From be597d7ce973541d1d75477310bff1093e73463c Mon Sep 17 00:00:00 2001
From: Andy Wilkinson <awilkinson@pivotal.io>
Date: Mon, 24 Oct 2016 11:44:01 +0100
Subject: [PATCH] Fix handling of cyrillic characters in AsciiBytes hashCode
 method

Closes gh-7202
---
 .../boot/loader/jar/AsciiBytes.java           | 41 +++++++++++++------
 .../boot/loader/jar/AsciiBytesTests.java      | 23 ++++++++---
 2 files changed, 45 insertions(+), 19 deletions(-)

diff --git a/spring-boot-tools/spring-boot-loader/src/main/java/org/springframework/boot/loader/jar/AsciiBytes.java b/spring-boot-tools/spring-boot-loader/src/main/java/org/springframework/boot/loader/jar/AsciiBytes.java
index 83e29a7abec..62f26cedb23 100644
--- a/spring-boot-tools/spring-boot-loader/src/main/java/org/springframework/boot/loader/jar/AsciiBytes.java
+++ b/spring-boot-tools/spring-boot-loader/src/main/java/org/springframework/boot/loader/jar/AsciiBytes.java
@@ -23,6 +23,7 @@ import java.nio.charset.Charset;
  * reasons to save constructing Strings for ZIP data.
  *
  * @author Phillip Webb
+ * @author Andy Wilkinson
  */
 final class AsciiBytes {
 
@@ -156,21 +157,35 @@ final class AsciiBytes {
 		int hash = this.hash;
 		if (hash == 0 && this.bytes.length > 0) {
 			for (int i = this.offset; i < this.offset + this.length; i++) {
-				int b = this.bytes[i] & 0xff;
-				if (b > 0x7F) {
-					// Decode multi-byte UTF
-					for (int size = 0; size < 3; size++) {
-						if ((b & (0x40 >> size)) == 0) {
-							b = b & (0x1F >> size);
-							for (int j = 0; j < size; j++) {
-								b <<= 6;
-								b |= this.bytes[++i] & 0x3F;
-							}
-							break;
-						}
+				int b = this.bytes[i];
+				if (b < 0) {
+					b = b & 0x7F;
+					int limit;
+					int excess = 0x80;
+					if (b < 96) {
+						limit = 1;
+						excess += 0x40 << 6;
 					}
+					else if (b < 112) {
+						limit = 2;
+						excess += (0x60 << 12) + (0x80 << 6);
+					}
+					else {
+						limit = 3;
+						excess += (0x70 << 18) + (0x80 << 12) + (0x80 << 6);
+					}
+					for (int j = 0; j < limit; j++) {
+						b = (b << 6) + (this.bytes[++i] & 0xFF);
+					}
+					b -= excess;
+				}
+				if (b <= 0xFFFF) {
+					hash = 31 * hash + b;
+				}
+				else {
+					hash = 31 * hash + ((b >> 0xA) + 0xD7C0);
+					hash = 31 * hash + ((b & 0x3FF) + 0xDC00);
 				}
-				hash = 31 * hash + b;
 			}
 			this.hash = hash;
 		}
diff --git a/spring-boot-tools/spring-boot-loader/src/test/java/org/springframework/boot/loader/jar/AsciiBytesTests.java b/spring-boot-tools/spring-boot-loader/src/test/java/org/springframework/boot/loader/jar/AsciiBytesTests.java
index f1c5458c6d9..03c13c8658e 100644
--- a/spring-boot-tools/spring-boot-loader/src/test/java/org/springframework/boot/loader/jar/AsciiBytesTests.java
+++ b/spring-boot-tools/spring-boot-loader/src/test/java/org/springframework/boot/loader/jar/AsciiBytesTests.java
@@ -26,6 +26,7 @@ import static org.assertj.core.api.Assertions.assertThat;
  * Tests for {@link AsciiBytes}.
  *
  * @author Phillip Webb
+ * @author Andy Wilkinson
  */
 public class AsciiBytesTests {
 
@@ -140,16 +141,26 @@ public class AsciiBytesTests {
 
 	@Test
 	public void hashCodeSameAsString() throws Exception {
-		String s = "abcABC123xyz!";
-		AsciiBytes a = new AsciiBytes(s);
-		assertThat(s.hashCode()).isEqualTo(a.hashCode());
+		hashCodeSameAsString("abcABC123xyz!");
 	}
 
 	@Test
 	public void hashCodeSameAsStringWithSpecial() throws Exception {
-		String s = "special/\u00EB.dat";
-		AsciiBytes a = new AsciiBytes(s);
-		assertThat(s.hashCode()).isEqualTo(a.hashCode());
+		hashCodeSameAsString("special/\u00EB.dat");
+	}
+
+	@Test
+	public void hashCodeSameAsStringWithCyrillicCharacters() throws Exception {
+		hashCodeSameAsString("\u0432\u0435\u0441\u043D\u0430");
+	}
+
+	@Test
+	public void hashCodeSameAsStringWithEmoji() throws Exception {
+		hashCodeSameAsString("\ud83d\udca9");
+	}
+
+	private void hashCodeSameAsString(String input) {
+		assertThat(new AsciiBytes(input).hashCode()).isEqualTo(input.hashCode());
 	}
 
 }