/* * Copyright (c) 1996, 2017, Oracle and/or its affiliates. All rights reserved. * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. * * This code is free software; you can redistribute it and/or modify it * under the terms of the GNU General Public License version 2 only, as * published by the Free Software Foundation. Oracle designates this * particular file as subject to the "Classpath" exception as provided * by Oracle in the LICENSE file that accompanied this code. * * This code is distributed in the hope that it will be useful, but WITHOUT * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License * version 2 for more details (a copy is included in the LICENSE file that * accompanied this code). * * You should have received a copy of the GNU General Public License version * 2 along with this work; if not, write to the Free Software Foundation, * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. * * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA * or visit www.oracle.com if you need additional information or have any * questions. */ package java.util.zip; import java.lang.ref.Cleaner.Cleanable; import jdk.internal.ref.CleanerFactory; /** * This class provides support for general purpose compression using the * popular ZLIB compression library. The ZLIB compression library was * initially developed as part of the PNG graphics standard and is not * protected by patents. It is fully described in the specifications at * the java.util.zip * package description. * *

The following code fragment demonstrates a trivial compression * and decompression of a string using {@code Deflater} and * {@code Inflater}. * *

 * try {
 *     // Encode a String into bytes
 *     String inputString = "blahblahblah";
 *     byte[] input = inputString.getBytes("UTF-8");
 *
 *     // Compress the bytes
 *     byte[] output = new byte[100];
 *     Deflater compresser = new Deflater();
 *     compresser.setInput(input);
 *     compresser.finish();
 *     int compressedDataLength = compresser.deflate(output);
 *     compresser.end();
 *
 *     // Decompress the bytes
 *     Inflater decompresser = new Inflater();
 *     decompresser.setInput(output, 0, compressedDataLength);
 *     byte[] result = new byte[100];
 *     int resultLength = decompresser.inflate(result);
 *     decompresser.end();
 *
 *     // Decode the bytes into a String
 *     String outputString = new String(result, 0, resultLength, "UTF-8");
 * } catch(java.io.UnsupportedEncodingException ex) {
 *     // handle
 * } catch (java.util.zip.DataFormatException ex) {
 *     // handle
 * }
 * 
* * @apiNote * To release resources used by this {@code Deflater}, the {@link #end()} method * should be called explicitly. Subclasses are responsible for the cleanup of resources * acquired by the subclass. Subclasses that override {@link #finalize()} in order * to perform cleanup should be modified to use alternative cleanup mechanisms such * as {@link java.lang.ref.Cleaner} and remove the overriding {@code finalize} method. * * @implSpec * If this {@code Deflater} has been subclassed and the {@code end} method has been * overridden, the {@code end} method will be called by the finalization when the * deflater is unreachable. But the subclasses should not depend on this specific * implementation; the finalization is not reliable and the {@code finalize} method * is deprecated to be removed. * * @see Inflater * @author David Connelly * @since 1.1 */ public class Deflater { private final DeflaterZStreamRef zsRef; private byte[] buf = new byte[0]; private int off, len; private int level, strategy; private boolean setParams; private boolean finish, finished; private long bytesRead; private long bytesWritten; /** * Compression method for the deflate algorithm (the only one currently * supported). */ public static final int DEFLATED = 8; /** * Compression level for no compression. */ public static final int NO_COMPRESSION = 0; /** * Compression level for fastest compression. */ public static final int BEST_SPEED = 1; /** * Compression level for best compression. */ public static final int BEST_COMPRESSION = 9; /** * Default compression level. */ public static final int DEFAULT_COMPRESSION = -1; /** * Compression strategy best used for data consisting mostly of small * values with a somewhat random distribution. Forces more Huffman coding * and less string matching. */ public static final int FILTERED = 1; /** * Compression strategy for Huffman coding only. */ public static final int HUFFMAN_ONLY = 