Fix frozenset deserialization.

[pchanial]

Closes #341

Note that there is another unit test failure : the deserialization of BigInt is not a BigInt instance. For this test case, I've explicitly disabled the new type check since it is another issue.

[wyfo]

There is no need to rewrite a complete method for frozenset; ListMethod should be reused, as it's already the case for variadic tuple

[wyfo]

By the way, I think super() was not supported (in Cython generation) when I wrote VariadicTupleMethod, but this one, as well as FrozenSetMethod should simply inherit ListMethod and use for example return tuple(super().deserialize(data))

[pchanial]

After inspection, it is not straightforward, since VariadicTupleMethod could inherit from ListCheckOnlyMethod or ListMethod.
It looks like all collection deserialization methods could benefit from only checking their item values, so I'd say that it would be better to do some refactoring here.

[wyfo]

You're right, it was not a question of super() support … I was too quick in my response. So composition should be used the same way VariadicTupleMethod does.

[codecov-commenter]

Codecov Report

Merging #342 (7c47d6d) into master (af1bf4a) will decrease coverage by 0.02%.
The diff coverage is 80.95%.

@@            Coverage Diff             @@
##           master     #342      +/-   ##
==========================================
- Coverage   86.13%   86.11%   -0.03%     
==========================================
  Files          67       67              
  Lines        5879     5898      +19     
  Branches     1218     1222       +4     
==========================================
+ Hits         5064     5079      +15     
- Misses        596      599       +3     
- Partials      219      220       +1     

Impacted Files	Coverage Δ
apischema/deserialization/methods.py	84.51% <76.47%> (-0.21%)	⬇️
apischema/deserialization/__init__.py	98.98% <100.00%> (+<0.01%)	⬆️

---

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 af1bf4a...7c47d6d. Read the comment docs.

[wyfo]

@pchanial
Actually, I can merge this PR like this, and do the small refactoring myself if you're ok with that. I would like to release the patch tomorrow morning (I don't like having a bug more than 24h in the issues).
If you want to make the change yourself, I will wait, no worry.

[pchanial]

I had made the change before your last comment, so I just pushed it. I have started looking at ways to refactor this part but the tests fail with cython. I'll make another PR so we can discuss it.