feat(neon_framework): use logging framework instead of print

[Leptopoda]

implements: #608

[provokateurin]

I guess we should have a script that checks that print and debugPrint are never used in any of the neon_* packages?

[Leptopoda]

Not sure if you already had a look at the latest commit but I've tried to address the suggestion from your other #1642 (comment) I've opted to use the custom_lint package which provides an abstraction to write custom lint rules using the analyzer api (similar to what the analyzer itself uses).

I had to bump the project sdk version to 3.2.0 (not the for the packages though) and the analyzer plugin api is still experimental. You can read more about it here https://dart.dev/tools/analysis#plugins.
There are some limitations like recommended 16GB of RAM and and not more than 10 packages in a monorepo. I'd say let's see how far we can go with it and if desired we can still switch it to regexes.

A benefit of this approach is that we can write custom quickFixes for things like dart:io=> universal_io, higher accuracy as we can use the analyzer directly and direct feedback while developing (I never run these scripts before submit 😅)

[provokateurin]

Yeah I really liked that you added this, it looks like a pretty cool feature to use. I don't like that we have to use a custom script to do the analyzing, but I guess that's a trade off worth it.

[provokateurin]

Two things:

1. In a lot of places where errors are caught only the info or warning levels are used, but they should be severe instead. I know that some places can stay info or warning, but most should really be severe.
2. My IntelliJ doesn't pick up the custom linting rule, do I need to change something so it works? (Yes I restarted the analyzer and the whole IDE).

[Leptopoda]

I don't get any linting in my IDEA (not set up correctly) but in android studio it works.
Have you also activated the custom_lint package globally? Maybe run the setup script again.

[Leptopoda]

In a lot of places where errors are caught only the info or warning levels are used, but they should be severe instead. I know that some places can stay info or warning, but most should really be severe.

This is intentional. We did handle the error in the catch clause so just a warning that a process went wrong. I'd only use severe in the global error handler for uncaught errors.
I quite like the graphic in this post https://stackoverflow.com/a/64806781.

[provokateurin]

Have you also activated the custom_lint package globally? Maybe run the setup script again.

I'm not sure what I did before, but now I ran the setup script, restarted IntelliJ and still nothing.

[provokateurin]

I quite like the graphic in this post stackoverflow.com/a/64806781.

Can you add this to our docs? I like the graphic as well.

[Leptopoda]

where would this fit?
Maybe a "how to write a client" where we explain package choices and such conventions? Or just a code_style.md?

[provokateurin]

LGTM, just some minor things and you added the requested changes to the wrong commit.

[provokateurin]

Thanks a lot ❤️

Can you add the simple test for the lint?