&#104;&#46;&#110;&#111;&#45;&#105;&#112;&=
#46;&#111;&#114;&#103;</a>
  Reviewed-by: Amit Kapila <a style=3D"color: #3498db; text-decoration: und=
erline">&#97;&#109;&#105;&#116;&#46;&#107;&#97;&#112;&#105;&#108;&#97;&#49;=
&#54;&#64;&#103;&#109;&#97;&#105;&#108;&#46;&#99;&#111;&#109;</a> Discussio=
n:
  <a href=3D"https://postgr.es/m/YMRlmB3/lZw8YBH+@paquier.xyz" style=3D"col=
or: #3498db; text-decoration: underline">https://postgr.es/m/YMRlmB3/lZw8YB=
H+@paquier.xyz</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/33c509956704e1d918077b5=
1ccd954f2e201a8f5" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/33c509956704e1d918077b51ccd954f2e201a=
8f5</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Add test case for obsoleting slot with acti=
ve walsender. The code to signal a
  running walsender when its reserved WAL size grows too large is completely
  uncovered before this commit; this adds coverage for that case.  This test
  involves sending SIGSTOP to walsender and walreceiver and running a check=
point
  while advancing WAL, then sending SIGCONT.  There's no precedent for this
  coding in Perl tests, and my reading of relevant manpages says it's likel=
y to
  fail on Windows.  Because of this, this test is always skipped on that
  platform.  Author: =C3=81lvaro Herrera <a style=3D"color: #3498db; text-d=
ecoration: underline">&#97;&#108;&#118;&#104;&#101;&#114;&#114;&#101;&#64;&=
#97;&#108;&#118;&#104;&#46;&#110;&#111;&#45;&#105;&#112;&#46;&#111;&#114;&#=
103;</a> Discussion:
  <a href=3D"https://postgr.es/m/202106102202.mjw4huiix7lo@alvherre.pgsql" =
style=3D"color: #3498db; text-decoration: underline">https://postgr.es/m/20=
2106102202.mjw4huiix7lo@alvherre.pgsql</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/09126984a2631db8dd6299f=
23954e0dede69735f" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/09126984a2631db8dd6299f23954e0dede697=
35f</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Revert "Add test case for obsoleting slot w=
ith active walsender". This reverts
  commit 09126984a263; the test case added there failed once in circumstanc=
es
  that remain mysterious.  It seems better to remove the test for now so th=
at
  14beta2 doesn't have random failures built in.
  <a href=3D"https://git.postgresql.org/pg/commitdiff/96795176810b979a2bf1f=
4563bdcb161679d5b95" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/96795176810b979a2bf1f4563bdcb161679=
d5b95</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Add test case for obsoleting slot with acti=
ve walsender, take 2. The code to
  signal a running walsender when its reserved WAL size grows too large is
  completely uncovered before this commit; this adds coverage for that case.
  This test involves sending SIGSTOP to walsender and walreceiver, then
  advancing enough WAL for a checkpoint to trigger, then sending SIGCONT.
  There's no precedent for STOP signalling in Perl tests, and my reading of
  relevant manpages says it's likely to fail on Windows.  Because of this, =
this
  test is always skipped on that platform.  This version fixes a couple of
  rarely hit race conditions in the previous attempt 09126984a263; most not=
ably,
  both LOG string searches are loops, not just the second one; we acquire t=
he
  start-of-log position before STOP-signalling; and reference the correct
  process name in the test description.  All per Tom Lane.  Author: =C3=81l=
varo
  Herrera <a style=3D"color: #3498db; text-decoration: underline">&#97;&#10=
8;&#118;&#104;&#101;&#114;&#114;&#101;&#64;&#97;&#108;&#118;&#104;&#46;&#11=
0;&#111;&#45;&#105;&#112;&#46;&#111;&#114;&#103;</a> Discussion:
  <a href=3D"https://postgr.es/m/202106102202.mjw4huiix7lo@alvherre.pgsql" =
style=3D"color: #3498db; text-decoration: underline">https://postgr.es/m/20=
2106102202.mjw4huiix7lo@alvherre.pgsql</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/24043c27b46f873211177e3=
460ab09dc011802a5" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/24043c27b46f873211177e3460ab09dc01180=
2a5</a></p>
</li>
</ul>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Noah Misch pushed:</p>
<ul style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal;=
 margin: 0; margin-bottom: 15px">
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Remove pg_wait_for_backend_termination(). I=
t was unable to wait on a backend
  that had already left the procarray. Users tolerant of that limitation can
  poll pg_stat_activity.  Other users can employ the "timeout" argument of
  pg_terminate_backend().  Reviewed by Bharath Rupireddy.  Discussion:
  <a href=3D"https://postgr.es/m/20210605013236.GA208701@rfd.leadboat.com" =
style=3D"color: #3498db; text-decoration: underline">https://postgr.es/m/20=
210605013236.GA208701@rfd.leadboat.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/5f1df62a459b51ab3bb625a=
0ee5e655429254257" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/5f1df62a459b51ab3bb625a0ee5e655429254=
257</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Copy-edit text for the pg_terminate_backend=
() "timeout" parameter. Revert the
  pg_description entry to its v13 form, since those messages usually remain
  shorter and don't discuss individual parameters.  No catversion bump, sin=
ce
  pg_description content does not impair backend compatibility or applicati=
on
  compatibility.  Justin Pryzby  Discussion:
  <a href=3D"https://postgr.es/m/20210612182743.GY16435@telsasoft.com" styl=
e=3D"color: #3498db; text-decoration: underline">https://postgr.es/m/202106=
12182743.GY16435@telsasoft.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/0aac73e6a2602696d23aa7a=
9686204965f9093dc" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/0aac73e6a2602696d23aa7a9686204965f909=
3dc</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Finish rename of PQtraceSetFlags() to PQset=
TraceFlags(). Jie Zhang
  Discussion:
  <a href=3D"https://postgr.es/m/TYWPR01MB767844835390EDD8DB276D75F90A9@TYW=
PR01MB7678.jpnprd01.prod.outlook.com" style=3D"color: #3498db; text-decorat=
ion: underline">https://postgr.es/m/TYWPR01MB767844835390EDD8DB276D75F90A9@=
TYWPR01MB7678.jpnprd01.prod.outlook.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/047a259e35b9dde2dad5fd0=
e5d5d784bb327b848" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/047a259e35b9dde2dad5fd0e5d5d784bb327b=
848</a></p>
</li>
</ul>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Amit Kapila pushed:</p>
<ul style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal;=
 margin: 0; margin-bottom: 15px">
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Fix decoding of speculative aborts. During =
decoding for speculative inserts,
  we were relying for cleaning toast hash on confirmation records or next c=
hange
  records. But that could lead to multiple problems (a) memory leak if ther=
e is
  neither a confirmation record nor any other record after toast insertion =
for a
  speculative insert in the transaction, (b) error and assertion failures i=
f the
  next operation is not an insert/update on the same table.  The fix is to =
start
  queuing spec abort change and clean up toast hash and change record durin=
g its
  processing. Currently, we are queuing the spec aborts for both toast and =
main
  table even though we perform cleanup while processing the main table's sp=
ec
  abort record. Later, if we have a way to distinguish between the spec abo=
rt
  record of toast and the main table, we can avoid queuing the change for s=
pec
  aborts of toast tables.  Reported-by: Ashutosh Bapat Author: Dilip Kumar
  Reviewed-by: Amit Kapila Backpatch-through: 9.6, where it was introduced
  Discussion:
  <a href=3D"https://postgr.es/m/CAExHW5sPKF-Oovx_qZe4p5oM6Dvof7_P+XgsNAViu=
g15Fm99jA@mail.gmail.com" style=3D"color: #3498db; text-decoration: underli=
ne">https://postgr.es/m/CAExHW5sPKF-Oovx_qZe4p5oM6Dvof7_P+XgsNAViug15Fm99jA=
@mail.gmail.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/4daa140a2f50e9a160bc180=
c3997ee13c7aabf9e" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/4daa140a2f50e9a160bc180c3997ee13c7aab=
f9e</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Document a few caveats in synchronous logic=
al replication. In a synchronous
  logical setup, locking [user] catalog tables can cause deadlock. This is
  because logical decoding of transactions can lock catalog tables to access
  them so exclusively locking those in transactions can lead to deadlock. To
  avoid this users must refrain from having exclusive locks on catalog tabl=
es.
  Author: Takamichi Osumi Reviewed-by: Vignesh C, Amit Kapila Backpatch-thr=
ough:
  9.6 Discussion:
  <a href=3D"https://www.postgresql.org/message-id/20210222222847.tpnb6eg3y=
iykzpky%40alap3.anarazel.de" style=3D"color: #3498db; text-decoration: unde=
rline">https://www.postgresql.org/message-id/20210222222847.tpnb6eg3yiykzpk=
y%40alap3.anarazel.de</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/3cb828dbe26087e7754f49f=
3cfe3ed036d5af439" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/3cb828dbe26087e7754f49f3cfe3ed036d5af=
439</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Handle no replica identity index case in Re=
lationGetIdentityKeyBitmap. Commit
  e7eea52b2d has introduced a new function RelationGetIdentityKeyBitmap whi=
ch
  omits to handle the case where there is no replica identity index on a
  relation.  Author: Mark Dilger Reviewed-by: Takamichi Osumi, Amit Kapila
  Discussion:
  <a href=3D"https://www.postgresql.org/message-id/4C99A862-69C8-431F-960A-=
81B1151F1B89@enterprisedb.com" style=3D"color: #3498db; text-decoration: un=
derline">https://www.postgresql.org/message-id/4C99A862-69C8-431F-960A-81B1=
151F1B89@enterprisedb.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/2731ce1bd550d08f3fdd7bc=
b1497af4b95170976" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/2731ce1bd550d08f3fdd7bcb1497af4b95170=
976</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Doc: Update caveats in synchronous logical =
replication. Reported-by: Simon
  Riggs Author: Takamichi Osumi Reviewed-by: Amit Kapila Backpatch-through:=
 9.6
  Discussion:
  <a href=3D"https://www.postgresql.org/message-id/20210222222847.tpnb6eg3y=
iykzpky@alap3.anarazel.de" style=3D"color: #3498db; text-decoration: underl=
ine">https://www.postgresql.org/message-id/20210222222847.tpnb6eg3yiykzpky@=
alap3.anarazel.de</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/c66fb78ebb4f473bb4fd877=
3de3cda9339e14f81" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/c66fb78ebb4f473bb4fd8773de3cda9339e14=
f81</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Doc: Update logical replication message for=
mats. Commits 9de77b5453 and
  ac4645c015 missed to update the logical replication message formats secti=
on in
  the docs.  Author: Brar Piening Reviewed-by: Amit Kapila Discussion:
  <a href=3D"https://www.postgresql.org/message-id/cc70956c-e578-e54f-49e6-=
b5d68c89576f@gmx.de" style=3D"color: #3498db; text-decoration: underline">h=
ttps://www.postgresql.org/message-id/cc70956c-e578-e54f-49e6-b5d68c89576f@g=
mx.de</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/f08722cf83ab1fabee948a4=
e5754bf6236ad700b" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/f08722cf83ab1fabee948a4e5754bf6236ad7=
00b</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Doc: Fix minor formatting issue in logicald=
ecoding.sgml. Author: Guillaume
  Lelarge Discussion:
  <a href=3D"https://www.postgresql.org/message-id/CAECtzeXf3_oZoU6mgFCOy5+=
pDZ5n4XtH0Da4a5n_KacraVWiHQ@mail.gmail.com" style=3D"color: #3498db; text-d=
ecoration: underline">https://www.postgresql.org/message-id/CAECtzeXf3_oZoU=
6mgFCOy5+pDZ5n4XtH0Da4a5n_KacraVWiHQ@mail.gmail.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/847c62ee76cbc237b3a204c=
a94b1b12811d698e3" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/847c62ee76cbc237b3a204ca94b1b12811d69=
8e3</a></p>
</li>
</ul>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Alexander Korotkov pushed:</p>
<ul style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal;=
 margin: 0; margin-bottom: 15px">
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Support for unnest(multirange) and cast mul=
tirange as an array of ranges. It
  has been spotted that multiranges lack of ability to decompose them into
  individual ranges.  Subscription and proper expanded object representation
  require substantial work, and it's too late for v14.  This commit provide=
s the
  implementation of unnest(multirange) and cast multirange as an array of
  ranges, which is quite trivial.  unnest(multirange) is defined as a
  polymorphic procedure.  The catalog description of the cast underlying
  procedure is duplicated for each multirange type because we don't have
  anyrangearray polymorphic type to use here.  Catversion is bumped.
