[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: X fonts selection weirdness

From: Kenichi Handa
Subject: Re: X fonts selection weirdness
Date: Wed, 25 Jun 2008 20:31:56 +0900

In article <address@hidden>, Miles Bader <address@hidden> writes:

> Kenichi Handa <address@hidden> writes:
> > The above should work, but I'm now thinking about
> > introducing non-default-font-script-list:
> > Documentation:
> > List of scripts not to display by the default font.
> > Usually, the font specified for ASCII characters is also used
> > to display a non-ASCII character if the font supports that character.
> > But, if a character belongs to one of these scripts, a font for it is
> > determined by looking up the current fontset and the default fontset.

> Why would you do that though?

> If the user choose a default font capable of displaying one of those
> charsets, it seems like he would _want_ the default font to be used!

Not for non-Latin scripts, especially CJK and those require
CTL (Complex Text Layout).  Usually the default font, even
if it claims that it has the specific character, the design
is not good or doesn't contain proper OTF tables.

Jason Rumney <address@hidden> writes:

> Conversely, if the user has explicitly specified fonts in a fontset,
> then they expect those fonts to be used, not just for specific scripts.

Yes.  As written in the docstring,
non-default-font-script-list doesn't prevent such a case.

Currently, if a user specifies a font (not fontset), a new
fontset is created, and in it, all Latin characters are
mapped to the specified font, and the iso10646-1 version of
that font is set in the fallback font of the fontset.

So, if a user specifies XXX-iso8859-1, XXX-iso10646-1 is
used for all characters supported by that font.  This works
well for most characters except for those in the above
scripts.  Current idea of mine is to, instead of setting
XXX-iso10646-1 to the fallback font, map all characters
(except for latin-1 and those belongs to the above scripts)
to XXX-iso10646-1.

Kenichi Handa

reply via email to

[Prev in Thread] Current Thread [Next in Thread]