On Fri, Feb 27, 2015 at 5:47 PM, Juri Lelli juri.lelli@arm.com wrote:
Hi Amit,
On 19/02/15 10:40, Amit Kucheria wrote:
Hi Juri,
Not sure what update you were waiting for, but Alex has gone ahead and split the current patch stack into stable/sched-upstream and stable/sched-core branches. So any updates from you should apply on top of the stable/sched-core branch. If you've rebased the patchset, please let Alex know in your pull request so we can replace the entire branch[1].
Thanks for the information, I've found the new branches and used them as a basis for the new backport. However, I also noticed that the wikipage is not yet updated. Can we put some information on how we intend to manage the thing there?
I was hoping to do a couple of rounds of merging new code in to iron our any kinks in the process before publishing the tree maintenance process. Besides, IMHO the document we currently have is too detailed and targeted to contributors to the code line, not to the users who just want to take the branches and test it. Do you agree?
The other thing that is still left unresolved is what will be the acceptance criteria for patches in the future? Here are a few different vectors: - Benchmark improvements for a defined set of benchmarks (Start with WAv2 bundled-benchmarks?) - Ack from other contributors - What test platform to consider as representative for our testing (Juno?)
Regarding the EASv3 backport, I had also to backport 5 additional patches from mainline that should go on top of the upstream branch. The EASv3 backport patches apply on top of upstream (with those 5 patches). I attach an archive containing them. Is there a proper way to issue a pull request to Alex or is this way just fine? I could post them on the list as a patch series maybe.
As Alex suggested, let us use well-known git workflows. In this case, I'd send two pull-requests to eas-dev and Alex - one to update the sched-upstream branch, the other to upstream sched-core. There should also be some comments on what sort of testing has been done with these patches. I believe that eventually you and Alex will be the people to decide what code should be merged and what should be rejected.
Thanks a lot.
Best,
- Juri
Regards, Amit