More LZ4 tests.

git-svn-id: https://svn.apache.org/repos/asf/lucene/dev/trunk@1417227 13f79535-47bb-0310-9956-ffa450edef68

lucene/core/src/java/org/apache/lucene/codecs/compressing/LZ4.java

@@ -121,7 +121,7 @@ class LZ4 {
       matchLen += MIN_MATCH;
 
       // copying a multiple of 8 bytes can make decompression from 5% to 10% faster
-      final int fastLen = ((matchLen - 1) & 0xFFFFFFF8) + 8;
+      final int fastLen = (matchLen + 7) & 0xFFFFFFF8;
       if (matchDec < matchLen || dOff + fastLen > destEnd) {
         // overlap -> naive incremental copy
         for (int ref = dOff - matchDec, end = dOff + matchLen; dOff < end; ++ref, ++dOff) {
@@ -222,7 +222,7 @@ class LZ4 {
         }
 
         // compute match length
-        final int matchLen = MIN_MATCH + commonBytes(bytes, ref + 4, off + 4, limit);
+        final int matchLen = MIN_MATCH + commonBytes(bytes, ref + MIN_MATCH, off + MIN_MATCH, limit);
 
         encodeSequence(bytes, anchor, ref, off, matchLen, out);
         off += matchLen;

lucene/core/src/test/org/apache/lucene/codecs/compressing/AbstractTestCompressionMode.java

@@ -91,10 +91,13 @@ public abstract class AbstractTestCompressionMode extends LuceneTestCase {
   }
 
   public void testDecompress() throws IOException {
-    final byte[] decompressed = randomArray();
-    final byte[] compressed = compress(decompressed);
-    final byte[] restored = decompress(compressed, decompressed.length);
-    assertArrayEquals(decompressed, restored);
+    final int iterations = atLeast(10);
+    for (int i = 0; i < iterations; ++i) {
+      final byte[] decompressed = randomArray();
+      final byte[] compressed = compress(decompressed);
+      final byte[] restored = decompress(compressed, decompressed.length);
+      assertArrayEquals(decompressed, restored);
+    }
   }
 
   public void testPartialDecompress() throws IOException {
@@ -120,11 +123,12 @@ public abstract class AbstractTestCompressionMode extends LuceneTestCase {
     assertArrayEquals(compressed, copyCompressedData(compressed, decompressed.length));
   }
 
-  public void test(byte[] decompressed) throws IOException {
+  public byte[] test(byte[] decompressed) throws IOException {
     final byte[] compressed = compress(decompressed);
     final byte[] restored = decompress(compressed, decompressed.length);
     assertEquals(decompressed.length, restored.length);
     assertArrayEquals(compressed, copyCompressedData(compressed, decompressed.length));
+    return compressed;
   }
 
   public void testEmptySequence() throws IOException {
@@ -143,31 +147,4 @@ public abstract class AbstractTestCompressionMode extends LuceneTestCase {
     test(decompressed);
   }
 
-  // for LZ compression
-
-  public void testShortLiteralsAndMatchs() throws IOException {
-    // literals and matchs lengths <= 15
-    final byte[] decompressed = "1234562345673456745678910123".getBytes("UTF-8");
-    test(decompressed);
-  }
-
-  public void testLongMatchs() throws IOException {
-    // match length > 16
-    final byte[] decompressed = new byte[RandomInts.randomIntBetween(random(), 300, 1024)];
-    for (int i = 0; i < decompressed.length; ++i) {
-      decompressed[i] = (byte) i;
-    }
-    test(decompressed);
-  }
-
-  public void testLongLiterals() throws IOException {
-    // long literals (length > 16) which are not the last literals
-    final byte[] decompressed = randomArray(RandomInts.randomIntBetween(random(), 400, 1024), 256);
-    final int matchRef = random().nextInt(30);
-    final int matchOff = RandomInts.randomIntBetween(random(), decompressed.length - 40, decompressed.length - 20);
-    final int matchLength = RandomInts.randomIntBetween(random(), 4, 10);
-    System.arraycopy(decompressed, matchRef, decompressed, matchOff, matchLength);
-    test(decompressed);
-  }
-
 }

lucene/core/src/test/org/apache/lucene/codecs/compressing/AbstractTestLZ4CompressionMode.java

