Florian Bruhin

The teardown part of fixtures is always called, even if there was an exception in the test - so I'm afraid I don't really see how this changes anything over...

Taking the freedom to close this, as I believe the current code already works correctly, and I don't see any benefit in changing it.

That seems like it would now make the error message confusing for the far more common case, when there indeed is a test which wasn't found...

Superseded.

One thing that seems somewhat suspicious is how IPython's `pytest_ipdoctest.py` seems to have some custom code around importing conftest files: https://github.com/ipython/ipython/blob/d7cc335a36cdfe4159c9b88c94311866c923a547/IPython/testing/plugin/pytest_ipdoctest.py#L635-L646

Not sure if this should be closed, @Zac-HD - it seems like an useful proposal at first sight.

Why a comment rather than a "needs backport" label? Would be nice IMHO if the backport was done automatically if there was such a label, and then the label gets...

I suppose we could also use an existing GitHub Action, such as ~~https://github.com/Gaurav0/backport~~ (requires to make your API token public, wut...) https://github.com/tibdex/backport

> https://github.com/tibdex/backport seems like a good fit, although it doesn't seem to require a public API token, it uses a secret, so it seems it would work with our existing...

@thedrow the problem is that Hypothesis won't know how many tests it's going to run during the collection phase, it'll only know while running the tests. See https://github.com/pytest-dev/pytest/issues/916