Should no-redeclare also check functions?

[dominicbarnes]

I just got burned by a stupid mistake, but I think it should've been caught by no-redeclare.

Given the following code:

var a = true;
function a() {}

If I turn on no-redeclare: 2, this is not treated as an error, but I think it should.

To be clear, this was a stupid mistake on my part. I added a var to the top of a file that was named the same as a function declared much further down in my code. Nonetheless, I think no-redeclare should take hoisted functions into account as well.

[gyandeeps]

I know there have been conversations around this. Have you tried using no-shadow rule.
Related discussions happened here: #2645

[dominicbarnes]

This is not an issue no-shadow will catch, since this all happened in the "global scope" (not truly global, but it wasn't in a nested function of my module)

[gyandeeps]

What all env are you using in your eslint config?

[dominicbarnes]

In this particular case, I'm using browser (plus all ES6 features)

[nzakas]

Yeah, most likely a bug even though the docs would indicate otherwise. We should fix this and update the docs to be more accurate.

[mysticatea]

I tried:

> eslint -v
v0.24.0
> echo "var a = true; function a() {}" | eslint --stdin --env browser --env es6 --rule "no-redeclare:2" --reset --no-eslintrc

<text>
  1:24  error  a is already defined  no-redeclare

✖ 1 problem (1 error, 0 warnings)

If you have turned ecmaFeatures: {modules: true} on, yes, the no-redeclare rule had been missing global declarations. And it was fixed on master branch now. (I want to turn modules on at one-liner...)

Related: #2903

[dominicbarnes]

I can verify that master has resolved this. Should I close this issue?

[gyandeeps]

If you have verified this then you can close.

[dominicbarnes]

Actually, could we add tests for no-redeclare that include checking named/hoisted functions? (just noticed all the examples there are either var or class) For example:

var a; function a() {}