@@ -0,0 +1,106 @@
+package org.apache.lucene.codecs.compressing;
+
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements.  See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License.  You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+import java.io.IOException;
+
+import com.carrotsearch.randomizedtesting.generators.RandomInts;
+
+public abstract class AbstractTestLZ4CompressionMode extends AbstractTestCompressionMode {
+
+  public byte[] test(byte[] decompressed) throws IOException {
+    final byte[] compressed = super.test(decompressed);
+    int off = 0;
+    int decompressedOff = 0;
+    for (;;) {
+      final int token = compressed[off++] & 0xFF;
+      int literalLen = token >>> 4;
+      if (literalLen == 0x0F) {
+        while (compressed[off] == (byte) 0xFF) {
+          literalLen += 0xFF;
+          ++off;
+        }
+        literalLen += compressed[off++] & 0xFF;
+      }
+      // skip literals
+      off += literalLen;
+      decompressedOff += literalLen;
+
+      // check that the stream ends with literals and that there are at least
+      // 5 of them
+      if (off == compressed.length) {
+        assertEquals(decompressed.length, decompressedOff);
+        assertTrue("lastLiterals=" + literalLen + ", bytes=" + decompressed.length,
+            literalLen >= LZ4.LAST_LITERALS || literalLen == decompressed.length);
+        break;
+      }
+
+      final int matchDec = (compressed[off++] & 0xFF) | ((compressed[off++] & 0xFF) << 8);
+      // check that match dec is not 0
+      assertTrue(matchDec + " " + decompressedOff, matchDec > 0 && matchDec <= decompressedOff);
+
+      int matchLen = token & 0x0F;
+      if (matchLen == 0x0F) {
+        while (compressed[off] == (byte) 0xFF) {
+          matchLen += 0xFF;
+          ++off;
+        }
+        matchLen += compressed[off++] & 0xFF;
+      }
+      matchLen += LZ4.MIN_MATCH;
+
+      // if the match ends prematurely, the next sequence should not have
+      // literals or this means we are wasting space
+      if (decompressedOff + matchLen < decompressed.length - LZ4.LAST_LITERALS) {
+        final boolean moreCommonBytes = decompressed[decompressedOff + matchLen] == decompressed[decompressedOff - matchDec + matchLen];
+        final boolean nextSequenceHasLiterals = ((compressed[off] & 0xFF) >>> 4) != 0;
+        assertTrue(!moreCommonBytes || !nextSequenceHasLiterals);
+      }      
+
+      decompressedOff += matchLen;
+    }
+    assertEquals(decompressed.length, decompressedOff);
+    return compressed;
+  }
+
+  public void testShortLiteralsAndMatchs() throws IOException {
+    // literals and matchs lengths <= 15
+    final byte[] decompressed = "1234562345673456745678910123".getBytes("UTF-8");
+    test(decompressed);
+  }
+
+  public void testLongMatchs() throws IOException {
+    // match length >= 20
+    final byte[] decompressed = new byte[RandomInts.randomIntBetween(random(), 300, 1024)];
+    for (int i = 0; i < decompressed.length; ++i) {
+      decompressed[i] = (byte) i;
+    }
+    test(decompressed);
+  }
+
+  public void testLongLiterals() throws IOException {
+    // long literals (length >= 16) which are not the last literals
+    final byte[] decompressed = randomArray(RandomInts.randomIntBetween(random(), 400, 1024), 256);
+    final int matchRef = random().nextInt(30);
+    final int matchOff = RandomInts.randomIntBetween(random(), decompressed.length - 40, decompressed.length - 20);
+    final int matchLength = RandomInts.randomIntBetween(random(), 4, 10);
+    System.arraycopy(decompressed, matchRef, decompressed, matchOff, matchLength);
+    test(decompressed);
+  }
+
+}

lucene/core/src/test/org/apache/lucene/codecs/compressing/TestFastCompressionMode.java

@@ -17,7 +17,7 @@ package org.apache.lucene.codecs.compressing;
  * limitations under the License.
  */
 
-public class TestFastCompressionMode extends AbstractTestCompressionMode {
+public class TestFastCompressionMode extends AbstractTestLZ4CompressionMode {
 
   public void setUp() throws Exception {
     super.setUp();

lucene/core/src/test/org/apache/lucene/codecs/compressing/TestFastDecompressionMode.java

@@ -17,12 +17,23 @@ package org.apache.lucene.codecs.compressing;
  * limitations under the License.
  */
 
+import java.io.IOException;
 
-public class TestFastDecompressionMode extends AbstractTestCompressionMode {
+public class TestFastDecompressionMode extends AbstractTestLZ4CompressionMode {
 
   public void setUp() throws Exception {
     super.setUp();
     mode = CompressionMode.FAST_DECOMPRESSION;
   }
 
+  @Override
+  public byte[] test(byte[] decompressed) throws IOException {
+    final byte[] compressed = super.test(decompressed);
+    final byte[] compressed2 = compress(CompressionMode.FAST.newCompressor(), decompressed);
+    // because of the way this compression mode works, its output is necessarily
+    // smaller than the output of CompressionMode.FAST
+    assertTrue(compressed.length <= compressed2.length);
+    return compressed;
+  }
+
 }