Re: emacs-29 1ef8b90ae06: Simplify imenu setup for {cmake, dockerfile}-ts-modes

[Stefan Kangas]

Po Lu <luangruo@yahoo.com> writes:

> Yes, please.  Furthermore, if such bugfix cannot be described as
> anything but a simplification (with a correspondingly large diff), there
> should be another approach safer for the release branch.

We must not get overly rigid about that.  It takes extra work, and it
will not always be clear that it's worth the effort.  Often it is, but
sometimes it isn't.  We tend to decide on a case-by-case basis.

Furthermore, please keep in mind that Yuan is both the principal author
and maintainer of our treesitter support.  He's closer to the code than
anyone else, and is thus in the best position to make such judgement
calls.  I see no reason to second-guess him here, myself.

If there are any specific technical arguments for why this particular
change must not be installed on emacs-29, then let's hear them.  Neither
generalizations nor administrative arguments will cut it, I think.

> Our basic criterion for installing patches on emacs-29 is not whether a
> change contains a bugfix, but whether it runs a risk of destablizing the
> branch.  It is more important that a bugfix be the only component of a
> change, since that reduces the probability of unanticipated disruption.

(Speaking here as the co-maintainer of GNU Emacs, responsible for
leading Emacs development together with Eli.)

The above does not necessarily reflect the official line of the project.

It is true that we have preferred to avoid bug fixes that might risk
destabilizing the branch, but this has usually been applied in an ad-hoc
and not in a rigid fashion.  I'm not myself prepared to commit to much
more rigid guidelines at this point.

On a personal note, if we believe that some changed guidelines are
necessary, I'd probably be more inclined to make things more relaxed
rather than more strict.  But please note that I'm _not_ proposing to
open a discussion about that at the present time.