Re: v18 GUC in v17 translation

On 2025-Nov-14, Daniel Gustafsson wrote:

> The Japanese initdb translation contains strings which are only present in v18
> and onwards, which seems seems incorrect. While I know that we don't apply
> changes directly to the in-tree .po files I attached a diff for that to
> illustrate. Am I missing something or should these be removed from the source
> translation repo?

Hmm, this is pretty harmless -- it just means the message catalog has an
entry that will never be used. But it won't have any bad consequences.
I don't think we need to go out of our ways to clean it up: the
automated merge workflow should eventually remove that string from the
source file.

It looks like this string came in with commit faf514a1ccf9 in the
pgtranslation/messages repository. I don't understand what could have
happened here, other than the translator downloading the file for the
wrong branch. Maybe this explains why the stats for Japanese in 17 are
much lower than for all other branches?

I think our branch handling is really confusing and could easily induce
people to do the wrong thing. I'm not sure what would be better,
though. I think we could just get rid of branch master altogether, and
create the next one from the previous (e.g., create REL_19_STABLE from
REL_18_STABLE), rather than branching the previous from master each
time.

--
Álvaro Herrera PostgreSQL Developer — https://www.EnterpriseDB.com/
"Tiene valor aquel que admite que es un cobarde" (Fernandel)