Remove snappy

dump/pom.xml

@@ -23,11 +23,6 @@
          <artifactId>trove4j</artifactId>
          <version>3.0.3</version>
       </dependency>
-      <dependency>
-         <groupId>org.iq80.snappy</groupId>
-         <artifactId>snappy</artifactId>
-         <version>0.4</version>
-      </dependency>
       <dependency>
          <groupId>org.lz4</groupId>
          <artifactId>lz4-java</artifactId>

dump/src/util/dump/stream/Compression.java

@@ -28,7 +28,7 @@
 /**
  * Using Compression enum values in StreamProviders you can compress your dumps transparently.
  *
- * With <code>Compression.LZ4</code> and <code>Compression.Snappy</code> there are two options for very fast
+ * With <code>Compression.LZ4</code> there is an option for very fast
  * compression, where one might expect, that performance improves overall, simply because you do less IO.
  * Unfortunately, in a single-threaded use-case with no other IO load this is not the case, even when using 
  * the fastest option, LZ4. Externalization creates high load on CPU, compression increases that load. 
@@ -49,7 +49,6 @@ public enum Compression implements ByteArrayPacker {
    GZipLevel7,
    GZipLevel8,
    GZipLevel9,
-   Snappy,
    LZ4,
    Zstd1,
    Zstd5,
@@ -101,7 +100,6 @@ public <E extends Externalizable> byte[] initDictionary( Iterable<E> dictInputPr
    @Override
    public boolean isPackedSizeInFirstFourBytes() {
       switch ( this ) {
-      case Snappy:
       case LZ4:
       case Zstd1:
       case Zstd5:
@@ -136,8 +134,6 @@ public byte[] pack( byte[] bytes, int bytesLength, @Nullable byte[] target, @Nul
          return gzip(8, bytes, bytesLength);
       case GZipLevel9:
          return gzip(9, bytes, bytesLength);
-      case Snappy:
-         return snappy(bytes, bytesLength, target);
       case LZ4:
          return lz4(bytes, bytesLength, target);
       case Zstd1:
@@ -168,8 +164,6 @@ public byte[] unpack( byte[] source, int sourceLength, @Nullable byte[] target,
       case GZipLevel8:
       case GZipLevel9:
          return gunzip(source, sourceLength);
-      case Snappy:
-         return unsnappy(source, sourceLength, target);
       case LZ4:
          return unLZ4(source, target);
       case Zstd1:
@@ -256,19 +250,6 @@ private byte[] lz4( byte[] bytes, int bytesLength, byte[] target ) {
       return target;
    }
 
-   private byte[] snappy( byte[] data, int dataLength, byte[] target ) {
-      int length = org.iq80.snappy.Snappy.maxCompressedLength(dataLength) + 4;
-      if ( target == null || target.length < length ) {
-         target = new byte[length];
-      }
-      int compressedSize = org.iq80.snappy.Snappy.compress(data, 0, dataLength, target, 4);
-      target[0] = (byte)((compressedSize >>> 24) & 0xFF);
-      target[1] = (byte)((compressedSize >>> 16) & 0xFF);
-      target[2] = (byte)((compressedSize >>> 8) & 0xFF);
-      target[3] = (byte)((compressedSize) & 0xFF);
-      return target;
-   }
-
    private byte[] unLZ4( byte[] bytes, byte[] target ) {
       LZ4FastDecompressor lz4Decompressor = getLZ4Decompressor();
       int length = (((bytes[0] & 0xff) << 24) + ((bytes[1] & 0xff) << 16) + ((bytes[2] & 0xff) << 8) + (bytes[3] & 0xff));
@@ -298,18 +279,6 @@ private byte[] unZstd( byte[] source, int sourceLength, byte[] target, byte[] di
       return target;
    }
 
-   private byte[] unsnappy( byte[] bytes, int sourceLength, byte[] target ) {
-      int length = org.iq80.snappy.Snappy.getUncompressedLength(bytes, 0);
-      if ( length > 100_000_000 ) {
-         throw new RuntimeException("insane size for decompressed length:" + length + " - failing now to prevent OutOfMemoryErrors");
-      }
-      if ( target == null || target.length < length ) {
-         target = new byte[length];
-      }
-      org.iq80.snappy.Snappy.uncompress(bytes, 0, sourceLength, target, 0);
-      return target;
-   }
-
    private byte[] zstd( int compressionLevel, byte[] source, int sourceLength, byte[] target, byte[] dict ) {
       int length = sourceLength * 2;
       if ( target == null || target.length < length ) {

dump/src/util/dump/stream/SingleTypeObjectStreamProvider.java

@@ -15,7 +15,7 @@
  *  - the {@link Externalizable}s may not use readObject() during readExternal() or writeObject(o) during writeExternal()<br><br>
  *  - if you put an instance twice into the dump, you will have two instances after deserialization in memory<br><br>
  *
- * This ObjectStreamProvider can compress the streams using Gzip, Snappy, LZ4 or ZStd. Use the appropriate constructor with a CompressionType.
+ * This ObjectStreamProvider can compress the streams using Gzip, LZ4 or ZStd. Use the appropriate constructor with a CompressionType.
  * Use compression only if you have limited storage space on your server, an IO bottleneck on your server, or if you access the dumps via
  * network and have a network bottleneck.
  *

dump/test/util/dump/stream/ExternalizableObjectStreamProviderTest.java

@@ -51,8 +51,6 @@ public void testCompression() throws Exception {
       test(provider, testBeans);
       provider = new ExternalizableObjectStreamProvider(Compression.GZipLevel9);
       test(provider, testBeans);
-      provider = new ExternalizableObjectStreamProvider(Compression.Snappy);
-      test(provider, testBeans);
       provider = new ExternalizableObjectStreamProvider(Compression.LZ4);
       test(provider, testBeans);
 
@@ -83,7 +81,6 @@ public void testCompression() throws Exception {
       test(stProvider, testExternalizableBeans);
       stProvider = new SingleTypeObjectStreamProvider<>(TestExternalizableBean.class, Compression.GZipLevel9);
       test(stProvider, testExternalizableBeans);
-      stProvider = new SingleTypeObjectStreamProvider<>(TestExternalizableBean.class, Compression.Snappy);
       test(stProvider, testExternalizableBeans);
       stProvider = new SingleTypeObjectStreamProvider<>(TestExternalizableBean.class, Compression.LZ4);
       test(stProvider, testExternalizableBeans);