Re: [PATCH v6 00/10] commit-reach: terminate merge-base walk when one side is exhausted

From: Kristofer Karlsson <krka@spotify.com>
Date: Jul 11, 2026, 21:41 UTC
Message-ID: <CAL71e4O5=ZJoPD4dnPmh8mjsTKtugx05-8d83VeQdBNOjp=bFw@mail.gmail.com>
In-Reply-To: <xmqqv7al9rbj.fsf@gitster.g>

On Sat, 11 Jul 2026 at 22:58, Junio C Hamano <gitster@pobox.com> wrote:

Show 10 quoted lines

>
> As always, do *not* base your patches on 'next'.  I cannot apply
> such a patch series to my tree, as merging the resulting topic down
> to 'master' will pull _all_ the other topics, including those that
> are not ready, plus commits that merge these topics into 'next',
> into 'master'.
>
> Instead, choose the topics that you do depend on, prepare a merge of
> these branches into a stable base (like v2.55.0 or master), and then
> build your series on top.

Ah I think I phrased it poorly in the cover letter. When I said that it's based on next, I meant that it is verified to work against next but I also confirmed it works against what you suggested earlier, e.g. a synthetic base:

    git checkout -b synthetic-base origin/master
    git merge --no-ff kk/commit-reach-find-all-fix

There is one textual conflict in commit-reach.c; the resolution combines both:

    if (!min_generation && !corrected_commit_dates_enabled(r)) {
        queue.pq.compare = compare_commits_by_commit_date;
        gen_ordered = 0;
    }

After that, all ten patches apply cleanly with git am -3.

I should have stated this more clearly in the cover letter instead of mentioning next at all.

Thanks, Kristofer