Re: [agc%pkgsrc.org@localhost: Re: Google SoC Proposal Ideas]

[Thor Lancelot Simon <tls%rek.tjls.com@localhost>]

On Tue, Mar 24, 2009 at 03:05:43PM +0000, Alistair Crooks wrote:
> 
> I think future file systems design and implementation is beyond the
> scope of the file system resizer project - and I'd really like to
> see us having the ability to resize FFSv2 file systems well.

But I think there's almost no content to this project -- it's a trivial
amount of datastructure manipulation on an existing program, and some
testing and bugfixing -- if it is just to produce a utility which can
resize FFS filesystems (even v2) which are not mounted.

And resizing while the filesystems _are_ mounted will require either
extensive kernel support that works approximately like the support
for the LFS cleaner, with all its complexity and security issues (ugh!)
or a completely in-kernel implementation.  Either means quite a bit of
design and implementation work on new filesystem code, though that code
will, admittedly, read and write the existing FFSv1 and FFSv2 filesystem
formats.

Still and all, I think it's a manageable project for a summer.  But I
don't think it's useful if no new filesystem code is designed nor
written, because we already _have_ an offline resizer which just needs
a little bit of touch-up to be finished.

Thor