Paul Eggert <[EMAIL PROTECTED]> wrote:
> That sounds too strong. Sometimes 'tar' should ignore the
> nanoseconds, and sometimes not. That is why CVS tar has both a
> timespec_cmp function and a tar_timespec_cmp function, for example.
I have no good idea how to deal with nanoseconds correctly since I introduced
them 10 years ago in star.
Nsecs have been activatd 5 years ago with POSIX.1-2001 support
While it is no problem to set time stamps from the nsec fields in the archive,
it is unclear how to deal with them when doing time stamp compares.
The main problem seems to be missing feature in POSIX as you need to know the
time stam granularity of the local filesystem if you like to implement
comparisons correctly.
Note that the low-carb (aehm... FAT) filesytem only implements a granularity
of 2 seconds, so it is not even correct to believe in a one second granularity.
Jörg
--
EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin
[EMAIL PROTECTED] (uni)
[EMAIL PROTECTED] (work) Blog: http://schily.blogspot.com/
URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily
