Hack: Enabling Unicode in XScreenSaver

When I tried to display Arabic text in some of the hacks, like noseguy, I failed miserably. Panicked, I posted to the Ubuntu forum asking for help. With a little help from my friends, and some code digging, I found that XScreenSaver hacks are not Unicode-ready at all.

So I started to convert noseguy to Unicode as a proof of concept. This little experiment succeeded, and you can find attached a modified version of noseguy.c (from XScreenSaver 5.03) that handles Unicode non-English chars. The code uses a Unicode-enabling library for Xlib originally found on Debian.org, mixed with some Arabic-shaping code that's part of Mohammed Sameer's Katoob.

To run this fix:

• Replace hacks/noseguy.c with the attached one
• Place attached unicode.c and unicode.h in utils/
• Edit your utils/Makefile to include unicode.c in the compilation (SRCS, OBJS, HDRS)
• Edit your hacks/Makefile to include unicode.o in the compilation of noseguy.

Next steps are:

• Get jwz's approval before applying this fix to all hacks
• Fix the xscreensaver-text Perl script to return UTF-8
• Fix Makefile.in to properly include unicode.c
• Fix the Arabic-shaping code
• Abstract the Arabic-handling code into a language-independent mechanism

{{wiki:20070834-ar-utf8_0.png}}

Stay tuned!