Invalid assertion for UTF16Encoding with template unicode escapes

[isiahmeadows]

Something I found while implementing templates for a new parser.

"\u{FFFFFFFF}"
`\u{FFFFFFFF}`

The string is statically determined to be invalid, but the template version is not, due to the most recent changes to the grammar (relaxing the template literal restriction). In 11.8.6.1, it states the following:

• The TV of TemplateCharacter :: \ EscapeSequence is the SV of EscapeSequence.

In 11.8.4.3, it states the following:

• The SV of EscapeSequence :: UnicodeEscapeSequence is the SV of the UnicodeEscapeSequence.
• The SV of UnicodeEscapeSequence :: u { HexDigits } is the UTF16Encoding of the MV of HexDigits.

In 10.1.1, UTF16Encoding makes the following assertion:

1. Assert: 0 ≤ cp ≤ 0x10FFFF.

This is invalid for templates when parsing the escape \u{FFFFFFFF}, which is perfectly valid within template literals.

---

Here's how I feel it might be best resolved: change UTF16Encoding to do the following initially:

1. Assert: cp ≤ 0.
2. If cp > 0x10FFFF, let cp be 0x10FFFF.

[isiahmeadows]

Note: this bug exists independent of the template literal revision.

[isiahmeadows]

Note: this bug exists independent of the template literal revision.

[isiahmeadows]

Actually, the template literal revision fixes this, due to it using CodePoint and NotCodePoint to restrict interpretation.

[isiahmeadows]

Actually, the template literal revision fixes this, due to it using CodePoint and NotCodePoint to restrict interpretation.