hello!
I noticed that the glyph data is store in a group of surfaces, and the
glyph info structures are stored in a tree, Every time I look for glyph data,
first look for the tree, if can not find, create the structure and insert into
the tree.
This architecture is no problem with ASCII or IS08859-1 which don't
have
too much characters. but for chinese(GB18030) or japanese, GB18030 has more
then 20,000 characters, it's real a waste of memory. because with so much
characters, the freetype file is more the 4M and sometimes more than 10M,
so it's not easy to use freetype, in our former implementation we use 16*16
bitmap font in nor flash, which is more than 800K, so we don't need to load
into ram, but when we switch to directfb, we find we must use more 1M-1.5M ram
for this font, it's really a big chunk.
Anybody have some good idea to help me solve this problem?
swin
[EMAIL PROTECTED]
2006-03-22
_______________________________________________
directfb-users mailing list
[email protected]
http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-users