  Reported-by: Jonathan S. Katz Discussion:
  <a href=3D"https://postgr.es/m/flat/60258efe-bd7e-4886-82e1-196e0cac5433%=
40postgresql.org" style=3D"color: #3498db; text-decoration: underline">http=
s://postgr.es/m/flat/60258efe-bd7e-4886-82e1-196e0cac5433%40postgresql.org<=
/a>
  Author: Alexander Korotkov Reviewed-by: Justin Pryzby, Jonathan S. Katz,
  Zhihong Yu
  <a href=3D"https://git.postgresql.org/pg/commitdiff/29854ee8d1ca4a46adb7e=
84deb17e6fb18e531cc" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/29854ee8d1ca4a46adb7e84deb17e6fb18e=
531cc</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Add missing type name "multirange" in docs =
chapter title. Discussion:
  <a href=3D"https://postgr.es/m/CAFj8pRDioOxiJgmgw9TqQqZ3CxnJC4P5B2Oospf2e=
MgAjJuewA%40mail.gmail.com" style=3D"color: #3498db; text-decoration: under=
line">https://postgr.es/m/CAFj8pRDioOxiJgmgw9TqQqZ3CxnJC4P5B2Oospf2eMgAjJue=
wA%40mail.gmail.com</a>
  Author: Pavel Stehule, Alexander Korotkov Reviewed-by: Justin Pryzby, Tom=
 Lane
  <a href=3D"https://git.postgresql.org/pg/commitdiff/ad2da246c69dd5ec55764=
d4b6066f3b0c0d24ca2" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/ad2da246c69dd5ec55764d4b6066f3b0c0d=
24ca2</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Revert 29854ee8d1 due to buildfarm failures=
. Reported-by: Tom Lane Discussion:
  <a href=3D"https://postgr.es/m/CAPpHfdvcnw3x7jdV3r52p4%3D5S4WUxBCzcQKB3Ju=
kQHoicv1LSQ%40mail.gmail.com" style=3D"color: #3498db; text-decoration: und=
erline">https://postgr.es/m/CAPpHfdvcnw3x7jdV3r52p4%3D5S4WUxBCzcQKB3JukQHoi=
cv1LSQ%40mail.gmail.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/817bb0a7d1e02df4643d373=
04aed8574cf43f629" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/817bb0a7d1e02df4643d37304aed8574cf43f=
629</a></p>
</li>
</ul>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Peter Geoghegan pushed:</p>
<ul style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal;=
 margin: 0; margin-bottom: 15px">
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Remove unneeded field from VACUUM state. Bu=
gfix commit 5fc89376 effectively
  made the lock_waiter_detected field from vacuumlazy.c's global state stru=
ct
  into private state owned by lazy_truncate_heap().  Finish this off by
  replacing the struct field with a local variable.
  <a href=3D"https://git.postgresql.org/pg/commitdiff/958cfbcf2dd338e3179c2=
d8a35f48bde020eba60" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/958cfbcf2dd338e3179c2d8a35f48bde020=
eba60</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Support disabling index bypassing by VACUUM=
. Generalize the INDEX_CLEANUP
  VACUUM parameter (and the corresponding reloption): make it into a ternary
  style boolean parameter.  It now exposes a third option, "auto".  The "au=
to"
  option (which is now the default) enables the "bypass index vacuuming"
  optimization added by commit 1e55e7d1.  "VACUUM (INDEX_CLEANUP TRUE)" is
  redefined to once again make VACUUM simply do any required index vacuumin=
g,
  regardless of how few dead tuples are encountered during the first scan o=
f the
  target heap relation (unless there are exactly zero).  This gives users a=
 way
  of opting out of the "bypass index vacuuming" optimization, if for whatev=
er
  reason that proves necessary.  It is also expected to be used by PostgreS=
QL
  developers as a testing option from time to time.  "VACUUM (INDEX_CLEANUP
  FALSE)" does the same thing as it always has: it forcibly disables both i=
ndex
  vacuuming and index cleanup.  It's not expected to be used much in Postgr=
eSQL</p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">The failsafe mechanism added by commit 1e55=
e7d1 addresses the same
  problem in a simpler way. INDEX_CLEANUP can now be thought of as a testin=
g and
  compatibility option.  Author: Peter Geoghegan <a style=3D"color: #3498db=
; text-decoration: underline">&#112;&#103;&#64;&#98;&#111;&#119;&#116;&#46;=
&#105;&#101;</a> Reviewed-By:
  Masahiko Sawada <a style=3D"color: #3498db; text-decoration: underline">&=
#115;&#97;&#119;&#97;&#100;&#97;&#46;&#109;&#115;&#104;&#107;&#64;&#103;=
09;&#97;&#105;&#108;&#46;&#99;&#111;&#109;</a> Reviewed-By: Justin Pryzby
  <a style=3D"color: #3498db; text-decoration: underline">&#112;&#114;&#121=
;&#122;&#98;&#121;&#64;&#116;&#101;&#108;&#115;&#97;&#115;&#111;&#102;&#116=
;&#46;&#99;&#111;&#109;</a> Discussion:
  <a href=3D"https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXc=
rhqsv6OHQ@mail.gmail.com" style=3D"color: #3498db; text-decoration: underli=
ne">https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXcrhqsv6OHQ=
@mail.gmail.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/3499df0dee8c4ea51d264a6=
74df5b5e31991319a" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/3499df0dee8c4ea51d264a674df5b5e319913=
19a</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Remove overzealous VACUUM failsafe assertio=
ns. The failsafe can trigger when
  index processing is already disabled. This can happen when VACUUM's
  INDEX_CLEANUP parameter is "off" and the failsafe happens to trigger.  Re=
move
  assertions that assume that index processing is directly tied to the fail=
safe.
