7.1 says function expressions are always unnamed, but this changed in ES6

[Arnavion]

7.1 contains this statement:

Function declarations are named, so they're easier to identify in call stacks.

Anonymous function expressions assigned to a binding are also named, per ES6 semantics (1.e.iii). That is, the "bad" example const foo = function () { }; is the same as const foo = function foo() { };, which is equivalent to the "good" example function foo() { } in that respect.

Should the statement be qualified that it's only a distinction for ES5 and below ?

[ljharb]

Function name inference is implemented in hardly any of the browsers we support. In addition, relying on function name inference hinders greppability, so even in a pure ES6 environment, we wouldn't recommend that.

(other subtler differences are that a function with an inferred name still doesn't have a lexical name binding in the function body, but that's not relevant here)

[Arnavion]

Do you want to clarify that in the document itself? Via footnote, etc.

[ljharb]

Sure, that's a great idea for a footnote!

[tandav]

... frome Style Guide:

// bad
const foo = function () {
  // ...
};

// good
const foo = function bar() {
  // ...
};

So, you say, there should be 2 different names for one function?

For example,

// Is it worse
const sum = function(a, b) {
  return a + b;
};

// than this?
const my_sum = function sum(a, b) {
  return a + b;
};

I'm a begginer in JS/ES, can you tell, please, what are advantages of second method?

[ljharb]

@tandav In your last example, the first function doesn't have a .name - which means in debugging, it's likely to show up as an anonymous function. In the latter example, it will show up with the name sum - and when you grep for it, all the uses in the file will use my_sum, so the actual function will be easier to locate.