feat: added the prefer-strict-boolean-matchers rule

[marekdedic]

I am explicitly not reporting for .toEqual(true) or .toStrictEqual(true), there's already prefer-to-be for that.

Closes #639

[marekdedic]

Additionally, IMO, this should be in the recommended ruleset, but that should probably come in a major version down the line

[veritem]

thanks!

[FloEdelmann]

Nice, thanks for adding this! Unfortunately, the rule is not documented in docs/rules or in the readme. Could you please add this?

[veritem]

aah, I forgot about that. @marekdedic, could you please add documentation for this rule? the docs generator is broken due to dependency issues. Feel free to look at how other rules are documented as a reference.

[marekdedic]

Ok, will add it in a separate PR :)

[marekdedic]

Hmm, I just realized I added an auto-fix for this rule, but it just replaces .toBeTruthy with .toBe(true) - which can make some tests fail if they rely on type coercion. Is that OK?

[marekdedic]

See #653

[veritem]

Hmm, I just realized I added an auto-fix for this rule, but it just replaces .toBeTruthy with .toBe(true) - which can make some tests fail if they rely on type coercion. Is that OK?

i think that's okay, did you want todo otherwise?

[marekdedic]

i think that's okay, did you want todo otherwise?

No, I'm fine with it, just wanted to make sure it's ok for the autofix to potentially break tests :) If so (I think it's warranted here), then everything's ok :)

[gtbuchanan]

The introduction of this rule in the all config causes an infinite recursive fix situation. Using toBe(true) triggers prefer-to-be-truthy, but changing it to toBeTruthy() results in prefer-strict-boolean-matchers.

[marekdedic]

Hi, yes, that is true, they are basically opposites of each other. @veritem it seems to me that the all config is automatically updated, so I don't know how to fix this :/

[veritem]

@marekdedic
we can disable this rule by default and have users enable it when they want. check in src/index.ts and change this rule from warn to off

[marekdedic]

Ok, that makes sense. One question though: Wouldn't it be better to have this rule on by default instead of prefer-to-be-truthy? This rule forces the users to be more precise with their matchers, lowering the chance of a misbehaving test. The other one does the exact opposite and I honestly see it as bad advice and a bit of an antipattern...

[FloEdelmann]

I agree, prefer-to-be-truthy and prefer-to-be-falsy are less useful than this new rule.

[veritem]

prefer-to-be-truthy has been turned off by default in favor of this rule

[marekdedic]

Aaah, sorry, prefer-to-be-falsy has to be turned off as well, sorry, I didn't write that down exactly...

[veritem]

done