Re: Plans for 1.3/1.4

[Owen Taylor <otaylor redhat com>]

Damon Chaplin <damon@karuna.freeserve.co.uk> writes:

> Owen Taylor wrote:
> 
> >  1) Unicode support. This would involve moving all interfaces
> >     to Unicode, with support for western language, the
> >     CJK Asian languages, Hebrew, and possibly Arabic.
> >     The complex text scripts (in particular, those of
> >     South Asia) would not be included in this release.
> > 
> >  2) A completely rewritten text widget (this is closely
> >     tied to 1)
> 
> 
> Does the font selection widget need to be updated to handle i18n
> better?
> 
> Currently it displays the same preview text for all fonts -
> 'abcdefgh ABCDEFGH', which isn't useful for all fonts. Do we need
> a different preview string for all charsets? Or an option in the
> rc file?
> 
> And does it need to support fontsets somehow?
> 
> I'm willing to write the code, if people tell me what it should do.

Yes, it does require some updating. Actually there are two
different things going on here.

It would be nice if the font selector could handle the current
X-based i18n mechanisms for fontsets. I looked at this
some, and decided it was pretty difficult to do this
without making a UI that would be extremely intimidating
for the end user.

What I thought would be a good compromise would to support
only font sets of the form:

 "-*-fixed-medium-r-normal--16-*-*-*-*-*-*-*,*"

that is, a xlfd with a star for the character-set/encoding,
and an extra ,* at the end to catch up with miscellaneous
fonts that aren't caught by that.

and enable them with a single checkbutton, that would also
disable the bits of the UI for character set selection

But I suspect that for 1.4 we are going to need a more
powerful font abstraction on top of X's basic rendering
stuff. The details of that abstraction will determine what
the font selector looks like.

(The new font abstraction basically needs to do the equivalent 
of fontsets, but within GDK - since we will be rendering
Unicode => multiple X fonts ourselves, and probably should also
at the same time try to work aroudn some of the deficiencies
of the X font model - for instance, the inability of a program
to take a font, and get a different font which is a bolded
variant of the first).

Regards,
                                        Owen