On Thursday 11 October 2001 05:14, you wrote:
> On Thu, 11 Oct 2001 at 11:00, Seth Mos wrote:
> > Maybe we can bribe Alan with a few DVD's or CD's ;-)
> I know Alan's in RedHat's payroll (and RedHat is rather ext3-centric, it
> seems, but this is just my opinion), but perhaps if we could get XFS on
> his good side he'd manage to keep his tree XFS-enabled? That would be
> really great (except I don't know what he thinks about all the internal
> changes that XFS needs done). SGI can still keep a CVS tree of XFS, where
> bug squashing can be handled. When a patch looks like it fixes things
> decently, it can be fed to Alan.
> But of course the decision of having an XFS-enabled tree is entirely up to
> Alan, so ...
Alan is on RedHat's payroll but he seems to put some considerable effort
into being impartial, at least on kernel issues. He's not even in charge of
packaging up RedHat's release kernels, and at least once he's refused certain
jobs that could be viewed as a conflict of interest.
Alan and the other kernel developers have stated their reasons why XFS is
not in any of the official kernels -- largely because the code duplicates too
many functions already present in Linux. SGI has their reasons for wanting
to keep it that way (it's well tested, both on IRIX and now on Linux); the
kernel developers have their own for not allowing it in the official kernels
(it's ugly and not the Right Thing.)
The problems with 2.95.x compilers is another stumbling block -- though
some documents may say these compilers are unsupported for Linux, they are
"unofficially" supported for 2.[4|5], as is, to a lesser extent, 3.0.x.
Anything that won't build (and work) with at least 2.95.x is broken.
I still think XFS will make it into the official kernels for 2.6, but
it's not looking like SGI's developers are going to be the ones to get it
there; there just isn't enough incentive on SGI's part to devote the
resources to do it "the Linux way". But then, that debate died as soon as it
started; both parties have bigger eggs to crack right now.
(Whether Andrea's VM will live on in 2.4 is another interesting
question... I have little doubt it'll be the 2.[5|6] VM.)
This is not intended as a flame at SGI or any of the SGI developers...
IMHO they're all doing great work, and I'm glad they've chosen to GPL XFS,
and to devote so many other resources to Linux.
Heh. Someone should make me editor of Kernel Traffic or something. ;-)
"Bull in pure form is rare; there is usually some contamination by data."
-- William Graves Perry Jr.