Add template literal tests for supplementary-plane unicode escape sequences

[hexbinoct]

Closes #3001.

Template literals had no coverage for \u{} escapes above U+FFFF, so the surrogate pair branch of UTF16EncodeCodePoint (steps 3 to 5) was never exercised. The only astral escapes in these tests today are the invalid seven-digit \u{10FFFFF} cases.

The new file checks the cooked and raw values of tagged templates for U+10000 (the first code point that needs a surrogate pair), U+10437 (the character the spec itself uses as its UTF-16 example), U+10FFFF (the largest code point), and a leading-zeros form, plus the untagged path and concatenation with adjacent template characters. Expected values are written as explicit \uD800\uDC00 style pairs so they do not depend on the construct under test.

Following jugglinmike's suggestion in the issue, the assertions live in a dedicated file next to tv-utf16-escape-sequence.js. Thanks to p-bakker for reporting the gap from their work on Rhino.

The copyright line says hexbinoct. Happy to swap in the legal name from the CLA if that is preferred.

[ptomato]

Thanks!

Totally separate from a code review, I don't know the answer to the copyright name question.

[bakkot]

My understanding is that the copyright notice isn't actually the thing that creates the copyright; it's mainly to track provenance as things get copied to new places, and to ensure you get credit. So your username should be OK if that's what you'd prefer. We already have some tests with just usernames.

Incidentally US law explicitly allows using "a generally known alternative designation of the owner" in a copyright notice, though I don't know about other countries.

[hexbinoct]

Thanks both, I will leave the copyright line as hexbinoct, and the prop-desc.js precedent is a useful one to have on record.