Horton

Re: abi-compliance-check failure due to recent changes to pg_{clear,restore}_{attribute,relation}_stats()

Mankirat Singh <mankiratsingh1315@gmail.com>

View thread View on postgr.es
From: Mankirat Singh <mankiratsingh1315@gmail.com>
To: "David E. Wheeler" <david@justatheory.com>, Tom Lane <tgl@sss.pgh.pa.us>
Cc: nathandbossart@gmail.com, pg@bowt.ie, andrew@dunslane.net, pgsql-hackers@lists.postgresql.org
Date: 2025-10-18T20:48:58Z
Lists: pgsql-hackers 
On Sat, 18 Oct 2025 at 21:17, David E. Wheeler <david@justatheory.com>
wrote:
>
> On Oct 18, 2025, at 11:14, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>
> > I am strongly against relying on suppression files.  I think that
> > is a very imprecise technology, and it is certainly harder to use
> > than the "choose a blessed reference point" approach.
>
> Seconded. I’m also not keen on using something specific to libabigail if,
long term, we want to enable similar checks on other platforms using other
tools that may not support it suppression file format.
>
> Let’s do the baseline SHA file and see how it goes.

Oh right, I didn't think about the other platforms.

I've implemented the support for .abi-compliance-history file. If the file
is present in a STABLE branch, it will be used by default; otherwise, the
latest tag will be used. You can view the changes here[1].

Regards,
Mankirat

[1] -
https://github.com/PGBuildFarm/client-code/pull/38/commits/f88873625ba0466e9609225947d40092748ff555

Commits

  1. Add commit 24f6c1bd4 to v17 .abi-compliance-history.

  2. Add .abi-compliance-history to back-branches.

  3. Add reminder to create .abi-compliance-history.

  4. Update .abi-compliance-history file.

  5. Add .abi-compliance-history to v18 branch.

  6. Fix thinko in commit 7d129ba54.

  7. Fix lookups in pg_{clear,restore}_{attribute,relation}_stats().

  8. Add defenses against unexpected changes in the NodeTag enum list.