On Tue, Sep 17, 2013 at 09:35:58AM -0400, Jim Fulton wrote:
On Tue, Sep 17, 2013 at 8:58 AM, Marius Gedminas <marius@gedmin.as> wrote:
FWIW the only reason I'm in favour of self-merges is that this short-circuits the "have you signed the ZF committer agreement?" dance. Only people who have can merge.
Sorry, I don't understand the point you're making.
I'll try to explain better.
I also feel silly when I ask this question from people with very familiar names. (I feel that I have to do when I don't see ZF membership on their GitHub profile.)
So are you saying you don't merge other people's code because you don't want to ask if they're contributors?
No. I'm saying the reviewer-merges workflow looks like this: 1. Somebody creates a pull request 2. A reviewer reviews 3. OP fixes 4. Reviewer asks "have you signed the agreement? if not please sign" 5. OP says he/she signed it 6. Reviewer trusts OP's word and merges (I don't have a good short word for "somebody who created the pull request", so I abused Original Poster.) Whereas I'd prefer 1. Somebody creates a pull request 2. A reviewer reviews 3. OP fixes 4. Reviewer says "looks good to me, feel free to merge (if you're not a committer already, see http://foundation.zope.org/agreements)" 5. OP merges It's not a very strong preference. I can feel myself changing my mind already ;-) OTOH the implicit trust in step 6 makes me a bit uneasy, and I'm not quite sure how to verify the fact of the signing. Wait for the user to show up in https://github.com/zopefoundation?tab=members ?
I can understand this, but I'd still try to encourage a more review-centric workflow.
+1
Also, if a change is trivial, the PR doesn't have to be from a contributor. I understand that triviality isn't always clear.
It's completely unclear to me, once the PR goes beyond "typo fixes". :( Marius Gedminas -- http://pov.lt/ -- Zope 3/BlueBream consulting and development