1.142.0rc1 regression fix: Allow coercing a str to a FilePath in MasConfigModel

[anoadragon453]

Otherwise MasConfigModel would always fail to validate, as the yaml parser would not produce a FilePath instance.

Could not validate Matrix Authentication Service configuration: Input should be an instance of Path

Missed in #19071.

Pull Request Checklist

• Pull request is based on the develop branch
• Pull request includes a changelog file. The entry should:
  • Be a short description of your change which makes sense to users. "Fixed a bug that prevented receiving messages from other servers." instead of "Moved X method from EventStore to EventWorkerStore.".
  • Use markdown where necessary, mostly for code blocks.
  • End with either a period (.) or an exclamation mark (!).
  • Start with a capital letter.
  • Feel free to credit yourself, by adding a sentence "Contributed by @github_username." or "Contributed by [Your Name]." to the end of the entry.
• Code style is correct (run the linters)

[MadLittleMods]

How was this problem caught?

Can we be sure it doesn't happen elsewhere? (feels like no without manually auditing things or test coverage to stress these pieces)

When having a cursory look, only this stood out to me,

synapse/synapse/config/matrixrtc.py

Lines 44 to 45 in 6790312

	class MatrixRtcConfigModel(ParseModel):
	transports: list = []

[benbz]

How was this problem caught?

I raised this after the CI for element-hq/ess-helm#853 started failing on RC3

[MadLittleMods]

This kinda feels like a bug with the pydantic.FilePath type 🤔. Of course it has to be parsed from a string. Feels like it should act like Annotated[StrictStr, AfterValidator(validate_file_path)]

May be worth finding existing issues tracking this or creating one.

[anoadragon453]

I can see both sides of the argument. On one hand, it would be a bit odd to allow coercion for one specific type (str -> FilePath) in strict mode. On the other, most people will be taking in paths as str types. There should be a simple way to validate that the file exists etc. (which is what people expect to do when setting the type of the field to FilePath).

So I agree that there's something either wrong/missing. I'll raise it as an issue in a bit while waiting for the release assets to build.

[anoadragon453]

Failing CI tests are flake and are unrelated to this PR.