Re: linux-next: manual merge of the tip tree with the arm tree

[Ingo Molnar <mingo@elte.hu>]

* Russell King - ARM Linux <linux@arm.linux.org.uk> wrote:

> On Fri, May 13, 2011 at 11:26:46AM +0200, Ingo Molnar wrote:
> > Had you asked us before committing it one day after it was posted, or had you 
> > *noticed* that those files are not in your tree and are already modified in 
> > linux-next, you'd have gotten a response like:
> 
> Please also don't read anything into the commit date - it merely shows
> when the last update happened.
>
> My workflow for patch series involves keeping them in git right from the 
> start.  So actually they've been in git since _before_ they were posted. In 
> fact, the emails which I send out for any patch series are always generated 
> from the git commits.
> 
> So, all my patches live in git _first_ before being mailed out.

It is not a problem at all if you commit it to some non-permanent development 
branch of your own - we all do it.

The commit date i pointed out was of the *final* commit, which got into 
linux-next. That showed a timestamp of just a day after the patch was sent
out: presumably you rebased it to add John's Acked-by.

The step where your workflow failed was to take upon yourself to maintain a 
file you do not normally maintain *and* messing up doing that:

 - you did not ask the maintainers who maintain it (which is fine as long as 
   you do not mess up)

 - you did not realize that the file you modified is already modified in that
   tree, almost two months ago (it's not that hard to fetch linux-next once 
   every week or so)

 - you did not even notify them that you committed something so when the bug
   happened in linux-next they had no idea what was going on

Had you done any of those steps differently we'd have a better outcome.

It's not a big problem all and we can resolve it, but you need to stop 
pretending that your workflow was just fine - it sucked here.

Thanks,

	Ingo