Re: broken tables-big.in pango/modules/basic/

[Owen Taylor <otaylor redhat com>]

suzukis@flavor1.ipc.hiroshima-u.ac.jp writes:

> After the remaking the tables-big.i following to Mr. Kusano's report,
> I could make testtext (of gtk+-1.3.1) display Japanese text in HELLO.utf8
> correctly (at present, anon-cvs server seems to be busy, and I could
> not check out new source.)
> 
> I tried to display HELLO.utf8 by UCS-2 encoded fonts, but it was not
> accepted as usable fonts (I've installed old "unifont" package, which
> provides fonts downloaded from http://czyborra.com/unifont/).
> If I specify the unifont only in ~/.pangox_aliases, I receive warning:
> 
> 	Gtk-ERROR **: Unable to load default font.
> 	aborting...
> 	zsh: trace trap  ./testtext

I have no problem using iso-10646 encoded fonts such as unifont.
My guess is that there probably is something a little wrong in your
.pangox-aliases.

Can you post what you are using?

(Another approach is to, instead of changing the alias for Sans,
have GTK+ use unifont instead. To do this, create a ~/.gtkrc-2.0
files with something like:

======
style "user-font"
{
  font_name="unifont 12"
}
widget_class "*" style "user-font"
======

which saves you from having to deal with XLFDS at all.
 
> Taking a glance on tables-big.i, char_mask_map[] has no mentions about
> ISO-10646 encodings, and the most parts of char_masks[] are filled by 0.
> This "0" means "ISO-10646 fonts can display this character"?
>      or, means "no fonts can be display this character?".

ISO-10646 is implicitely added for all characters - from basic.c:

      mask = char_mask_map[mask_index] | ENC_ISO_10646;

> However, I don't know if giving XLFD name with "ISO-10646" to the UCS-2
> encoded fonts.

the XLFD should end in iso10646-1.

(This parallels iso8859-1 - the registery is iso10646, while the 
encoding is 1.)

Regards,
                                        Owen