Index: trunk/src/org/tukaani/xz/ArrayCache.java =================================================================== --- trunk/src/org/tukaani/xz/ArrayCache.java (revision 16021) +++ (revision ) @@ -1,172 +1,0 @@ -/* - * ArrayCache - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -/** - * Caches large arrays for reuse (base class and a dummy cache implementation). - *

- * When compressing or decompressing many (very) small files in a row, the - * time spent in construction of new compressor or decompressor objects - * can be longer than the time spent in actual compression or decompression. - * A large part of this initialization overhead comes from allocation and - * garbage collection of large arrays. - *

- * The {@code ArrayCache} API provides a way to cache large array allocations - * for reuse. It can give a major performance improvement when compressing or - * decompressing many tiny files. If you are only (de)compressing one or two - * files or the files a very big, array caching won't improve anything, - * although it won't make anything slower either. - *

- * Important: The users of ArrayCache don't return the allocated arrays - * back to the cache in all situations. - * This a reason why it's called a cache instead of a pool. - * If it is important to be able to return every array back to a cache, - * {@link ResettableArrayCache} can be useful. - *

- * In compressors (OutputStreams) the arrays are returned to the cache - * when a call to {@code finish()} or {@code close()} returns - * successfully (no exceptions are thrown). - *

- * In decompressors (InputStreams) the arrays are returned to the cache when - * the decompression is successfully finished ({@code read} returns {@code -1}) - * or {@code close()} or {@code close(boolean)} is called. This is true even - * if closing throws an exception. - *

- * Raw decompressors don't support {@code close(boolean)}. With raw - * decompressors, if one wants to put the arrays back to the cache without - * closing the underlying {@code InputStream}, one can wrap the - * {@code InputStream} into {@link CloseIgnoringInputStream} when creating - * the decompressor instance. Then one can use {@code close()}. - *

- * Different cache implementations can be extended from this base class. - * All cache implementations must be thread safe. - *

- * This class also works as a dummy cache that simply calls {@code new} - * to allocate new arrays and doesn't try to cache anything. A statically - * allocated dummy cache is available via {@link #getDummyCache()}. - *

- * If no {@code ArrayCache} is specified when constructing a compressor or - * decompressor, the default {@code ArrayCache} implementation is used. - * See {@link #getDefaultCache()} and {@link #setDefaultCache(ArrayCache)}. - *

- * This is a class instead of an interface because it's possible that in the - * future we may want to cache other array types too. New methods can be - * added to this class without breaking existing cache implementations. - * - * @since 1.7 - * - * @see BasicArrayCache - */ -public class ArrayCache { - /** - * Global dummy cache instance that is returned by {@code getDummyCache()}. - */ - private static final ArrayCache dummyCache = new ArrayCache(); - - /** - * Global default {@code ArrayCache} that is used when no other cache has - * been specified. - */ - private static volatile ArrayCache defaultCache = dummyCache; - - /** - * Returns a statically-allocated {@code ArrayCache} instance. - * It can be shared by all code that needs a dummy cache. - */ - public static ArrayCache getDummyCache() { - return dummyCache; - } - - /** - * Gets the default {@code ArrayCache} instance. - * This is a global cache that is used when the application - * specifies nothing else. The default is a dummy cache - * (see {@link #getDummyCache()}). - */ - public static ArrayCache getDefaultCache() { - // It's volatile so no need for synchronization. - return defaultCache; - } - - /** - * Sets the default {@code ArrayCache} instance. - * Use with care. Other libraries using this package probably shouldn't - * call this function as libraries cannot know if there are other users - * of the xz package in the same application. - */ - public static void setDefaultCache(ArrayCache arrayCache) { - if (arrayCache == null) - throw new NullPointerException(); - - // It's volatile so no need for synchronization. - defaultCache = arrayCache; - } - - /** - * Creates a new {@code ArrayCache} that does no caching - * (a dummy cache). If you need a dummy cache, you may want to call - * {@link #getDummyCache()} instead. - */ - public ArrayCache() {} - - /** - * Allocates a new byte array. - *

- * This implementation simply returns {@code new byte[size]}. - * - * @param size the minimum size of the array to allocate; - * an implementation may return an array that - * is larger than the given {@code size} - * - * @param fillWithZeros if true, the caller expects that the first - * {@code size} elements in the array are zero; - * if false, the array contents can be anything, - * which speeds things up when reusing a cached - * array - */ - public byte[] getByteArray(int size, boolean fillWithZeros) { - return new byte[size]; - } - - /** - * Puts the given byte array to the cache. The caller must no longer - * use the array. - *

- * This implementation does nothing. - */ - public void putArray(byte[] array) {} - - /** - * Allocates a new int array. - *

- * This implementation simply returns {@code new int[size]}. - * - * @param size the minimum size of the array to allocate; - * an implementation may return an array that - * is larger than the given {@code size} - * - * @param fillWithZeros if true, the caller expects that the first - * {@code size} elements in the array are zero; - * if false, the array contents can be anything, - * which speeds things up when reusing a cached - * array - */ - public int[] getIntArray(int size, boolean fillWithZeros) { - return new int[size]; - } - - /** - * Puts the given int array to the cache. The caller must no longer - * use the array. - *

- * This implementation does nothing. - */ - public void putArray(int[] array) {} -} Index: trunk/src/org/tukaani/xz/BCJCoder.java =================================================================== --- trunk/src/org/tukaani/xz/BCJCoder.java (revision 16021) +++ (revision ) @@ -1,35 +1,0 @@ -/* - * BCJCoder - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -abstract class BCJCoder implements FilterCoder { - public static final long X86_FILTER_ID = 0x04; - public static final long POWERPC_FILTER_ID = 0x05; - public static final long IA64_FILTER_ID = 0x06; - public static final long ARM_FILTER_ID = 0x07; - public static final long ARMTHUMB_FILTER_ID = 0x08; - public static final long SPARC_FILTER_ID = 0x09; - - public static boolean isBCJFilterID(long filterID) { - return filterID >= 0x04 && filterID <= 0x09; - } - - public boolean changesSize() { - return false; - } - - public boolean nonLastOK() { - return true; - } - - public boolean lastOK() { - return false; - } -} Index: trunk/src/org/tukaani/xz/BCJDecoder.java =================================================================== --- trunk/src/org/tukaani/xz/BCJDecoder.java (revision 16021) +++ (revision ) @@ -1,62 +1,0 @@ -/* - * BCJDecoder - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import java.io.InputStream; -import org.tukaani.xz.simple.*; - -class BCJDecoder extends BCJCoder implements FilterDecoder { - private final long filterID; - private final int startOffset; - - BCJDecoder(long filterID, byte[] props) - throws UnsupportedOptionsException { - assert isBCJFilterID(filterID); - this.filterID = filterID; - - if (props.length == 0) { - startOffset = 0; - } else if (props.length == 4) { - int n = 0; - for (int i = 0; i < 4; ++i) - n |= (props[i] & 0xFF) << (i * 8); - - startOffset = n; - } else { - throw new UnsupportedOptionsException( - "Unsupported BCJ filter properties"); - } - } - - public int getMemoryUsage() { - return SimpleInputStream.getMemoryUsage(); - } - - public InputStream getInputStream(InputStream in, ArrayCache arrayCache) { - SimpleFilter simpleFilter = null; - - if (filterID == X86_FILTER_ID) - simpleFilter = new X86(false, startOffset); - else if (filterID == POWERPC_FILTER_ID) - simpleFilter = new PowerPC(false, startOffset); - else if (filterID == IA64_FILTER_ID) - simpleFilter = new IA64(false, startOffset); - else if (filterID == ARM_FILTER_ID) - simpleFilter = new ARM(false, startOffset); - else if (filterID == ARMTHUMB_FILTER_ID) - simpleFilter = new ARMThumb(false, startOffset); - else if (filterID == SPARC_FILTER_ID) - simpleFilter = new SPARC(false, startOffset); - else - assert false; - - return new SimpleInputStream(in, simpleFilter); - } -} Index: trunk/src/org/tukaani/xz/BasicArrayCache.java =================================================================== --- trunk/src/org/tukaani/xz/BasicArrayCache.java (revision 16021) +++ (revision ) @@ -1,281 +1,0 @@ -/* - * BasicArrayCache - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import java.lang.ref.Reference; -import java.lang.ref.SoftReference; -import java.util.Arrays; -import java.util.LinkedHashMap; -import java.util.Map; - -/** - * A basic {@link ArrayCache} implementation. - *

- * This caches exact array sizes, that is, {@code getByteArray} will return - * an array whose size is exactly the requested size. A limited number - * of different array sizes are cached at the same time; least recently used - * sizes will be dropped from the cache if needed (can happen if several - * different (de)compression options are used with the same cache). - *

- * The current implementation uses - * {@link java.util.LinkedHashMap LinkedHashMap} to map different array sizes - * to separate array-based data structures which hold - * {@link java.lang.ref.SoftReference SoftReferences} to the cached arrays. - * In the common case this should give good performance and fairly low - * memory usage overhead. - *

- * A statically allocated global {@code BasicArrayCache} instance is - * available via {@link #getInstance()} which is a good choice in most - * situations where caching is wanted. - * - * @since 1.7 - */ -public class BasicArrayCache extends ArrayCache { - /** - * Arrays smaller than this many elements will not be cached. - */ - private static final int CACHEABLE_SIZE_MIN = 32 << 10; - - /** - * Number of stacks i.e. how many different array sizes to cache. - */ - private static final int STACKS_MAX = 32; - - /** - * Number of arrays of the same type and size to keep in the cache. - * (ELEMENTS_PER_STACK - 1) is used as a bit mask so ELEMENTS_PER_STACK - * must be a power of two! - */ - private static final int ELEMENTS_PER_STACK = 512; - - /** - * A thread-safe stack-like data structure whose {@code push} method - * overwrites the oldest element in the stack if the stack is full. - */ - private static class CyclicStack { - /** - * Array that holds the elements in the cyclic stack. - */ - @SuppressWarnings("unchecked") - private final T[] elements = (T[])new Object[ELEMENTS_PER_STACK]; - - /** - * Read-write position in the {@code refs} array. - * The most recently added element is in {@code refs[pos]}. - * If it is {@code null}, then the stack is empty and all - * elements in {@code refs} are {@code null}. - *

- * Note that {@code pop()} always modifies {@code pos}, even if - * the stack is empty. This means that when the first element is - * added by {@code push(T)}, it can get added in any position in - * {@code refs} and the stack will start growing from there. - */ - private int pos = 0; - - /** - * Gets the most recently added element from the stack. - * If the stack is empty, {@code null} is returned. - */ - public synchronized T pop() { - T e = elements[pos]; - elements[pos] = null; - pos = (pos - 1) & (ELEMENTS_PER_STACK - 1); - return e; - } - - /** - * Adds a new element to the stack. If the stack is full, the oldest - * element is overwritten. - */ - public synchronized void push(T e) { - pos = (pos + 1) & (ELEMENTS_PER_STACK - 1); - elements[pos] = e; - } - } - - /** - * Maps Integer (array size) to stacks of references to arrays. At most - * STACKS_MAX number of stacks are kept in the map (LRU cache). - */ - private static class CacheMap - extends LinkedHashMap>> { - /** - * This class won't be serialized but this is needed - * to silence a compiler warning. - */ - private static final long serialVersionUID = 1L; - - /** - * Creates a new CacheMap. - */ - public CacheMap() { - // The map may momentarily have at most STACKS_MAX + 1 entries - // when put(K,V) has inserted a new entry but hasn't called - // removeEldestEntry yet. Using 2 * STACKS_MAX as the initial - // (and the final) capacity should give good performance. 0.75 is - // the default load factor and in this case it guarantees that - // the map will never need to be rehashed because - // (STACKS_MAX + 1) / 0.75 < 2 * STACKS_MAX. - // - // That last argument is true to get LRU cache behavior together - // with the overridden removeEldestEntry method. - super(2 * STACKS_MAX, 0.75f, true); - } - - /** - * Returns true if the map is full and the least recently used stack - * should be removed. - */ - protected boolean removeEldestEntry( - Map.Entry>> eldest) { - return size() > STACKS_MAX; - } - } - - /** - * Helper class for the singleton instance. - * This is allocated only if {@code getInstance()} is called. - */ - private static final class LazyHolder { - static final BasicArrayCache INSTANCE = new BasicArrayCache(); - } - - /** - * Returns a statically-allocated {@code BasicArrayCache} instance. - * This is often a good choice when a cache is needed. - */ - public static BasicArrayCache getInstance() { - return LazyHolder.INSTANCE; - } - - /** - * Stacks for cached byte arrays. - */ - private final CacheMap byteArrayCache = new CacheMap(); - - /** - * Stacks for cached int arrays. - */ - private final CacheMap intArrayCache = new CacheMap(); - - /** - * Gets {@code T[size]} from the given {@code cache}. - * If no such array is found, {@code null} is returned. - */ - private static T getArray(CacheMap cache, int size) { - // putArray doesn't add small arrays to the cache and so it's - // pointless to look for small arrays here. - if (size < CACHEABLE_SIZE_MIN) - return null; - - // Try to find a stack that holds arrays of T[size]. - CyclicStack> stack; - synchronized(cache) { - stack = cache.get(size); - } - - if (stack == null) - return null; - - // Try to find a non-cleared Reference from the stack. - T array; - do { - Reference r = stack.pop(); - if (r == null) - return null; - - array = r.get(); - } while (array == null); - - return array; - } - - /** - * Puts the {@code array} of {@code size} elements long into - * the {@code cache}. - */ - private static void putArray(CacheMap cache, T array, int size) { - // Small arrays aren't cached. - if (size < CACHEABLE_SIZE_MIN) - return; - - CyclicStack> stack; - - synchronized(cache) { - // Get a stack that holds arrays of T[size]. If no such stack - // exists, allocate a new one. If the cache already had STACKS_MAX - // number of stacks, the least recently used stack is removed by - // cache.put (it calls removeEldestEntry). - stack = cache.get(size); - if (stack == null) { - stack = new CyclicStack>(); - cache.put(size, stack); - } - } - - stack.push(new SoftReference(array)); - } - - /** - * Allocates a new byte array, hopefully reusing an existing - * array from the cache. - * - * @param size size of the array to allocate - * - * @param fillWithZeros - * if true, all the elements of the returned - * array will be zero; if false, the contents - * of the returned array is undefined - */ - public byte[] getByteArray(int size, boolean fillWithZeros) { - byte[] array = getArray(byteArrayCache, size); - - if (array == null) - array = new byte[size]; - else if (fillWithZeros) - Arrays.fill(array, (byte)0x00); - - return array; - } - - /** - * Puts the given byte array to the cache. The caller must no longer - * use the array. - *

- * Small arrays aren't cached and will be ignored by this method. - */ - public void putArray(byte[] array) { - putArray(byteArrayCache, array, array.length); - } - - /** - * This is like getByteArray but for int arrays. - */ - public int[] getIntArray(int size, boolean fillWithZeros) { - int[] array = getArray(intArrayCache, size); - - if (array == null) - array = new int[size]; - else if (fillWithZeros) - Arrays.fill(array, 0); - - return array; - } - - /** - * Puts the given int array to the cache. The caller must no longer - * use the array. - *

