Re: [RFC/PATCH] Nonotify - A simplistic way to determine directory content changes

[nf <nf2 scheinwelt at>]

On Mon, 2004-06-07 at 18:52, Manuel Amador (Rudd-O) wrote:
> El sáb, 05-06-2004 a las 08:39, John McCutchan escribió:
> 
> > Yes it does. When you call stat() there is a bunch of activity on the
> > filesystem that will stop cdrom ejection during the stat call. Assuming
> > that nonotify is stat()ing every X ms, there is a race between the
> > stat() call and the unmount. So it could happen, but it would be rare.
> 
> Being racy it means nonotify cannot be considered as a long term
> solution.

True - i started nonotify to provide a pragmatic and simple solution
until someone comes up with something better (inotify?). 

And remember probability theory: Lets say the chance that umount gets
blocked by a stat-call is 1% (1e-2) *). When the GUI would repeat the
umount three times before giving up - the likelihood that umount does
not succeed drops to (1e-2 ^ 3 = 1e-6) - one in a million! Therefore
retrying umount solves this problem.

*) Just as an example - It will be a lot lower in reality - when polling
every second.

But there is another more important showstopper for a polling approch
like nonotify: It might not work with with "supermount". Read the
"supermount" FAQ:
http://sourceforge.net/docman/display_doc.php?docid=19736&group_id=79609

Norbert