Refactor checker error messages

[tushar-deepsource]

Description

Migrates checker.py to use the ErrorMessage class.

This is a continuation from the discussion over at #10947

Some notes:

• Since the type signature of fail cannot be changed altogether, it has been overloaded using a Union to support ErrorMessage objects as well.
• Some fail calls are still using strings instead of ErrorMessage objects. This is because some of the strings were re-used in other modules, which will be refactored in later PRs.
• Other PRs will be smaller than this one :)

[TH3CHARLie]

Can you fix the style error (and also in the other PR)? Then I can review both PRs once all builds are passed.

[tusharsadhwani]

@TH3CHARLie I'm unsure why the mypyc build is failing. For some reason it fails to determine the type of ErrorCode objects being passed inside message_registry.py and I don't know how to fix it. Any clues?

[TH3CHARLie]

How about providing explicit type like Final[ErrorMessage]?

[tushar-deepsource]

Oh wow, so that didn't work but setting the ErrorCode objects as Final[ErrorCode] worked. What?

Anyway, thanks for that.

[TH3CHARLie]

Generally LGTM. I notice that some ErrorMessages have no associated codes at this moment, will you plan to provide them later?

[tushar-deepsource]

I notice that some ErrorMessages have no associated codes at this moment, will you plan to provide them later?

Yup. Current set of PRs are pure refactors. After this is done, I'll be adding unique error codes for each and every error, like most other linters have.

[TH3CHARLie]

Cool, looking forward to that

[emmatyping]

Could you please resolve the merge conflict?

[tushar-deepsource]

Done :)

[tushar-deepsource]

I have run mypy_primer locally, and there are no differences in the latest version of this branch :)

Command ran:
mypy_primer --repo https://github.com/deepsourcelabs/mypy.git -o diff --new checker-messages --old master

[tusharsadhwani]

@ethanhs @TH3CHARLie Please take another look, anything left for me to do?

[emmatyping]

The mypy primer output doesn't seem right if this is just a refactor, especially since graphql-core seems to have a bunch of type ignores become valid again...

[tusharsadhwani]

@ethanhs I added a few more commits which I believe fixes the output, checked by running mypy-primer locally. is there some way to re trigger the action?

[tushar-deepsource]

Alright, tidied up the final few bits.
@ethanhs hopefully this is the last time you'll have to trigger the workflows again

[tushar-deepsource]

I was 45 commits behind master... does that explain the primer difference?
Plus I missed an explicit annotation for TRUTHY_BOOL. mypyc should run fine now.

[tusharsadhwani]

@ethanhs PTAL?