From: Raphael Finkel (raphael@cs.uky.edu)
Date: Mon Sep 09 2002 - 08:49:52 EDT
A few weeks ago I got a copy of Abiword source and compiled it under Linux.
With some fiddling, I got a Unicode font (MS Arial) to show me Hebrew/Yiddish
characters acceptably. Then last week I used CVS to update the copy, and now
everything I type in or open in Hebrew/Yiddish is wrong. My setup:
LANG=he_IL.utf-8; export LANG
I have a utf-8 subdirectory of AbiSuite/fonts, in which:
* I have a copy of arial.ttf that I call arigl.ttf just to
disambiguate.
* fonts.scale has a line:
arigl.ttf -monotype-Arigl-medium-r-normal--0-0-0-0-p-0-iso10646-1
* fonts.dir is properly built from fonts.scale (looks identical)
* arigl.t42 and arigl.u2g have been built; both appear to
include codes for Hebrew/Yiddish. (Actually, these are symlinks to
arial.t42 and arial.u2g.)
When I start AbiWord, it offers me arigl as one of the fonts it knows
about. So AbiWord is finding the fonts/utf-8 subdirectory. If my fonts.scale
only includes the iso10646-1 line, I get no characters at all for
Hebrew/Yiddish in the Arigl font. If I add an iso8859-1 line, AbiWord
generates what looks like Latin-1 128-255 characters on the screen instead of
Hebrew/Yiddish. If I change that to iso8859-8, then I get some Hebrew/Yiddish
characters, but they are all wrong. It looks like the multi-byte utf8 codes
are not being fed to the font displayer properly. They are being treated as
multiple independent bytes.
Raphael Finkel
This archive was generated by hypermail 2.1.4 : Mon Sep 09 2002 - 08:52:57 EDT