- * Small arrays aren't cached and will be ignored by this method. - */ - public void putArray(int[] array) { - putArray(intArrayCache, array, array.length); - } -} Index: trunk/src/org/tukaani/xz/BlockInputStream.java =================================================================== --- trunk/src/org/tukaani/xz/BlockInputStream.java (revision 16021) +++ (revision ) @@ -1,305 +1,0 @@ -/* - * BlockInputStream - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import java.io.InputStream; -import java.io.DataInputStream; -import java.io.ByteArrayInputStream; -import java.io.IOException; -import java.util.Arrays; -import org.tukaani.xz.common.DecoderUtil; -import org.tukaani.xz.check.Check; - -class BlockInputStream extends InputStream { - private final DataInputStream inData; - private final CountingInputStream inCounted; - private InputStream filterChain; - private final Check check; - private final boolean verifyCheck; - - private long uncompressedSizeInHeader = -1; - private long compressedSizeInHeader = -1; - private long compressedSizeLimit; - private final int headerSize; - private long uncompressedSize = 0; - private boolean endReached = false; - - private final byte[] tempBuf = new byte[1]; - - public BlockInputStream(InputStream in, - Check check, boolean verifyCheck, - int memoryLimit, - long unpaddedSizeInIndex, - long uncompressedSizeInIndex, - ArrayCache arrayCache) - throws IOException, IndexIndicatorException { - this.check = check; - this.verifyCheck = verifyCheck; - inData = new DataInputStream(in); - - byte[] buf = new byte[DecoderUtil.BLOCK_HEADER_SIZE_MAX]; - - // Block Header Size or Index Indicator - inData.readFully(buf, 0, 1); - - // See if this begins the Index field. - if (buf[0] == 0x00) - throw new IndexIndicatorException(); - - // Read the rest of the Block Header. - headerSize = 4 * ((buf[0] & 0xFF) + 1); - inData.readFully(buf, 1, headerSize - 1); - - // Validate the CRC32. - if (!DecoderUtil.isCRC32Valid(buf, 0, headerSize - 4, headerSize - 4)) - throw new CorruptedInputException("XZ Block Header is corrupt"); - - // Check for reserved bits in Block Flags. - if ((buf[1] & 0x3C) != 0) - throw new UnsupportedOptionsException( - "Unsupported options in XZ Block Header"); - - // Memory for the Filter Flags field - int filterCount = (buf[1] & 0x03) + 1; - long[] filterIDs = new long[filterCount]; - byte[][] filterProps = new byte[filterCount][]; - - // Use a stream to parse the fields after the Block Flags field. - // Exclude the CRC32 field at the end. - ByteArrayInputStream bufStream = new ByteArrayInputStream( - buf, 2, headerSize - 6); - - try { - // Set the maximum valid compressed size. This is overridden - // by the value from the Compressed Size field if it is present. - compressedSizeLimit = (DecoderUtil.VLI_MAX & ~3) - - headerSize - check.getSize(); - - // Decode and validate Compressed Size if the relevant flag - // is set in Block Flags. - if ((buf[1] & 0x40) != 0x00) { - compressedSizeInHeader = DecoderUtil.decodeVLI(bufStream); - - if (compressedSizeInHeader == 0 - || compressedSizeInHeader > compressedSizeLimit) - throw new CorruptedInputException(); - - compressedSizeLimit = compressedSizeInHeader; - } - - // Decode Uncompressed Size if the relevant flag is set - // in Block Flags. - if ((buf[1] & 0x80) != 0x00) - uncompressedSizeInHeader = DecoderUtil.decodeVLI(bufStream); - - // Decode Filter Flags. - for (int i = 0; i < filterCount; ++i) { - filterIDs[i] = DecoderUtil.decodeVLI(bufStream); - - long filterPropsSize = DecoderUtil.decodeVLI(bufStream); - if (filterPropsSize > bufStream.available()) - throw new CorruptedInputException(); - - filterProps[i] = new byte[(int)filterPropsSize]; - bufStream.read(filterProps[i]); - } - - } catch (IOException e) { - throw new CorruptedInputException("XZ Block Header is corrupt"); - } - - // Check that the remaining bytes are zero. - for (int i = bufStream.available(); i > 0; --i) - if (bufStream.read() != 0x00) - throw new UnsupportedOptionsException( - "Unsupported options in XZ Block Header"); - - // Validate the Blcok Header against the Index when doing - // random access reading. - if (unpaddedSizeInIndex != -1) { - // Compressed Data must be at least one byte, so if Block Header - // and Check alone take as much or more space than the size - // stored in the Index, the file is corrupt. - int headerAndCheckSize = headerSize + check.getSize(); - if (headerAndCheckSize >= unpaddedSizeInIndex) - throw new CorruptedInputException( - "XZ Index does not match a Block Header"); - - // The compressed size calculated from Unpadded Size must - // match the value stored in the Compressed Size field in - // the Block Header. - long compressedSizeFromIndex - = unpaddedSizeInIndex - headerAndCheckSize; - if (compressedSizeFromIndex > compressedSizeLimit - || (compressedSizeInHeader != -1 - && compressedSizeInHeader != compressedSizeFromIndex)) - throw new CorruptedInputException( - "XZ Index does not match a Block Header"); - - // The uncompressed size stored in the Index must match - // the value stored in the Uncompressed Size field in - // the Block Header. - if (uncompressedSizeInHeader != -1 - && uncompressedSizeInHeader != uncompressedSizeInIndex) - throw new CorruptedInputException( - "XZ Index does not match a Block Header"); - - // For further validation, pretend that the values from the Index - // were stored in the Block Header. - compressedSizeLimit = compressedSizeFromIndex; - compressedSizeInHeader = compressedSizeFromIndex; - uncompressedSizeInHeader = uncompressedSizeInIndex; - } - - // Check if the Filter IDs are supported, decode - // the Filter Properties, and check that they are - // supported by this decoder implementation. - FilterDecoder[] filters = new FilterDecoder[filterIDs.length]; - - for (int i = 0; i < filters.length; ++i) { - if (filterIDs[i] == LZMA2Coder.FILTER_ID) - filters[i] = new LZMA2Decoder(filterProps[i]); - - else if (filterIDs[i] == DeltaCoder.FILTER_ID) - filters[i] = new DeltaDecoder(filterProps[i]); - - else if (BCJDecoder.isBCJFilterID(filterIDs[i])) - filters[i] = new BCJDecoder(filterIDs[i], filterProps[i]); - - else - throw new UnsupportedOptionsException( - "Unknown Filter ID " + filterIDs[i]); - } - - RawCoder.validate(filters); - - // Check the memory usage limit. - if (memoryLimit >= 0) { - int memoryNeeded = 0; - for (int i = 0; i < filters.length; ++i) - memoryNeeded += filters[i].getMemoryUsage(); - - if (memoryNeeded > memoryLimit) - throw new MemoryLimitException(memoryNeeded, memoryLimit); - } - - // Use an input size counter to calculate - // the size of the Compressed Data field. - inCounted = new CountingInputStream(in); - - // Initialize the filter chain. - filterChain = inCounted; - for (int i = filters.length - 1; i >= 0; --i) - filterChain = filters[i].getInputStream(filterChain, arrayCache); - } - - public int read() throws IOException { - return read(tempBuf, 0, 1) == -1 ? -1 : (tempBuf[0] & 0xFF); - } - - public int read(byte[] buf, int off, int len) throws IOException { - if (endReached) - return -1; - - int ret = filterChain.read(buf, off, len); - - if (ret > 0) { - if (verifyCheck) - check.update(buf, off, ret); - - uncompressedSize += ret; - - // Catch invalid values. - long compressedSize = inCounted.getSize(); - if (compressedSize < 0 - || compressedSize > compressedSizeLimit - || uncompressedSize < 0 - || (uncompressedSizeInHeader != -1 - && uncompressedSize > uncompressedSizeInHeader)) - throw new CorruptedInputException(); - - // Check the Block integrity as soon as possible: - // - The filter chain shouldn't return less than requested - // unless it hit the end of the input. - // - If the uncompressed size is known, we know when there - // shouldn't be more data coming. We still need to read - // one byte to let the filter chain catch errors and to - // let it read end of payload marker(s). - if (ret < len || uncompressedSize == uncompressedSizeInHeader) { - if (filterChain.read() != -1) - throw new CorruptedInputException(); - - validate(); - endReached = true; - } - } else if (ret == -1) { - validate(); - endReached = true; - } - - return ret; - } - - private void validate() throws IOException { - long compressedSize = inCounted.getSize(); - - // Validate Compressed Size and Uncompressed Size if they were - // present in Block Header. - if ((compressedSizeInHeader != -1 - && compressedSizeInHeader != compressedSize) - || (uncompressedSizeInHeader != -1 - && uncompressedSizeInHeader != uncompressedSize)) - throw new CorruptedInputException(); - - // Block Padding bytes must be zeros. - while ((compressedSize++ & 3) != 0) - if (inData.readUnsignedByte() != 0x00) - throw new CorruptedInputException(); - - // Validate the integrity check if verifyCheck is true. - byte[] storedCheck = new byte[check.getSize()]; - inData.readFully(storedCheck); - if (verifyCheck && !Arrays.equals(check.finish(), storedCheck)) - throw new CorruptedInputException("Integrity check (" - + check.getName() + ") does not match"); - } - - public int available() throws IOException { - return filterChain.available(); - } - - public void close() { - // This puts all arrays, that were allocated from ArrayCache, - // back to the ArrayCache. The last filter in the chain will - // call inCounted.close() which, being an instance of - // CloseIgnoringInputStream, won't close() the InputStream that - // was provided by the application. - try { - filterChain.close(); - } catch (IOException e) { - // It's a bug if we get here. The InputStreams that we are closing - // are all from this package and they are known to not throw - // IOException. (They could throw an IOException if we were - // closing the application-supplied InputStream, but - // inCounted.close() doesn't do that.) - assert false; - } - - filterChain = null; - } - - public long getUnpaddedSize() { - return headerSize + inCounted.getSize() + check.getSize(); - } - - public long getUncompressedSize() { - return uncompressedSize; - } -} Index: trunk/src/org/tukaani/xz/BlockOutputStream.java =================================================================== --- trunk/src/org/tukaani/xz/BlockOutputStream.java (revision 16021) +++ (revision ) @@ -1,135 +1,0 @@ -/* - * BlockOutputStream - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import java.io.OutputStream; -import java.io.ByteArrayOutputStream; -import java.io.IOException; -import org.tukaani.xz.common.EncoderUtil; -import org.tukaani.xz.check.Check; - -class BlockOutputStream extends FinishableOutputStream { - private final OutputStream out; - private final CountingOutputStream outCounted; - private FinishableOutputStream filterChain; - private final Check check; - - private final int headerSize; - private final long compressedSizeLimit; - private long uncompressedSize = 0; - - private final byte[] tempBuf = new byte[1]; - - public BlockOutputStream(OutputStream out, FilterEncoder[] filters, - Check check, ArrayCache arrayCache) - throws IOException { - this.out = out; - this.check = check; - - // Initialize the filter chain. - outCounted = new CountingOutputStream(out); - filterChain = outCounted; - for (int i = filters.length - 1; i >= 0; --i) - filterChain = filters[i].getOutputStream(filterChain, arrayCache); - - // Prepare to encode the Block Header field. - ByteArrayOutputStream bufStream = new ByteArrayOutputStream(); - - // Write a dummy Block Header Size field. The real value is written - // once everything else except CRC32 has been written. - bufStream.write(0x00); - - // Write Block Flags. Storing Compressed Size or Uncompressed Size - // isn't supported for now. - bufStream.write(filters.length - 1); - - // List of Filter Flags - for (int i = 0; i < filters.length; ++i) { - EncoderUtil.encodeVLI(bufStream, filters[i].getFilterID()); - byte[] filterProps = filters[i].getFilterProps(); - EncoderUtil.encodeVLI(bufStream, filterProps.length); - bufStream.write(filterProps); - } - - // Header Padding - while ((bufStream.size() & 3) != 0) - bufStream.write(0x00); - - byte[] buf = bufStream.toByteArray(); - - // Total size of the Block Header: Take the size of the CRC32 field - // into account. - headerSize = buf.length + 4; - - // This is just a sanity check. - if (headerSize > EncoderUtil.BLOCK_HEADER_SIZE_MAX) - throw new UnsupportedOptionsException(); - - // Block Header Size - buf[0] = (byte)(buf.length / 4); - - // Write the Block Header field to the output stream. - out.write(buf); - EncoderUtil.writeCRC32(out, buf); - - // Calculate the maximum allowed size of the Compressed Data field. - // It is hard to exceed it so this is mostly to be pedantic. - compressedSizeLimit = (EncoderUtil.VLI_MAX & ~3) - - headerSize - check.getSize(); - } - - public void write(int b) throws IOException { - tempBuf[0] = (byte)b; - write(tempBuf, 0, 1); - } - - public void write(byte[] buf, int off, int len) throws IOException { - filterChain.write(buf, off, len); - check.update(buf, off, len); - uncompressedSize += len; - validate(); - } - - public void flush() throws IOException { - filterChain.flush(); - validate(); - } - - public void finish() throws IOException { - // Finish the Compressed Data field. - filterChain.finish(); - validate(); - - // Block Padding - for (long i = outCounted.getSize(); (i & 3) != 0; ++i) - out.write(0x00); - - // Check - out.write(check.finish()); - } - - private void validate() throws IOException { - long compressedSize = outCounted.getSize(); - - // It is very hard to trigger this exception. - // This is just to be pedantic. - if (compressedSize < 0 || compressedSize > compressedSizeLimit - || uncompressedSize < 0) - throw new XZIOException("XZ Stream has grown too big"); - } - - public long getUnpaddedSize() { - return headerSize + outCounted.getSize() + check.getSize(); - } - - public long getUncompressedSize() { - return uncompressedSize; - } -} Index: trunk/src/org/tukaani/xz/CloseIgnoringInputStream.java =================================================================== --- trunk/src/org/tukaani/xz/CloseIgnoringInputStream.java (revision 16021) +++ (revision ) @@ -1,48 +1,0 @@ -/* - * CloseIgnoringInputStream - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import java.io.InputStream; -import java.io.FilterInputStream; - -/** - * An {@code InputStream} wrapper whose {@code close()} does nothing. - * This is useful with raw decompressors if you want to call - * {@code close()} to release memory allocated from an {@link ArrayCache} - * but don't want to close the underlying {@code InputStream}. - * For example: - * 

