GNU bug report logs - #67430
29.1; <Multi_key> is undefined

Previous Next

Full log

View this message in rfc822 format

From: Francesco Potortì <pot <at> gnu.org>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: Po Lu <luangruo <at> yahoo.com>, 67430 <at> debbugs.gnu.org
Subject: bug#67430: 29.1; <Multi_key> is undefined
Date: Sat, 25 Nov 2023 14:18:54 +0100

>> From: Po Lu <luangruo <at> yahoo.com>
>> Cc: pot <at> gnu.org,  67430 <at> debbugs.gnu.org
>> Date: Sat, 25 Nov 2023 19:10:15 +0800
>> 
>> Eli Zaretskii <eliz <at> gnu.org> writes:
>> 
>> > What other roles does this key play?  And how frequent is each role?
>> 
>> I don't know.  There is generally no other use for it under X, except
>> perhaps as a modifier key, which if true Emacs won't register key
>> presses at all.
>> 
>> > Also, if we bind this key by default as Francesco suggests, what
>> > adverse results could this cause on the systems where this binding is
>> > wrong?
>> 
>> Nothing beyond the obvious, to wit: Emacs will react to pressing the
>> Multi_key as though it were bound to iso-transl-ctl-x-8-map.
>
>Then maybe we should behave by default as Francesco suggested?  users
>which don't like the results could always unbind/rebind the key.
>WDYT?

In mormal usage, Emacs never even sees <Multi_key>.  I suppose that Emacs, like usual X applications, relies on X to interpret it as a modifier key, and X uses its composition rules.  But Xpra apparently is another piece of cake (at least the version I am using) and it passes the key as-is.  I think that this is unusual behaviour.

If Emacs happens to see <Multi_key>, maybe in principle it should ask X to interpret the subsequent characters using X's composition rules.  I don't know if that's possible in a clean way, I suppose that would not be obvious to do.  Lacking such functionality, I think the only reasonable thing to do is to interpretet is as C-x 8, which has composition rules mostly compatible with those of X.  So in most cases this will be transparent to the user.  I don't see how throwing an error instead can be better.  And yes, in principle users can rebind it.

An alternative would be to bind <Multi_key> to a command which is disabled by default (like for example narrow-to-page), and when enabled does the same as C-x 8.  But I think that binding it by default is simpler and has so rare and technical drawbacks that people caring it about are aware of the issue and can solve it.

Previous Next

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