Date.UTC(x) [single-argument case] semantics update

[dilijev]

Behavior implemented in #1333 is that Date.UTC(x) and Date.UTC() where typeof x === number returns NaN.

Spec semantics will be updated (see tc39/ecma262#642) so that Date.UTC() returns NaN and Date.UTC(2017) === Date.UTC(2017, 0).

• 0-argument case works according to spec.
• 1-argument case needs to be updated.

See:
tc39/ecma262#642
#1318 (issue)
#1333 (PR)

/cc @bterlson

[bterlson]

I think it's the reverse isn't it? Date.UTC() currently returns Date.UTC(0, 0, ...) and should return NaN. Date.UTC(2017) currently returns the right thing.

[dilijev]

The patch I previously submitted made it so Date.UTC() and Date.UTC(2017) would both return NaN.
It's possible I misspoke above anyway, so let's just make it explicit with tests:

assert(isNaN(Date.UTC()), "expected: Date.UTC() is NaN");
assert(!isNaN(Date.UTC(2017)), "expected: Date.UTC(2017) is not NaN");
assert(!isNaN(Date.UTC(2017, 0)), "expected: Date.UTC(2017, 0) is not NaN;");
assert(!isNaN(Date.UTC(2017, 0, 1)), "expected: Date.UTC(2017, 0, 1) is not NaN;");
assert(isNaN(Date.UTC(2017, undefined)), "expected: Date.UTC(2017, undefined) is NaN");
assert(Date.UTC(2017) === Date.UTC(2017, 0), "expected: Date.UTC(2017) === Date.UTC(2017, 0)");
assert(Date.UTC(2017, 0) === Date.UTC(2017, 0, 1), "expected: Date.UTC(2017, 0) === Date.UTC(2017, 0, 1)");

Current behavior:

PASS
FAIL: expected: Date.UTC(2017) is not NaN
PASS
PASS
PASS
FAIL: expected: Date.UTC(2017) === Date.UTC(2017, 0)
PASS

Expected behavior:
all PASS

[bterlson]

Yes ok sorry. Seems good, thank you!