[tahoe-dev] Tahoe-LAFS is widely misunderstood

[Scott Dial]

On 2/2/2011 6:51 PM, Zooko O'Whielacronx wrote:
> You haven't posted before, have you? Welcome!

Not quite my first time, but close. Thanks for the welcome.

> Your suggestion about using extended attributes to give users access
> to caps through POSIX sounds intriguing! Could you please open a
> ticket to keep track of it so that we don't forget?

I'll open a ticket, but with 3 (!) FUSE implementations laying about in
the source tree, I am skeptical this will get worked into something I
can use soon.

On the other hand, the recent traffic on ticket #1111 (PyFilesystem
backend) caught my attention and I have started using that interface
with much success (both on windows/dokan and linux/fuse). It took little
modification to support the extended attributes there[1]. I would put
this into a ticket over on pyfilesystem, but I don't want to create a
conflict if Tahoe-LAFS decides to do something different with xattrs.

So, I'll open a ticket just to start a discussion about what xattrs
should mean to a Tahoe-LAFS mount-point.

[1] http://scottdial.com/tahoelafs/getxattr-r624.patch

-- 
Scott Dial
scott at scottdial.com
scodial at cs.indiana.edu

-- 
Scott Dial
scott at scottdial.com
scodial at cs.indiana.edu