  Oversight in commit c242baa4, which made it possible for the failsafe to
  trigger in a two-pass strategy VACUUM that has yet to make its first call=
 to
  lazy_vacuum_all_indexes().
  <a href=3D"https://git.postgresql.org/pg/commitdiff/e8f201ab82be234b2f103=
234cf9f262f9afeaeba" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/e8f201ab82be234b2f103234cf9f262f9af=
eaeba</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Add list of ignorable pgindent commits for =
git-blame. Add a
  .git-blame-ignore-revs file with a list of pgindent, pgperlyidy, and
  reformat-dat-files commit hashes.  Postgres hackers that configure git to=
 use
  the ignore file will get git-blame output that avoids attributing line ch=
anges
  to the ignored indent commits.  This makes git-blame output much easier to
  work with in practice.  Author: Peter Geoghegan <a style=3D"color: #3498d=
b; text-decoration: underline">&#112;&#103;&#64;&#98;&#111;&#119;&#116;&#46=
;&#105;&#101;</a> Reviewed-By: Tom
  Lane <a style=3D"color: #3498db; text-decoration: underline">&#116;&#103;=
&#108;&#64;&#115;&#115;&#115;&#46;&#112;&#103;&#104;&#46;&#112;&#97;&#46;&#=
117;&#115;</a> Discussion:
  <a href=3D"https://postgr.es/m/CAH2-Wz=3DcVh3GHTP6SdLU-Gnmt2zRdF8vZkcrFdS=
zXQ=3DWhbWm9Q@mail.gmail.com" style=3D"color: #3498db; text-decoration: und=
erline">https://postgr.es/m/CAH2-Wz=3DcVh3GHTP6SdLU-Gnmt2zRdF8vZkcrFdSzXQ=
=3DWhbWm9Q@mail.gmail.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/8e638845ff6bb255ad1dea1=
5831089a234535391" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/8e638845ff6bb255ad1dea15831089a234535=
391</a></p>
</li>
</ul>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Andrew Dunstan pushed:</p>
<ul style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal;=
 margin: 0; margin-bottom: 15px">
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Further refinement of stuck_on_old_timeline=
 recovery test. TestLib::perl2host
  can take a file argument as well as a directory argument, so that code be=
comes
  substantially simpler. Also add comments on why we're using forward slash=
es,
  and why we're setting PERL_BADLANG=3D0.  Discussion:
  <a href=3D"https://postgr.es/m/e9947bcd-20ee-027c-f0fe-01f736b7e345@dunsl=
ane.net" style=3D"color: #3498db; text-decoration: underline">https://postg=
r.es/m/e9947bcd-20ee-027c-f0fe-01f736b7e345@dunslane.net</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/54a5ed22016940d7ad5060e=
d62d23473924756a1" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/54a5ed22016940d7ad5060ed62d2347392475=
6a1</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Don't set a fast default for anything but a=
 plain table. The fast default code
  added in Release 11 omitted to check that the table a fast default was be=
ing
  added to was a plain table. Thus one could be added to a foreign table, w=
hich
  predicably blows up. Here we perform that check.  In addition, on the back
  branches, since some of these might have escaped into the wild, if we
  encounter a missing value for an attribute of something other than a plain
  table we ignore it.  Fixes bug #17056  Backpatch to release 11,  Reviewed=
 by:
  Andres Freund, =C3=81lvaro Herrera and Tom Lane
  <a href=3D"https://git.postgresql.org/pg/commitdiff/0a4efdc7ebf2584257b16=
6c87e82797eb92815b5" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/0a4efdc7ebf2584257b166c87e82797eb92=
815b5</a></p>
</li>
</ul>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Heikki Linnakangas pushed:</p>
<ul style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal;=
 margin: 0; margin-bottom: 15px">
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Fix outdated comment that talked about seek=
 position of WAL file. Since commit
  c24dcd0cfd, we have been using pg_pread() to read the WAL file, which doe=
sn't
  change the seek position (unless we fall back to the implementation in
  src/port/pread.c). Update comment accordingly.  Backpatch-through: 12, wh=
ere
  we started to use pg_pread()
  <a href=3D"https://git.postgresql.org/pg/commitdiff/d0303bc8d2670d11c9df9=
220aa08a2c33529e100" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/d0303bc8d2670d11c9df9220aa08a2c3352=
9e100</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Tidy up GetMultiXactIdMembers()'s behavior =
on error. One of the error paths
  left <code>*members</code> uninitialized. That's not a live bug, because =
most callers don't
  look at <code>*members</code> when the function returns -1, but let's be =
tidy. One caller,
  in heap_lock_tuple(), does "if (members !=3D NULL) pfree(members)", but A=
FAICS
  it never passes an invalid 'multi' value so it should not reach that error
  case.  The callers are also a bit inconsistent in their expectations.
  heap_lock_tuple() pfrees the 'members' array if it's not-NULL, others pfr=
ee()
  it if "nmembers &gt;=3D 0", and others if "nmembers &gt; 0". That's not a=
 live bug
  either, because the function should never return 0, but add an Assert for=
 that
  to make it more clear. I left the callers alone for now.  I also moved the
  line where we set <code>*nmembers</code>. It wasn't wrong before, but I l=
ike to do that
  right next to the 'return' statement, to make it clear that it's always s=
et on
  return.  Also remove one unreachable return statement after ereport(ERROR=
),
  for brevity and for consistency with the similar if-block right after it.
  Author: Greg Nancarrow with the additional changes by me Backpatch-throug=
h:
  9.6, all supported versions
  <a href=3D"https://git.postgresql.org/pg/commitdiff/d24c5658a80c8f5037e9e=
1948de311d3f3350f12" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/d24c5658a80c8f5037e9e1948de311d3f33=
50f12</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Prevent race condition while reading relmap=
per file. Contrary to the comment
  here, POSIX does not guarantee atomicity of a read(), if another process =
calls
  write() concurrently. Or at least Linux does not. Add locking to
  load_relmap_file() to avoid the race condition.  Fixes bug #17064. Thanks=
 to
  Alexander Lakhin for the report and test case.  Backpatch-through: 9.6, a=
ll
  supported versions. Discussion:
  <a href=3D"https://www.postgresql.org/message-id/17064-bb0d7904ef72add3@p=
ostgresql.org" style=3D"color: #3498db; text-decoration: underline">https:/=
/www.postgresql.org/message-id/17064-bb0d7904ef72add3@postgresql.org</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/b6d8d2073f228d9f7198f1f=
9826d8f6ab643c219" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/b6d8d2073f228d9f7198f1f9826d8f6ab643c=
219</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Another fix to relmapper race condition. In=
 previous commit, I missed that
  relmap_redo() was also not acquiring the RelationMappingLock. Thanks to T=
homas
  Munro for pointing that out.  Backpatch-through: 9.6, like previous commi=
t.
