Hello everybody,
I'm not sure about a small point in the Tux3 design; please excuse me if
that's written somewhere I just didn't look.
There are some hacks to sort the directory contents by inode number before
doing the stat() calls on the names; it's been proposed in kernel, done via
LD_PRELOAD, and explicitly coded in some applications.
[ I'm interested in that because my pet project (fsvs.tigris.org) uses
a per-inode-number globally-sorted list of entries of whole directory
hierarchies, which makes querying them faster than "find" on cold
caches. ]
Now, if I understand Tux3 design correctly, it's no longer the case that the
inode numbers have any meaning regarding the on-disk location (because of
snapshots, versioning, and other strategies), so this hack (or "feature", if
you like) wouldn't work anymore, would it?
Thank you for your patience and all answers.
Regards,
Phil
_______________________________________________
Tux3 mailing list
Tux3@tux3.org
http://mailman.tux3.org/cgi-bin/mailman/listinfo/tux3
