Re: CORE, newest CVS
>>>> "Y" == Yoshiki Hayashi
<yoshiki(a)xemacs.org> writes:
Y> Norbert Koch <nk(a)LF.net> writes:
> After an update I encounter the following dump when trying to
open
> certain mails in gnus (it doesn't happen in .36):
> #0 0x185063d0 in kill () from /usr/lib/libc.so.3
> #1 0x80bb853 in fatal_error_signal (sig=6) at emacs.c:531
> #2 0xbfbfdfdc in ?? ()
> #3 0x80be7fa in assert_failed (file=0x82ad0e8 "buffer.h", line=1712,
> expr=0x82ad18f "IN_TRT_TABLE_DOMAIN (c)") at emacs.c:3192
Y> Mea culpa. I naively believed existing code did the right
Y> thing. But it did not. This means even with non MULE
Y> latin-1 environment, case-fold-search hasn't been working
Y> for bytes > 127. Here's the fix.
The ChangeLog entry and the fix use different types.
Y> 2000-11-16 Yoshiki Hayashi <yoshiki(a)xemacs.org>
Y> * buffer.h (IN_TRT_TABLE_DOMAIN): Cast to unsigned int.
Y> -#define IN_TRT_TABLE_DOMAIN(c) (((EMACS_UINT) (c)) <= 255)
Y> +#define IN_TRT_TABLE_DOMAIN(c) (((unsigned char) (c)) <= 255)
I don't understand TRT_TABLEs, but this must be wrong.
(((unsigned char) (c)) <= 255)
is always true.