CI: fix feather test

[jbrockmendel]

No description provided.

[jorisvandenbossche]

Or we can just stop testing the message here?

[jbrockmendel]

i would be on board for that, cc @MomIsBestFriend looks like there is a new external_error_raised; is that for roughly this purpose?

[ShaharNaveh]

i would be on board for that, cc @MomIsBestFriend looks like there is a new external_error_raised; is that for roughly this purpose?

@jbrockmendel Yes, exactly!

[jbrockmendel]

updated to use external_error_raised. Do we want to check anything about the message being reasonable?

[ShaharNaveh]

updated to use external_error_raised. Do we want to check anything about the message being reasonable?

I don't think we do, as this is an external error message.
But I would like to hear others as well.

[WillAyd]

Thanks @jbrockmendel yea I don't think we need to test for 8 different messages!

[lumberbot-app]

Owee, I'm MrMeeseeks, Look at me.

There seem to be a conflict, please backport manually. Here are approximate instructions:

1. Checkout backport branch and update it.

$ git checkout 1.0.x
$ git pull

2. Cherry pick the first parent branch of the this PR on top of the older branch:

$ git cherry-pick -m1 6d3cc14fc52685511329cedf1e11d15651ac6a8e

3. You will likely have some merge/cherry-pick conflict here, fix them and commit:

$ git commit -am 'Backport PR #31791: CI: fix feather test'

4. Push to a named branch :

git push YOURFORK 1.0.x:auto-backport-of-pr-31791-on-1.0.x

5. Create a PR against branch 1.0.x, I would have named this PR:

"Backport PR #31791 on branch 1.0.x"

And apply the correct labels and milestones.

Congratulation you did some good work ! Hopefully your backport PR will be tested by the continuous integration and merged soon!

If these instruction are inaccurate, feel free to suggest an improvement.

[WillAyd]

Hmm I guess that helper isn't available on the 1.0.x branch

[jreback]

why exactly was this merged @WillAyd @jbrockmendel @jorisvandenbossche was this fixing a CI issue?

it is not about testing for the error messages, but the fact that the error messages are different that is extremely troubling. This is de-facto removing legitimate tests.

[WillAyd]

Yea looks to have failed CI

[WillAyd]

I think alternately would have to add a 9th error message to test for with current design; maybe should break up and test error message specifically by reader?

[simonjayhawkins]

msg4 doesn't look like an external message. should external_error_raised fail this? @MomIsBestFriend

[simonjayhawkins]

not at a computer to check though. so could be external. just looks like one of ours.

[jorisvandenbossche]

msg4 doesn't look like an external message.

Yes, I think that is the only one raised by ourselves (it's "path_or_buf needs to be a string file path or file-like"). But it's actually also a wrong error message we were testing .. as what is passed is a string file path)

[ShaharNaveh]

msg4 doesn't look like an external message. should external_error_raised fail this? @MomIsBestFriend

@simonjayhawkins tm.external_error_raised is an alias for
pytest.raises(<An exception class>, match=None).

It does not check for whether or not this is an actual external error message.