CFbot failed on Windows platform

Hi,

On Thu, 11 Jul 2024 at 11:07, Tatsuo Ishii <ishii@postgresql.org> wrote:

Hi,

A cfbot test suddenly failed with regression test error.

http://cfbot.cputube.org/highlights/all.html#4460

Following the link "regress" shows:
https://api.cirrus-ci.com/v1/artifact/task/5428792720621568/testrun/build/testrun/recovery/027_stream_regress/data/regression.diffs

diff --strip-trailing-cr -U3 C:/cirrus/src/test/regress/expected/collate.windows.win1252.out C:/cirrus/build/testrun/recovery/027_stream_regress/data/results/collate.windows.win1252.out
--- C:/cirrus/src/test/regress/expected/collate.windows.win1252.out     2024-07-11 02:44:08.385966100 +0000
+++ C:/cirrus/build/testrun/recovery/027_stream_regress/data/results/collate.windows.win1252.out        2024-07-11 02:49:51.280212100 +0000
@@ -21,10 +21,10 @@
);
\d collate_test1
Table "collate_tests.collate_test1"
- Column |  Type   | Collation | Nullable | Default
+ Column |  Type   | Collation | Nullable | Default
--------+---------+-----------+----------+---------
- a      | integer |           |          |
- b      | text    | en_US     | not null |
+ a      | integer |           |          |
+ b      | text    | en_US     | not null |
:
:

The differences are that the result has an extra space at the end of
line. I am not familiar with Windows and has no idea why this could
happen (I haven't changed the patch set since May 24. The last
succeeded test was on July 9 14:58:44 (I am not sure about time zone).
Also I see exactly the same test failures in some other tests (for
example http://cfbot.cputube.org/highlights/all.html#4337)

Any idea?

I think It is related to the '628c1d1f2c' commit. This commit changed
the output of the regress test in Windows.

--
Regards,
Nazir Bilal Yavuz
Microsoft

The differences are that the result has an extra space at the end of
line. I am not familiar with Windows and has no idea why this could
happen (I haven't changed the patch set since May 24. The last
succeeded test was on July 9 14:58:44 (I am not sure about time zone).
Also I see exactly the same test failures in some other tests (for
example http://cfbot.cputube.org/highlights/all.html#4337)

Any idea?

I think It is related to the '628c1d1f2c' commit. This commit changed
the output of the regress test in Windows.

Yeah, it seems that explains. I see few buildfarm window animals
complain too.

Best reagards,
--
Tatsuo Ishii
SRA OSS LLC
English: http://www.sraoss.co.jp/index_en/
Japanese:http://www.sraoss.co.jp

Tatsuo Ishii <ishii@postgresql.org> writes:

I think It is related to the '628c1d1f2c' commit. This commit changed
the output of the regress test in Windows.

Yeah, it seems that explains. I see few buildfarm window animals
complain too.

I think that the contents of
src/test/regress/expected/collate.windows.win1252.out are actually
wrong, and we'd not noticed because it was only checked with diff -w.
psql does put an extra trailing space in some lines of table output,
but that space isn't there in collate.windows.win1252.out.

regards, tom lane

On 2024-07-11 Th 5:34 AM, Tatsuo Ishii wrote:

The differences are that the result has an extra space at the end of
line. I am not familiar with Windows and has no idea why this could
happen (I haven't changed the patch set since May 24. The last
succeeded test was on July 9 14:58:44 (I am not sure about time zone).
Also I see exactly the same test failures in some other tests (for
example http://cfbot.cputube.org/highlights/all.html#4337)

Any idea?

I think It is related to the '628c1d1f2c' commit. This commit changed
the output of the regress test in Windows.

Yeah, it seems that explains. I see few buildfarm window animals
complain too.

I have partially reverted that patch. Thanks for the report.

cheers

andrew

--
Andrew Dunstan
EDB: https://www.enterprisedb.com

On 2024-07-11 Th 9:50 AM, Tom Lane wrote:

Tatsuo Ishii <ishii@postgresql.org> writes:

I think It is related to the '628c1d1f2c' commit. This commit changed
the output of the regress test in Windows.

Yeah, it seems that explains. I see few buildfarm window animals
complain too.

I think that the contents of
src/test/regress/expected/collate.windows.win1252.out are actually
wrong, and we'd not noticed because it was only checked with diff -w.
psql does put an extra trailing space in some lines of table output,
but that space isn't there in collate.windows.win1252.out.

Yeah, makes sense I will produce an updated output file and then
re-enable --strip-trailing-cr (after a full test run).

cheers

andrew

--
Andrew Dunstan
EDB: https://www.enterprisedb.com

On 2024-07-11 Th 5:34 AM, Tatsuo Ishii wrote:

The differences are that the result has an extra space at the end of
line. I am not familiar with Windows and has no idea why this could
happen (I haven't changed the patch set since May 24. The last
succeeded test was on July 9 14:58:44 (I am not sure about time zone).
Also I see exactly the same test failures in some other tests (for
example http://cfbot.cputube.org/highlights/all.html#4337)

Any idea?

I think It is related to the '628c1d1f2c' commit. This commit changed
the output of the regress test in Windows.

Yeah, it seems that explains. I see few buildfarm window animals
complain too.

I have partially reverted that patch. Thanks for the report.

Thank you for taking care of this!
--
Tatsuo Ishii
SRA OSS LLC
English: http://www.sraoss.co.jp/index_en/
Japanese:http://www.sraoss.co.jp