2; /** * Default compression strategy. */ public static final int DEFAULT_STRATEGY = 0; /** * Compression flush mode used to achieve best compression result. * * @see Deflater#deflate(byte[], int, int, int) * @since 1.7 */ public static final int NO_FLUSH = 0; /** * Compression flush mode used to flush out all pending output; may * degrade compression for some compression algorithms. * * @see Deflater#deflate(byte[], int, int, int) * @since 1.7 */ public static final int SYNC_FLUSH = 2; /** * Compression flush mode used to flush out all pending output and * reset the deflater. Using this mode too often can seriously degrade * compression. * * @see Deflater#deflate(byte[], int, int, int) * @since 1.7 */ public static final int FULL_FLUSH = 3; static { ZipUtils.loadLibrary(); initIDs(); } /** * Creates a new compressor using the specified compression level. * If 'nowrap' is true then the ZLIB header and checksum fields will * not be used in order to support the compression format used in * both GZIP and PKZIP. * @param level the compression level (0-9) * @param nowrap if true then use GZIP compatible compression */ public Deflater(int level, boolean nowrap) { this.level = level; this.strategy = DEFAULT_STRATEGY; this.zsRef = DeflaterZStreamRef.get(this, init(level, DEFAULT_STRATEGY, nowrap)); } /** * Creates a new compressor using the specified compression level. * Compressed data will be generated in ZLIB format. * @param level the compression level (0-9) */ public Deflater(int level) { this(level, false); } /** * Creates a new compressor with the default compression level. * Compressed data will be generated in ZLIB format. */ public Deflater() { this(DEFAULT_COMPRESSION, false); } /** * Sets input data for compression. This should be called whenever * needsInput() returns true indicating that more input data is required. * @param b the input data bytes * @param off the start offset of the data * @param len the length of the data * @see Deflater#needsInput */ public void setInput(byte[] b, int off, int len) { if (b== null) { throw new NullPointerException(); } if (off < 0 || len < 0 || off > b.length - len) { throw new ArrayIndexOutOfBoundsException(); } synchronized (zsRef) { this.buf = b; this.off = off; this.len = len; } } /** * Sets input data for compression. This should be called whenever * needsInput() returns true indicating that more input data is required. * @param b the input data bytes * @see Deflater#needsInput */ public void setInput(byte[] b) { setInput(b, 0, b.length); } /** * Sets preset dictionary for compression. A preset dictionary is used * when the history buffer can be predetermined. When the data is later * uncompressed with Inflater.inflate(), Inflater.getAdler() can be called * in order to get the Adler-32 value of the dictionary required for * decompression. * @param b the dictionary data bytes * @param off the start offset of the data * @param len the length of the data * @see Inflater#inflate * @see Inflater#getAdler */ public void setDictionary(byte[] b, int off, int len) { if (b == null) { throw new NullPointerException(); } if (off < 0 || len < 0 || off > b.length - len) { throw new ArrayIndexOutOfBoundsException(); } synchronized (zsRef) { ensureOpen(); setDictionary(zsRef.address(), b, off, len); } } /** * Sets preset dictionary for compression. A preset dictionary is used * when the history buffer can be predetermined. When the data is later * uncompressed with Inflater.inflate(), Inflater.getAdler() can be called * in order to get the Adler-32 value of the dictionary required for * decompression. * @param b the dictionary data bytes * @see Inflater#inflate * @see Inflater#getAdler */ public void setDictionary(byte[] b) { setDictionary(b, 0, b.length); } /** * Sets the compression strategy to the specified value. * *

If the compression strategy is changed, the next invocation * of {@code deflate} will compress the input available so far with * the old strategy (and may be flushed); the new strategy will take * effect only after that invocation. * * @param strategy the new compression strategy * @exception IllegalArgumentException if the compression strategy is * invalid */ public void setStrategy(int strategy) { switch (strategy) { case DEFAULT_STRATEGY: case FILTERED: case HUFFMAN_ONLY: break; default: throw new IllegalArgumentException(); } synchronized (zsRef) { if (this.strategy != strategy) { this.strategy = strategy; setParams = true; } } } /** * Sets the compression level to the specified value. * *

