Should grouping parentheses as in (a?.b) interrupt short-circuiting?

[claudepache]

That is, should

(a?.b).c

be equivalent to

(1) (a == null ? undefined : a.b).c

rather than

(2) a?.b.c, or (a == null ? undefined : a.b.c) (as currently specced)?

As currently specced, grouping parentheses are transparent w.r.t. Nil references: this is consistent with how References work, but it is not necessarily a desirable consistency, as the Nil reference is a spec mechanism that the average user does not need to know.

One may rather want to make the semantics more similar to how short-circuiting or conditional evaluation works with other operators that have such a feature, namely ||, &&, and ? :.

Any feedback is welcome. What is more intuitive, (1) or (2)?

[Mouvedia]

In that peculiar example I think that (a?.b).c and a?.b.c are equivalent.

What is more intuitive, (1) or (2)?

2

[MaienM]

1

[saschanaz]

1 for me.

[Wedvich]

To me 1 is the most intuitive - I would expect it to work the same way as the other short-circuiting operators mentioned.

[claudepache]

I have chosen option 1.

[Mouvedia]

So you prefer to throw a TypeError over returning undefined. I wonder why would you do that example in the first place then.

[claudepache]

@Mouvedia Regardless of the semantics, there is hardly a practical reason to use parentheses in this situation anyway.