Re: Font names

[Juliusz Chroboczek]

Hidetoshi Yamano-uchi:

>> For example, the Microsoft "Shift_JIS" encoding (PlatformID=3,
>> EncodingID=2), it is not a wide-charcter encoding set, but a multibyte
>> encoding set.  It uses 8-bit(JIS X0201) and 16-bit(JIS X0208) in the
>> same name table.

Antoine Leca <address@hidden>:

AL> The bottom line is that both SJIS and EUC-xx (DBCS 8/16-bit, as described
AL> by Yamanoi-uchi-san) *and* uniform 16-bit wide characters appear to exist.
AL> Consequently, robust code may have to take care of that.

Note, however, that Microsoft encodings other than (3,0) (Symbol) and
(3,1) (Unicode) have been deprecated.  While there still are large
number of fonts that use them, new fonts for East-Asian languages will
use the Microsoft Unicode encoding (3,1).

Depending on you user base, it might or might not be suitable to
simply refuse supporting fonts with encodings other than Apple
Unicode, Apple Roman, and Microsoft Unicode.

(For example, the ``xfsft'' TrueType backend in XFree86 does a
half-serious job of supporting such ``legacy'' East-Asian fonts: it
allows them to be used, but doesn't even attempt to parse their name.
In two years, I've only received one complaint.  But then, most of the
Japanese user base use the ``X-TT'' backend anyhow.)

                                        J.