Re: Avoid a possible null pointer (src/backend/utils/adt/pg_locale.c)

Ranier Vilela <ranier.vf@gmail.com>

From: Ranier Vilela <ranier.vf@gmail.com>
To: Michael Paquier <michael@paquier.xyz>
Cc: Robert Haas <robertmhaas@gmail.com>, Pg Hackers <pgsql-hackers@postgresql.org>
Date: 2023-09-06T10:57:03Z
Lists: pgsql-hackers

Attachments

Em seg., 4 de set. de 2023 às 11:27, Ranier Vilela <ranier.vf@gmail.com>
escreveu:

> Em dom., 3 de set. de 2023 às 22:01, Michael Paquier <michael@paquier.xyz>
> escreveu:
>
>> On Sat, Sep 02, 2023 at 09:13:11AM -0300, Ranier Vilela wrote:
>> > I tried to keep the same behavior as before.
>> > Note that if the locale equals COLLPROVIDER_LIBC,
>> > the message to the user will be the same.
>>
>> -    /* shouldn't happen */
>> -    elog(ERROR, "unsupported collprovider: %c", locale->provider);
>> +    elog(ERROR, "collprovider '%c' does not support
>> pg_strnxfrm_prefix()",
>> +             locale->provider);
>>
>> Based on what's written here, these messages could be better because
>> full sentences are not encouraged in error messages, even for
>> non-translated elogs:
>> https://www.postgresql.org/docs/current/error-style-guide.html
>>
>> Perhaps something like "could not use collprovider %c: no support for
>> %s", where the function name is used, would be more consistent.
>>
> Sure.
> I have no objection to the wording of the message.
> If there is consensus, I can send another patch.
>
I think no one objected.

v1 attached.

best regards,
Ranier Vilela

Commits

  1. Refactor error messages for unsupported providers in pg_locale.c