Improve static analysis tests

[drupol]

This PR:

• Has static analysis tests (psalm, phpstan)

This PR depends on #161

[AlexandruGG]

👀

[AlexandruGG]

There has been a phpstan release today which is causing some new errors, I'll see if I can sort them out

[AlexandruGG]

There has been a phpstan release today which is causing some new errors, I'll see if I can sort them out

Psalm and PHPStan are having some disagreements now on this whole non-empty-string thing 🙄. I bet there's going to be a fix/tweak

[AlexandruGG]

could you use this commit ff9c4b6 as an example to add some "failure cases" for the other operations? If we are not "stress-testing" the type annotations I'm not confident that they are correct or that the static analysis tests are adding any value

[drupol]

@AlexandruGG I fixed the issue with Reduction, however, regarding the non-empty-string with phpstan I have some issues.
When I fix it, psalm complains... and when I don't phpstan complains!

Do you think we should fix the version of PHPStan until this is fixed upstream?

[AlexandruGG]

@AlexandruGG I fixed the issue with Reduction, however, regarding the non-empty-string with phpstan I have some issues.
When I fix it, psalm complains... and when I don't phpstan complains!

Do you think we should fix the version of PHPStan until this is fixed upstream?

Yeah I had that issue as well, yeah let's fix the version for a bit and see what happens

[drupol]

@AlexandruGG I fixed the PHPStan issues, now Psalm complains.

Which version of PHPStan do you advice to use?

[AlexandruGG]

I don't understand why this is different from Reduceable.

Same of ScanLeftable and ScanRightable, shouldn't they all be the same?

[drupol]

You're right, there was an inconsistency, I just fixed it.

However, ScanLeft and ScanRight might return the initial value if the Iterator is empty, this is why they might return V or W but Reduce and Reduction will not return anything, this is why it only returns W.
ScanLeft1 and ScanRight1 uses the first value of the iterator as initial value, so the signature has small differences.

[drupol]

Best news of the day! Thanks for the review!