What to do on issue opinion divergence?

(Mickaël Schoentgen) #1

I was doing some triage on opened PRs and after almost finishing one, I saw that a core developer asked to close the BPO after the reporter created the PR.
Should I ping the core dev on the PR to clarify what action should be done or still review the PR and let the process continue?

FTR I am talking about this one: https://github.com/python/cpython/pull/6191. There is nothing alarmant, just a general question when such cases happen.

(Inada Naoki) #2

Ah, it’s very hard problem to reduce issues.

It is not only usability problem of one method.
It needs design decision about missing_ok option for all removing operations.

I think there are three ways:

  • If there are module owners, delegate to them.
  • If it seems really worth, start discussion: invite some core devs who contributes the module to the issue, or post new thread on python-dev ML, or “Users” category of discuss.python.org.
  • If you don’t think it’s worth enough to discuss, close the issue. And say “If you really want this feature, please post proposal to python-ideas or python-dev ML”.

Rejecting proposal is a hard decision. But it’s not end of the world.
Python has many features which were rejected once or several times.
And reducing opening issues makes it easy to focus more important issues.

(Mickaël Schoentgen) #3

Those are good pointers!

This is valuable information for everyone, it would be a good add to the documentation (if it is not already written somewhere).

Thank you :slight_smile:

(Brett Cannon) #4

Please feel free to propose a PR for the devguide!

(Guido van Rossum) #5

Also, I would think that if the BPO issue is closed, that means the PR should be rejected – or if you strongly disagree, you should lobby to reopen the issue first before spending more time on the PR.

The reason for this hard stance is that in the normal workflow we never close issues until the PR has landed. Some trivial PRs (e.g., typo fixes) may exist without an issue, but once a PR is linked to an issue, decisions on the issue override the PR.

(Serhiy Storchaka) #6

In this particular case the BPO issue has not be not closed yet. I just suggested to close it and added the pathlib author (Antoine) to the nosy list. It was a half year ago, and since no other core developer challenged it, I am going to close this issue and the corresponding PR.

(Carol Willing) #7

Thanks @storchaka.