1887834 - Don't mark a patch as WIP when there are reviewers assigned in phabricator

Status: RESOLVED FIXED

(Conduit :: moz-phab, enhancement, P3)

Description

[Dave Townsend [:mossop]]

Sometimes I prefer to add reviewers directly in phabricator, maybe I can't remember their names or maybe I know a reviewer will be automatically assigned by a herald rule. So I do this:

1. Commit my patch with the description Bug XXXXXXX: Do the thing.
2. moz-phab
3. Go to the patch in phabricator and check/add reviewers.
4. Request review.

Later after a first review:

1. Update the patch.
2. moz-phab

At this point moz-phab re-marks the patch as WIP even though there are reviewers assigned to the patch. It would be nice if when no reviewers are marked in the commit it would check to see if there are reviewers in phabricator and not mark as WIP in this case.

(It would be even nicer if reviewers are going to be automatically assigned that the first submission didn't get marked WIP but I assume that is somewhat more difficult to achieve!)

Comment 1

[Mathew Hodson]

Attachment: submit: don't mark revisions with reviewers as WIP (Bug 1887834)

Comment 2

[Mathew Hodson]

Attachment: linting: fix type problems in conduit.py and submit.py (Bug 1887834) r=sheehan

Reported by Pylance extension in VS Code.

This fixes errors like the following:

• "None" is not assignable to "int"
• Object of type "None" is not subscriptable

Comment 3

[Mathew Hodson]

Attachment: submit: fix `test_submit_update_revision_not_found` (Bug 1887834)

In some scenarios that were exposed by the next patch, get_diffs
was called with an empty list in show_commit_stack, which raises an
error. This can happen when a commit is submitted with a revision ID
that doesn't exist.

Fix assert in test_show_commit_stack that depended on previous state.

Comment 4

[Connor Sheehan [:sheehan]]

This will go out in the next release.

Comment 6

[Mathew Hodson]

https://github.com/mozilla-conduit/review/commit/46cce99b6c72bf4cb2c8cc72a7eb9628e316f792