[Fontconfig] Font encoding problem

Eronen, Jussi (GE Healthcare) jussi.eronen at med.ge.com
Tue May 10 18:22:32 EST 2005

I have a small X application, that displays 16-bit Unicode strings using Xlib function

Fonts used in application are bdf fonts having only a subset of character codes from Unicode,
some greek chars, arrows and a few special symbols in PUA area. Encoding is ISO10646-1.

For some reason, application displays only some of the available glyphs. 
1. With original font I only get PUA area symbols.
2. After creating an iso8859-1 front from original with ucs2any, I get all characters from Latin1 set + PUA area, but only 1/3 of other glyphs. For example, code 0x3a6 (Phi) displays as two
characters '&' and '5'. Another example is that code 0x2665 (heart) displays as 'quotedbl' (0x22) and '>' (0x3e), but code 0x2666 (diamond) displays OK.

I have tried to convert bdf toc pcf.gz, but that does not help.

Some posting here suggests that encapsulating bitmaps to truetype font might help. Any suggestions how to do this (fontforge and
fonttosfnt greated unusable ttf files)?

Is some layer here doing clever encoding tricks? And if so, could it be disabled?

My runtime systems is:
FC3 w/ 2.6.11
Xorg 6.8.2
fontconfig 2.3.2
freetype 2.1.9
locale en_US.UTF-8


More information about the Fontconfig mailing list