Hi,
On Thu, Nov 27, 2025 at 03:07:17PM +0300, Nazir Bilal Yavuz wrote:
> On Thu, 27 Nov 2025 at 11:57, Bertrand Drouvot
> <bertranddrouvot.pg@gmail.com> wrote:
> >
> > The patches are split by modules to ease the review.
>
> I applied all patches and built/tested without any errors. So I guess
> that the changes are correct.
Thanks for looking at it!
> I skimmed through 0001-0003 to see if I can find more parameters to
> delete and found that we can remove the 'tinfo' parameter from the
> gbt_var_node_truncate() function.
Yeah, also found it with an improved version of the coccinelle script. Also
there is one more in postgres_fdw: unused rte parameter in create_foreign_modify()
is oversight in commit a61b1f74823.
And also one in pg_walinspect.
All added in the attached.
> > For 0005 (collid not used in internal_citext_pattern_cmp()) I think that the
> > unused column should be used (as done in 0005) due to an oversight in commit
> > f2464997644c.
>
> I do not have a context on this but there is a comment in the
> citextcmp() function, which might be related to this context:
>
> /*
> * We must do our str_tolower calls with DEFAULT_COLLATION_OID, not the
> * input collation as you might expect. This is so that the behavior of
> * citext's equality and hashing functions is not collation-dependent. We
> * should change this once the core infrastructure is able to cope with
> * collation-dependent equality and hashing functions.
> */
Interesting... So maybe we should just remove collid in
internal_citext_pattern_cmp() instead: that's what 0005 is now doing.
Tha said, I'm not familiar with this code area, so better to wait other feedbacks
for 0005 then.
Regards,
--
Bertrand Drouvot
PostgreSQL Contributors Team
RDS Open Source Databases
Amazon Web Services: https://aws.amazon.com
