From: | Michael Paquier <michael(at)paquier(dot)xyz> |
---|---|
To: | Andrew Dunstan <andrew(at)dunslane(dot)net> |
Cc: | Noah Misch <noah(at)leadboat(dot)com>, PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org> |
Subject: | Re: enable certain TAP tests for MSVC builds |
Date: | 2021-12-08 08:14:14 |
Message-ID: | YbBpViH/EIJ2ZgBO@paquier.xyz |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Tue, Dec 07, 2021 at 03:40:29PM -0500, Andrew Dunstan wrote:
> All done.
bowerbird is complaining here with the tests of pg_basebackup:
https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=bowerbird&dt=2021-12-08%2004%3A52%3A27
tar: Cannot execute remote shell: No such file or directory
tar: H\\:/prog/bf/root/HEAD/pgsql.build/src/bin/pg_basebackup/tmp_check/t_010_pg_basebackup_main_data/backup/tarbackup2/base.tar:
Cannot open: I/O error
tar: Error is not recoverable: exiting now
This comes from init_from_backup() via the tar_command. I am not sure
what to think about this error. Perhaps this is related to the
environment.
+$ENV{TAR} ||= 'tar';
+$ENV{LZ4} ||= 'lz4';
+$ENV{GZIP_PROGRAM} ||= 'gzip';
This means that the default is to assume that those commands will be
always present, no? But that may not be the case, which would cause
the tests of pg_basebackup to fail in 010_pg_basebackup.pl. Shouldn't
we at least check that the command can be executed if defined and skip
the tests if not, in the same fashion as LZ4 and GZIP?
--
Michael
From | Date | Subject | |
---|---|---|---|
Next Message | Bharath Rupireddy | 2021-12-08 08:30:12 | Re: add recovery, backup, archive, streaming etc. activity messages to server logs along with ps display |
Previous Message | Masahiko Sawada | 2021-12-08 08:03:30 | Re: Make pg_waldump report replication origin ID, LSN, and timestamp. |