Hi Paul,
Paul Eggert <[email protected]> writes:
> Yes, this is an iffy area in the mktime spec. I installed the attached
> further coreutils patch to try to avoid the iffy area.
Thanks, the test passes with that patch.
I guess non-existent times due to daylight savings were probably an
after thought when writing the mktime spec, and were only considered
when people started testing edge cases.
>From skimming the austin-group mailing list I think I found the POSIX
correction you mentioned [1]. Which is already a part of POSIX.1-2024
[2]:
In order to allow applications to distinguish between a successful
return of (time_t)-1 and an [EOVERFLOW] error, mktime() is required
not to change tm_wday on error.
Just in case you think it is worth documenting somewhere.
Thanks,
Collin
[1] https://austingroupbugs.net/view.php?id=1614
[2] https://pubs.opengroup.org/onlinepubs/9799919799/functions/mktime.html
