[MRG] Str(e) to str(e.value)

[Matthew-Jennings]

No description provided.

[SimonBiggs]

I feel bad for taking you away from your holiday...

[SimonBiggs]

I still don't get what caused this... Maybe you can fill me in when we manage to catch each other on the phone :).

Oh? A breaking change in pytest? We may need to pin pytest >= the current version of that's the case, as that doesn't look backwards compatible.

[SimonBiggs]

There we go, that's the difference:

pytest version 5.0.0:

Changelog:

https://docs.pytest.org/en/latest/changelog.html#removals

Relevant issue:

pytest-dev/pytest#5412

[SimonBiggs]

@pchlap, @Matthew-Jennings has fixed this issue. It was due to an upgrade in pytest to 5.0.0. see the above comment. This can now be merged into your branch.

[Matthew-Jennings]

Yeah, I was thinking it had to be an update to a dependency. I was actually thinking Python itself, but hadn’t gotten around to checking the logs between passing and failing tests to interrogate package versions! Anyway, nice find :)

[SimonBiggs]

Oh? You didn't find that? How did you fix it without finding that?

[SimonBiggs]

Ahh, your prints :)

[Matthew-Jennings]

I re-checked the pytest.raises() example code and noticed it had changed from asserting x in str(excinfo) to asserting x in str(excinfo.value). But I’d seen examples of the former before, so I printed away, haha.

I hadn’t made the leap to recognising pytest had updated, I’d actually thought it was Python itself. I’d mistakenly thought excinfo Referred to Python Exception info, but now I see that was silly, it’s actually part of pytest.

[SimonBiggs]

I'll let @pchlap do the merge given it's his branch