missleading error on getfuncargvalue for parameterized fixture #460
Comments
Original comment by Matthias Geier (BitBucket: geier, GitHub: geier): I get the same error when trying to create a fixture that concatenates two or more other fixtures (see my question on stackoverflow). It's possible to have a fixture using a fixture using a fixture and so on, but it doesn't seem to be possible if parametrized fixtures are involved. |
Original comment by Matthias Geier (BitBucket: geier, GitHub: geier): Here's my code example (without classes) that shows the same error:
Is there a way to make this work? |
Original comment by Fayaz Khan (BitBucket: fayaz, GitHub: fayaz): In my case, the fixture seems to be failing because it shares name with a parameter on the test.
Try changing the name |
@RonnyPfannschmidt any reason this is listed as backward compatibility? Is request.getfuncargvalue deprecated? I'm hitting this issue with an on/off settings flipper fixture being used by a fixture that uses request.getfuncargvalue. A minimal test case, equivalent of the one above:
Result is same as above: |
that kind of usage should raise an error |
this is a rather bad case, |
Yeah I can see that using getfuncargvalue in a fixture can't start multiplying the cases, like params do. In fact it seems as difficult as supporting unittest subtests (#1367). I think a better error than |
its a completely different dimension than subtests ^^ subtests happen by design within a test this case is requesting something that should have been parametrized before the test within the test and without parameters |
Fair enough. But we're agreed on needing a better error message? |
correct |
Fixed by #1620 Please close. (I'll do the magic commit message that closes issues next time) |
thanks |
Originally reported by: Nikolaus Rath (BitBucket: nikratio, GitHub: nikratio)
The attached meta-test-case results in
with pytest from Mercurial tip.
The text was updated successfully, but these errors were encountered: