[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Sat, 03 Jun 2006 16:22:56 -0700
Gnus/5.110004 (No Gnus v0.4) Emacs/21.4 (gnu/linux)
Paul Eggert <address@hidden> writes:
> Ben Pfaff <address@hidden> writes:
>> I really meant ssize_max.h, analogous to size_max.h. Or they
>> could be combined into one header file.
> I think size_max.h is there only because SIZE_MAX is in different
> headers on different systems. Is that also true for SSIZE_MAX?
> I don't offhand recall SSIZE_MAX being defined anywhere other than its
> standard location <limits.h>. If so, then we wouldn't need a
> ssize_max.h header, and we wouldn't need to overload size_max.h; we
> could just tell people to include <limits.h>.
The ssize_t module is there, I think, because ssize_t isn't
necessarily defined at all on all systems. ssize_t.m4 defines it
as "int" on systems that don't have it:
if test $gt_cv_ssize_t = no; then
[Define as a signed type of the same size as size_t.])
If ssize_t doesn't exist then I wouldn't expect SSIZE_MAX to
exist either. The current ssize_t module doesn't define
SSIZE_MAX, but, in my opinion, it should. That's really what I'm
trying to suggest, although I suppose suggesting a mechanism
instead of describing the actual goal is confusing.
"Then, I came to my senses, and slunk away, hoping no one overheard my
--Steve McAndrewSmith in the Monastery