Am Mittwoch, 14. Dezember 2005 10:35 schrieb [EMAIL PROTECTED]:
> Wilhelm Meier:
> > O.k., I see the point. But: why does unionfs lookup a whiteout in a
> > read-only branch?
>
> I don't know the reason and I expect the other person will reply.
>
> But your question showed me a good point. If I replace the word
> 'read-only' into 'lowest branch', yes, the number of looking up
> the whiteouted entry can be reduced. Of course it is based upon that
> unionfs is working correctly, eg. whiteout never exists on the same
> branch of the actual entry.
>
> And currect unionfs lookup code checks the passed filename ('fileB' or
> '.wh.fileB') first, by is_validname(), before any actual lookup.

Thank you for this advice. 

On the server-side unionfs-module I commented the is_validname(). Now, the 
lookup of whiteout-entries doesn't return ENOPERM and it works as it should. 
But with the drawback, that you can create .wh.foo whiteouts, which are not 
even listed ...

Would it be possible to integrate this "feature" into the main-line as stated 
above, only for ro-branches? Possibly this should be done in 
unionfs_lookup_backend(), when iterating over the branches. Does this break 
other things?

>
>
> Junjiro Okajima
> _______________________________________________
> unionfs mailing list
> [email protected]
> http://www.fsl.cs.sunysb.edu/mailman/listinfo/unionfs

-- 
--
Wilhelm Meier
email: [EMAIL PROTECTED]
_______________________________________________
unionfs mailing list
[email protected]
http://www.fsl.cs.sunysb.edu/mailman/listinfo/unionfs

Reply via email to