Richard Zidlicky writes:
I have not seen [Xft rendering artifacts] in any other program
using Xft so I would think that it is something special about
xemacs. Will look at some other programs doing Xft rendering
directly.
Very few programs do their own redisplay any more; they delegate to a
redisplay like Gecko or webkit, or simply use trivial text areas. The
(X)Emacs redisplay is perhaps even more complex than Gecko, since it
not only handles redisplay on graphical displays, but also on TTYs.
In fact, if you look carefully at Gecko-based programs, you'll see
that they do the opposite from XEmacs: when clearing the background
for a rectangle, they tend to clip (and not redisplay) surrounding
content.
It's not clear to me that looking at how other programs do it is going
to be all that useful at this point. The problem is the complexity of
the XEmacs redisplay. We can't just replace it with something else,
because it's intertwined with everything in XEmacs. Instead, IMO it
would be best to document, refactor, and simplify the XEmacs
redisplay. Hoping that we don't lose performance....
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta