[PATCH 01 of 35] largefiles: declare commands using decorator

[Gregory Szorc]

On 5/5/2014 12:38 AM, Pierre-Yves David wrote:
>> [PATCH 01 of 35] largefiles: declare commands using decorator
> 
> While I appreciate your contributions, the areas you clean up, the
> quality of your patches, this is still terribly wrong.
> 
> You cannot blindly past bomb such gigantic series. Please start talking
> with reviewer to have a more efficient way to do this. You are currently
> creating massive disruption to the review process, impacting all other
> contributors

OK, so I grok that the run-tests patch series at >100 patches is
excessive. But I'm kinda scratching my head as to what's wrong here.

While this series is 35 patches, it is mostly cosmetic and IMO should
involve lots of rubber stamping. I was very liberal about using multiple
patches because fold is easier than split. I didn't want to risk someone
rejecting the patches because they didn't like changing multiple files
in a single patch. I believe the patch guidelines are on my side here.
Furthermore, I'm pretty sure I've seen series with fewer parts but more
actual code to review. I wouldn't get too held up on patch count without
looking at actual diff size or the nature of the diff (e.g. new code vs
copy) (although this is hard to identify in vanilla diffs - tools like
Phabricator make it easier, but you know that).

It's also not clear to me what "talking with [a] reviewer" entails.
Please clarify.

It is fair for you to take issue with me not following procedure. I just
ask that the procedure be better documented so future contributors don't
fall into unmarked pits. AFAICT
http://mercurial.selenic.com/wiki/ContributingChanges says nothing on
the subject of series length and disrupting review processes.

I apologize. I'm just trying to contribute.