Re: ON CONFLICT with constraint name doesn't work

On 2017-03-16 13:56:27 -0700, Nikolay Samokhvalov wrote:
> On Thu, Mar 16, 2017 at 1:23 PM, Andres Freund <andres(at)anarazel(dot)de> wrote:
> >
> > Seems like a good idea to improve that message. I wouldn't vote for
> > backpatching it, however. Could you propose a patch for that?
> >
>
> OK. Here it is.

I don't think that's an appropriate fix. ISTM we should say 'violates
unique index' when it's just an index and 'violates unique constraint'
when the index is backing a constraint.

> But:
> 1) what's next with all the i18n stuff? I've changed RU version as well,
> but unfortunately I don't the other languages.

You don't need to patch any (including RU). They're maintained
separately and the in-core stuff is periodically refreshed by the
translators.

> 2) it will definitely break many regression tests, should I patch them as
> well?

Yes, after above adaption, that'll probably reduce the size of the diff.

> Also, documentation explaining ON CONFLICT might be still not clear, at
> least for some readers. Do you want me to propose a patch for that as well?

Please feel free to give it a try.

Greetings,

Andres Freund