> Alexander Viro wrote:
> > > OTOH, the fastest solution of all would be one that's as good as locate
> > > but always up to date. Why do you think I keep mentioning ways to
> > > propagate "directory changed" information up directory trees?
> >
> > Oh, lovely. So instead of slow find we are getting helluva lot of inodes
> > updated almost constantly. Wonderful. Especially if you plan to propagate
> > it across the mounpoints. You know, I'll take slow find.
>
> Nonsense.
>
> If `a/b' is updated many times, `a' is updated just once between tree
> searching operations. That's once a day if your operation is updatedb.
> It applies recursively, therefore the filesystem very quickly settles
> into a state where all the paths that need flagging are already flagged.
> The path inodes don't have to remain in memory.
Oh, please. How the heck will tree search tell the kernel what it is? How
will you deal with changes that happen _during_ the search? What
privileges will you need for 'cleaning' directory?
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/