Penelope Doesn't Skip "Wrong" Lint Rules

I marked them wrong in the Critic Browser, which now skips them, but the validation report still shows them as Warnings. Anyone know why? I see the versions in the inbox references by the slice have their manifest classes...

The monkey does a first run of tests/critics with the current pharo image ans store failed tests/rules. Then, it loads the slice and does a second run. At the end, issues reported are only the new failed tests/ rules.

----- Sean P. DeNigris <[hidden email]> a écrit :

demarey wrote

At the end, issues reported are only the new failed tests/ rules.

Yes, but it's reporting new "failed" tests that are marked wrong (and are properly skipped by the Critic Browser)

I think the manifest is not read at all by the monkey. It runs a predefined set of rules whatever is in the manifest.

----- Sean P. DeNigris <[hidden email]> a écrit :

Ben Coman wrote

How about we open an issue for this, and temporarily introduce a new package CaseNNNN for
testing this, where Sean could reproduce his case.

I have also experienced what you're talking about - that Lint rules were failing that were not introduced with my fix; but I also was unable to properly document them. Just to be clear, this is a different problem - namely, that manifests declarations of wrong rules are not honored i.e. they are still reported as errors/warnings. As Nicolai pointed out in a comment in issue 15507, maybe this is a good thing because human reviewers may want to check that they were properly marked "wrong". But ideally, they would be reported as two separate categories - 1. new failures not marked "wrong" and 2. new failures so marked