Support use of TypeVar on generic subclasses

[zpencerq]

This change was originally introduced to support generic functions on generic types defined for pydantic.

Notably, passing a TypeVar on a GenericModel subclass raises a TypeError (presumably?) meant for assisting users in defining their generic model classes.

Example:

import typing
from pydantic.generics import GenericModel

T = typing.TypeVar('T')
R = typing.TypeVar('R')

class Model(GenericModel, typing.Generic[T]):
  items: typing.List[T]

# Next line raises: 
# TypeError: Type parameters should be placed on typing.Generic, not GenericModel
def convert(model: Model[T], mapper: typing.Callable[[T], R]) -> Model[R]: 
  return Model[R](items=list([mapper(item) for item in model.items]))

Change Summary

• Allow use of TypeVar on GenericModel subclasses.

Checklist

• Unit tests for the changes exist
• Tests pass on CI and coverage remains at 100%
• Documentation reflects the changes where applicable
• changes/<pull request or issue id>-<github username>.rst file added describing change
  (see changes/README.md for details)

[codecov]

Codecov Report

Merging #842 into master will not change coverage.
The diff coverage is n/a.

@@          Coverage Diff          @@
##           master   #842   +/-   ##
=====================================
  Coverage     100%   100%           
=====================================
  Files          17     17           
  Lines        3024   3024           
  Branches      582    582           
=====================================
  Hits         3024   3024

---

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 6320d8c...ea31f58. Read the comment docs.

[dmontagu]

@zpencerq Yes, I noticed recently that this also prevents you from using bounds to restrict the TypeVars in subclasses. Thanks for contributing this.

You should be able to fix the formatting issues currently causing tests to fail by running make format from the pydantic root folder.

You can also run a plain make to check that the CI checks will pass.

(More detail here: https://pydantic-docs.helpmanual.io/#contributing-to-pydantic)

[zpencerq]

Should I open this against (or change base to) v0.32.x as well?

[dmontagu]

@zpencerq I think just against master is fine — I think @samuelcolvin is hoping to release a beta this week.

[samuelcolvin]

thanks so much.