Add explicit serialization / deserialization to the JsonReporter

[Pierre-Sassoulas]

Type of Changes

	Type
✓	✨ New feature

Description

Necessary for Refs #7000

[coveralls]

Pull Request Test Coverage Report for Build 2581426689

• 25 of 25 (100.0%) changed or added relevant lines in 2 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.006%) to 95.331%

---

Totals
Change from base Build 2580923231:	0.006%
Covered Lines:	16680
Relevant Lines:	17497

---

💛 - Coveralls

[DanielNoord]

@Pierre-Sassoulas What is the bug/issue we're trying to solve here?

[Pierre-Sassoulas]

The goal is to be able to use the message object in the primer instead of a complicated to type dict from the JSON reporter. I identified the issue of confidence and abspath that were lost in the reporting thanks to this.

[DanielNoord]

Should we consider adding those? Adding new keys to a JSON result shouldn't really break anything right?

[Pierre-Sassoulas]

It's a breaking change depending on the implementation ? Say if someone loop on it and try to assign each value to something and it error on unknown value.

[DanielNoord]

It's a breaking change depending on the implementation ? Say if someone loop on it and try to assign each value to something and it error on unknown value.

Hmm yeah. We really should provide a json-version-2, but I haven't gotten to it yet.

[Pierre-Sassoulas]

We really should provide a json-version-2, but I haven't gotten to it yet.

Silly me, we just have to add a JsonReporter and deprecate the old one ?

[DanielNoord]

Silly me, we just have to add a JsonReporter and deprecate the old one ?

Yeah, but we also need to fix some of the issues that are currently in the beta of 3.0. That's why I haven't gotten to it.

[Pierre-Sassoulas]

Hmm ok. Seems we can deal with it later and merge this one then ?

[DanielNoord]

Hmm ok. Seems we can deal with it later and merge this one then ?

Well, this is a bit of a handicapped solution. I'd prefer tackling the actual problem, but if you don't have the time for that now I guess this should would be okay?

[Pierre-Sassoulas]

I'm not trying to fix the missing confidence / abspath in the json right now, it's just an issue I realized existed while doing what's necessary to use pylint.message.Message instance in the primer (which is itself an issue I realized existed when trying to do a generic comparator, which is a refactor I'm doing because I want to have a better diff in the primer when messages did not move 😄).

[Pierre-Sassoulas]

@DanielNoord can you review this again please ? I added the typed dict :) I have multiple PR ready as follow-up but I don't want to put too much into this one.

[Pierre-Sassoulas]

TypedDict is not available in some python version this is why i added it in the type checking guard.

[DanielNoord]

I pushed some changes which were a bit hard to suggest using the interface.

Let's make the types of the serialisation more explicit and add types of we need to.

[DanielNoord]

TypedDict is not available in some python version this is why i added it in the type checking guard.

Whenever typing stuff is unavailable on older versions we tend to put those under sys.version guards as then we'll remember to remove them when the guard is no longer needed.

[github-actions]

🤖 According to the primer, this change has no effect on the checked open source code. 🤖🎉

This comment was generated for commit 0464370