If the compression level is changed, the next invocation * of {@code deflate} will compress the input available so far * with the old level (and may be flushed); the new level will * take effect only after that invocation. * * @param level the new compression level (0-9) * @exception IllegalArgumentException if the compression level is invalid */ public void setLevel(int level) { if ((level < 0 || level > 9) && level != DEFAULT_COMPRESSION) { throw new IllegalArgumentException("invalid compression level"); } synchronized (zsRef) { if (this.level != level) { this.level = level; setParams = true; } } } /** * Returns true if the input data buffer is empty and setInput() * should be called in order to provide more input. * @return true if the input data buffer is empty and setInput() * should be called in order to provide more input */ public boolean needsInput() { synchronized (zsRef) { return len <= 0; } } /** * When called, indicates that compression should end with the current * contents of the input buffer. */ public void finish() { synchronized (zsRef) { finish = true; } } /** * Returns true if the end of the compressed data output stream has * been reached. * @return true if the end of the compressed data output stream has * been reached */ public boolean finished() { synchronized (zsRef) { return finished; } } /** * Compresses the input data and fills specified buffer with compressed * data. Returns actual number of bytes of compressed data. A return value * of 0 indicates that {@link #needsInput() needsInput} should be called * in order to determine if more input data is required. * *

This method uses {@link #NO_FLUSH} as its compression flush mode. * An invocation of this method of the form {@code deflater.deflate(b, off, len)} * yields the same result as the invocation of * {@code deflater.deflate(b, off, len, Deflater.NO_FLUSH)}. * * @param b the buffer for the compressed data * @param off the start offset of the data * @param len the maximum number of bytes of compressed data * @return the actual number of bytes of compressed data written to the * output buffer */ public int deflate(byte[] b, int off, int len) { return deflate(b, off, len, NO_FLUSH); } /** * Compresses the input data and fills specified buffer with compressed * data. Returns actual number of bytes of compressed data. A return value * of 0 indicates that {@link #needsInput() needsInput} should be called * in order to determine if more input data is required. * *

This method uses {@link #NO_FLUSH} as its compression flush mode. * An invocation of this method of the form {@code deflater.deflate(b)} * yields the same result as the invocation of * {@code deflater.deflate(b, 0, b.length, Deflater.NO_FLUSH)}. * * @param b the buffer for the compressed data * @return the actual number of bytes of compressed data written to the * output buffer */ public int deflate(byte[] b) { return deflate(b, 0, b.length, NO_FLUSH); } /** * Compresses the input data and fills the specified buffer with compressed * data. Returns actual number of bytes of data compressed. * *

Compression flush mode is one of the following three modes: * *

* *

In the case of {@link #FULL_FLUSH} or {@link #SYNC_FLUSH}, if * the return value is {@code len}, the space available in output * buffer {@code b}, this method should be invoked again with the same * {@code flush} parameter and more output space. Make sure that * {@code len} is greater than 6 to avoid flush marker (5 bytes) being * repeatedly output to the output buffer every time this method is * invoked. * * @param b the buffer for the compressed data * @param off the start offset of the data * @param len the maximum number of bytes of compressed data * @param flush the compression flush mode * @return the actual number of bytes of compressed data written to * the output buffer * * @throws IllegalArgumentException if the flush mode is invalid * @since 1.7 */ public int deflate(byte[] b, int off, int len, int flush) { if (b == null) { throw new NullPointerException(); } if (off < 0 || len < 0 || off > b.length - len) { throw new ArrayIndexOutOfBoundsException(); } synchronized (zsRef) { ensureOpen(); if (flush == NO_FLUSH || flush == SYNC_FLUSH || flush == FULL_FLUSH) { int thisLen = this.len; int n = deflateBytes(zsRef.address(), b, off, len, flush); bytesWritten += n; bytesRead += (thisLen - this.len); return n; } throw new IllegalArgumentException(); } } /** * Returns the ADLER-32 value of the uncompressed data. * @return the ADLER-32 value of the uncompressed data */ public int getAdler() { synchronized (zsRef) { ensureOpen(); return getAdler(zsRef.address()); } } /** * Returns the total number of uncompressed bytes input so far. * *

Since the number of bytes may be greater than * Integer.MAX_VALUE, the {@link #getBytesRead()} method is now * the preferred means of obtaining this information.

