Thomas Hood wrote: > $ for F in 1 2 3 ; do :>$F ; done > $ ls -l --full-time > total 0 > -rw-r--r-- 1 jdthood jdthood 0 2011-05-24 17:32:19.022224195 +0200 1 > -rw-r--r-- 1 jdthood jdthood 0 2011-05-24 17:32:19.022224195 +0200 2 > -rw-r--r-- 1 jdthood jdthood 0 2011-05-24 17:32:19.022224195 +0200 3 > > > Hmm, it seems that ls -l --full-time does not actually illustrate > nanosecond mtime resolution. :)
I guess it rather depends on the resolution of the kernel time-source. > > But still, the problem of multiple-updates-within-a-second needs > to be solved otherwise than by making the resolvconf update > sleep one second, every time. > > It would already be better if the script slept for one second > when (and only when) the current time is the same as the > mtime as /var/run/dnsmasq/resolv.conf. It should be possible > to test for this condition without too much difficulty. Shall > I submit a patch for this? That sounds good, please do. Cheers, Simon. -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
