Fix #165: empty parameter set error

[plammens]

The only public way to access the correct config object seems to be through the pytestconfig fixture, however I don't know how to get a hold of that outside a test function. So I used get_config from _pytest.config. No idea whether this is the correct approach but at least the test passes 😅

---

Update:

After taking a walk through pytest code with the company of a debugger, it seems the point at which the config object is constructed is here:

https://github.com/pytest-dev/pytest/blob/ad65e816e4612af4413711200f9e58d74e765a3a/src/_pytest/config/__init__.py#L143

Hence probably _prepare_config is the way to go? For now this PR adds a cached lazy _get_config static method to MiniMetafunc that calls _prepare_config. Ideally though it would be best to reuse the config object that pytest has already computed, but I have no idea how to get that from the MiniMetafunc code.

Fixes #165

[codecov-io]

Codecov Report

Merging #166 (246ba8a) into master (600b294) will decrease coverage by 0.02%.
The diff coverage is 75.00%.

@@            Coverage Diff             @@
##           master     #166      +/-   ##
==========================================
- Coverage   91.22%   91.19%   -0.03%     
==========================================
  Files         130      131       +1     
  Lines        5025     5031       +6     
==========================================
+ Hits         4584     4588       +4     
- Misses        441      443       +2     

Impacted Files	Coverage Δ
pytest_cases/tests/cases/issues/test_issue_165.py	60.00% <60.00%> (ø)
pytest_cases/common_pytest.py	88.46% <100.00%> (+0.03%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 600b294...56330d7. Read the comment docs.

[smarie]

Thanks @plammens for this PR !

I did not yet bother using the config object for now, it should be actually used everywhere where I generate ids instead of pytest (my own mini_idvalset and mini_idval methods in common_pytest.py).

So I wonder, maybe there is a way to get this globally instead of hacking in the pytest internals ?

• the easiest way to get the config seems to simply add this hook in our plugin.py: https://docs.pytest.org/en/stable/reference.html#pytest.hookspec.pytest_configure . Actually I already added it :) so it can be as simple as storing the object in a module variable when the hook is called, so as to be accessed later by MiniMetaFunc. Do you wish to try this ?

• there seem to be a fixture available https://docs.pytest.org/en/stable/reference.html#pytestconfig, but not sure it was there in pytest 2 so we'll need to backport it (or not support it in old versions). EDIT: actually it is not possible to access a fixture from the inside of a decorator call (since fixtures are not yet initialized), so this is probably not the way to go.

[plammens]

• the easiest way to get the config seems to simply add this hook in our plugin.py: https://docs.pytest.org/en/stable/reference.html#pytest.hookspec.pytest_configure . Actually I already added it :) so it can be as simple as storing the object in a module variable when the hook is called, so as to be accessed later by MiniMetaFunc. Do you wish to try this ?

That seems way better, I'll take a look!

[plammens]

Allright, I used your suggestion above: when the pytest_configure hook is called, I set a global module attribute PYTEST_CONFIG which is later retrieved by MiniMetafunc. Take a look to see if this is what you intended, @smarie! Also rebased onto the latest master.

[smarie]

Perfect ! Nicely done, thanks @plammens