Re: IPC::Run::time[r|out] vs our TAP tests
Tom Lane <tgl@sss.pgh.pa.us>
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Dagfinn Ilmari Mannsåker <ilmari@ilmari.org>
Cc: Erik Wienhold <ewie@ewie.name>, Michael Paquier <michael@paquier.xyz>,
pgsql-hackers@lists.postgresql.org,
Alexander Korotkov <aekorotkov@gmail.com>
Date: 2024-04-05T21:18:51Z
Lists: pgsql-hackers
=?utf-8?Q?Dagfinn_Ilmari_Manns=C3=A5ker?= <ilmari@ilmari.org> writes: > Tom Lane <tgl@sss.pgh.pa.us> writes: >> So it seems like the bug does not exist in any currently-supported >> NetBSD release. Debian has been known to ship obsolete libedit >> versions, though. > Both the current (bokworm/12) and previous (bullseye/11) versions of > Debian have new enough libedits to not be affected by this bug: > ... > But in bullseye they decided that OpenSSL is a system library as far as > the GPL is concerned, so are linking directly to readline. > And even before then their psql wrapper would LD_PRELOAD readline > instead if installed, so approximately nobody actually ever used psql > with libedit on Debian. Based on this info, I'm disinclined to put work into trying to make the case behave correctly with that old libedit version, or even to lobotomize the test case enough so it would pass. What I suggest Michael do with tanager is install the OS-version-appropriate version of GNU readline, so that the animal will test what ilmari describes as the actually common use-case. (I see that what he did for the moment is add --without-readline. Perhaps that's a decent long-term choice too, because I think we have rather little coverage of that option except on Windows.) regards, tom lane
Commits
-
Restart BackgroundPsql's timer more nicely.
- 53c6bd0aa3de 19 (unreleased) landed
-
Implement pg_wal_replay_wait() stored procedure
- 3c5db1d6b016 18.0 cited
-
psql: fix variable existence tab completion
- 927332b95e77 17.0 cited