Per Abrahamsen <abraham(a)dina.kvl.dk> writes:
If so, why does it work perfectly well on Emacs 20.3?
Because the custom font-model is based on the FSF font model and
ignores the specifier system completely. There is code to construct a
specifier from Custom specs, but not the other way around. In FSF
Emacs this is trivial since customize exactly matches the FSF font
model. Adapting customize to more powerful font models (whether for
XEmacs or for FSF 20.5+) without overly complicating the interface
will be a challenge. I would be cool to have M-x edit-faces back for
real.
XEmacs reimplementation[1] of the Customize face support. Things
are
initialized in the wrong sequence. MULE fonts are screwed. And now
this.
I think I know how to fix the MULE problems. (By doing a partial port
to specifiers. This fixes the throw-away-what-I-don't-understand
attitude the Customize takes on XEmacs). What would be cool is a
_true_ port to specifiers, but I am afraid this also needs some
hacking on the specifiers for the light-vs-dark isse.
Will there be time for any hacking sessions in Japan? I think it
would be useful if some of us
I sure hope so. I am bringing my new notebook and I'll put a late
model XEmacs on it, so we can just lock us up in our hotel room and
hack at night (which also gives us a good Solution to jet-lag).
Jan