Virtualization - Apr 2007 - numbered patches in the paravirtops patch series

On Mon, 2006-07-24 at 08:56 -0500, Jeremy Fitzhardinge
wrote:
> I don't think the numbered patch scheme we're using in the
paravirtops
> patch series is going to work very well.  It assumes that we've got the
> patch order of all the existing patches right, and that we don't need
to
> fit in any new patches between them.  I think we'll need the
flexibility
> of rearranging/grouping patches to make them most suited for submission, 
> but if the patch names contain their (original) order encoded into their 
> names, it will just be a confusing mess.
Series file does control order, I just like ls -l to look sane.

Feel free to break/alter/move/change!
Rusty.
-- 
Help! Save Australia from the worst of the DMCA: http://linux.org.au/law

I don't think the numbered patch scheme we're using in the paravirtops 
patch series is going to work very well.  It assumes that we've got the 
patch order of all the existing patches right, and that we don't need to 
fit in any new patches between them.  I think we'll need the flexibility 
of rearranging/grouping patches to make them most suited for submission, 
but if the patch names contain their (original) order encoded into their 
names, it will just be a confusing mess.

We could rename the existing patches, but since they're probably going 
to go upstream pretty quickly it doesn't matter all that much.  But for 
future patches, I think we should just give them meaningful names, and 
use the series file to control the order.

Comments?

Also, I'm assuming that virtualization@osdl is the best place to get all 
these low-level paravirtops messages public, so we save lkml from 
basically admin-noise.  We should cc:lkml for any substantial technical 
discussion, of course.

    J