[Mlterm-dev-en] a patch for GBK encoding handling

[Wang W. <wa...@ma...>]

Hi, there:

There is a bug where handling GBK characters in mkf/lib/mkf_euccn_parser.c
My patch is based on mlterm-2.6.1.cvs20020927(debian sid), and I believe
that the
bug also exist in the newest 2.6.2 version(not checked yet). The patch
file is in
the attachment.

Bug details: I found this bug when I am using it on the BBS in our
campus. When a half GBK character following an ANSI escape sequence(such
as 0xBF ESC [ m) appears on the terminal, mlterm is confused. It takes
"0xBF ESC" as one GBK char and left others ordinary
ascii characters, and the screen layout is wrong, a blank line jump out.
I've look it up
in an official GBK standard, it saying that such "0xBF ESC" cannot be a
GBK char. In fact,
THE SECOND CHARACTER OF A GBK CHAR IS ALWAYS GREATER OR EQUAL THAN 0x40.
So, the parser
should swallow only ONE character in this case.

I have test my code for some time and it works well. :) I like mlterm
very much and
be interested improving it. I have made a patch for saving/restoring
cursor ANSI sequence
two month ago. Wish this one would help.

Wang WenRui