Re: history damage in linux.git

From: Linus Torvalds <torvalds@linux-foundation.org>
Date: Apr 21, 2016, 16:36 UTC
Message-ID: <CA+55aFx8hPKKcuwe-HHoO7LHVYLmJ6khndd-OtQotMs3EJzZ0w@mail.gmail.com>
In-Reply-To: <87lh473xic.fsf@linux-m68k.org>

On Thu, Apr 21, 2016 at 6:24 AM, Andreas Schwab <schwab@linux-m68k.org> wrote:

Show 5 quoted lines

>
> The branches recently pulled by Linus contain commits with rather old
> author dates, eg 8cad489261c564d4ee1db4de4ac365af56807d8a or
> 281baf7a702693deaa45c98ef0c5161006b48257.  That will probably cause git
> describe --contains to take a different path through the history.

Nothing in git should look at author dates (except for the obvious code that then shows the date).

The committer date is in fact used for the traversal heuristics for some cases, but those are different and recent in the examples you note.

This seems to be a git bug.

That commit aed06b9 can also be described as

    v3.13-rc7~9^2~14^2~42

so describing it as 'v4.6-rc1~9^2~792' is clearly not closer in any way.

However did git decide to use that further-away name? That looks odd.

I'm wondering if it's because of the new root commit we have, and that confuses the distance calculation somehow? That came in somewhat recently (mid March) with the GPIO pull.

       Linus