- * InputStream rawdec = new LZMA2InputStream(
- *         new CloseIgnoringInputStream(myInputStream),
- *         myDictSize, null, myArrayCache);
- * doSomething(rawdec);
- * rawdec.close(); // This doesn't close myInputStream.
- *
- *

- * With {@link XZInputStream}, {@link SingleXZInputStream}, and - * {@link SeekableXZInputStream} you can use their {@code close(boolean)} - * method to avoid closing the underlying {@code InputStream}; with - * those classes {@code CloseIgnoringInputStream} isn't needed. - * - * @since 1.7 - */ -public class CloseIgnoringInputStream extends FilterInputStream { - /** - * Creates a new {@code CloseIgnoringInputStream}. - */ - public CloseIgnoringInputStream(InputStream in) { - super(in); - } - - /** - * This does nothing (doesn't call {@code in.close()}). - */ - public void close() {} -} Index: trunk/src/org/tukaani/xz/CorruptedInputException.java =================================================================== --- trunk/src/org/tukaani/xz/CorruptedInputException.java (revision 16021) +++ (revision ) @@ -1,37 +1,0 @@ -/* - * CorruptedInputException - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -/** - * Thrown when the compressed input data is corrupt. - * However, it is possible that some or all of the data - * already read from the input stream was corrupt too. - */ -public class CorruptedInputException extends XZIOException { - private static final long serialVersionUID = 3L; - - /** - * Creates a new CorruptedInputException with - * the default error detail message. - */ - public CorruptedInputException() { - super("Compressed data is corrupt"); - } - - /** - * Creates a new CorruptedInputException with - * the specified error detail message. - * - * @param s error detail message - */ - public CorruptedInputException(String s) { - super(s); - } -} Index: trunk/src/org/tukaani/xz/CountingInputStream.java =================================================================== --- trunk/src/org/tukaani/xz/CountingInputStream.java (revision 16021) +++ (revision ) @@ -1,47 +1,0 @@ -/* - * CountingInputStream - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import java.io.FilterInputStream; -import java.io.InputStream; -import java.io.IOException; - -/** - * Counts the number of bytes read from an input stream. - * The close() method does nothing, that is, the underlying - * InputStream isn't closed. - */ -class CountingInputStream extends CloseIgnoringInputStream { - private long size = 0; - - public CountingInputStream(InputStream in) { - super(in); - } - - public int read() throws IOException { - int ret = in.read(); - if (ret != -1 && size >= 0) - ++size; - - return ret; - } - - public int read(byte[] b, int off, int len) throws IOException { - int ret = in.read(b, off, len); - if (ret > 0 && size >= 0) - size += ret; - - return ret; - } - - public long getSize() { - return size; - } -} Index: trunk/src/org/tukaani/xz/CountingOutputStream.java =================================================================== --- trunk/src/org/tukaani/xz/CountingOutputStream.java (revision 16021) +++ (revision ) @@ -1,54 +1,0 @@ -/* - * CountingOutputStream - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import java.io.OutputStream; -import java.io.IOException; - -/** - * Counts the number of bytes written to an output stream. - *

- * The finish method does nothing. - * This is FinishableOutputStream instead - * of OutputStream solely because it allows - * using this as the output stream for a chain of raw filters. - */ -class CountingOutputStream extends FinishableOutputStream { - private final OutputStream out; - private long size = 0; - - public CountingOutputStream(OutputStream out) { - this.out = out; - } - - public void write(int b) throws IOException { - out.write(b); - if (size >= 0) - ++size; - } - - public void write(byte[] b, int off, int len) throws IOException { - out.write(b, off, len); - if (size >= 0) - size += len; - } - - public void flush() throws IOException { - out.flush(); - } - - public void close() throws IOException { - out.close(); - } - - public long getSize() { - return size; - } -} Index: trunk/src/org/tukaani/xz/DeltaCoder.java =================================================================== --- trunk/src/org/tukaani/xz/DeltaCoder.java (revision 16021) +++ (revision ) @@ -1,26 +1,0 @@ -/* - * DeltaCoder - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -abstract class DeltaCoder implements FilterCoder { - public static final long FILTER_ID = 0x03; - - public boolean changesSize() { - return false; - } - - public boolean nonLastOK() { - return true; - } - - public boolean lastOK() { - return false; - } -} Index: trunk/src/org/tukaani/xz/DeltaDecoder.java =================================================================== --- trunk/src/org/tukaani/xz/DeltaDecoder.java (revision 16021) +++ (revision ) @@ -1,32 +1,0 @@ -/* - * DeltaDecoder - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import java.io.InputStream; - -class DeltaDecoder extends DeltaCoder implements FilterDecoder { - private final int distance; - - DeltaDecoder(byte[] props) throws UnsupportedOptionsException { - if (props.length != 1) - throw new UnsupportedOptionsException( - "Unsupported Delta filter properties"); - - distance = (props[0] & 0xFF) + 1; - } - - public int getMemoryUsage() { - return 1; - } - - public InputStream getInputStream(InputStream in, ArrayCache arrayCache) { - return new DeltaInputStream(in, distance); - } -} Index: trunk/src/org/tukaani/xz/DeltaEncoder.java =================================================================== --- trunk/src/org/tukaani/xz/DeltaEncoder.java (revision 16021) +++ (revision ) @@ -1,37 +1,0 @@ -/* - * DeltaEncoder - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -class DeltaEncoder extends DeltaCoder implements FilterEncoder { - private final DeltaOptions options; - private final byte[] props = new byte[1]; - - DeltaEncoder(DeltaOptions options) { - props[0] = (byte)(options.getDistance() - 1); - this.options = (DeltaOptions)options.clone(); - } - - public long getFilterID() { - return FILTER_ID; - } - - public byte[] getFilterProps() { - return props; - } - - public boolean supportsFlushing() { - return true; - } - - public FinishableOutputStream getOutputStream(FinishableOutputStream out, - ArrayCache arrayCache) { - return options.getOutputStream(out, arrayCache); - } -} Index: trunk/src/org/tukaani/xz/DeltaInputStream.java =================================================================== --- trunk/src/org/tukaani/xz/DeltaInputStream.java (revision 16021) +++ (revision ) @@ -1,146 +1,0 @@ -/* - * DeltaInputStream - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import java.io.InputStream; -import java.io.IOException; -import org.tukaani.xz.delta.DeltaDecoder; - -/** - * Decodes raw Delta-filtered data (no XZ headers). - *

