Re: gitk regression in version 2.36.0

From: René Scharfe <l.s.r@web.de>
Date: Apr 23, 2022, 10:13 UTC
Message-ID: <737bd2f8-dfd7-6a51-b7f5-33eefa33e975@web.de>
In-Reply-To: <xmqqpml82vkd.fsf@gitster.g>

Am 23.04.22 um 07:54 schrieb Junio C Hamano:

Show 14 quoted lines

> Matthias Aßhauer <mha1993@live.de> writes:
>
>> Git 2.36.0 (or more precisely 244c27242f (diff.[ch]: have diff_free()
>> call clear_pathspec(opts.pathspec), 2022-02-16)) introduced some
>> change in behaviour that causes gitks highlight feature not to work
>> correctly anymore.
>
> I am reasonably sure that reverting that commit will be the right
> thing to do.  It is somewhat unfortunate that it would reintroduce
> resource leaks that having clear_pathspec() in a wrong place (i.e.
> diff_free()) was covering up.  We should instead need to find the
> place where a diff_opt struct goes out of scope (after being used
> for zero or more times, calling diff_free() after each iteration to
> release resources consumed per-iteration) and call clear_pathspec().

Right; a small memory leak is better than wrong output.

Finding those places is a bit complicated by diff_options often being embedded in struct rev_info, though.

René

PS: And I need to learn to download new posts before hitting Send
(or become faster); sorry for my near-duplicate reply.