  Discussion:
  <a href=3D"https://www.postgresql.org/message-id/CA%2BhUKGLev%3DPpOSaL3WR=
ZgOvgk217et%2BbxeJcRr4eR-NttP1F6Q%40mail.gmail.com" style=3D"color: #3498db=
; text-decoration: underline">https://www.postgresql.org/message-id/CA%2BhU=
KGLev%3DPpOSaL3WRZgOvgk217et%2BbxeJcRr4eR-NttP1F6Q%40mail.gmail.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/9b8ed0f52bffceaccf6fa65=
0ffe482e7da20fbf2" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/9b8ed0f52bffceaccf6fa650ffe482e7da20f=
bf2</a></p>
</li>
</ul>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Tom=C3=A1=C5=A1 Vondra pushed:</p>
<ul style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal;=
 margin: 0; margin-bottom: 15px">
<li style=3D"list-style-position: inside; margin-left: 5px">Fix copying dat=
a into slots with FDW batching. Commit b676ac443b optimized
  handling of tuple slots with bulk inserts into foreign tables, so that the
  slots are initialized only once and reused for all batches. The data was
  however copied into the slots only after the initialization, inserting
  duplicate values when the slot gets reused. Fixed by moving the ExecCopyS=
lot
  outside the init branch.  The existing postgres_fdw tests failed to catch=
 this
  due to inserting data into foreign tables without unique indexes, and then
  checking only the number of inserted rows. This adds a new test with both=
 a
  unique index and a check of inserted values.  Reported-by: Alexander Pyha=
lov
  Discussion:
  <a href=3D"https://postgr.es/m/7a8cf8d56b3d18e5c0bccd6cd42d04ac%40postgre=
spro.ru" style=3D"color: #3498db; text-decoration: underline">https://postg=
r.es/m/7a8cf8d56b3d18e5c0bccd6cd42d04ac%40postgrespro.ru</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/99cea49d6525e30bc3768e4=
ffb95377e7584dea1" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/99cea49d6525e30bc3768e4ffb95377e7584d=
ea1</a></li>
</ul>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Fujii Masao pushed:</p>
<ul style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal;=
 margin: 0; margin-bottom: 15px">
<li style=3D"list-style-position: inside; margin-left: 5px">Make archiver p=
rocess handle barrier events. Commit d75288fb27 made WAL
  archiver process an auxiliary process. An auxiliary process needs to hand=
le
  barrier events but the commit forgot to make archiver process do that.
  Reported-by: Thomas Munro Author: Fujii Masao Reviewed-by: Thomas Munro
  Discussion:
  <a href=3D"https://postgr.es/m/CA+hUKGLah2w1pWKHonZP_+EQw69=3Dq56AHYwCgEN=
8GDzsRG_Hgw@mail.gmail.com" style=3D"color: #3498db; text-decoration: under=
line">https://postgr.es/m/CA+hUKGLah2w1pWKHonZP_+EQw69=3Dq56AHYwCgEN8GDzsRG=
_Hgw@mail.gmail.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/981524d2e3aa3f28d364c47=
2e34f5386be846811" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/981524d2e3aa3f28d364c472e34f5386be846=
811</a></li>
</ul>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Peter Eisentraut pushed:</p>
<ul style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal;=
 margin: 0; margin-bottom: 15px">
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">amcheck: Fix code comments. Code comments w=
ere claiming that verify_heapam()
  was checking privileges on the relation it was operating on, but it didn't
  actually do that.  Perhaps earlier versions of the patch did that, but no=
w the
  access is regulated by privileges on the function.  Remove the wrong comm=
ents.
  <a href=3D"https://git.postgresql.org/pg/commitdiff/97b7134186490b36e01ef=
c9d2feaf7038c666457" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/97b7134186490b36e01efc9d2feaf7038c6=
66457</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Translation updates. Source-Git-URL:
  git://git.postgresql.org/git/pgtranslation/messages.git Source-Git-Hash:
  70796ae860c444c764bb591c885f22cac1c168ec
  <a href=3D"https://git.postgresql.org/pg/commitdiff/a7bb0ce58f56ee8907c3f=
49c52d99f502536c796" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/a7bb0ce58f56ee8907c3f49c52d99f50253=
6c796</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Remove redundant variable pageSize in gisti=
nitpage. In gistinitpage, pageSize
  variable looks redundant, instead just pass BLCKSZ. This will be consiste=
nt
  with its peers BloomInitPage, brin_page_init and SpGistInitPage.  Author:
  Bharath Rupireddy <a style=3D"color: #3498db; text-decoration: underline"=
>&#98;&#104;&#97;&#114;&#97;&#116;&#104;&#46;&#114;&#117;&#112;&#105;&#114;=
&#101;&#100;&#100;&#121;&#64;&#101;&#110;&#116;&#101;&#114;&#112;&#114;&#10=
5;&#115;&#101;&#100;&#98;&#46;&#99;&#111;&#109;</a> Discussion:
  <a href=3D"https://www.postgresql.org/message-id/flat/CALj2ACWj=3DV1k5591=
eeZK2sOg2FYuBUp6azFO8tMkBtGfXf8PMQ@mail.gmail.com" style=3D"color: #3498db;=
 text-decoration: underline">https://www.postgresql.org/message-id/flat/CAL=
j2ACWj=3DV1k5591eeZK2sOg2FYuBUp6azFO8tMkBtGfXf8PMQ@mail.gmail.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/a60c4c5c1a99746485123ae=
93fbd3e58c78e5d62" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/a60c4c5c1a99746485123ae93fbd3e58c78e5=
