Allow '==' to check for null and undefined in one go.

[ludofischer]

Using == against null is a common pattern (see here for example). Since ESLint caters to it, should it not be allowed? Relevant ESLint documentation: http://eslint.org/docs/rules/eqeqeq

[ljharb]

@ludovicofischer are you still interested in landing this PR? if so, please rebase it on top of latest master. I'd be happy to land it.

[ludofischer]

Hi, thanks for the heads up.

[ljharb]

@ludovicofischer I ended up committing it because I'm hoping to release soon :-) thanks for the contribution!

[spikebrehm]

Thanks @ljharb!

[mhofman]

Any reason not to go all the way and allow smart equals? basically same as allow-null with relaxation when the type is known, e.g. literals comparison or more importantly typeof

[ljharb]

@mhofman yes - that setting presents a refactoring hazard when one is changing a literal to a variable. == null is special, and otherwise, === should be used without exception.

[mhofman]

Well the lint rule should still catch it if the refactoring change didn't also change the == to === ;)

I don't care much for literals, but I find === overkill for typeof use cases.

[ljharb]

yes that is a fair point. but == null makes me uncomfortable enough; I don't want to train developers that == is ever OK.