Compare strings with strict equality operators

[thefourtheye]

Previous effort: #2392

CI Run: https://jenkins-iojs.nodesource.com/job/node-test-pull-request/189/

[seishun]

Why? If it doesn't change the behavior, then it's pointless code churn. If it does, then it's a breaking change.

[rvagg]

agreed, I'm somewhere between -1 and -0 on this, unless there's a measurable performance gain from doing this then I'm not a fan, sorry for the second negative comment in a row @thefourtheye

[thefourtheye]

@seishun@rvagg No problem :-)

[ronkorving]

lgtm, it's one step in the direction of unification (style, if nothing else).

[seishun]

I'm not convinced that the majority agrees on the "use === unconditionally everywhere" style. Either way, it seems very unlikely it will ever be enforced throughout the code base anyway, so why apply it partially?

[pmq20]

@seishun if it were a breaking change then it would be picked up by the CI. And I think the code churn is not pointless, because == i.e. comparison w/ type conversions could easily introduce bugs and therefore should be avoid.

[ronkorving]

Also, he applied it partially because last time he did it not-so-partially and got bad feedback on that.

[rvagg]

@seishun if it were a breaking change then it would be picked up by the CI.

Not true, we don't have 100% coverage and we certainly don't test anything near the number of weird ways people use Node and are expecting it to behave. We have a diverse ecosystem and are always running in to edge-cases and subtle bugs encountered by novel uses.

[pmq20]

@rvagg Agreed with you. 👎 on the PR then.

[cjihrig]

Also -1. Just make changes like this when you're modifying the code for more meaningful reasons.

[Fishrock123]

Yeah... it's hard to agree to change these once they are there. We should just make sure we always land with strict equality, unless explicitly sating why otherwise.