parametrize fails when values are unhashable

[pytestbot]

Originally reported by: Ronny Pfannschmidt (BitBucket: RonnyPfannschmidt, GitHub: RonnyPfannschmidt)

---

seems like a recent change in parametrizing disallows hashable items
this broke the hgdistver testsuite

---

• Bitbucket: https://bitbucket.org/pytest-dev/pytest/issue/244

[pytestbot]

Original comment by holger krekel (BitBucket: hpk42, GitHub: hpk42):

---

Can you create a little example?

[pytestbot]

Original comment by Ronny Pfannschmidt (BitBucket: RonnyPfannschmidt, GitHub: RonnyPfannschmidt):

---

removed, mistake

[pytestbot]

Original comment by Ronny Pfannschmidt (BitBucket: RonnyPfannschmidt, GitHub: RonnyPfannschmidt):

---

#!python

archival_mapping = {
    '1.0': {'tag': '1.0'},
    '1.2.2a1': {'tag': 'release-1.2.2a1'},
}
import pytest
@pytest.mark.parametrize('expected data'.split(),
                         archival_mapping.items(),
                         archival_mapping)
def test_archival_to_version(expected, data):
    pass

[pytestbot]

Original comment by Mark Tsuchida (BitBucket: marktsuchida, GitHub: marktsuchida):

---

This might not add anything, but I've encountered the same problem with pytest 2.3.4 and Python 3.3.0, using the following example.

#!python

import pytest

@pytest.fixture(params=[[1, 2], [3, 4]])
def myfixture(request): 
    pass

def test_nothing(myfixture):
    pass

[pytestbot]

Original comment by Anatoly Bubenkov (BitBucket: bubenkoff, GitHub: bubenkoff):

---

working on this

[pytestbot]

Original comment by holger krekel (BitBucket: hpk42, GitHub: hpk42):

---

@bubenkoff any progress?

[pytestbot]

Original comment by Anatoly Bubenkov (BitBucket: bubenkoff, GitHub: bubenkoff):

---

I mentioned this before, the case worked, so test just passes
test was merged i think

[pytestbot]

Original comment by holger krekel (BitBucket: hpk42, GitHub: hpk42):

---

my memory is partially outsourced, didn't you know?

More seriously, someone showed me a failure at PyconDE that looked much like this issue. This reproduces it:

#!python

import pytest

def pytest_generate_tests(metafunc):
    metafunc.parametrize("x", [{},{}], indirect=True)

@pytest.fixture
def x(request):
    return request.param

def test_hello(x):
    assert x

[pytestbot]

Original comment by Anatoly Bubenkov (BitBucket: bubenkoff, GitHub: bubenkoff):

---

here's the PR i did to reproduce
https://bitbucket.org/hpk42/pytest/pull-request/68/parametrize-fails-when-values-are/diff

[pytestbot]

Original comment by Anatoly Bubenkov (BitBucket: bubenkoff, GitHub: bubenkoff):

---

https://bitbucket.org/hpk42/pytest/pull-request/68/parametrize-fails-when-values-are/diff
here's the MR i did to reproduce

[pytestbot]

Original comment by holger krekel (BitBucket: hpk42, GitHub: hpk42):

---

it's not using the "indirect" parameter which seems to be the failing case currently.

[pytestbot]

Original comment by Anatoly Bubenkov (BitBucket: bubenkoff, GitHub: bubenkoff):

---

but indirect wasn't mentioned in this issue at all

[pytestbot]

Original comment by holger krekel (BitBucket: hpk42, GitHub: hpk42):

---

sure, to be clear, i am not blaming you for anythying. Just noting that there still is an issue that fits under the title. If you can't/don't want to care i can take a look.

[pytestbot]

Original comment by Anatoly Bubenkov (BitBucket: bubenkoff, GitHub: bubenkoff):

---

no, im not trying to quit, i just wasn't aware that it's an issue only when using indirect
actually i even wrote a fix (before running test)
but then i reverted it as test passed anyway :)

[pytestbot]

Original comment by Anatoly Bubenkov (BitBucket: bubenkoff, GitHub: bubenkoff):

---

so in conclusion, then i'll continue on this :)

[pytestbot]

Original comment by holger krekel (BitBucket: hpk42, GitHub: hpk42):

---

great :)

[pytestbot]

Original comment by holger krekel (BitBucket: hpk42, GitHub: hpk42):

---

@bubenkoff are you still up for the issue? If not please drop a quick note here.

[pytestbot]

Original comment by Anatoly Bubenkov (BitBucket: bubenkoff, GitHub: bubenkoff):

---

sorry, still in todo
will try to finish this week

[pytestbot]

Original comment by holger krekel (BitBucket: hpk42, GitHub: hpk42):

---

@bubenkoff can you try to resolve this issue or un-assign yourselve until thursday?

[pytestbot]

Original comment by Anatoly Bubenkov (BitBucket: bubenkoff, GitHub: bubenkoff):

---

https://bitbucket.org/hpk42/pytest/pull-request/88/parametrize-fails-when-values-are/diff closes this

[pytestbot]

Original comment by holger krekel (BitBucket: hpk42, GitHub: hpk42):

---

This was fixed by Anatoly and me on trunk.