when to backport

[Michael Meeks]

Hi Luke,

On Wed, 2012-12-12 at 20:00 +1100, Luke Deller wrote:
> Newbie developer question: what is the convention around here for 
> backporting fixes to branches?

	Good question :-)

> I see that the 3.6 release plan in wiki says "only important bug fixes", 
> but what counts as important ?  If a fix seems moderately important to me 
> do I just repost my patch to gerrit against the release branch (perhaps 
> a few days after it was committed to master to let it settle), and see 
> if the reviewer agrees it is important?

	Sounds reasonable to me - of course there is always a balance of risk
vs. importance. If the risk is tiny then assuming someone wants to
review it it doesn't matter how mind-burningly useful it is - that you
want it included is prolly enough. Conversely if it is a -300+1000 line
patch that is mostly a random re-factor then - we think harder :-)

> What about if my fix is not burningly important, but it seems small and 
> safe - is that something I should resubmit for a release branch too?

	Why not; of course for the -3-6 branch you'd need a review by someone
else to get it included. For 4.0 bug fixes can be committed directly
without review; if you're a gerrit user anyway there should be ~no
difference :-)

	Does that help ?

	Thanks !

		Michael.

-- 
michael.meeks at suse.com  <><, Pseudo Engineer, itinerant idiot