Allow lone & in Unicode sets regexp classes

[Ijtihed Kilani (Ijtihed)]

Fixes microsoft/TypeScript#62707

port of microsoft/TypeScript#63535.

This updates the native scanner to allow a lone & in Unicode Sets regexp character classes while also preserving && handling for class set intersection.

AI assistance disclosure: I used Codex to help port and test this change. Of course went over everything myself afterwards.

Validation:

• go test ./internal/testrunner -run 'TestLocal/regularExpressionUnicodeSetsLoneAmpersand'
• go test ./internal/testrunner -run 'TestSubmodule/regularExpressionScanning'
• npx hereby format
• npx hereby check:format
• npx hereby lint
• npx hereby build
• npx hereby test failed only in internal/fswatch because local Windows symlink privileges are unavailable
• go test all packages except internal/fswatch passed locally

[Copilot]

Pull request overview

Note

Copilot was unable to run its full agentic suite in this review.

Adds a compiler regression test ensuring lone & characters are accepted inside Unicode Sets character classes (/v), and updates the regexp scanner to treat & as an operator only when it appears as &&.

Changes:

• Added a new compiler test case covering lone & in /v character classes.
• Added baseline outputs (types/symbols/js) for the new test.
• Updated the regexp scanner’s handling of & to only special-case &&.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
testdata/tests/cases/compiler/regularExpressionUnicodeSetsLoneAmpersand.ts	New regression test for lone & in /v character classes.
testdata/baselines/reference/compiler/regularExpressionUnicodeSetsLoneAmpersand.types	Baseline type output for the new test.
testdata/baselines/reference/compiler/regularExpressionUnicodeSetsLoneAmpersand.symbols	Baseline symbol output for the new test.
testdata/baselines/reference/compiler/regularExpressionUnicodeSetsLoneAmpersand.js	Baseline JS emit output for the new test.
internal/scanner/regexp.go	Scanner change to only treat && as special and allow lone &.

[Ijtihed Kilani (Ijtihed)]

Ryan Cavanaugh (@RyanCavanaugh) I love the label lmfao

edit: nvm. i'd love an explanation though..

[Ryan Cavanaugh (RyanCavanaugh)]

Ijtihed Kilani (@Ijtihed) label gets applied when someone seems to have a suspiciously high count of PRs and has possibly run afoul of our "no bulk contributors" policy. What made you decide to fix that issue in particular?

[Ijtihed Kilani (Ijtihed)]

Ijtihed Kilani (@Ijtihed) label gets applied when someone seems to have a suspiciously high count of PRs and has possibly run afoul of our "no bulk contributors" policy. What made you decide to fix that issue in particular?

tbch I'm a somewhat junior swe and was browsing issues and saw that one and thought it would be cool to learn on. I didn't run into the issue myself or anything beforehand

[Ryan Cavanaugh (RyanCavanaugh)]

graphemecluster (@graphemecluster) in case you want to see it

[graphemecluster (graphemecluster)]

Ryan Cavanaugh (@RyanCavanaugh) Looks fine by itself for this small fix, though I actually planned to port microsoft/TypeScript#62716 but I haven't done so. I would prefer porting that PR if its scope is not wider than your expectation.

[Ijtihed Kilani (Ijtihed)]

so what's the consensus here? :) I can close if needed

[Ijtihed Kilani (Ijtihed)]

anything left for me to do here? thanks!
/