incubator-cloudstack-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sheng Yang <sh...@yasker.org>
Subject Re: [ACS41] Concerns about where development has happened
Date Wed, 09 Jan 2013 22:28:34 GMT
On Wed, Jan 9, 2013 at 12:51 PM, John Kinsella <jlk@stratosec.co> wrote:

> So, we have nearly 1500 lines of code changes committed in CLOUDSTACK-306.
>
> There's 10 people related to the issues listed below, I believe they're
> all from the same employer.
>
> Folks, this is hindering our process to graduate to a top level Apache
> project. The first few times it's an honest mistake, but now i begin to
> wonder exactly what steps we need to take to ensure it stops happening.
>
> I'm hoping (expecting, really) somebody can shed some light for us why
> this is happening.
>
> Lastly, while not a threat, I would like to remind people that being a
> committer is a privilege, and that when the PPMC considers people for
> either committer or PPMC status, we do review their participation within
> the community.
>
> John
>
> On Jan 9, 2013, at 7:06 AM, Chip Childers <chip.childers@sungard.com>
> wrote:
>
> > Hi all,
> >
> > In reviewing the feature proposals for 4.1.0, David and I have found
> > many problems that indicate that development has happened outside of
> > the community.  While I can't be sure that we've found all of the
> > issues, it's certainly problematic to see this many.
> >
> > Please review and let me know if I'm misinterpreting the state of things.
> >
> > I'm not sure where to go from here.  I guess we have 2 options: we
> > re-write the code from scratch as CloudStack code, or Citrix donates
> > the code produced for CloudPlatform (and it gets taken through the IP
> > clearance process).
> >
> > The following features are potentially issues:
> >
> > CLOUDSTACK-297
> > This is in the CloudPlatform 3.0.6 release
> > Discussion occurred in October
> > I don't believe that the code is in the ASF repo
> >
> > CLOUDSTACK-299
> > This is in the CloudPlatform 3.0.6 release
> > The UI code appears to be in our repo, but the backend does not.
> > Example, grep for: createEgressFirewallRule
> >
> > CLOUDSTACK-306 (CLOUDSTACK-775 is a duplicate)
> > This is in the CloudPlatform 3.0.6 release
> > Commits went into master on Jan 4 (there are 3 commits)
> > Discussion happened in October
> >
> > CLOUDSTACK-737
> > This is in the CloudPlatform 3.0.6 release
> > UI work completed (CLOUDSTACK-537) in the asf repo, starting in november
> > I can't find any commits for the backend work in our repo
> > The requirements wiki page and jira record were created on Jan 3
> > Dev list discussion started in November, but there were outstanding
> > questions that were not addressed in that thread.  Unsure if consensus
> > was achieved.
> >
> > CLOUDSTACK-774
> > Frank identified that this was a "Byron feature" and that all "Byron
> > features should be merged to ASF repo", but I'm unable to find in the
> > CloudPlatform release notes
> > Unable to find dev list discussion
> >
> > CLOUDSTACK-777
> > This is in the CloudPlatform 3.0.6 release
> > Docs already submitted, but no FS available.
> > Unable to find dev list discussion
> >
> > CLOUDSTACK-778
> > This is in the CloudPlatform 3.0.6 release
> > Docs are done, but feature doesn't exist in CloudStack
> > Unable to find dev list discussion
> >
> > Also, generally all documentation originally created for CloudPlatform
> > 3.0.6 features, but not created in the CloudStack git repo or
> > submitted prior to publication will need to go through IP clearance.
> >
> > Also: CLOUDSTACK-873 is not proposed for 4.1.0, but appears to be in
> > CloudPlatform 3.0.6.  I may be misinterpreting this, but it appears to
> > be something that will need to go through IP clearance.
> >
> > -chip
> >
>
> I am the developer of CLOUDSTACK-306(SRX & F5 inline mode). It's a feature
of CloudPlatform 3.0.6 release.

The discussion happened on October last year.

Here is what's happened later:

Basically I kept synchronized apache tree and out internal tree till
October, then I failed to so do because the code dependence issues become
more common, and the development of CloudPlatform is somehow overwhelming.
I know that's not a good reason, but that's what's happened.

And after CP 3.0.6 release, I want to merge back the feature ASAP(so does
our team), so it took me about one week to get this done.

Since it involved too many commits in our internal tree, I get this done
through file compare, and check every commit manually for about 30
commits(including all the fixes after the original feature in 3.0.6
release), then do the test. That's what makes the 3 commits.

The reason I merge it to master directly because I think even the LoC is
big, the impact is comparably low. Biggest modification are in nonoss part(
SRX and F5). There are some modifications on NetworkManager, but it mostly
didn't touch the existing code. So as a committer, I give the estimate of
impact is low, and there are already discussion in the mailing list. So I
checked them in.

I think what I should have done is give a notification to the mailing list
before the merging code, that would make thing more smoothly. I am sorry I
didn't do it.

And I do recognize the committer is a privilege.

Thank you.

--Sheng

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message