- * The delta filter doesn't change the size of the data and thus it - * cannot have an end-of-payload marker. It will simply decode until - * its input stream indicates end of input. - */ -public class DeltaInputStream extends InputStream { - /** - * Smallest supported delta calculation distance. - */ - public static final int DISTANCE_MIN = 1; - - /** - * Largest supported delta calculation distance. - */ - public static final int DISTANCE_MAX = 256; - - private InputStream in; - private final DeltaDecoder delta; - - private IOException exception = null; - - private final byte[] tempBuf = new byte[1]; - - /** - * Creates a new Delta decoder with the given delta calculation distance. - * - * @param in input stream from which Delta filtered data - * is read - * - * @param distance delta calculation distance, must be in the - * range [DISTANCE_MIN, - * DISTANCE_MAX] - */ - public DeltaInputStream(InputStream in, int distance) { - // Check for null because otherwise null isn't detect - // in this constructor. - if (in == null) - throw new NullPointerException(); - - this.in = in; - this.delta = new DeltaDecoder(distance); - } - - /** - * Decode the next byte from this input stream. - * - * @return the next decoded byte, or -1 to indicate - * the end of input on the input stream in - * - * @throws IOException may be thrown by in - */ - public int read() throws IOException { - return read(tempBuf, 0, 1) == -1 ? -1 : (tempBuf[0] & 0xFF); - } - - /** - * Decode into an array of bytes. - *

- * This calls in.read(buf, off, len) and defilters the - * returned data. - * - * @param buf target buffer for decoded data - * @param off start offset in buf - * @param len maximum number of bytes to read - * - * @return number of bytes read, or -1 to indicate - * the end of the input stream in - * - * @throws XZIOException if the stream has been closed - * - * @throws IOException may be thrown by underlaying input - * stream in - */ - public int read(byte[] buf, int off, int len) throws IOException { - if (len == 0) - return 0; - - if (in == null) - throw new XZIOException("Stream closed"); - - if (exception != null) - throw exception; - - int size; - try { - size = in.read(buf, off, len); - } catch (IOException e) { - exception = e; - throw e; - } - - if (size == -1) - return -1; - - delta.decode(buf, off, size); - return size; - } - - /** - * Calls in.available(). - * - * @return the value returned by in.available() - */ - public int available() throws IOException { - if (in == null) - throw new XZIOException("Stream closed"); - - if (exception != null) - throw exception; - - return in.available(); - } - - /** - * Closes the stream and calls in.close(). - * If the stream was already closed, this does nothing. - * - * @throws IOException if thrown by in.close() - */ - public void close() throws IOException { - if (in != null) { - try { - in.close(); - } finally { - in = null; - } - } - } -} Index: trunk/src/org/tukaani/xz/DeltaOptions.java =================================================================== --- trunk/src/org/tukaani/xz/DeltaOptions.java (revision 16021) +++ (revision ) @@ -1,103 +1,0 @@ -/* - * DeltaOptions - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import java.io.InputStream; - -/** - * Delta filter options. The Delta filter can be used only as a non-last - * filter in the chain, for example Delta + LZMA2. - *

- * Currently only simple byte-wise delta is supported. The only option - * is the delta distance, which you should set to match your data. - * It's not possible to provide a generic default value for it. - *

- * For example, with distance = 2 and eight-byte input - * A1 B1 A2 B3 A3 B5 A4 B7, the output will be A1 B1 01 02 01 02 01 02. - *

