nanogui: Problems with chinese font implementation
Subject:
Problems with chinese font implementation
From:
"Greg Haerr" ####@####.####
Date:
16 Jun 2000 02:11:08 -0000
Message-Id: <004101bfd738$23217980$15320cd0@gregh>
Tanghao/Jauming, and others:
I'm very pleased with the recent support for chinese character
in Microwindows. I added most everything submitted and have
released it as 0.88pre9. However, there are a few concerns I
have that I would like to see as further patches. Also, since
I don't read chinese, I've not fully tested the implementation.
Following are some of my suggestions.
In engine/devfont.c, the KU mapping files are loaded and
unloaded every time a font is requested. It would probably
be better to load the tables the first time a font is used, and
then unloaded in the hzk_destroyfont function.
After Jauming's Big5 patch, there are still uses of the cfont_width
static global variables that have now moved to the HZKFONT
structure. These global statics should be removed.
There seems to be a bug in t1demo such that when using HZK
fonts, they stop displaying after a while. I was thinking it was
a file descriptor issue, but haven't debugged it.
I removed the HZK_FONT_SIZE config option, since it's
no longer needed. The xconfigure script should be revised
accordingly, as well.
It's possible we might want to leave in the GB2312 decoding
support that Guo added (HAVE_GB2312_SUPPORT) rather
than doing the decoding in the hzk_* functions. In other words,
we should separate the GR_ASCII decoding functions from
the font rendering, since they are different. What do you think?
Regards,
Greg
