Created attachment 15003 [details] [review]
fix patch

I think the problem is not that the fontmap is not loading it but that
gnome-print can't parse them. I tried adding the extension to the
fontmap but gnome-print couln't parse them, I opened bug 103856 for it.

Yeah, gnome-print has the handling of face issue for TrueType font.
but gnome-print won't handles even face 0 in .ttc without this patch.
at least it should be needed.

I don't think gnome-print can't even handle face 0 at this point,
which is why a larger patch is needed. Have you tested and embeded a
.ttc in a PS or PDF job?

Yeah. I've tested this with gedit. actual problem is
gnome_font_family_list() doesn't return the fontname in .ttc files,
because it's ignored since fcpattern_to_gp_font_entry() returns NULL
for unknown extension.

Okay. I made a patch to fix the handling of multiple faces.

Created attachment 15023 [details] [review]
fixes the handling of the multiple faces

*** Bug 103856 has been marked as a duplicate of this bug. ***

updated patch will also supports the TTF embed code. when I tested
that with Japanese TTC file, I got three warnings like "Too big table
in font". I think it means gnome-print could read 'glyf' table,
because basically TTF for CJK always gets this warning right now.
Unfortunately I have no TTC except for CJK. so someone needs to test
gnome-print, which applies first patch and updated patch anyway.

Created attachment 15032 [details] [review]
updated patch (contains second patch and supporting the TTF embed code)

Can this go into 2.2.x?

By adding a field to the _GPFontEntry struct, this seems to break the
API and ABI. Am I wrong?

Yes or no. because basically nobody can't access GPFontEntry itself
directly so that it's a closed structure. and changing the number of
arguments for gp_tt_split_file() should be no problem so that it's
also a closed API. so I think this changes is globally no API changes
and there are ABI breakage, but it shouldn't be affected. I think
soname should be bumped to be safe, though.

Akira,

do you have a changelog patch for the last patch attached to this report?

patch committed to cvs HEAD (post 2.5.0.1)

we will see haow many complaints we get about the slight interface changes