This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
RE: 64 bit time_t, when?
- To: "'Ellis, John E (Edward)'" <edward dot ellis at pnl dot gov>, 'Gianni Mariani' <gianni_ at uluru dot mariani dot ws>, aoliva at redhat dot com
- Subject: RE: 64 bit time_t, when?
- From: Donn Terry <donnte at microsoft dot com>
- Date: Fri, 1 Sep 2000 16:18:09 -0700
- Cc: gcc at gcc dot gnu dot org
This is actually a rather messier problem than just time_t being
extended to 64 bits (because the current relatively crude accuracy
of time_t (1 second) is a source of a long litany of problems).
In another life I'm involved in the latest revision to the POSIX
standard, and one of the topics that is being discussed (inactively,
at the moment, but it will ressurect itself when mental bandwidth
is available) is what to do about the 2038 problem before it becomes
another Y2K last minute panic.
I'm not proposing that this be discussed further here, because it IS
an OS problem, not a language problem. However, if you're interested
start by looking at http://www.opengroup.org/austin/. There's a
time_t study group within that activity.
Donn
P.S. It's probably better to respond to this by private mail if you need
more info (unless there's a lot of interest, anyway :-) ).
> -----Original Message-----
> From: Ellis, John E (Edward) [mailto:edward.ellis@pnl.gov]
> Sent: Thursday, August 24, 2000 1:27 PM
> To: 'Gianni Mariani'; aoliva@redhat.com
> Cc: gcc@gcc.gnu.org
> Subject: RE: 64 bit time_t, when?
>
>
> Are you suggesting that we create a time64.h that duplicates
> the whole set of
> ANSI/POSIX time functions? Why is that better? I have
> gotten a message that
> told me this already works on systems with 64-bit processors.
> Is it so hard to
> do
>
>
> J. Edward Ellis
>
> Battelle, Pacific Northwest National Laboratory
> (509) 375-3627 voice
> (509) 375-2379 FAX
> mailto:edward.ellis@pnl.gov
> http://electronics.pnl.gov:2080/ellis.html
>
> -----Original Message-----
> From: Gianni Mariani [mailto:gianni_@uluru.mariani.ws]
> Sent: Thursday, August 24, 2000 10:42 AM
> To: aoliva@redhat.com; Ellis, John E (Edward)
> Cc: gcc@gcc.gnu.org
> Subject: Re: 64 bit time_t, when?
>
>
> You could probably start by proposing system functions for
> extended time_t functions ;like for exmaple:
>
> time64_t time64(time64_t *)
>
> That way libc developers may actually get motivated.
>