👍 Pretty sure this is taught as a common best practice anyway. I'm all for disallowing && and || together only.

I agree with @sotojuan. Programmers who know what they're doing are probably already parenthesizing their statements. IMO it's naive to assume that someone writing low-level code is more likely to know what they're doing.

Even with relatively simple operations parenthesizing helps a ton.

// This is difficult to read, not complicated.
const foo = 3 + 4 / 5 + 3 + 5 * 4 + 6

// This is much easier to read.
const foo = 3 + (4 / 5) + 3 + (5 * 4) + 6

Agreed, if mixing: parenthesis.

Yup, I still mess this up every now and then haha

Targeting this for standard v9.

This will be part of the standard v9 beta.

We will check all operators except for bitwise operators, since it's very common in low-level modules to omit parens when dealing with binary operators. Maybe we can consider making this stricter in a future version, but I doubt we'll want to. The rule is already very aggressive.

    "no-mixed-operators": ["error", {
        "groups": [
            ["+", "-", "*", "/", "%", "**"],
            ["==", "!=", "===", "!==", ">", ">=", "<", "<="],
            ["&&", "||"],
            ["in", "instanceof"]
        ],
        "allowSamePrecedence": true
    }],

Ecosystem effect is pretty large, but I looked at each failing example and agree that parens ought to be used in each instance to improve clarity. Many of the errors were in my own repos.

1..422
# tests 422
# pass  395
# fail  27

After fixing repos I control, ecosystem impact looks better, but still substantial.

1..422
# tests 422
# pass  402
# fail  20

I do use this pattern a bit: var a = obj && obj.prop || defaultValue

Edit: just read this more closely:

We will check all operators except for binary operators

So IIUC the above pattern would not be affected. If so, LGTM

Edit 2: lol you wrote "binary" and I read "boolean" but you meant "bitwise"

Oh, sorry, I meant bitwise operators, like |, &, etc. (corrected!). With this rule, you'd have to write: var a = (obj && obj.prop) || defaultValue

Personally, I find it confusing to read that line without the parens, even though I know I've written lines like that before. It's harder to visually parse what's going on when reading it, IMO

Yeah ok, the parens aren't too bad.

I note some irony since in the past I've seen it argued, possibly by me, that explicit semicolons everywhere is as superfluous as explicit parens everywhere. "Just learn these simple operator precedence rules, it's fine. Really!" I realise now that it's a false equivalence though, operator precedence has a great many more rules to remember than ASI.

operator precedence has a great many more rules to remember than ASI.

This is so true. 😄

Btw, this is just for the beta release, so worst case if it sucks we can revert before release.

Not being able to use simple math with +, -, /, * without having to force parens seems annoying (i write a lot of math code). Bit-wise ops are confusing though in terms of preference!

I can already hear my first grade math teacher yelling at me for not knowing the basic operation order

@mafintosh We could turn it off for +, -, /, * since I agree that basic arithmetic order is well-understood. But on longer lines it's still helpful, e.g.

 return value - Math.floor((value - min) / range) * range

vs.

 return value - (Math.floor((value - min) / range) * range)

I don't personally find that helpful but i get your point.

Stuff like this is incredible hard to parse though and i mess up ALL the time so I'd def be okay with enforcing it there

if (a & 1 === 0) { // same as a & (1 === 0)
}

@feross A more common mistake I've seen is something like

return value - a + b

Where someone ment

return value - (a + b)

So just forcing them all over the place for the classic ops makes it harder to question if someone ment to add them or not imo

@mafintosh The rule currently doesn't warn about operators that have the exact same precedence, so return value - a + b would be allowed. Not sure I followed your point -- does this make it better/worse for you?