Linters, anyone?

[dead-claudia]

I think it would be fantastic if Mithril was actually covered by a linter. I have no religious objection to any linter, as long as it parses JavaScript, but things like #662 are easily preventable with a linter.

Other things I've noticed:

• Lines spanning 150+ columns (This line literally can't fit on my screen, for example)
• Very inconsistent semicolon usage (~50% of the statements have them)
• A few long and/or complex boolean expressions that probably could stand to be wrapped (ties in with the first point)
• A few cases where the code is doing the work of the minifier. UglifyJS is smart enough to know to minify many of these kinds of things (among others):

// In
if (foo) {
  var bar = doA();
} else if (baz) {
  var bar = doB(baz);
} else {
  var bar = doC();
}

// Out
var bar=foo?doA():baz?doB(baz):doC();

[kaukas]

+1.

Actually +10! Automatic JSHint and JSCS checking helps clean up code and simplifies contributions. And please add CodeClimate to that.

I do not think suggestions by these tools should be followed religiously. But having some agreed conventions for a widely used library is way better than having none.

Is help needed to set these up?

[dead-claudia]

And another complete WTF: a ternary using logical operators instead of a ternary operator. I'm at a loss for words on this one...

[dead-claudia]

@kaukas It would result in a long series of commits + quite a few PRs running into merge conflicts for the short term, and it's not a small operation. I'll ping @lhorie here.

[dead-claudia]

And these two tests fail if a triple-equals is used.

[elektronik2k5]

+1

[dead-claudia]

@lhorie said in the gitter room that he was planning on hooking up his rewrite (which will feature multiple breaking API changes) with JSHint.

[maranomynet]

+1 for linting!

[dead-claudia]

Fixed with #822, #828.