Hi Jonathan,

> pax is incorrectly using utimes() to attempt to set the timestamps of symbolic
> links, rather than utimensat() with AT_SYMLINK_NOFOLLOW.

thanks for the preliminary analysis. I’ve noticed an error (in the
native BSD environment) which I believe to be related, if not the
same, but haven’t gotten around to debugging yet.

I’ll have to find a way to do this on systems without the *at()
syscalls, however. Since you already looked into this, perhaps
you have an idea for me ;)

“It is inappropriate to require that a time represented as
 seconds since the Epoch precisely represent the number of
 seconds between the referenced time and the Epoch.”
        -- IEEE Std 1003.1b-1993 (POSIX) Section B.2.2.2

Reply via email to