I've finished the last piece of code required to obtain feature
parity with the xft/fontconfig backend in Mozilla. This means
that you can use pango to render web pages, including pages that
contain MathML. It's interesting to see the backend rendering all
of the languages listed on the bbc's languages
page. Tamil? Yep.
Urdu? Got it. Arabic?
Of course.
From here I have to work on the pango to printing code, so you can
use pango's font selection and glyph choices to generate
postscript from Mozilla's backend. This shouldn't be too hard
since all of the important pieces are already done. jshin and
some of the Sun folks did the work to get the fonts into the
postscript output as well as using fontconfig/xft for the font and
glyph selection. I just have to build some easy-to-use callbacks
from the pango font rendering layer to the postscript bits, and
then push those glyphs down into the freetype code that's already
in the code.
The biggest problem is that there's no way to currently exclude
bitmap fonts from the font list that fontconfig generates, without
postprocessing the list. Owen says he's going to work on that
problem and will give me some kind of easy-to-use solution. He
needs it for gnome-print anyway, since it suffers from the same
problem. (Right now if you happen to print something that pulls
in bitmap fonts, it won't draw the glyphs. This will be a much
more common occurance with Mozilla because font specifications in
web pages often include proprietary bitmap family names, and
people do tend to include those in their path.) If I'm super
lucky I'll never have to open the postscript language
reference that's sitting on the bookshelf above my head.
In the next day or so this code is going to make its way into
rawhide. So expect some bugs.