bug in EUC-CN charmap

hashao hashao@china.com
Mon Sep 4 08:08:00 GMT 2000


Hello Bruno,

Monday, September 04, 2000, 8:51:42 PM, you wrote:

BH> The GB2312 charmap contains identity mappings for 0x80..0xA0, 0xAA..0xAF,
BH> 0xF8..FF, but iconv doesn't implement them in either direction and other
BH> systems don't have these mappings either.

BH> Here is a patch to remove them from the charmap.

Does this mean that a single byte in the range 0x80 - 0xAF is not
valid under GB2312 locale? Because I just had this conversation about
situation like this. Mutt cannot display Chinese chars without a work
around in mutt because it use isprint() which is buggie under glibc
2.1.x.

We has this 'fix' to add chars in the range of 0x80 - 0xFF into the
print class of locale definition. So if 0x80 - 0xFF are removed from
charmap, will all the reference to those chars in locale definition
unnecessary?

The similar situation exists in zh_TW.Big5 locale too.

BH> 2000-09-03  Bruno Haible  <haible@clisp.cons.org>

BH>         * charmaps/GB2312: Remove 0x80..0xA0, 0xAA..0xAF, 0xF8..FF.

BH> *** glibc-20000831/localedata/charmaps/GB2312.bak       Wed Aug 30 23:44:19 2000
BH> --- glibc-20000831/localedata/charmaps/GB2312   Sun Sep  3 04:26:34 2000
BH> ***************
BH> *** 138,190 ****
BH>   <U007D>     /x7d         RIGHT CURLY BRACKET
BH>   <U007E>     /x7e         TILDE
BH>   <U007F>     /x7f         DELETE (DEL)
BH> - <U0080>     /x80         PADDING CHARACTER (PAD)
BH> - <U0081>     /x81         HIGH OCTET PRESET (HOP)



-- 
Best regards,
 hashao                            mailto:hashao@telebot.com.no.spam.no




More information about the Libc-alpha mailing list