Fix infinite hang in regex parser for invalid UTF-8 with unicode flag (#3857)

Copilot · jakebailey · web-flow · commit 95223255de20 · 2026-05-15T20:41:19.000Z
Co-authored-by: copilot-swe-agent[bot] &lt;198982749+Copilot@users.noreply.github.com&gt;
Co-authored-by: jakebailey &lt;5341706+jakebailey@users.noreply.github.com&gt;
diff --git a/internal/scanner/regexp.go b/internal/scanner/regexp.go
@@ -1015,7 +1015,12 @@ func (p *regExpParser) scanSourceCharacter() string {
 		return string(ch)
 	}
 	ch, size := utf8.DecodeRuneInString(p.text()[p.pos():])
-	if size == 0 || ch == utf8.RuneError {
+	if size == 0 {
+		return ""
+	}
+	if ch == utf8.RuneError {
+		// Invalid UTF-8; consume the byte to avoid infinite loops.
+		p.incPos(size)
 		return ""
 	}
 	p.incPos(size)
diff --git a/testdata/baselines/reference/compiler/regexInvalidUtf8WithUnicodeFlag.js b/testdata/baselines/reference/compiler/regexInvalidUtf8WithUnicodeFlag.js
@@ -0,0 +1,9 @@
+//// [tests/cases/compiler/regexInvalidUtf8WithUnicodeFlag.ts] ////
+
+//// [regexInvalidUtf8WithUnicodeFlag.ts]
+/�/u
+
+
+//// [regexInvalidUtf8WithUnicodeFlag.js]
+"use strict";
+/�/u;
diff --git a/testdata/baselines/reference/compiler/regexInvalidUtf8WithUnicodeFlag.symbols b/testdata/baselines/reference/compiler/regexInvalidUtf8WithUnicodeFlag.symbols
@@ -0,0 +1,6 @@
+//// [tests/cases/compiler/regexInvalidUtf8WithUnicodeFlag.ts] ////
+
+=== regexInvalidUtf8WithUnicodeFlag.ts ===
+
+/�/u
+
diff --git a/testdata/baselines/reference/compiler/regexInvalidUtf8WithUnicodeFlag.types b/testdata/baselines/reference/compiler/regexInvalidUtf8WithUnicodeFlag.types
@@ -0,0 +1,6 @@
+//// [tests/cases/compiler/regexInvalidUtf8WithUnicodeFlag.ts] ////
+
+=== regexInvalidUtf8WithUnicodeFlag.ts ===
+/�/u
+>/�/u : RegExp
+
diff --git a/testdata/tests/cases/compiler/regexInvalidUtf8WithUnicodeFlag.ts b/testdata/tests/cases/compiler/regexInvalidUtf8WithUnicodeFlag.ts
@@ -0,0 +1,2 @@
+// @target: esnext
+/�/u

Original file line number	Diff line number	Diff line change
`@@ -1015,7 +1015,12 @@ func (p *regExpParser) scanSourceCharacter() string {`
`1015`	`1015`	`return string(ch)`
`1016`	`1016`	`}`
`1017`	`1017`	`ch, size := utf8.DecodeRuneInString(p.text()[p.pos():])`
`1018`		`- if size == 0 \|\| ch == utf8.RuneError {`
	`1018`	`+ if size == 0 {`
	`1019`	`+ return ""`
	`1020`	`+ }`
	`1021`	`+ if ch == utf8.RuneError {`
	`1022`	`+ // Invalid UTF-8; consume the byte to avoid infinite loops.`
	`1023`	`+ p.incPos(size)`
`1019`	`1024`	`return ""`
`1020`	`1025`	`}`
`1021`	`1026`	`p.incPos(size)`
-Original file line number
+Diff line change
@@ @@ -0,0 +1,6 @@ @@
 +//// [tests/cases/compiler/regexInvalidUtf8WithUnicodeFlag.ts] ////
++
 +=== regexInvalidUtf8WithUnicodeFlag.ts ===
++
 +/�/u
++