* * @return the total number of uncompressed bytes input so far */ public int getTotalIn() { return (int) getBytesRead(); } /** * Returns the total number of uncompressed bytes input so far. * * @return the total (non-negative) number of uncompressed bytes input so far * @since 1.5 */ public long getBytesRead() { synchronized (zsRef) { ensureOpen(); return bytesRead; } } /** * Returns the total number of compressed bytes output so far. * *

Since the number of bytes may be greater than * Integer.MAX_VALUE, the {@link #getBytesWritten()} method is now * the preferred means of obtaining this information.

* * @return the total number of compressed bytes output so far */ public int getTotalOut() { return (int) getBytesWritten(); } /** * Returns the total number of compressed bytes output so far. * * @return the total (non-negative) number of compressed bytes output so far * @since 1.5 */ public long getBytesWritten() { synchronized (zsRef) { ensureOpen(); return bytesWritten; } } /** * Resets deflater so that a new set of input data can be processed. * Keeps current compression level and strategy settings. */ public void reset() { synchronized (zsRef) { ensureOpen(); reset(zsRef.address()); finish = false; finished = false; off = len = 0; bytesRead = bytesWritten = 0; } } /** * Closes the compressor and discards any unprocessed input. * * This method should be called when the compressor is no longer * being used. Once this method is called, the behavior of the * Deflater object is undefined. */ public void end() { synchronized (zsRef) { zsRef.clean(); buf = null; } } /** * Closes the compressor when garbage is collected. * * @deprecated The {@code finalize} method has been deprecated and will be * removed. It is implemented as a no-op. Subclasses that override * {@code finalize} in order to perform cleanup should be modified to use * alternative cleanup mechanisms and to remove the overriding {@code finalize} * method. The recommended cleanup for compressor is to explicitly call * {@code end} method when it is no longer in use. If the {@code end} is * not invoked explicitly the resource of the compressor will be released * when the instance becomes unreachable. */ @Deprecated(since="9", forRemoval=true) protected void finalize() {} private void ensureOpen() { assert Thread.holdsLock(zsRef); if (zsRef.address() == 0) throw new NullPointerException("Deflater has been closed"); } private static native void initIDs(); private static native long init(int level, int strategy, boolean nowrap); private static native void setDictionary(long addr, byte[] b, int off, int len); private native int deflateBytes(long addr, byte[] b, int off, int len, int flush); private static native int getAdler(long addr); private static native void reset(long addr); private static native void end(long addr); /** * A reference to the native zlib's z_stream structure. It also * serves as the "cleaner" to clean up the native resource when * the Deflater is ended, closed or cleaned. */ static class DeflaterZStreamRef implements Runnable { private long address; private final Cleanable cleanable; private DeflaterZStreamRef(Deflater owner, long addr) { this.cleanable = (owner != null) ? CleanerFactory.cleaner().register(owner, this) : null; this.address = addr; } long address() { return address; } void clean() { cleanable.clean(); } public synchronized void run() { long addr = address; address = 0; if (addr != 0) { end(addr); } } /* * If {@code Deflater} has been subclassed and the {@code end} method is * overridden, uses {@code finalizer} mechanism for resource cleanup. So * {@code end} method can be called when the {@code Deflater} is unreachable. * This mechanism will be removed when the {@code finalize} method is * removed from {@code Deflater}. */ static DeflaterZStreamRef get(Deflater owner, long addr) { Class clz = owner.getClass(); while (clz != Deflater.class) { try { clz.getDeclaredMethod("end"); return new FinalizableZStreamRef(owner, addr); } catch (NoSuchMethodException nsme) {} clz = clz.getSuperclass(); } return new DeflaterZStreamRef(owner, addr); } private static class FinalizableZStreamRef extends DeflaterZStreamRef { final Deflater owner; FinalizableZStreamRef (Deflater owner, long addr) { super(null, addr); this.owner = owner; } @Override void clean() { run(); } @Override @SuppressWarnings("deprecation") protected void finalize() { owner.end(); } } } }