Re: vcs_info backports? (was: test release: 5.9.1.2-test (will be 5.9.2))

["Daniel Shahaf" <d.s@xxxxxxxxxxxxxxxxxx>]

dana wrote on Sun, 05 Jul 2026 22:58 +00:00:
> (this was probably meant for -workers, right? sorry if not)
>

Cc'ing -users@ as opposed to -workers@ wasn't an oversight.  I kept the
Cc that way for several reasons, one them being to make the question at
hand visible to any stg(1) users who may lurk on -users@ (cf. below about 51144).

That's not a suggestion to move the thread _back_, though :-)

> On Sun 5 Jul 2026, at 17:14, Daniel Shahaf wrote:
>> workers/51138: Does this patch meet our criteria for backporting to a
>> patch release?
>
> yeah, it sounds like that can't do much harm. i'll back-port
>

Thanks.

> On Sun 5 Jul 2026, at 17:14, Daniel Shahaf wrote:
>> workers/51142 seems like it could easily break people's setups ("I
>> upgraded and vcs_info stopped showing information from git-rebase(1)'s
>> todo list"), so not proposing it.  If anything, it should perhaps be
>> mentioned in 5.10's list of changes in README?
>
> i can mention it yeah
>

Thanks!  Saves me having to do the work.

> On Sun 5 Jul 2026, at 17:14, Daniel Shahaf wrote:
>> workers/51144+51146... Not sure.
>
> me either, i know nothing about stg and not much more about vcs_info. i
> guess i lean towards being conservative with it but if you think it'll
> solve more problems than it creates maybe it's fine
>

I'm not sure I'd call the lack of patch subject extraction a "problem" that's being solved.
The previous code was correct, operating within the documented hook API
contract, but, essentially, wasn't passing in an optional parameter.
I was just saying that for vcs_info to start passing the "optional
parameter" into people's hooks in a patch release would be unlikely to
break anything _in the particular circumstances_, as users' hooks which
would be affected, have probably been run against another codepath which
does supply the "optional parameter".

At the end of the day I guess I'd vote to "Not backport, but open to be
convinced otherwise".

> On Sun 5 Jul 2026, at 17:14, Daniel Shahaf wrote:
>> If I've missed the 5.9.2 release train then please consider the
>> questions for 5.9.3, should there be one (even if not soon).
>
> i was going to release today but i've decided to delay a couple of days,
> maybe even a week. there are some more fixes (like 54909) i'd like to go
> in, and although i'd rather not do another test release we should at
> least give time for review

Sounds good.  And thanks for adding that unposted regression test!

Daniel