JS: Add ECMAScript 2024 v Flag Operators for Regex Parsing

[Napalys]

This pull request adds support for parsing ECMAScript 2024 v flag operators, including:

• Nested Classes: Enables using nested character classes in regexes.
  Example: /[[abc][cz]]/v
• Intersection (&&): Matches characters common to both sets.
  Example: /[[abc]&&[cz]]/v
• Subtraction (--): Removes characters from a set.
  Example: /[[abc]--[cz]]/v
  Mixing operations at the same level is not allowed:
  • Invalid: /[[abc]&&[cz]--[zz]]/v
  • Valid: /[[abc]&&[[cz]--[zz]]]/v
• Union: Combines multiple sets.
  Example: /[[abc][cz]]/v
• Quoted Strings (\q{}): Allows matching exact sequences.
  Example: /[\q{ab|cb|db}]/v

Commit by commit review encouraged.

Useful links:

• https://v8.dev/features/regexp-v-flag
• https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/RegExp/unicodeSets

With correct parsing, this no longer produces an false positive in Closes #18854.

[Copilot]

PR Overview

This pull request introduces support for ECMAScript 2024 regex constructs under the new "v" flag. Key changes include:

• New AST node classes for character class operations (Subtraction, QuotedString, Intersection, Union)
• Enhancements to RegExpParser to conditionally enable nested character classes, new operators, and quoted string parsing with a fallback mechanism when errors are encountered
• New test inputs covering quoted strings, unions, intersections, subtractions, and nested character classes

Reviewed Changes

File	Description
javascript/extractor/src/com/semmle/js/ast/regexp/CharacterClassSubtraction.java	New AST node for subtraction operator in character classes
javascript/extractor/src/com/semmle/js/ast/regexp/CharacterClassQuotedString.java	New AST node for handling quoted string escapes
javascript/extractor/src/com/semmle/js/ast/regexp/CharacterClassIntersection.java	New AST node for intersection operator in character classes
javascript/extractor/src/com/semmle/js/ast/regexp/CharacterClassUnion.java	New AST node for union operator in character classes
javascript/extractor/src/com/semmle/js/parser/RegExpParser.java	Extended parser functionality to support the new "v" flag and corresponding regex operations
javascript/extractor/src/com/semmle/js/extractor/ASTExtractor.java and RegExpExtractor.java	Updated extraction logic to accommodate new AST node types and conditional flag handling

Copilot reviewed 31 out of 31 changed files in this pull request and generated 2 comments.

Tip: Leave feedback on Copilot's review comments with the 👎 and 👍 buttons to help improve review quality. Learn more

[asgerf]

Excellent work! I have a couple of comments to keep you busy during the week 😄

[erik-krogh]

Nice work 👍

I didn't look through it thoroughly, I assume Asger did that.

Did you run database creation on the latest main of https://github.com/babel/babel and https://github.com/tc39/test262?
Those projects contain all kinds of valid and invalid syntax, so it's a nice test of whether something is horribly wrong.

[Napalys]

Nice work 👍

I didn't look through it thoroughly, I assume Asger did that.

Did you run database creation on the latest main of https://github.com/babel/babel and https://github.com/tc39/test262? Those projects contain all kinds of valid and invalid syntax, so it's a nice test of whether something is horribly wrong.

For Babel, the extraction failed catastrophically on this Invalid Syntax File. However, after deleting it, the database was successfully created. I assume this is expected since the file contains invalid syntax?

The test262 database was created successfully without any issues.

[erik-krogh]

For Babel, the extraction failed catastrophically on this Invalid Syntax File. However, after deleting it, the database was successfully created. I assume this is expected since the file contains invalid syntax?

No, that is not expected.
Database creation should succeed, but with some extracted syntax errors.

However, that seems to be unrelated to this PR.
But maybe you could look into fixing that crash later? (And bump the SHA for babel/babel in DCA in the process).

[asgerf]

🎉