[Reviewers] dropped a couple changesets and flushed the queue

[Augie Fackler]

On Dec 2, 2016, at 10:53 PM, Pierre-Yves David <pierre-yves.david at ens-lyon.org> wrote:

[snip extensive discussion of circumstances about which we’re generally in violent agreement, which aren’t relevant to the broader policy discussion I was trying to have]

> I did got time to spend on Mercurial until today, almost one week later.

What level of time commitment should we be expecting from you, out of curiosity? It occurs to me I don’t really know how much we should be expecting.

> About rational for dropping Mads patch
> --------------------------------------
> 
> My initial email amongst other thing contains two important data about
> this.
> 
> 1) a sort explanation of the reason why I dropped mads patch:
> 
>>> Mads histedit series got dropped because it change existing behavior
>>> in way inconsistent with rebase and potentially harmful for user
>>> relying on backward compatibility.
> 
> 2) an explicitly mention I had detailed explanation about these drops
> coming on the public mailing list threads. Verbatim below:
> 
>>> Long explanation for each drop are on the way on relevant thread. But
> here is a short version:
> 
> I do not see these two elements taken in account in the main part of
> your email replying to mine. Your objection to the drop of Mads patch
> does not make any mention of elements in the "short explanation" and
> later in that paragraph you requested a long version I already said to
> be on the way.
> 
> I'm mentioning and highlighting this because this is becoming a recurring feeling that your reply does not takes in account important
> elements of email you are replying to. If we leave this un-addressed this will evolve in a major issue.

I didn’t ever say it was unreasonable to have dropped these patches. I just haven’t personally observed any attempt to contact me before discarding a patch I’ve queued, so I wanted to bring it up as a potential policy point for discussion.


> On policy around dropping patch
> -------------------------------
> 
> I agree that getting a green light from another reviewers before
> dropping is a good thing. Ideally, you get the approval of reviewers who
> initially push the change (and any additional reviewer who approved it).
> 
> However I trust other reviewer to be thoughtful and flexible on that.
> If there is a legitimate reason to bypass it after careful
> consideration, they should got for it.

Note that I agree in this case the patches should have been discarded, but would have appreciated even a cursory attempt to contact me and ask before doing so. It’s a bad feeling for me to have my review work discarded without me having a chance understand why or even know it’s about to happen.

> Dropping a patches is not a big deal, it is easy to requeue it as draft
> on the top of the start and resume the discussion. Right
> now, the main pain around dropping is our tooling.

Is dropping any more complicated than a prune, stabilize and push at this point? (With a warning and/or helping hand to the contributor so they don’t lose their change)?

Thanks!
Augie