d62</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">doc: Change reloption data type spelling fo=
r consistency. Use "floating point"
  rather than "float4", like everywhere else in this context.  Author:
  Shinya11.Kato@nttdata.com Discussion:
  <a href=3D"https://www.postgresql.org/message-id/flat/TYAPR01MB28965989AF=
84B57FC351B97BC40F9@TYAPR01MB2896.jpnprd01.prod.outlook.com" style=3D"color=
: #3498db; text-decoration: underline">https://www.postgresql.org/message-i=
d/flat/TYAPR01MB28965989AF84B57FC351B97BC40F9@TYAPR01MB2896.jpnprd01.prod.o=
utlook.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/63e6d05bf34da58eef7cd1e=
d461b9c8315177a38" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/63e6d05bf34da58eef7cd1ed461b9c8315177=
a38</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Fixes in ALTER SUBSCRIPTION DROP PUBLICATIO=
N code. ALTER SUBSCRIPTION DROP
  PUBLICATION does not actually support copy_data option, so remove it from=
 tab
  completion.  Also, reword the error message that is thrown when all the
  publications from a subscription are specified to be dropped.  Also, made=
 few
  doc and cosmetic adjustments.  Author: Vignesh C <a style=3D"color: #3498=
db; text-decoration: underline">&#118;&#105;&#103;&#110;&#101;&#115;&#104;&=
#50;&#49;&#64;&#103;&#109;&#97;&#105;&#108;&#46;&#99;&#111;&#109;</a>
  Reviewed-by: Bharath Rupireddy <a style=3D"color: #3498db; text-decoratio=
n: underline">&#98;&#104;&#97;&#114;&#97;&#116;&#104;&#46;&#114;&#117;&#112=
;&#105;&#114;&#101;&#100;&#100;&#121;&#64;&#101;&#110;&#116;&#101;&#114;=
12;&#114;&#105;&#115;&#101;&#100;&#98;&#46;&#99;&#111;&#109;</a>
  Reviewed-by: Japin Li <a style=3D"color: #3498db; text-decoration: underl=
ine">&#106;&#97;&#112;&#105;&#110;&#108;&#105;&#64;&#104;&#111;&#116;&#109;=
&#97;&#105;&#108;&#46;&#99;&#111;&#109;</a> Discussion:
  <a href=3D"https://www.postgresql.org/message-id/flat/CALDaNm21RwsDzs4xj1=
4ApteAF7auyyomHNnp+NEL-sH8m-jMvQ@mail.gmail.com" style=3D"color: #3498db; t=
ext-decoration: underline">https://www.postgresql.org/message-id/flat/CALDa=
Nm21RwsDzs4xj14ApteAF7auyyomHNnp+NEL-sH8m-jMvQ@mail.gmail.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/e59d428f34297cd0eb67e3b=
4e4b8c8bc58504921" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/e59d428f34297cd0eb67e3b4e4b8c8bc58504=
921</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Put option listing back into alphabetical o=
rder.
  <a href=3D"https://git.postgresql.org/pg/commitdiff/3af10943ce21450e299b3=
915b9cad47cd90369e9" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/3af10943ce21450e299b3915b9cad47cd90=
369e9</a></p>
</li>
<li style=3D"list-style-position: inside; margin-left: 5px">
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Error message refactoring. Take some untran=
slatable things out of the message
  and replace by format placeholders, to reduce translatable strings and re=
duce
  translation mistakes.
  <a href=3D"https://git.postgresql.org/pg/commitdiff/c302a6139072fed941020=
4fa9e751d95930e05ff" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/c302a6139072fed9410204fa9e751d95930=
e05ff</a></p>
</li>
</ul>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">David Rowley pushed:</p>
<ul style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal;=
 margin: 0; margin-bottom: 15px">
<li style=3D"list-style-position: inside; margin-left: 5px">Fix assert fail=
ure in expand_grouping_sets. linitial_node() fails in assert
  enabled builds if the given pointer is not of the specified type.  Here t=
he
  type is IntList.  The code thought it should be expecting List, but it was
  wrong.  In the existing tests which run this code the initial list elemen=
t is
  always NIL.  Since linitial_node() allows NULL, we didn't trigger any ass=
ert
  failures in the existing regression tests.  There is still some discussio=
n as
  to whether we need a few more tests in this area, but for now, since beta=
2 is
  looming, fix the bug first.  Bug: #17067 Discussion:
  <a href=3D"https://postgr.es/m/17067-665d50fa321f79e0@postgresql.org" sty=
le=3D"color: #3498db; text-decoration: underline">https://postgr.es/m/17067=
-665d50fa321f79e0@postgresql.org</a>
  Reported-by: Yaoguang Chen
  <a href=3D"https://git.postgresql.org/pg/commitdiff/8d29d45d9b3cab95a866e=
fbcdd9138b3d76741b3" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/8d29d45d9b3cab95a866efbcdd9138b3d76=
741b3</a></li>
</ul>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Andres Freund pushed:</p>
<ul style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal;=
 margin: 0; margin-bottom: 15px">
<li style=3D"list-style-position: inside; margin-left: 5px">Use correct hor=
izon when vacuuming catalog relations. In dc7420c2c92 I
  (Andres) accidentally used RelationIsAccessibleInLogicalDecoding() as the=
 sole
  condition to use the non-shared catalog horizon in
  GetOldestNonRemovableTransactionId(). That is incorrect, as
  RelationIsAccessibleInLogicalDecoding() checks whether wal_level is logic=
al.
