| From: | Daniel Gustafsson <daniel(at)yesql(dot)se> |
|---|---|
| To: | Álvaro Herrera <alvherre(at)kurilemu(dot)de> |
| Cc: | "Zhijie Hou (Fujitsu)" <houzj(dot)fnst(at)fujitsu(dot)com>, PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org> |
| Subject: | Re: Few untranslated error messages in OAuth |
| Date: | 2025-11-13 12:19:30 |
| Message-ID: | CB611A15-5033-473F-ABE0-E07449714A3F@yesql.se |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
> On 13 Nov 2025, at 09:13, Álvaro Herrera <alvherre(at)kurilemu(dot)de> wrote:
> On 2025-Nov-13, Zhijie Hou (Fujitsu) wrote:
>> During testing of libpq, I noticed several error messages not processed by
>> libpq_gettext, resulting in their exclusion from *.pot files. The attached patch
>> wraps these messages.
Thanks for the report!
> Yeah, this roughly makes sense. I think the documentation for errctx in
> async_ctx needs to be updated though, because currently it says "it'll
> be translated for you" but this patch makes it the other way around (it
> must come translated).
+1
> Alternatively, we could mark the strings with gettext_noop(), which only
> adds the messages to the catalog without translating them; so the
> responsibility of translating them would remain where it is today. This
> might be easier.
I was pondering that since auth error messages in backend libpq does this, but
since we don't use gettext_noop anywhere else in frontend libpq today it makes
sense to go with libpq_gettext and update the docs instead.
Unless you, who has more translation/nls insights than me, feel strongly about
gettext_noop I'll go ahead with the current patch.
--
Daniel Gustafsson
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Mihail Nikalayeu | 2025-11-13 12:22:38 | Re: [BUG?] check_exclusion_or_unique_constraint false negative |
| Previous Message | Yugo Nagata | 2025-11-13 12:00:33 | Re: Suggestion to add --continue-client-on-abort option to pgbench |