RFC: 2.6.release.patchlevel: Patch against 2.6.release[.0] ?

["L. A. Walsh" <lkml@tlinx.org>]

Given the frequency with which stabilization patches may be released, it
may not be practical to expect users to catch each release announcement
and download each patch.

Especially if small patches are released for stability, as one might
(hopefully) expect.  Assuming that stability and "fix-it" patches will
generally be small (I'd hope).  Seeing that the latest "fix-it" patch
is already at ".6", I'd have to load multiple patches to catch up from
2.6.11.  I blinked my eyes and missed a few or 5 previous stability
patches, so I just downloaded the entire bzip...not a biggie, but
might create less load on servers if I didn't need to go through 6
patch applications to get current.

What do people think?  Would it be desirable to have the stability
patchsets based against the base release (2.6.11 in this case)?  I'll
already have downloaded 2.6.11 or the previous base release, but
with the frequency of patch releases, it might be more reasonable to
have patch revisions all patch against a base release rather than
having to download and apply what may grow to be a large number (but
small diff) against a base release?

Do people think patch-releases will get too big, or might it not
be easier to apply them to a constant downloaded copy of the base?

It's a bit amusing since I was one of those that complained about the
kernel stability, but 2.6.11 has been fairly solid for me, so, of course,
I'm 6 patches behind -- I don't think the patch release notifications
are getting as wide-spread press (or at least not reaching "/." :-)) as
the main releases get.

Linda