Re: BMP chars lacking fonts

[Owen Taylor <otaylor redhat com>]

On Fri, 2003-08-22 at 22:13, James H. Cloos Jr. wrote:
> >>>>> "Owen" == Owen Taylor <otaylor redhat com> writes:
> 
> >> There are now only 244 BMP chars listed in UnicodeData.txt that
> >> pango attempts to paint and for which I have no free / libre font.
> 
> Owen> But how many characters are in the BMP that you have fonts that
> Owen> cover, but not to render correctly?
> 
> That is a good point, and one I cannot answer w/o learning several
> new scripts. ;-/
> 
> It is I think cool, though, that coverage is close to complete.

The "checklist of glyphs" approach bothers me a bit because it
leads to fonts like FreeSans which have the "Unicode table" form
of the  characters without the necessary shaped forms of the glyphs
and without any mechanism (OpenType, Silf, whatever) 
for proper shaping of the glyphs.

We're going to need fairly complicated mechanisms in fontconfig
and Pango to weed out such fonts and prevent them from being
selected in place of the right fonts.

> Owen> Are the VARIATION SElECTOR characters really printable
> Owen> characters?
> 
> Only as combining characters following (if you want to follow the
> unicode std strictly) a rather small set of math characters as well
> as several mongolian characters.
> 
> Most of the characters that are "illegal" show up as <F00B>, but
> the variation selectors (after a non-variable character or after
> nothing) show up as a hexed box.

Not sure what you mean by <FOOB>.

Regards,
						Owen