perf: String#getBytes(Charset) vs getBytes(String)

4 files changed, 76 insertions, 0 deletions

diff --git a/java/src/test/java/com/google/protobuf/BoundedByteStringTest.java b/java/src/test/java/com/google/protobuf/BoundedByteStringTest.java
index 6c9596ca..a11bef2e 100644
--- a/java/src/test/java/com/google/protobuf/BoundedByteStringTest.java
+++ b/java/src/test/java/com/google/protobuf/BoundedByteStringTest.java
@@ -72,6 +72,19 @@ public class BoundedByteStringTest extends LiteralByteStringTest {
         testString.substring(2, testString.length() - 6), roundTripString);
   }
 
+  @Override
+  public void testCharsetToString() throws UnsupportedEncodingException {
+    String testString = "I love unicode \u1234\u5678 characters";
+    LiteralByteString unicode = new LiteralByteString(testString.getBytes(ByteString.UTF_8));
+    ByteString chopped = unicode.substring(2, unicode.size() - 6);
+    assertEquals(classUnderTest + ".substring() must have the expected type",
+        classUnderTest, getActualClassName(chopped));
+
+    String roundTripString = chopped.toString(ByteString.UTF_8);
+    assertEquals(classUnderTest + " unicode bytes must match",
+        testString.substring(2, testString.length() - 6), roundTripString);
+  }
+
   public void testJavaSerialization() throws Exception {
     ByteArrayOutputStream out = new ByteArrayOutputStream();
     ObjectOutputStream oos = new ObjectOutputStream(out);
diff --git a/java/src/test/java/com/google/protobuf/LiteralByteStringTest.java b/java/src/test/java/com/google/protobuf/LiteralByteStringTest.java
index f3ad774f..8607040e 100644
--- a/java/src/test/java/com/google/protobuf/LiteralByteStringTest.java
+++ b/java/src/test/java/com/google/protobuf/LiteralByteStringTest.java
@@ -298,6 +298,13 @@ public class LiteralByteStringTest extends TestCase {
     assertEquals(classUnderTest + " unicode must match", testString, roundTripString);
   }
 
+  public void testCharsetToString() throws UnsupportedEncodingException {
+    String testString = "I love unicode \u1234\u5678 characters";
+    LiteralByteString unicode = new LiteralByteString(testString.getBytes(ByteString.UTF_8));
+    String roundTripString = unicode.toString(ByteString.UTF_8);
+    assertEquals(classUnderTest + " unicode must match", testString, roundTripString);
+  }
+
   public void testToString_returnsCanonicalEmptyString() throws UnsupportedEncodingException{
     assertSame(classUnderTest + " must be the same string references",
         ByteString.EMPTY.toString(UTF_8), new LiteralByteString(new byte[]{}).toString(UTF_8));
diff --git a/java/src/test/java/com/google/protobuf/RopeByteStringSubstringTest.java b/java/src/test/java/com/google/protobuf/RopeByteStringSubstringTest.java
index 8106201d..43872d1d 100644
--- a/java/src/test/java/com/google/protobuf/RopeByteStringSubstringTest.java
+++ b/java/src/test/java/com/google/protobuf/RopeByteStringSubstringTest.java
@@ -94,4 +94,34 @@ public class RopeByteStringSubstringTest extends LiteralByteStringTest {
     assertEquals(classUnderTest + " string must must have same hashCode as the flat string",
         flatString.hashCode(), unicode.hashCode());
   }
+
+  @Override
+  public void testCharsetToString() throws UnsupportedEncodingException {
+    String sourceString = "I love unicode \u1234\u5678 characters";
+    ByteString sourceByteString = ByteString.copyFromUtf8(sourceString);
+    int copies = 250;
+
+    // By building the RopeByteString by concatenating, this is actually a fairly strenuous test.
+    StringBuilder builder = new StringBuilder(copies * sourceString.length());
+    ByteString unicode = ByteString.EMPTY;
+    for (int i = 0; i < copies; ++i) {
+      builder.append(sourceString);
+      unicode = RopeByteString.concatenate(unicode, sourceByteString);
+    }
+    String testString = builder.toString();
+
+    // Do the substring part
+    testString = testString.substring(2, testString.length() - 6);
+    unicode = unicode.substring(2, unicode.size() - 6);
+
+    assertEquals(classUnderTest + " from string must have the expected type",
+        classUnderTest, getActualClassName(unicode));
+    String roundTripString = unicode.toString(ByteString.UTF_8);
+    assertEquals(classUnderTest + " unicode bytes must match",
+        testString, roundTripString);
+    ByteString flatString = ByteString.copyFromUtf8(testString);
+    assertEquals(classUnderTest + " string must equal the flat string", flatString, unicode);
+    assertEquals(classUnderTest + " string must must have same hashCode as the flat string",
+        flatString.hashCode(), unicode.hashCode());
+  }
 }
diff --git a/java/src/test/java/com/google/protobuf/RopeByteStringTest.java b/java/src/test/java/com/google/protobuf/RopeByteStringTest.java
index 4775f03a..54eb9683 100644
--- a/java/src/test/java/com/google/protobuf/RopeByteStringTest.java
+++ b/java/src/test/java/com/google/protobuf/RopeByteStringTest.java
@@ -119,6 +119,32 @@ public class RopeByteStringTest extends LiteralByteStringTest {
   }
 
   @Override
+  public void testCharsetToString() throws UnsupportedEncodingException {
+    String sourceString = "I love unicode \u1234\u5678 characters";
+    ByteString sourceByteString = ByteString.copyFromUtf8(sourceString);
+    int copies = 250;
+
+    // By building the RopeByteString by concatenating, this is actually a fairly strenuous test.
+    StringBuilder builder = new StringBuilder(copies * sourceString.length());
+    ByteString unicode = ByteString.EMPTY;
+    for (int i = 0; i < copies; ++i) {
+      builder.append(sourceString);
+      unicode = RopeByteString.concatenate(unicode, sourceByteString);
+    }
+    String testString = builder.toString();
+
+    assertEquals(classUnderTest + " from string must have the expected type",
+        classUnderTest, getActualClassName(unicode));
+    String roundTripString = unicode.toString(ByteString.UTF_8);
+    assertEquals(classUnderTest + " unicode bytes must match",
+        testString, roundTripString);
+    ByteString flatString = ByteString.copyFromUtf8(testString);
+    assertEquals(classUnderTest + " string must equal the flat string", flatString, unicode);
+    assertEquals(classUnderTest + " string must must have same hashCode as the flat string",
+        flatString.hashCode(), unicode.hashCode());
+  }
+
+  @Override
   public void testToString_returnsCanonicalEmptyString() throws UnsupportedEncodingException {
     RopeByteString ropeByteString =
         RopeByteString.newInstanceForTest(ByteString.EMPTY, ByteString.EMPTY);