Re: [PATCH v6 00/18] reftable: stop using "git-compat-util.h"

[Patrick Steinhardt <ps@pks.im>]

On Wed, Mar 12, 2025 at 07:51:41AM +0100, Patrick Steinhardt wrote:
> On Tue, Mar 11, 2025 at 04:29:27PM -0700, Junio C Hamano wrote:
> > Junio C Hamano <gitster@pobox.com> writes:
> > 
> > > Patrick Steinhardt <ps@pks.im> writes:
> > >
> > >> Changes in v6:
> > >>   - I have rebased the patch series on top of "master" at 03944513488
> > >>     (The eleventh batch, 2025-02-14). On the one hand this gets rid of
> > >>     the conflict with the zlib-ng changes, which have meanwhile been
> > >>     merged to that branch. And on the other hand it pulls in the new CI
> > >>     build for MSVC+Meson, which allows me to have better test coverage
> > >>     on Windows to better detect breakage there.
> > >>   - Stop moving around Win32 files so that we do the minimum viable
> > >>     change in this context. Hopefully, this makes it easier for Dscho to
> > >>     integrate into Git for Windows. If it doesn't I don't really have
> > >>     any other good ideas for how to do it, so the only alternative would
> > >>     be to just ignore Windows altogether.
> > >>   - Link to v5: https://lore.kernel.org/r/20250207-pks-reftable-drop-git-compat-util-v5-0-ba2adc79110f@pks.im
> > >
> > > Thanks.  I think the "unlink" thing on Windows I have was queued
> > > depended on the previous iteration of this series so let me make
> > > sure that I'd unentangle it while replacing the topic with this
> > > round.
> > >
> > > Will queue.  Thanks.
> > 
> > So, sorry I lost track.  I think I saw you ask Dscho about the
> > Windows specific bits once in a thread on What's cooking report,
> > but I do not recall the outcome.  Is this topic (and the unlink fix
> > that depends on this series) now OK, or would we see further updates
> > once 2.49 is tagged?
> 
> There was a class of test failures caused by mismatching allocators in
> Git for Windows for which I've sent a proposed fix for downstream [1].
> I've been talking to Johannes on Monday and he mentioned that he'll
> likely end up taking that patch as the pragmatic fix. Which to the best
> of my knowledge would mean that this series here is unblocked and can
> start to move forward again. Johannes, please correct me in case I'm
> misstating anything here.
> 
> So based on that I don't plan to reroll and hope that we can move
> forward with the series once Git v2.49.0 is out.

Quick status update: Johannes mentioned in [1] that he is fine with
moving forward with this topic now. Thanks!

Patrick

[1]: https://github.com/gitgitgadget/git/pull/1874#issuecomment-2740342672