Hubert Feyrer wrote:
>Might be useful to know a bit about your experiences to judge if it's a 
>good project for SoC and _your_. :)
well, your right, I forgot to introduce myself. So, I'm a student in a
school of computer science near Paris. I already worked last semester on
graph and tree, and have good knowledge in C, but I have no real
experience with kernel hacking (it could be a chance for me to begin :) ).

As you said, I have almost two solutions for this project. The first one
is to use SGI's sources as done with the freebsd port of XFS. The
problem, I think, with this solution is that sources are under GPL, not
BSD as wanted by the NetBSD team. The second solution is to restart from
scratch, by choosing this solution, I'm about 99.9% sure that I won't
have a fully working implementation of XFS (also as I'll be alone to do
it) in given time. By fully working implementation, I mean having a
stable, improved XFS port with both reading and writing support. 

In spite of that, I think I can begin to work on implementing XFS and
having a read support in the given time (I hope so) even if I know that
the task will be huge ... After that, nothing forbid me to continue to
work on it for having a writing support. 

As someone said in the thread "Summer Of Code - Filesystem What do you
think ?", two month, it's really short to implement a FS, but lots of
thing can be started.

regards,

Arnaud