Fix unlabeled break escaping a labeled switch statement

[svenrog]

Summary

Absorb an unlabeled break inside a labeled switch statement.

Details

An unlabeled break inside a labeled switch was not absorbed by the switch:

function f(e) {
    var o = 0;
    outer: switch (e) {
        case 1: o = 8; break;
        default: o = 29;
    }
    return o;
}
f(1); // returned undefined, should be 8

JintSwitchStatement.HandleCompletion only converted a Break completion to Normal when the break's target equalled the switch's own label:

if (r.Type == CompletionType.Break
    && string.Equals(context.Target, _statement.LabelSet?.Name, StringComparison.Ordinal))

For an unlabeled break, context.Target is null. When the switch is unlabeled, LabelSet?.Name is also null, so null == null absorbed it. But once the switch carries a label, LabelSet.Name is non-null, so the unlabeled break (target null) failed the equality check and propagated out as a Break completion — skipping statements after the switch and, at function scope, making the body complete abnormally so the function returned undefined.

All loop statements (for, for-in/of, while, do-while) already guard this with context.Target == null || string.Equals(...); the switch was the only breakable missing the null-target arm. This adds it, so a switch always absorbs an unlabeled break regardless of its own label — matching the loops and the spec (a SwitchStatement absorbs break completions with an empty target).

Found running minified React 19 inside a pure-JS DOM: createFiberFromTypeAndProps compiles to e: switch (e) { case T: o = 8; l |= 24; break; ... } and returned undefined for StrictMode elements, which then blew up downstream as Cannot convert undefined or null to object.

Linked issue

No tracking issue — discovered while running minified React 19 on Jint. Refs #

Test plan

• Added or updated unit tests in Jint.Tests
• Ran dotnet test --configuration Release locally
• For ECMAScript spec changes: ran Jint.Tests.Test262 and confirmed no regressions
• For interop changes: covered in Jint.Tests/Runtime/Interop
• For perf changes: included before/after numbers from Jint.Benchmark

Added three regression tests to Jint.Tests/Runtime/SwitchTests.cs (unlabeled break absorbed by a labeled switch; labeled break still escapes correctly; unlabeled break in a nested switch stays innermost). 3168 unit tests pass, and 737 Test262 switch/labeled/break/continue conformance tests pass, 0 failures.

Breaking change?

No.

[lahma]

Thank you!