GNU bug report logs - #65451
30.0.50; `after-change-functions' are not triggered in the same order the changes are made

Previous Next

Full log

Message #80 received at 65451 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Monnier <monnier <at> iro.umontreal.ca>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: acm <at> muc.de, yantar92 <at> posteo.net, 65451 <at> debbugs.gnu.org
Subject: Re: bug#65451: 30.0.50; `after-change-functions' are not triggered
 in the same order the changes are made
Date: Sat, 30 Mar 2024 11:38:24 -0400

> I still stand by my opinion: Org is relying on something it cannot
> rely upon, not as long as a function that changes a buffer can be
> called from another function which changes the same buffer.

`*-change-functions` should not modify the buffer (we could try and
enforce this, tho in my experience those that do will get punished
pretty quickly already), so the only cases I can think of where "a
function that changes a buffer can be called from another function which
changes the same buffer" is when both of those functions are in our
C code and we should have enough control to fix those cases.

> I don't see how we can avoid breaking code which relies on such
> assumptions, not in general anyway.

All sophisticated-enough users of `*-change-functions` rely on this
(and come with sanity checks to detect the problem and fallback on an
expensive recovery when needed, because indeed we tend to break our
promises 🙁).

So we really should try and fix those.
Alan did convince me that we should treat them as bugs and that we
should try and fix them.


        Stefan

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.