Does not unmark when activity on PR when using keywords pulls: and issues:

[johlju]

In this PR dsccommunity/xSmbShare#20 (comment) it did not remove the stale label 'abandoned' when there was activity on the PR (see style.yml). Looking at the code it seems is not using the `this.getConfigValue(type, 'staleLabel')' or similar to get the correct value.

stale/index.js

Lines 36 to 37 in 628cbc9

	const staleLabelAdded = context.payload.action === 'labeled' &&
	context.payload.label.name === stale.config.staleLabel

@bkeepers Can I get your input on this one? Could this be the problem?

[hiimbex]

The line you linked to is checking to make sure the event wasn't stale bot adding a label that would otherwise count as activity. So in your case this context.payload.action === 'labeled' would be false since the actions would have been issue_comment.created.

The code that actually does the unmarking is here, but after looking through this for a bit, I don't immediately see the issue. Let me know if you can reproduce this anywhere.

[johlju]

@hiimbex Thanks for looking at this - example of it not being removed can be seen here johlju/DebugApps#16. Just commented on the PR but the label is not getting removed. Feel free to test/debug anyway you want in this repo, it is just a "lab" for testing/verifying Apps.

[hiimbex]

Hmmm just guessing, but looking at your stale.yml I see https://github.com/PowerShell/xSmbShare/blob/dev/.github/stale.yml#L7 in which you never plan to close the PR, so maybe that's causing this unmark run never to happen?

I'll try to look into this more shortly.

GitHub

PowerShell/xSmbShare

Contribute to xSmbShare development by creating an account on GitHub.

[johlju]

It could be, because it unmarks for issues, which is configured to be closed.

[johlju]

@hiimbex It looks like it actually works for pushes, but not other activity (what I have seen yet)? 🤔

I saw these below, where Stale actually correctly removed the abandoned label (both when user pushed).
dsccommunity/FailoverClusterDsc#177 (comment)
dsccommunity/SqlServerDsc#1052 (comment)

So isn't seeing the correct (all) events for example when commenting? 🤔

[johlju]

@hiimbex After seeing this issue a few times more, it seems Stale removes the "stale" label ('abanonded' in my case) when there is a push to a PR, but it seems it does not remove the "stale" label when there is a "pull request comment" event.

A comment was made here, but Stale did not remove the label.
dsccommunity/CertificateDsc#134 (comment)

[stale]

Is this still relevant? If so, what is blocking it? Is there anything you can do to help move it forward?

[johlju]

This is still an issue since the abandoned label is not removed when commenting on a pull request (as per my previous comment).

Updating so this issue is active.

[hiimbex]

Sorry I haven't had time to look into this further. If you'd like to investigate and open a PR, I'd be happy to review it, otherwise I don't think it's likely this will be prioritized in the near future.

[infotexture]

@hiimbex wrote:

Let me know if you can reproduce this anywhere.

I've seen similar behavior on issues in the DITA Open Toolkit repository where the Stale bot is enabled, but does not reliably remove labels on activity, so the problem is apparently not limited to PRs.

In dita-ot/dita-ot#2286 (comment), the Stale bot added a comment and label to an old issue as expected, but failed to remove the stale label a few days later when the original poster added a comment dita-ot/dita-ot#2286 (comment), which should have removed the label according to the Probot description:

If the Issue or Pull Request is updated, or anyone comments, then the stale label is removed and nothing further is done until it becomes stale again.

Since the label was not removed, the issue was later closed although it shouldn't have been.

One of the project maintainers subsequently re-opened the issue, which the Stale bot noticed and removed the stale label.

I double-checked the configuration, but don't see anything that would seem to prevent Probot from removing the the stale label when comments are posted.

— Any ideas?

[max-sixty]

We're seeing the same bug here: pydata/xarray#2797. It's happened on multiple issues.

I'm looking at the configurations between xarray and @infotexture to see if there's some similarity. We both use the stale label, which is not default. Could that be a cause?

Thanks for the awesome tool

[stale]

Is this still relevant? If so, what is blocking it? Is there anything you can do to help move it forward?

[johlju]

Still relevant. Someone in the community just need to help resolve this. 😊

[max-sixty]

We haven't seen this again FWIW

[stale]

Is this still relevant? If so, what is blocking it? Is there anything you can do to help move it forward?