Stage 3 Specification Review: Waldemar Horwat

[rbuckton]

This is a placeholder task for the Stage 3 Specification Review feedback from Waldemar Horwat (@waldemarhorwat).

[rbuckton]

@waldemarhorwat have you had an opportunity to review the specification text?

[rbuckton]

FYI, I am considering bringing this to plenary in June for Stage 3, pending reviews.

[waldemarhorwat]

Here's my review:

The Pattern Semantics section defines the following aliases:

DotAll is true if the RegExp object's [[OriginalFlags]] internal slot contains "s" and otherwise is false.
IgnoreCase is true if the RegExp object's [[OriginalFlags]] internal slot contains "i" and otherwise is false.
Multiline is true if the RegExp object's [[OriginalFlags]] internal slot contains "m" and otherwise is false.

These aliases shouldn't be needed any more because they're all replaced with accesses of modifiers objects being passed around. If any remain, that's probably a bug.

Extra "the" before "any" in

It is a Syntax Error if the any code point in the source text matched by the first RegularExpressionFlags is also contained in the source text matched by the second RegularExpressionFlags.

There is a dangling free variable modifiers in the definition of CompileToCharSet:

CharacterClassEscape :: w

1. Return GetWordCharacters(modifiers).

This one may require a significant amount of refactoring to fix.

There may also be tricky interactions between this proposal's spec text and the new character classes proposal's spec text in how they handle case-insensitive character sets.

[rbuckton]

I believe all of this feedback is addressed in #14, which updates the proposal specification text to align with the latest version of ECMA-262, which includes the RegExp v flag. Unicode character sets always leverage the value of rer.[[IgnoreCase]], which this proposal now updates when the i modifier is set or cleared, so it should be consistent.