Renaming issues/PRs incorrectly changes awaiting change review and awaiting merge to awaiting review

[arhadthedev]

While changing bpo- to gh- I noticed that Bedevere makes forbidden transitions in response:

• awaiting change review → awaiting review: gh-89532: Remove LibreSSL workarounds cpython#28728
• awaiting merge → awaiting review: gh-84443: SSLSocket.recv_into() shouldn't require __len__ cpython#20310

According to README, awaiting review has no incoming conditions:

flowchart TD
    A([Published PR]):::creator
    A_draft([New Draft PR]):::creator
    A_draft -- publish draft by contributor --> A:::creator
    A -- by contributor --> B[Awaiting review]:::anyone
    A -- by core dev --> C[Awaiting core review]:::coredev
    B & C -- new review by\nanother contributor --> C
    C & B & E  -- new core review\nrequests changes --> D[Awaiting changes]:::creator
    D -- changes by contributor --> E[Awaiting change review]:::coredev
    C & E & B -- new core review\napproves ---> F[Awaiting merge]:::coredev
    G[When a review is dismissed\nthe highest remaining state is restored\nbased on the remaining reviews]
classDef creator stroke:#CC0;
classDef anyone stroke:#00C;
classDef coredev stroke:#0C0;
classDef triager stroke:#C0C;
linkStyle 0,1,7 stroke:#CC0,color:auto;
linkStyle 2,3 stroke:#00C,color:auto;
linkStyle 4,5,6,8,9,10 stroke:#0C0,color:auto;

Loading

I suspect the bug is here, in a handler of PR/issue modifications introduced in gh-555:

bedevere/bedevere/stage.py

Lines 127 to 131 in 0f350ce

	if pull_request.get("draft"):
	await _remove_stage_labels(gh, issue)
	else:
	blocked_on = Blocker.core_review if await util.is_core_dev(gh, username) else Blocker.review
	await stage(gh, issue, blocked_on)

We need to check whether assignment of Blocker.core_review or Blocker.review is allowed before executing the else branch.

Ideally (for some distant future) we need to move tag management into a separate .py file that reads existing tags, checks if a new one is applicable, and removes the old ones. However, I have no free time to do it currently (because it requires massive modification of both reactors and their tests).

cc @itamaro

[itamaro]

oops, you're right, this handler is doing way more than it was intended!
that handler should respond only to PR edits that change the PR from draft to published or from published to draft.

[arhadthedev]

@itamaro Could you submit a PR, please?

[itamaro]

@itamaro Could you submit a PR, please?

of course