Implement importlib.abc.InspectLoader in AssertionRewritingHook

[ronaldoussoren]

This is a first take at a PR for issue #7804, at this time without tests and documentation updates.

At this time I'm primarily interested to hear if the code changes are acceptable.

closes #7804.

---

Here is a quick checklist that should be present in PRs.

• Include documentation when adding new features.
• Include new tests or update existing tests when applicable.
• Allow maintainers to push and squash when merging my commits. Please uncheck this if you prefer to squash the commits yourself.

If this change fixes an issue, please:

• Add text like closes #XYZW to the PR description and/or commits (where XYZW is the issue number). See the github docs for more information.

Unless your change is trivial or a small documentation fix (e.g., a typo or reword of a small section) please:

• Create a new changelog file in the changelog folder, with a name like <ISSUE NUMBER>.<TYPE>.rst. See changelog/README.rst for details.

  Write sentences in the past or present tense, examples:

  • Improved verbose diff output with sequences.
  • Terminal summary statistics now use multiple colors.

  Also make sure to end the sentence with a ..

• Add yourself to AUTHORS in alphabetical order.

[bluetech]

This would also fix #6643. In there, @iwanb suggested a different approach of subclassing importlib.abc.FileLoader and importlib.abc.SourceLoader. I didn't check what it actually means, but can you comment on that?

[ronaldoussoren]

IMHO it is easier to just implement is_package. I'm not entirely happy with my implementation, it would be better to just cache the value (similarly to the _fullname_to_path attribute). I'll push an update to the PR later.

[ronaldoussoren]

I'm pretty sure that subclassing FileLoader and SourceLoader would require splitting the loader functionality from the rewriting hook. That wouldn't be too hard, but wouldn't buy much and requires more code.

[ronaldoussoren]

Something else: I'm not sure what's the best way to add tests for the new methods. I'm not a pytest user myself.

[asottile]

this comment is unnecessary -- the type checker will enforce this for us (remove the other one as well)

[asottile]

is this error possible? this shouldn't be needed in a PEP451 loader (the code retrieval step doesn't happen if the finder isn't returned) -- this makes me think that we've reverted back to 302

[asottile]

yeah InspectLoader is the 302 interface which would regress us back into imp land