From 700c45f46f9a84b2e7683df2138e130abd209864 Mon Sep 17 00:00:00 2001
From: Chris Nokleberg <chrisn@google.com>
Date: Wed, 25 Oct 2023 17:46:08 +0000
Subject: [PATCH] Use an explicit "UTF-8" character set argument when creating
 Strings from bytes. The platform default character set is guaranteed to be
 UTF-8.

PiperOrigin-RevId: 576577338
---
 java/com/google/re2j/Matcher.java | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/java/com/google/re2j/Matcher.java b/java/com/google/re2j/Matcher.java
index c2d1a7d2..5c32b82c 100644
--- a/java/com/google/re2j/Matcher.java
+++ b/java/com/google/re2j/Matcher.java
@@ -7,6 +7,7 @@
 package com.google.re2j;
 
 import com.google.re2j.MatcherInput.Encoding;
+import java.io.UnsupportedEncodingException;
 import java.util.Map;
 
 /**
@@ -362,8 +363,12 @@ private boolean genMatch(int startByte, int anchor) {
   /** Helper: return substring for [start, end). */
   String substring(int start, int end) {
     // UTF_8 is matched in binary mode. So slice the bytes.
-    if (matcherInput.getEncoding() == Encoding.UTF_8) {
-      return new String(matcherInput.asBytes(), start, end - start);
+    try {
+      if (matcherInput.getEncoding() == Encoding.UTF_8) {
+        return new String(matcherInput.asBytes(), start, end - start, "UTF-8");
+      }
+    } catch (UnsupportedEncodingException e) {
+      throw new RuntimeException(e); // Not possible.
     }
 
     // This is fast for both StringBuilder and String.