Re: FYI: settime.c doesn't compile on OSF1 V5.1 due to stime

[Paul Eggert]

Jim Meyering <address@hidden> writes:

> [OSF1] does have options to switch between 32 and 64-bit modes,
> but I didn't think it'd be worthwhile, for code that might
> never be used.

Yes, but I'm more worried about date, ls, etc.  These all use time_t,
and if OSF1 can be cajoled into using 64-bit time_t then that would be
a win for them.

I'm not worried much about OSF1 (it's pretty much dead) but if Tru64
has the same problem it's worth addressing.

Casper H.S. Dik wrote "You can #define something and it will give you
64-bit time_t on Tru64, but it's still 32-bit by default."
<http://groups.google.com/group/comp.unix.solaris/msg/7a35c1c7d968accb>.

All we need to do is find out what that #define symbol is, and #define
it in config.h.  Presumably we can discover what the symbol is by
looking in a Tru64 <time.h> and looking for ifdefs.


> BTW, do you know what (if any) system ends up using that HAVE_STIME block?

System V hosts that lack BSD extensions.  There's not many left these
days.  POSIX has required gettimeofday since 2001, and settimeofday
usually comes along with it.

Google Groups reports that Dynix 4.2 (circa 1997) lacked gettimeofday
<http://groups.google.com/group/comp.security.ssh/msg/795d04ff06c8bdf3>
and that SCO 3.2.4.2 (circa 1996) also lacked it
<http://groups.google.com/group/gnu.emacs.bug/msg/ad28bc2f9f6cee2f>.