bug#2761: Tibetan font selection broken

[Kenichi Handa <handa@m17n.org>]

In article <m23ad97bag.fsf@igel.home>, Andreas Schwab <schwab@linux-m68k.org> writes:

> This change:

> 2009-03-05  Kenichi Handa  <handa@m17n.org>

> 	* font.c (font_select_entity): New function.
> 	(font_find_for_lface): Use font_select_entity to select a font.

> 	* fontset.c (fontset_find_font): If a font found without
> 	restricting of the characters C doen't support C, try to find a
> 	font with C restriction.

> has broken display of tibetan characters.  The characters are no longer
> found in the gnu-unicode font.  Instead a weird, one-pixel wide empty
> box is shown for each font (see the attached snapshot).  I have also
> attached the font logs from before and after that change.

Actually this change revealed two bugs.

> The problem appears to be that the font encoding muletibetan-0 is
> unknown to Emacs.  The only reference is in
> face-font-registry-alternatives, which is not used by
> find_font_encoding.

Yes.  That's one bug.  I added an entry for muletibetan-0 in
font-encoding-alist.

Another bug is the handling of unknown encoding in X font
backend, which caused that weird one-pixel wide empty box.
I fixed that too.

But, please note that these fixes results in selecting the
font:
  -tibmdxa-fixed-medium-r-normal--16-*-72-72-m-160-muletibetan-0
instead of the gnu-unicode font in your environment.

---
Kenichi Handa
handa@m17n.org