XEmacs Carbon icons and GPLv3.

[Stephen J. Turnbull]

Aidan Kehoe writes:

 >  > Personally, I think that if it's good enough for people using it now,
 >  > we should leave it, but that further effort should be directed
 >  > elsewhere (except for fixing reported bugs as time and expertise
 >  > allow).
 > 
 > Leave it, as in continue with the repository available, or leave it, as in
 > continue with the merges as I’ve been doing?

I had in mind the former.  But that's up to you.

 >  > I'd like to see a Cocoa port, but apparently (cf emacs-devel traffic)
 >  > that would be as much work to do well as Windows was.  The display and
 >  > event models of "modern" GUI systems seem designed to prevent Emacs
 >  > from being Emacs....
 > 
 > Emacs was developed on a TTY, where the editor redisplay has comparatively
 > little control over anything. I really don’t understand why the various
 > non-X11 GUI implementations don’t abdicate more of redisplay in favour of
 > the OS; Choi’s did, and that was a fine start.

I don't understand what you're talking about.  AIUI (I read *all* that
code closely, but superficially, as is required for copyright
analysis), Choi's redisplay was a quite faithful port of the Qt code
(which is why the license problem exists), and very similar to the
whole family of X11 redisplays.  The event-handling code (eg, keyboard
and mouse) was simpler but also had bugs (eg, situations where C-g was
ignored that wouldn't happen with X11).  I'm not sure how good GTK and
Qt are on that front.

Agreed, it would be nice if Pango could replace home-grown Mule code
in redisplay, but my experience with Havoc Pennington's code has been
way less than pleasant.  We'll see how that goes as Jeff's work
progresses.