-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Bruno Haible on 6/28/2006 11:01 AM:
> Eric Blake wrote:
>> By the way, is it safe to assume wchar_t exists
>
> No it isn't. Good point. Either the module description should have
>
...
>
> Yes, but that's not a reason for removing the de
Eric Blake wrote:
> By the way, is it safe to assume wchar_t exists
No it isn't. Good point. Either the module description should have
#if HAVE_WCHAR_T
# include "wcwidth.h"
#endif
or wcwidth.h should protect against this case itself. The latter is
obviously better.
> Also, on mingw, wher
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Bruno Haible on 6/27/2006 12:15 PM:
>> I note that mbswidth.c takes precautions to ensure wcwidth exists;
>> should we break that out into a wcwidth module and make both mbchar and
>> mbswidth depend on it?
>
> This would make sense, yes.
Eric Blake wrote:
> Using cygwin to cross-compile to mingw, I got:
>
> $ CFLAGS='-Wall -O2' CC='gcc -mno-cygwin' ./gnulib-tool --test mbchar
> ...
> gcc -mno-cygwin -DHAVE_CONFIG_H -I. -I../../lib -I.. -Wall -O2 -
> c ../../lib/mbchar.c
> In file included from ../../lib/mbchar.c:24:
> ../../li
