cvs commit: src/usr.bin/mklocale yacc.y src/lib/libc/locale Makefile.inc gb2312.c setrunelocale.c

[Tim Robbins]

On Thu, Nov 06, 2003 at 05:59:25PM +0800, David Xu wrote:

> Tim Robbins wrote:
> 
> >On Wed, Nov 05, 2003 at 02:52:51PM -0800, David Xu wrote:
> >
> > 
> >
> >>davidxu     2003/11/05 14:52:51 PST
> >>
> >> FreeBSD src repository
> >>
> >> Modified files:
> >>   usr.bin/mklocale     yacc.y 
> >>   lib/libc/locale      Makefile.inc setrunelocale.c 
> >> Added files:
> >>   lib/libc/locale      gb2312.c 
> >> Log:
> >> Add gb2312 encoding.
> >>   
> >>
> >
> >This looks like EUC. Is there any chance you could use the EUC encoding
> >code instead of adding another one? Also, out of curiosity, how is it
> >different from zh_CN.eucCN?
> >
> >
> >Tim
> >
> > 
> >
> I hear euc does not check error sequence which is not encoded by GB2312 
> , right ?
> Despite that,  most program ported from Linux only support GB2312 or GBK 
> , but not
> euc, I always has trouble with euc when running some Linux programs.

Right. Apple solved this in Darwin by calling isrune() on the wide character
before returning it from mbrtowc(). I'm considering doing the same thing
in FreeBSD, but I'm not sure whether it would break any applications
that relied on the old behaviour. If I did that, would it be possible
to use the EUC encoding code?


Tim