Xiyue Deng <manp...@gmail.com> writes:

> David Bremner <da...@tethera.net> writes:
>
>> Xiyue Deng <manp...@gmail.com> writes:
>>
>>>
>>> Friendly ping.  Should we merge and prepare for release (AIUI the
>>> current proposed fix has reached consensus)?
>>>
>>
>> In order to do request a freeze exception, we need version that has the
>> minimal set of changes to fix the RC bug; it seems the git regex bug is
>> not RC? Personally I'd make another branch off of 2.1.9 and work there.
>>
>
> Given that 2.1.10 has not happened yet, maybe we can revert the regex
> fix for 2.1.9 and wait until after Trixie to revert the revert?  Let me
> know.

That would work also, but in the event further updates are needed for
trixie, it makes sense to start a branch debian/trixie now (even though
we will upload to unstable).

>> I wasn't able to satisfactorially test this fix so far in trixie; I
>> guess I should setup a bookworm container/VM. I tried manually replacing
>> the directories in /usr/share/emacs/site-lisp/elpa/doom-themes with
>> symlinks, but the various ways I tried simulating the upgrade always
>> ended up removing the entire /usr/share/emacs/site-lisp/elpa/doom-themes
>> (via removing the package). That can be verified with "ls -id" before
>> and after.  Can you give me a recipe?
>>
>> d

> ,----
> | apt install -y --no-install-recommends elpa-doom-themes emacs-nox
> | sed -i 's/bookworm/trixie/' /etc/apt/sources.list.d/debian.sources
> | apt update
> | apt upgrade -y --without-new-pkgs
> `----

As you note in a followup message, this doesn't reproduce the bug in
trixie anymore. I think that in order to get the release team to allow a
freeze exception, we'll need to have a way to trigger the problem in
(upgrading to) current trixie. It isn't clear to me what guarantees we
have about ordering, but I guess to manifest this bug we would need the
postinst of elpa-foo to run before the preinst of emacsen-common.

None of this is to say the bug shouldn't be fixed; I'm only checking
whether it needs to be fixed for trixie, or at least that the risk of
triggering the bug outweighs the risk of a change during the freeze.

Reply via email to