Re: [bug-gnulib] Re: gettimeofday() for Win32

[Paul Eggert]

Bruno Haible <address@hidden> writes:

>> +  AC_CACHE_CHECK([for gettimeofday whose signature conforms to POSIX],
>
> What's the point of checking the signature? The only signatures of
> gettimeofday() I've seen are:
>
>     int gettimeofday (struct timeval * tp, ...);
>     int gettimeofday (struct timeval * tp, struct timezone *tzp);
>     int gettimeofday (struct timeval * tp, void *tzp);
>
> and they are all OK with the usual way to call it.

That rings an old bell.  Some old SVR4 systems put a one-argument
gettimeofday prototype into their <sys/time.h>.  Also, Solaris 2.4 has
a one-argument prototype for C++ only.  See my old Usenet post on the
subject here:

http://groups.google.com/group/gnu.gcc.bug/msg/1079ba9ae74dd606