ensure cythonized functions are left untouched

[kollmats]

Change Summary

Extending the check for untouched types to include a check for cython functions or methods.

Related issue number

Resolves #1943

[codecov]

Codecov Report

Merging #1944 into master will not change coverage.
The diff coverage is n/a.

@@            Coverage Diff            @@
##            master     #1944   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files           21        21           
  Lines         3909      3909           
  Branches       788       788           
=========================================
  Hits          3909      3909           

---

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 bf9cc4a...06b941e. Read the comment docs.

[samuelcolvin]

thanks so much for the contribution.

please can you add tests and a change description.

also, in future please fill in the template in pull requests.

[samuelcolvin]

bump

[samuelcolvin]

@kollmats please could you fixthis.

[kollmats]

Hi @samuelcolvin,

I'm sorry for this question since it may already be stated clearly somewhere, but where would be a good place to test this in your setup given that a proper test would require a cythonized application? Thanks.

[samuelcolvin]

all tests are run on pydantic when compiled with cython.

Just choose the file in tests/ that makes most sense.

[PrettyWood]

Made a test on #2208 just to see the CI fail with a test similar to the one in the issue and it didn't fail. I probably miss something

[samuelcolvin]

I see the problem, the point is that the code where get_double_a is defined (using the example from #2208) needs to be compiled.

It doesn't matter one way or the other whether or not pydantic itself is compiled.

I'm just seeing if I can add a simple test for this.

[samuelcolvin]

in the end I created another PR to fix this, see #2228. @kollmats I've thanked you in the change description, hope that is sufficient.

Thanks for this.