It's fine for me to skip spec review step if it is just a simple task to track or we have already submit a blueprint in upstream.
My idea is that when we have something complicated to implement in doctor. We'd better launch a spec review in gerrit. As the run.sh growing larger, it might become a high priority task to discuss how to refactor it in release D. So that we can introduce more use cases in doctor, with different monitors, inspectors and of course to be tested under different installers. What do you think? -- Yujun On Wed, Sep 7, 2016 at 5:18 PM Carlos Goncalves <carlos.goncal...@neclab.eu> wrote: > I didn’t mention earlier but we do use Gerrit for reviewing OpenStack > specs before submitting upstream to review.openstack.org. > > > > Still, the point you are raising is that you feel we need to introduce > specs for Doctor in general. Have you felt already the drawbacks you > mentioned of using Jira when proposing new items? Personally I did not fell > myself or noticed anyone having problems, but I don’t follow every single > Doctor-related issue so I could be missing something. > > > > Please understand that the main reason I see for trying to avoid spec > files at this moment is that they add lots of overhead to everyone. I don’t > see a big item justifying a spec file in Doctor, except items we need to > implement upstream (e.g. OpenStack). > > > > Wrt contributors visiting Gerrit more than Jira: the OPNFV team, and in > particular the release team, is making the effort of utilizing Jira as much > as possible for the purpose of documenting epics, tasks, bugs, etc. Going > against that should be avoided and contributors encouraged to use Jira as > well. > > > > Carlos > > > > *From:* Yujun Zhang [mailto:zhangyujun+...@gmail.com] > *Sent:* 07 September 2016 10:48 > *To:* Carlos Goncalves; TECH-DISCUSS OPNFV > *Subject:* Re: [opnfv-tech-discuss] [doctor] How to propose topics for > D-release > > > > Hi, Carlos > > > > The etherpad for release B looks good and I think we should continue using > it. > > > > On the working items, personally I don't think JIRA is a good system for > storing the spec for some reasons. > > 1. it is difficult to comment inline in JIRA issues > 2. it is difficult to track the change history of a spec in JIRA > 3. it is difficult to publish approved working items to a document > from JIRA > > Last but not least, contributors visits JIRA less often than gerrit, at > least on my side... > > > > We don't have to force a strict rule to approve a working item, but we > need a convenient tool to discuss about it. And I prefer to keep the spec > close to the source code so they won't deviate too much from each other. > > > > These are my starting point to propose gerrit instead of JIRA for specs. > Open for comments :-) > > > > -- > > Yujun > > > > On Wed, Sep 7, 2016 at 4:34 PM Carlos Goncalves < > carlos.goncal...@neclab.eu> wrote: > > We’ve been using etherpad, weekly meeting and F2F team meetings for > brainstorming. The team even strives to plan and align public presentations > (e.g. OpenStack and OPNFV Summit events). > > > Examples: > > https://etherpad.opnfv.org/p/doctor > > https://etherpad.opnfv.org/p/doctor_use_case_for_b_release > > https://etherpad.opnfv.org/p/doctor_meetings > > > > Due to large communities in e.g. Neutron and Nova, those projects felt the > need to stricter the reviewing process of work item proposals. I think > Doctor is not at that level, at least just yet. I believe we’re just fine > with the tools and flows we’ve been using so far, including JIRA. Unless we > have a compelling reason to change our workflow, I would suggest keeping as > is. > > > > I am open for suggestions/comments. > > > > Carlos > > > > *From:* opnfv-tech-discuss-boun...@lists.opnfv.org [mailto: > opnfv-tech-discuss-boun...@lists.opnfv.org] *On Behalf Of *Yujun Zhang > *Sent:* 07 September 2016 09:00 > *To:* TECH-DISCUSS OPNFV > *Subject:* [opnfv-tech-discuss] [doctor] How to propose topics for > D-release > > > > This topic was raised in the last meeting but one [1] which I was absent > from. I'm not sure if there is any conclusion out. > > > > Here are my proposals. Please feel free to comment > > 1. using etherpad for brainstorming, since it is lightweight and real > time > 2. submit detail proposals to gerrit for review like openstack > blueprint [2] > 3. track status of the approved spec in JIRA > > [1] > http://ircbot.wl.linuxfoundation.org/meetings/opnfv-doctor/2016/opnfv-doctor.2016-08-30-13.01.html > > [2] > https://wiki.openstack.org/wiki/Blueprints#Spec_.2B_Blueprints_lifecycle > >
_______________________________________________ opnfv-tech-discuss mailing list opnfv-tech-discuss@lists.opnfv.org https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss
