Re: 039_end_of_wal: error in "xl_tot_len zero" test - Mailing list pgsql-hackers

From Christoph Berg
Subject Re: 039_end_of_wal: error in "xl_tot_len zero" test
Date
Msg-id Zz8pJW9xVqIZvwdk@msg.df7cb.de
Whole thread Raw
In response to Re: 039_end_of_wal: error in "xl_tot_len zero" test  (Christoph Berg <myon@debian.org>)
List pgsql-hackers
Re: To Thomas Munro
> It is (was?) weird. The PG13 build for bullseye on apt.pg.o worked
> normally, but when I built for bullseye-security (Debian LTS), I had
> to remove the patch that removed the 039_end_of_wal.pl file because
> you had changed the file. The build then failed and I put the
> remove-that-file-patch back. I did not try to repeat the build.
> 
> I'll try again with today's releases.

******** build/src/test/recovery/tmp_check/log/regress_log_039_end_of_wal ********
No such file or directory at /home/myon/projects/postgresql/debian/13/build/../src/test/perl/TestLib.pm line 655.

It's still happening, but it's Debian's fault. We used to lobotomize
the relocatability of the binaries a bit to cater for our directory
layout, so in the PG13 in bullseye, includedir-server is
/usr/include/postgresql/13/server even when called from a temp tree.


https://salsa.debian.org/postgresql/postgresql/-/blob/13-bullseye/debian/patches/50-per-version-dirs.patch?ref_type=heads

In the meantime, I had noticed that the logic is no longer required
(if it ever was), so the newer branches (including 13 for apt.pg.o)
have a much shorter version of that patch:

https://salsa.debian.org/postgresql/postgresql/-/blob/13/debian/patches/50-per-version-dirs.patch?ref_type=heads

so tl;dr: it's been fixed in the meantime, sorry for the noise.

Christoph



pgsql-hackers by date:

Previous
From: Daniel Gustafsson
Date:
Subject: Re: Windows 2016 server crashed after changes in Postgres 15.8 pgAdmin
Next
From: Peter Eisentraut
Date:
Subject: Re: SQL:2011 application time