Re: fc-name-parse called in non-Xft XEmacs
Stephen J. Turnbull writes:
The display turds are way low on my list of things to do.
That's the kind of thing that somebody who's willing to face
a
couple crashes before they get the code right can improve by just
poking around in redisplay-x.c.
N.B. By "that" I mean the display turds.
AFAICT, the problem is that in the nature of things the extent of
redisplay of text displayed in an antialiased font is necessarily a
bit fuzzy. Xft seems to report optimistically (ie, the smallest
extent that could be needed).
I think that blanking an extra pixel around the extent reported by Xft
is what we need to do, but that can effectively clip the surrounding
text (and I have in fact observed such artifacts in Mozilla and
OpenOffice, so apparently that is what they do). The amount of code
in XEmacs that deals with it is pretty small and well-marked by #ifdef
XFT or similar. This would be a good project for somebody who wants a
useful task that is limited in scope.
I don't understand why the cursor is so likely to leave turds. There
may be an off-by-one error in the calculation of its size. Why that
doesn't matter for legacy X fonts I don't know, though.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta