Editorial: Remove ! from IsCallable invocation, be consistent with others.

[zhangenming]

NOW, I'm just wondering why other methods(forEach/map...) don't have this.
Should it be consistent ?
THANKS...

[jmdyck]

Hi. Normally, the opening comment in a PR should have a bit more explanation. For instance, in this case, it would be good to point out that IsCallable can't return an abrupt completion, so it's unnecessary to put ! before an invocation of IsCallable.

Also the title "fix flatMap" is somewhat misleading, because flatMap isn't 'broken': although the ! is unnecessary, it isn't causing flatMap to have incoherent or incorrect semantics. My suggestion for title would be "Remove ! from IsCallable invocation", as it gives a better idea of the nature of the PR's change.

Note that the spec's only other occurrence of ! IsCallable is 34 lines up, in FlattenIntoArray, so you might want to include that in this PR.

[zhangenming]

Thank you for your patient instruction : )

I'm just wondering why other methods(forEach/map...) don't have this
I'm not familiar with spec
I thought the ! means js Logical NOT.
so I made the mistake of thinking it was a mistake
thank you

[jmdyck]

I thought the ! means js Logical NOT.

Ah! That's why it's good to explain your thinking. The ! and ? prefixes are defined in 5.2.3.4 ReturnIfAbrupt Shorthands.

[claudepache]

About !: see also #568

[ljharb]

Currently, every AO call whose return value isn't an introspected Completion Record has a ? or a !. However, the editors have avoided adding them to predicates (like IsCallable), and #1796 will alter the conventions around Completion Records.

There are no plans I'm aware of to change the ! and ? notation.