Content-Type: multipart/mixed; boundary="------------F4lcAGcwSZq0CA2jm9m4Ca6g"
Message-ID: <60d434b9-53d9-9ea1-819b-efebdcf44e41@gmail.com>
Date: Sat, 25 Nov 2023 23:00:01 +0300
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.4.2
From: Alexander Lakhin <exclusion@gmail.com>
Subject: Test 002_pg_upgrade fails with olddump on Windows
To: pgsql-hackers <pgsql-hackers@postgresql.org>
Content-Language: en-US
Archived-At: 
 <https://www.postgresql.org/message-id/60d434b9-53d9-9ea1-819b-efebdcf44e41%40gmail.com>
Precedence: bulk

This is a multi-part message in MIME format.
--------------F4lcAGcwSZq0CA2jm9m4Ca6g
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit

Hello,

When trying to use a custom dump with the test pg_upgrade/002_pg_upgrade,
I observe the following test failure on Windows:
 >meson test --suite setup
 >echo create database regression>...\dump.sql
 >set olddump=...\dump.sql& set oldinstall=.../tmp_install/usr/local/pgsql& meson test pg_upgrade/002_pg_upgrade

1/1 postgresql:pg_upgrade / pg_upgrade/002_pg_upgrade ERROR            11.38s   exit status 1

regress_log_002_pg_upgrade.txt contains:
...
[09:07:06.704](3.793s) ok 11 - run of pg_upgrade for new instance
...
[09:07:07.301](0.001s) not ok 15 - old and new dumps match after pg_upgrade
[09:07:07.301](0.000s) #   Failed test 'old and new dumps match after pg_upgrade'
#   at .../src/bin/pg_upgrade/t/002_pg_upgrade.pl line 452.
[09:07:07.301](0.000s) #          got: '1'
#     expected: '0'
=== diff of ...\build\testrun\pg_upgrade\002_pg_upgrade\data\tmp_test_ifk8/dump1.sql and 
...\build\testrun\pg_upgrade\002_pg_upgrade\data\tmp_test_ifk8/dump2.sql
=== stdout ===
=== stderr ===
=== EOF ===


 >dir "testrun\pg_upgrade\002_pg_upgrade\data\tmp_test_ifk8/"
11/25/2023  09:06 AM             2,729 dump1.sql
11/25/2023  09:07 AM             2,590 dump2.sql

 >diff -s "testrun\pg_upgrade\002_pg_upgrade\data\tmp_test_ifk8\dump1.sql" 
"testrun\pg_upgrade\002_pg_upgrade\data\tmp_test_ifk8\dump2.sql"
Files testrun\pg_upgrade\002_pg_upgrade\data\tmp_test_ifk8\dump1.sql and 
testrun\pg_upgrade\002_pg_upgrade\data\tmp_test_ifk8\dump2.sql are identical

As I can see, dump1.sql contains line endings 0d 0a, while dump2.sql — 0a.

The attached patch fixes the issue for me.

Best regards,
Alexander
--------------F4lcAGcwSZq0CA2jm9m4Ca6g
Content-Type: text/x-patch; charset=UTF-8;
 name="fix_002_pg_upgrade-with-olddump-win.patch"
Content-Disposition: attachment;
 filename="fix_002_pg_upgrade-with-olddump-win.patch"
Content-Transfer-Encoding: base64

ZGlmZiAtLWdpdCBhL3NyYy9iaW4vcGdfdXBncmFkZS90LzAwMl9wZ191cGdyYWRlLnBsIGIv
c3JjL2Jpbi9wZ191cGdyYWRlL3QvMDAyX3BnX3VwZ3JhZGUucGwKaW5kZXggYzZkODNkM2My
MS4uZDM0YjQ1ZTM0NiAxMDA2NDQKLS0tIGEvc3JjL2Jpbi9wZ191cGdyYWRlL3QvMDAyX3Bn
X3VwZ3JhZGUucGwKKysrIGIvc3JjL2Jpbi9wZ191cGdyYWRlL3QvMDAyX3BnX3VwZ3JhZGUu
cGwKQEAgLTI5Myw2ICsyOTMsNyBAQCBpZiAoZGVmaW5lZCgkRU5We29sZGluc3RhbGx9KSkK
IAl9CiAKIAlvcGVuIG15ICRmaCwgIj4iLCAkZHVtcDFfZmlsZSBvciBkaWUgImNvdWxkIG5v
dCBvcGVuIGR1bXAgZmlsZSI7CisJYmlubW9kZSAkZmg7CiAJcHJpbnQgJGZoICRkdW1wX2Rh
dGE7CiAJY2xvc2UgJGZoOwogCg==

--------------F4lcAGcwSZq0CA2jm9m4Ca6g--