Re: tiny step toward threading: reduce dependence on setlocale()

From: Peter Eisentraut <peter(at)eisentraut(dot)org>
To: Jeff Davis <pgsql(at)j-davis(dot)com>, Andreas Karlsson <andreas(at)proxel(dot)se>, pgsql-hackers(at)postgresql(dot)org
Subject: Re: tiny step toward threading: reduce dependence on setlocale()
Date: 2024-09-09 13:37:23
Message-ID: 84d415fc-6780-419e-b16c-61a0ca819e2b@eisentraut.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On 08.08.24 22:00, Jeff Davis wrote:
> On Wed, 2024-08-07 at 22:44 +0200, Peter Eisentraut wrote:
>> But after this patch set, locale cannot be NULL anymore, so the third
>> branch is obsolete.
>
> ...
>
>> Second, there are a number of functions in like.c like the above that
>> take separate arguments like pg_locale_t locale, bool locale_is_c.
>> Because pg_locale_t now contains the locale_is_c information, these
>> can
>> be combined.
>
> I believe these patches are correct, but the reasoning is fairly
> complex:
>
> 1. Some MatchText variants are called with 0 for locale. But that's OK
> because ...
>
> 2. A MatchText variant only cares about the locale if MATCH_LOWER(t) is
> defined, and ...
>
> 3. Only one variant, SB_IMatchText() defines MATCH_LOWER(), and ...
>
> 4. SB_IMatchText() is called with a non-zero locale.
>
> All of these are a bit confusing to follow because it's generated code.
> #2 is particularly non-obvious, because "locale" is not even an
> argument of the MATCH_LOWER(t) or GETCHAR(t) macros, it's taken
> implicitly from the outer scope.
>
> I don't think your patches cause this confusion, but is there a way you
> can clarify some of this along the way?

In the end, I figured the best thing to do here is to add an explicit
locale argument to MATCH_LOWER() and GETCHAR() so one can actually
understand this code by reading it. New patch attached.

Attachment Content-Type Size
v2-0001-Remove-separate-locale_is_c-arguments.patch text/plain 6.5 KB

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Frédéric Yhuel 2024-09-09 13:47:49 Re: pgstattuple: fix free space calculation
Previous Message Jim Jones 2024-09-09 13:22:46 Re: Psql meta-command conninfo+