Re: [Devel] Fonts with glyphs containing zero contours

[Keith Packard]

Around 17 o'clock on Feb 9, Antoine Leca wrote:

> I can really guess that this is (loosely) trick to gain space.
> I believe the missing glyphs are the Unicode codepoints that does not
> exist in Shift-JIS encoding. Since they do not exist in the original
> encoding, the missing glyphs are not used (except by special programs),
> and users do not complain; on the other way, the fact there is only
> one big range does speed both loading time and the time to search for
> a glyphId given a charId.

The question is whether there's anything that distinguishes these "bogus" 
empty glyphs from "normal" blanks -- Unicode has a lot of glyphs which are 
supposed to be blank, so I obviously want to make sure I don't report 
"legit" characters as missing.  From my investigation, these glyphs appear 
identical to other blanks.

This leaves me with little choice but to enumerate the encoding values 
which are supposed to be blank and assume the remaining characters are 
broken.

> There were a discussion back around August last year, with Brian Stell
> raising the first question

Yes, I was picking up on that discussion to see if anyone had additional 
ideas; my problem relates to use of FreeType with X applications like 
Mozilla.

Keith Packard        XFree86 Core Team        Compaq Cambridge Research Lab