[XeTeX] Finding out if a font supports a particular Unicode character and using it

[Mike Maxwell]

R (Chandra) Chandrasekhar wrote:
> I stumbled upon this website from a Google search and it seems to list 
> font support for a block as a percentage:
> 
> http://www.fileformat.info/info/unicode/block/miscellaneous_symbols/fontsupport.htm 

I wonder how accurate it is, though.  For the Arabic block
http://www.fileformat.info/info/unicode/block/arabic/fontsupport.htm
it lists SIL's Scheherazade font as supporting 94% of the Unicode Arabic 
block.  I believe that in fact Scheherazade covers the whole thing, at 
least that's what I read here:
   http://scripts.sil.org/ArabicFonts
The page also doesn't list SIL's Lateef, which I think has similar 
coverage to Scheherazade.

But maybe the issue is what would count as 100%; it claims 250 code 
points in the Arabic block.  There are 256 code points in all (0600 
thorugh 06FF), but in Unicode 5.0, 21 of them are empty, meaning there 
are in fact only 235 code points that need coverage.  Given that figure, 
Scheherazade's coverage would go up to 100%.

Also, the "Unicode BMP Fallback Font" it lists as having 100% coverage 
is actually a debugging font that doesn't cover *any* characters in the 
sense of providing their normal glyphs.  Rather, it gives a square 
containing the Unicode code point.

I would certainly like to find an accurate web page with this kind of 
information (or failing that, for font providers to give the true code 
point coverage; few do).
-- 
    Mike Maxwell
    What good is a universe without somebody around to look at it?
    --Robert Dicke, Princeton physicist