Re: tosfs, faster hash function...

[Julian Reschke <reschke@GOEDEL.UNI-MUENSTER.DE>]

> 
> As Julian pointed out, CRCs cannot be guaranteed to be unique. For that
> reason, I don't think it's a good idea to use them for inodes in tosfs.
> The current scheme is inefficient for some purposes, but much safer,
> and I'll take safety over efficiency any day.
> 
> Regards,
> Eric
> 

Eric, the current implementation is NOT safer. Inodes have to be different
for different files, BUT they have to be identical for identical files,
too. The current implementation satisfies the first, but ignores the
second point.

I think the current implementation is useless. If you really want to keep
it, there should be a portable method to *inquire* whether the inodes
are 'real' or just a counter.

Juergen, re start cluster as inode: for empty files, one could use
(start cluster of directory << 16) | (index of file within directory)
(note that cluster numbers are always 16 bit, while the inode has 32 bits).

-- 
---------------------------------------------------
Julian F. Reschke, Hensenstr. 142, D-48161 Muenster
 eMail: reschke@math.uni-muenster.de jr@ms.maus.de
___________________________________________________