Re: state or future of LFS?

[David Holland <dholland-current%netbsd.org@localhost>]

On Mon, Apr 13, 2009 at 11:21:25PM +0000, Wouter Klouwen wrote:
 > AFAICT, the LFS implementation is suffering from trying to tie it into FFS.
 > Given the fact it's a completely different concept to FFS, that was probably
 > not a good thing to do.
 >
 > [...]
 > 
 > Is it not possible to have some sort of modular framework for file systems
 > (perhaps as LKM), which provides for some of the basic FS operations?
 > This could be used to maximise code sharing and minimise bugs, yet avoid
 > trying to tie different FSes into a commonality they don't have. 

That's what sys/ufs/ufs is *supposed* to be. Too bad it doesn't really
work that way.

(We also have sys/miscfs/genfs, which is another approximation from
the other end.)

This kind of structural stuff is hard to get right, and each failed
try makes the problem only that much worse...

-- 
David A. Holland
dholland%netbsd.org@localhost