Hussein Shafie wrote:
> The clean solution is to edit the declaration of Java's SansSerif font
> in order to map the Vrinda font to the bengali range of Unicode
> characters (0980?09FF). See
> http://java.sun.com/j2se/1.5.0/docs/guide/intl/fontconfig.html
OK, I did this and it still doesn't work. Specifically, here's what I
did to the fontconfig.properties file (the only fontconfig file that is
relevant on my PC, apart from the compiled fontconfig.bfc--but according
to the above page at Sun, fontconfig.properties takes priority over
fontconfig.bfc):
--Added a line at the top of the 'Component Font Mappings':
allfonts.bengali=Vrinda
--Modified the Default Search Sequence line to include Bengali
(unclear why this is necessary, but it is):
sequence.allfonts=alphabetic/default,bengali,dingbats,symbol
--Modified the UTF-8 Search Sequence line to include Bengali (ignore
spurious line break inserted by email):
sequence.allfonts.UTF-8.hi=alphabetic/
1252,devanagari,bengali,dingbats,symbol
--Added a line at the top of the 'Font File Names':
filename.Vrinda=VRINDA.TTF
(The vrinda.ttf file is indeed present in my Windows/fonts directory.)
I tested this on a Java applet that someone gave me; without the above
changes, the applet displays Bengali text as boxes. With the above
changes, the applet displays Bengali text correctly.
Unfortunately, this has no effect on XMLmind; the Bengali characters
continue to display as boxes.
Another method that was suggested to me is to put the Vrinda.ttf file in
Java's lib/fonts/fallback directory. I tried this as well, but it
likewise has no effect on XMLmind (it does work for the little Java
applet I was given). I am told that this method only works if the app
in question uses Java2D for displaying characters; presumably XMLmind
does not.
I also confirmed (using Microsoft's procman.exe) that when XMLmind
loads, Java reads both my modified fontconfig.properties file, and the
Vrinda.ttf file in my Windows/fonts directory.
Where do I go from here?
--
Mike Maxwell
maxwell at umiacs.umd.edu
