> 1. I have to set -f 3 to get KAIU.TTF and MINGLIU.TTC with big5-0
[...]
> 2. The default ``fuzz'' value can find the encodings correctly for all
> four arphic ttf.
[...]
> 3. For baekmuk ttf, it missed this font for 'ksc5601.1987-0' encoding with
> default ``fuzz'' value.
> hline.ttf -ibm-Baekmuk Headline-medium-r-normal--0-0-0-0-p-0-iso10646-1
The fuzz value is only used with East-Asian encodings (more precisely,
with linear encodings with more than 256 codepoint space and matrix
encodings, but not for iso10646-1). I have no objection to setting
its default value to whatever it is that people who actually use such
encodings think is best.
If fuzz is set too high, you may end up with e.g. JIS X 208 fonts
being exported as JIS X 212.
Altogether, I do not think that fuzz is the right solution. For
eight-bit fonts, mkfontscale uses some very simple codeset-dependent
heuristics (search for koi8) which work fine. I would have hoped that
people with intimate knowledge of East-Asian encodings would have come
up with analogous ideas (this doesn't seem too difficult, especially
for Big-5 which has a well-defined non-essential region).
> Very good but the spacing value still is 'p'.
That's not my fault: the font is incorrect (bad value in post table).
Does it have a post table in the first place?
And I don't want to use the Panose data, which is seldom accurate.
Juliusz
_______________________________________________
Fonts mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/fonts
