2.6.24 - time to switch

[Werner Almesberger]

Andy Green wrote:
> Are we talking about the same thing here?  Show me where the "2.6.0" or
> whatever kernel tree is in svn?  It ain't there if I understood it.

Ah, you mean across upstream changes. You're right, the upstream is
implicit and, we don't have the kernel.org revision history in SVN
anyway.

> Once we're set up with git, we should be able to revert to a whole tree
> of the kernel at any patchlevel, whatever the base kernel version was at
> that time we will revert to that too.

Hmm, but you'd still have to follow whatever branches are there.
So if we rebase only every once in a while, this wouldn't be too
different from what we have now.

> But the kind of branch we are looking at for this job won't ever do
> that... if you want to revert something you have to commit the
> "antipatch" for whatever it was you didn't like any more,

Okay, and stgit will find the antipatches or slight-change-patches
as well, and apply them to the chunk we're preparing for mainline ?

Example: would the _fiq_ipc change, after you've committed the _fiq_ipc
version to our public GIT repository, still result in a single
introduce-fiq-etc.patch for upstream, or would have to feed mainstream
two patches and/or combine them manually or through some mechanism
external to stgit ?

- Werner