Tim

My apologies for waiting so long to reply.

Regarding my case notation and implication I meant to suggest that
even in an RTC environment there is still a chance for review after
the commit.  Because of that, I feel the benefits of RTC are in play
and the safety net of CTR is still in play.  So for me i'd suggest
anyone with committer status be considered a valid reviewer.  I like
the spectrum you suggest though :-)

Thanks
Joe

On Mon, Jul 25, 2016 at 8:58 AM, Tim Ellison <[email protected]> wrote:
> On 24/07/16 16:10, Joe Witt wrote:
>> For sure.
>
> Members of the PMC are not necessarily committers, Suneel and I are
> proof positive.  However, I agree that in established projects the PMC
> is drawn from the technical community, and it would be unusual for
> someone to be invited to the PMC (to give technical direction and
> project management overview) without them also being committers.
>
>> Code can always evolve and if a commit happens that needs some refinement
>> all is fine.  In essence ctr is always available.  For us adopting RTC it
>> means, in my opinion, that you should obtain an independent opinion that it
>> is good to go.  As new folks contribute it stokes engagement and even
>> mentoring and as veterans of the project contribute it encourages shared
>> understanding.
>
> Agreed.  Without debating the criteria for committership, I'm still
> "acquiring my merit" in the project, and when the PMC deem fit, I hope
> to be invited to become a committer.
>
> In the meantime, I will continue to comment on PRs etc from all authors,
> but don't consider my +1 to be a sufficient review until it is backed by
> the merit of my committership.
>
>> I see it as a sort of rTCtr (case intentional).
>
> Does your case notation mean the reviews are not so important as the
> commits?
>
> I'm curious.  Presuming RTC only really makes sense of the reviewer
> groks the code base sufficiently well to understand the implications of
> the proposed change.  Where on the scale of "some passing dude we never
> heard of before" to "grizzled old committer" do you place 'reviewer'? ;-)
>
> As I wrote before, and you state here, we can be quite relaxed about it
> -- it's all in version control, mistakes can be fixed or rolled-back.
>
> Different projects acquire a personality about how they operate, and it
> is healthy for Pirk to be thinking about these and choosing a model that
> reflects the community best.
>
> Regards,
> Tim
>
>> Provably worth documenting the approach either way on a wiki.
>>
>> Thanks
>> Joe
>>
>> On Jul 24, 2016 10:59 AM, "Ellison Anne Williams" <[email protected]>
>> wrote:
>>
>>> This was raised on a recent PR commit; bringing it here for discussion:
>>>
>>> It's my understanding the members of the PMC (Tim, Suneel, etc) are also
>>> committers and able to perform a review in a RTC scenario.
>>>
>>> Thoughts?
>>>
>>> On Fri, Jul 22, 2016 at 10:05 PM, Josh Elser <[email protected]> wrote:
>>>
>>>> I lean towards just needing a +1 to commit and still doing that myself.
>>>>
>>>> Just need to be clear however is decided.
>>>>
>>>>
>>>> Ellison Anne Williams wrote:
>>>>
>>>>> I don't see point in having to get a +1 from another committer for
>>> fixing
>>>>> trivial items such as merge conflicts, build breaks, etc. I also would
>>>>> through trivial updates to the website like fixing typos, broken
>>>>> hyperlinks, adding hyperlinks, etc.
>>>>>
>>>>> My interpretation of RTC is that the author can commit their own changes
>>>>> after receiving +1 from a reviewer (another committer).
>>>>>
>>>>> On Fri, Jul 22, 2016 at 4:15 AM, Tim Ellison<[email protected]>
>>>>> wrote:
>>>>>
>>>>> On 22/07/16 05:50, Andy LoPresto wrote:
>>>>>>
>>>>>>> One of the nice effects of RTC is that the community gets an
>>>>>>> opportunity to gently enforce code convention and prevent rapid build
>>>>>>> up of technical debt. Something else I've witnessed is that as the
>>>>>>> community grows, non-committers feel more comfortable submitting PRs
>>>>>>> because they've seen the same review process applied regardless of
>>>>>>> the submitter's status.
>>>>>>>
>>>>>>> Apologies if I missed these points on an earlier thread.
>>>>>>>
>>>>>> Sure, I don't feel strongly either way -- though I would expect a
>>>>>> committer to be able to fix trivial items like a simple merge conflict,
>>>>>> or backout a build breaking change without always having to get a
>>>>>> backing +1 from elsewhere.
>>>>>>
>>>>>> Out of curiosity, does folks' interpretation of RTC mean that the
>>> author
>>>>>> never commits their own changes? i.e. the reviewer always commits (as
>>>>>> the original author)?
>>>>>>
>>>>>> Regards,
>>>>>> Tim
>>>>>>
>>>>>>
>>>>>
>>>
>>

Reply via email to