  The correct check, as done e.g. in GlobalVisTestFor(), is to check
  IsCatalogRelation() and RelationIsAccessibleInLogicalDecoding().  The obs=
erved
  misbehavior of this bug was that there could be an endless loop in
  lazy_scan_prune(), because the horizons used in heap_page_prune() and the
  individual tuple liveliness checks did not match. Likely there are other
  potential consequences as well.  A later commit will unify the determinat=
ion
  which horizon has to be used, and add additional assertions to make it ea=
sier
  to catch a bug like this.  Reported-By: Justin Pryzby <a style=3D"color: =
#3498db; text-decoration: underline">&#112;&#114;&#121;&#122;&#98;&#121;&#6=
4;&#116;&#101;&#108;&#115;&#97;&#115;&#111;&#102;&#116;&#46;&#99;&#111;&#10=
9;</a>
  Diagnosed-By: Matthias van de Meent <a style=3D"color: #3498db; text-deco=
ration: underline">&#98;&#111;&#101;&#107;&#101;&#119;&#117;&#114;&#109;=
3;&#112;&#111;&#115;&#116;&#103;&#114;&#101;&#115;&#64;&#103;&#109;&#97;=
05;&#108;&#46;&#99;&#111;&#109;</a> Author:
  Matthias van de Meent <a style=3D"color: #3498db; text-decoration: underl=
ine">&#98;&#111;&#101;&#107;&#101;&#119;&#117;&#114;&#109;&#43;&#112;&#111;=
&#115;&#116;&#103;&#114;&#101;&#115;&#64;&#103;&#109;&#97;&#105;&#108;&#46;=
&#99;&#111;&#109;</a> Discussion:
  <a href=3D"https://postgr.es/m/CAEze2Wg32Y9+WJfw=3DaofkRx1ZRFt_Ev6bNPc4PS=
az7PjSFtZgQ@mail.gmail.com" style=3D"color: #3498db; text-decoration: under=
line">https://postgr.es/m/CAEze2Wg32Y9+WJfw=3DaofkRx1ZRFt_Ev6bNPc4PSaz7PjSF=
tZgQ@mail.gmail.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/5a1e1d83022b976ebdec5cf=
a8f255c4278b75b8e" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/5a1e1d83022b976ebdec5cfa8f255c4278b75=
b8e</a></li>
</ul>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Joe Conway pushed:</p>
<ul style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal;=
 margin: 0; margin-bottom: 15px">
<li style=3D"list-style-position: inside; margin-left: 5px">Stamp 14beta2.
  <a href=3D"https://git.postgresql.org/pg/commitdiff/bafad2c5b261a1449bed6=
0ebac9e7918ebcc42b4" style=3D"color: #3498db; text-decoration: underline">h=
ttps://git.postgresql.org/pg/commitdiff/bafad2c5b261a1449bed60ebac9e7918ebc=
c42b4</a></li>
</ul>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Thomas Munro pushed:</p>
<ul style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal;=
 margin: 0; margin-bottom: 15px">
<li style=3D"list-style-position: inside; margin-left: 5px">Prepare for for=
thcoming LLVM 13 API change. LLVM 13 (due out in September) has
  changed the semantics of LLVMOrcAbsoluteSymbols(), so we need to bump some
  reference counts to avoid a double-free that causes crashes and bad query
  results.  A proactive change seems necessary to avoid having a window of =
time
  where our respective latest releases would interact badly.  It's possible=
 that
  the situation could change before then, though.  Thanks to Fabien Coelho =
for
  monitoring bleeding edge LLVM and Andres Freund for tracking down the cha=
nge.
  Back-patch to 11, where the JIT code arrived.  Discussion:
  <a href=3D"https://postgr.es/m/CA%2BhUKGLEy8mgtN7BNp0ooFAjUedDTJj5dME7NxL=
U-m91b85siA%40mail.gmail.com" style=3D"color: #3498db; text-decoration: und=
erline">https://postgr.es/m/CA%2BhUKGLEy8mgtN7BNp0ooFAjUedDTJj5dME7NxLU-m91=
b85siA%40mail.gmail.com</a>
<a href=3D"https://git.postgresql.org/pg/commitdiff/9b4e4cfe66ff133717c1b8b=
a3c2725d525c3e67c" style=3D"color: #3498db; text-decoration: underline">htt=
ps://git.postgresql.org/pg/commitdiff/9b4e4cfe66ff133717c1b8ba3c2725d525c3e=
67c</a></li>
</ul>
<h1 style=3D"color: #000; font-family: sans-serif; line-height: 1.4; margin=
: 0; margin-bottom: 30px; font-size: 25px; font-weight: 300; text-align: ce=
nter">Pending Patches</h1>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Nitin Jadhav sent in another revision of a =
patch to make a way to track the
progress of the operations performed during the startup process.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Yuzuko Hosoya sent in a patch to control wh=
ether autoanalyze runs for partition
operations via new GUCs autovacuum_analyze_attach_partition,
autovacuum_analyze_detach_partition, and autovacuum_analyze_drop_partition.=
</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Peter Eisentraut sent in a patch to add an =
index OID macro argument to
DECLARE_INDEX.  Instead of defining symbols such as AmOidIndexId explicitly,
include them as an argument of DECLARE_INDEX() and have genbki.pl generate =
the
way as the table OID symbols from the CATALOG() declaration</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Filip Gospodinov sent in a patch to intende=
d to fix a bug that manifested as the
shipped pkg-config file is broken for statically linking libpq because
libpgcommon and libpgport are missing. The fix adds those two missing priva=
te
dependencies.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Heikki Linnakangas sent in another revision=
 of a patch to split xlog.c into
xlog.c and xlogrecovery.c.  This moves the functions related to performing =
WAL
recovery into the new xlogrecovery.c source file, leaving xlog.c responsibl=
e for
maintaining the WAL buffers, coordinating the startup and switch from recov=
ery
to normal operations, and other miscellaneous stuff that have always been in
xlog.c</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Greg Nancarrow sent in two revisions of a p=
atch to remove the useless int64
range checks on BIGINT sequence MINVALUE/MAXVALUE values.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Peter Geoghegan sent in three revisions of =
a patch to add a list of ignorable
pgindent commits for "git blame".</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">David Rowley sent in another revision of a =
patch to add a new hash table type
which has stable pointers, and use same to speed up SMgr.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Peter Smith and Ajin Cherian traded patches=
 to add an option to support prepared
transactions to built-in logical replication.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Daniel Gustafsson and Micha=C3=ABl Paquier =
traded patches to document some SSL/TLS
related acronyms, and replace usage of SSL with SSL/TLS, the latter being m=
ore
accurate and up to date.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Simon Riggs sent in three revisions of a pa=
tch to add a documentation chapter on
hash indexes.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Atsushi Torikoshi sent in another revision =
of a patch to add a function to log
the complete query string and its plan for the query currently running on t=
he
backend with the specified process ID.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Li Japin sent in a patch to make a minor co=
de beautification in
RelationGetIdentityKeyBitmap.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Bertrand Drouvot sent in another revision o=
f a patch to enable minimal logical
decoding on standbys.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Alexander Pyhalov sent in another revision =
of a patch to allow pushing CASE
expressions to foreign servers.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Peter Eisentraut sent in a patch to Make th=
e Unicode makefile more
parallel-safe, Make UCS_to_most.pl process encodings in sorted order, remove
some whitespace in generated C output to get it in line with the coding sty=
le of
the rest of the project, simplify the code-generation code, and fix indenta=
tion
in generated output.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Maxim Orlov sent in another revision of a p=
atch to fix parallel worker failed
assertion and coredump.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Vigneshwaran C sent in two more revisions o=
f a patch to add schema level support
for PUBLICATIONs.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Mike Fiedler  sent in a patch to emit the n=
amespace in post-copy output.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Emre Hasegeli sent in a patch to decouple o=
perator classes from index access
methods.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Jacob Champion sent in two revisions of a p=
atch to decouple the SASL framework
from the SCRAM code, as it may be useful separately.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Yugo Nagata sent in another revision of a p=
atch to address some Pgbench errors
by using the Variables structure for client variables.  This is most import=
ant
when it is used to reset client variables during the repeating of transacti=
ons
after serialization/deadlock failures.  Don't allocate Variable structs one=
 by
one. Instead, add a constant margin each time it overflows. Also included, =
a fix
for pgbench errors and serialization/deadlock retries.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Jacob Champion, Micha=C3=ABl Paquier, and T=
om Lane traded patches to make it possible
to use jsonapi with libpq.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Gurjeet Singh sent in a patch to add automa=
tic notification for top transaction
IDs.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Haiying Tang sent in another revision of a =
patch to support tab completion with
a query result for upper case inputs in psql.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Peter Eisentraut sent in a patch to make ge=
nbki error handling more useful by
counting the errors and error out at the end instead of just writing warnin=
gs
for invalid cross-catalog lookups.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Daniel Gustafsson sent in another revision =
of a patch to support NSS as a TLS
backend for libpq.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Andrey V. Lepikhov sent in a patch to infor=
m genericcostestimate()'s index
selectivity value of the fact that in the case of an unique one-row btree
index, scan only one row can be returned.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Tom=C3=A1=C5=A1 Vondra sent in another revi=
sion of a patch to make it possible to
replicate sequences in logical decoding.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Ranier Vilela sent in a patch to fix an uni=
nitialized copy_data var in
src/backend/commands/subscriptioncmds.c.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Peter Eisentraut sent in another revision o=
f a patch to improve error messages
about mismatching relkind by making the primary error message shorter and s=
aying
more directly that what was attempted is not possible.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Peter Eisentraut sent in a patch to add tes=
ts for UNBOUNDED syntax ambiguity.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Simon Riggs sent in a patch to make pgbench=
 use COPY FREEZE.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Justin Pryzby sent in a patch to avoid doub=
le parentheses, and fix a comment in
tablefunc.c to refer to the correct comment.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Craig Ringer sent in a patch to add some mo=
re PGDLLIMPORTs to expose some of the
things heretofore unavailable on Windows.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Amit Kapila sent in a patch to allow stream=
ing the changes after speculative
aborts.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Hayato Kuroda sent in a patch to make ECPG'=
s new DECLARE STATEMENT work for
DEALLOCATE and DESCRIBE.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Bruce Momjian sent in another revision of a=
 patch to implement cluster file
encryption.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Jie Zhang sent in a patch to bring the gene=
rated sample postgresql.conf into
line with what's actually the default for <code>shared_buffers</code>.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Richard Guo sent in a patch to use each rel=
 as both outer and inner for anti
joins.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Tom Lane sent in a patch to prevent any cal=
ls of abort() or exit() from inside
libpq.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Alexander Korotkov sent in a patch to fix s=
ome small inconsistencies in the
catalog definition of multirange operators.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Julien Rouhaud sent in another revision of =
a patch to preserve param location
when generating a const node.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Andrey Borodin sent in another revision of =
a patch to reorganize the pglz
compression code.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Julien Rouhaud and Ranier Vilela traded pat=
ches to add a pg_get_query_def() to
deparse and print a rewritten SQL statement.</p>
<p style=3D"font-family: sans-serif; font-size: 14px; font-weight: normal; =
margin: 0; margin-bottom: 15px">Tom=C3=A1=C5=A1 Vondra sent in a PoC patch =
to do cardinality estimation using runtime
sampling.</p>

                      </td>
                    </tr>
                  </table>
                </td>
              </tr>

            </table>

            <div class=3D"footer" style=3D"clear: both; Margin-top: 10px; t=
ext-align: center; width: 100%;">
              <table border=3D"0" cellpadding=3D"0" cellspacing=3D"0" style=
=3D"border-collapse: separate; mso-table-lspace: 0pt; mso-table-rspace: 0pt=
; width: 100%;">
                <tr>
                  <td class=3D"content-block" style=3D"font-family: sans-se=
rif; vertical-align: top; padding-bottom: 10px; padding-top: 10px; font-siz=
e: 12px; color: #999999; text-align: center;">
                    <span class=3D"apple-link" style=3D"color: #999999; fon=
t-size: 12px; text-align: center;">
This email was sent to you from PWN. It was delivered on their behalf by
the PostgreSQL project. Any questions about the content of the message shou=
ld be
sent to PWN.
</span>
		    <br><br>
You were sent this email as a subscriber of the <em>pgsql-announce</em> mai=
linglist, for
the content tag PWN.
To unsubscribe from
further emails, or change which emails you want to receive, please click th=
e personal unsubscribe
link that you can find in the headers of this email, or visit
<a href=3D"https://lists.postgresql.org/unsubscribe/" style=3D"color: #3498=
db; text-decoration: underline">https://lists.postgresql.org/unsubscribe/</=
a>.

                  </td>
                </tr>
              </table>
            </div>

          </div>
        </td>
        <td style=3D"font-family: sans-serif; font-size: 14px; vertical-ali=
gn: top;">&nbsp;</td>
      </tr>
    </table>
  </body>
</html>

--===============0330074949510022916==--

--===============8052356165775126479==--