pending u-boot changes committed (GTA02)
werner at openmoko.org
Wed Jan 30 17:20:56 CET 2008
Andy Green wrote:
> Whatever way you look at it silence followed by a patchbomb isn't ideal,
> nobody has the time to go through all that at once so in fact there's no
> real review of stuff done that way.
If you look at my recent mails, you should actually find pretty much
all major changes mentioned, often with pointers to the corresponding
not-yet-merged patch file in SVN, or even the actual code, when I
didn't dare to commit the patch.
> We could really do with a positive culture of review outsourced to the
> list, it will definitely reduce the burden on you trying to do it for
> incoming patches alone (while your patches get no review).
Hmm, I'd view the mailing list more as one mechanism to maintain the
maintainer's patch queue. As a very welcome side-effect, it also
allows patches to be reviewed while waiting.
However, I hope we can move (back) to a more direct model for the
principal developers, i.e., those whom we can trust to show good
judgement and who will also be careful about not breaking things.
There's no problem with reviewing and improving reasonably clean
patches even after they've been committed to the tree. Also,
potentially controversial patches can always start their life in a
What is essential for this to work, though is notification mechanism
for new patches that works as a push model. That can be separate
mails to the openmoko-kernel list or it can be the commitlog. (By
the way, I had been subscribed for a long while to the commitlog
digest, because I thought that would clutter my mailbox less, but
I found it incredibly hard to get information out of it. This
improved drastically once I switched to individual mails.)
More information about the openmoko-kernel