- * The Delta filter can be good with uncompressed bitmap images. It can - * also help with PCM audio, although special-purpose compressors like - * FLAC will give much smaller result at much better compression speed. - */ -public class DeltaOptions extends FilterOptions { - /** - * Smallest supported delta calculation distance. - */ - public static final int DISTANCE_MIN = 1; - - /** - * Largest supported delta calculation distance. - */ - public static final int DISTANCE_MAX = 256; - - private int distance = DISTANCE_MIN; - - /** - * Creates new Delta options and sets the delta distance to 1 byte. - */ - public DeltaOptions() {} - - /** - * Creates new Delta options and sets the distance to the given value. - */ - public DeltaOptions(int distance) throws UnsupportedOptionsException { - setDistance(distance); - } - - /** - * Sets the delta distance in bytes. The new distance must be in - * the range [DISTANCE_MIN, DISTANCE_MAX]. - */ - public void setDistance(int distance) throws UnsupportedOptionsException { - if (distance < DISTANCE_MIN || distance > DISTANCE_MAX) - throw new UnsupportedOptionsException( - "Delta distance must be in the range [" + DISTANCE_MIN - + ", " + DISTANCE_MAX + "]: " + distance); - - this.distance = distance; - } - - /** - * Gets the delta distance. - */ - public int getDistance() { - return distance; - } - - public int getEncoderMemoryUsage() { - return DeltaOutputStream.getMemoryUsage(); - } - - public FinishableOutputStream getOutputStream(FinishableOutputStream out, - ArrayCache arrayCache) { - return new DeltaOutputStream(out, this); - } - - public int getDecoderMemoryUsage() { - return 1; - } - - public InputStream getInputStream(InputStream in, ArrayCache arrayCache) { - return new DeltaInputStream(in, distance); - } - - FilterEncoder getFilterEncoder() { - return new DeltaEncoder(this); - } - - public Object clone() { - try { - return super.clone(); - } catch (CloneNotSupportedException e) { - assert false; - throw new RuntimeException(); - } - } -} Index: trunk/src/org/tukaani/xz/DeltaOutputStream.java =================================================================== --- trunk/src/org/tukaani/xz/DeltaOutputStream.java (revision 16021) +++ (revision ) @@ -1,113 +1,0 @@ -/* - * DeltaOutputStream - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import java.io.IOException; -import org.tukaani.xz.delta.DeltaEncoder; - -class DeltaOutputStream extends FinishableOutputStream { - private static final int FILTER_BUF_SIZE = 4096; - - private FinishableOutputStream out; - private final DeltaEncoder delta; - private final byte[] filterBuf = new byte[FILTER_BUF_SIZE]; - - private boolean finished = false; - private IOException exception = null; - - private final byte[] tempBuf = new byte[1]; - - static int getMemoryUsage() { - return 1 + FILTER_BUF_SIZE / 1024; - } - - DeltaOutputStream(FinishableOutputStream out, DeltaOptions options) { - this.out = out; - delta = new DeltaEncoder(options.getDistance()); - } - - public void write(int b) throws IOException { - tempBuf[0] = (byte)b; - write(tempBuf, 0, 1); - } - - public void write(byte[] buf, int off, int len) throws IOException { - if (off < 0 || len < 0 || off + len < 0 || off + len > buf.length) - throw new IndexOutOfBoundsException(); - - if (exception != null) - throw exception; - - if (finished) - throw new XZIOException("Stream finished"); - - try { - while (len > FILTER_BUF_SIZE) { - delta.encode(buf, off, FILTER_BUF_SIZE, filterBuf); - out.write(filterBuf); - off += FILTER_BUF_SIZE; - len -= FILTER_BUF_SIZE; - } - - delta.encode(buf, off, len, filterBuf); - out.write(filterBuf, 0, len); - } catch (IOException e) { - exception = e; - throw e; - } - } - - public void flush() throws IOException { - if (exception != null) - throw exception; - - if (finished) - throw new XZIOException("Stream finished or closed"); - - try { - out.flush(); - } catch (IOException e) { - exception = e; - throw e; - } - } - - public void finish() throws IOException { - if (!finished) { - if (exception != null) - throw exception; - - try { - out.finish(); - } catch (IOException e) { - exception = e; - throw e; - } - - finished = true; - } - } - - public void close() throws IOException { - if (out != null) { - try { - out.close(); - } catch (IOException e) { - if (exception == null) - exception = e; - } - - out = null; - } - - if (exception != null) - throw exception; - } -} Index: trunk/src/org/tukaani/xz/FilterCoder.java =================================================================== --- trunk/src/org/tukaani/xz/FilterCoder.java (revision 16021) +++ (revision ) @@ -1,16 +1,0 @@ -/* - * FilterCoder - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -interface FilterCoder { - boolean changesSize(); - boolean nonLastOK(); - boolean lastOK(); -} Index: trunk/src/org/tukaani/xz/FilterDecoder.java =================================================================== --- trunk/src/org/tukaani/xz/FilterDecoder.java (revision 16021) +++ (revision ) @@ -1,17 +1,0 @@ -/* - * FilterDecoder - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import java.io.InputStream; - -interface FilterDecoder extends FilterCoder { - int getMemoryUsage(); - InputStream getInputStream(InputStream in, ArrayCache arrayCache); -} Index: trunk/src/org/tukaani/xz/FilterEncoder.java =================================================================== --- trunk/src/org/tukaani/xz/FilterEncoder.java (revision 16021) +++ (revision ) @@ -1,18 +1,0 @@ -/* - * FilterEncoder - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -interface FilterEncoder extends FilterCoder { - long getFilterID(); - byte[] getFilterProps(); - boolean supportsFlushing(); - FinishableOutputStream getOutputStream(FinishableOutputStream out, - ArrayCache arrayCache); -} Index: trunk/src/org/tukaani/xz/FilterOptions.java =================================================================== --- trunk/src/org/tukaani/xz/FilterOptions.java (revision 16021) +++ (revision ) @@ -1,104 +1,0 @@ -/* - * FilterOptions - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import java.io.InputStream; -import java.io.IOException; - -/** - * Base class for filter-specific options classes. - */ -public abstract class FilterOptions implements Cloneable { - /** - * Gets how much memory the encoder will need with - * the given filter chain. This function simply calls - * getEncoderMemoryUsage() for every filter - * in the array and returns the sum of the returned values. - */ - public static int getEncoderMemoryUsage(FilterOptions[] options) { - int m = 0; - - for (int i = 0; i < options.length; ++i) - m += options[i].getEncoderMemoryUsage(); - - return m; - } - - /** - * Gets how much memory the decoder will need with - * the given filter chain. This function simply calls - * getDecoderMemoryUsage() for every filter - * in the array and returns the sum of the returned values. - */ - public static int getDecoderMemoryUsage(FilterOptions[] options) { - int m = 0; - - for (int i = 0; i < options.length; ++i) - m += options[i].getDecoderMemoryUsage(); - - return m; - } - - /** - * Gets how much memory the encoder will need with these options. - */ - public abstract int getEncoderMemoryUsage(); - - /** - * Gets a raw (no XZ headers) encoder output stream using these options. - * Raw streams are an advanced feature. In most cases you want to store - * the compressed data in the .xz container format instead of using - * a raw stream. To use this filter in a .xz file, pass this object - * to XZOutputStream. - *

- * This is uses ArrayCache.getDefaultCache() as the ArrayCache. - */ - public FinishableOutputStream getOutputStream(FinishableOutputStream out) { - return getOutputStream(out, ArrayCache.getDefaultCache()); - } - - /** - * Gets a raw (no XZ headers) encoder output stream using these options - * and the given ArrayCache. - * Raw streams are an advanced feature. In most cases you want to store - * the compressed data in the .xz container format instead of using - * a raw stream. To use this filter in a .xz file, pass this object - * to XZOutputStream. - */ - public abstract FinishableOutputStream getOutputStream( - FinishableOutputStream out, ArrayCache arrayCache); - - /** - * Gets how much memory the decoder will need to decompress the data - * that was encoded with these options. - */ - public abstract int getDecoderMemoryUsage(); - - /** - * Gets a raw (no XZ headers) decoder input stream using these options. - *

- * This is uses ArrayCache.getDefaultCache() as the ArrayCache. - */ - public InputStream getInputStream(InputStream in) throws IOException { - return getInputStream(in, ArrayCache.getDefaultCache()); - } - - /** - * Gets a raw (no XZ headers) decoder input stream using these options - * and the given ArrayCache. - */ - public abstract InputStream getInputStream( - InputStream in, ArrayCache arrayCache) throws IOException; - - abstract FilterEncoder getFilterEncoder(); - - FilterOptions() {} -} Index: trunk/src/org/tukaani/xz/FinishableOutputStream.java =================================================================== --- trunk/src/org/tukaani/xz/FinishableOutputStream.java (revision 16021) +++ (revision ) @@ -1,31 +1,0 @@ -/* - * FinishableOutputStream - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import java.io.OutputStream; -import java.io.IOException; - -/** - * Output stream that supports finishing without closing - * the underlying stream. - */ -public abstract class FinishableOutputStream extends OutputStream { - /** - * Finish the stream without closing the underlying stream. - * No more data may be written to the stream after finishing. - *

- * The finish method of FinishableOutputStream - * does nothing. Subclasses should override it if they need finishing - * support, which is the case, for example, with compressors. - * - * @throws IOException - */ - public void finish() throws IOException {} -} Index: trunk/src/org/tukaani/xz/IndexIndicatorException.java =================================================================== --- trunk/src/org/tukaani/xz/IndexIndicatorException.java (revision 16021) +++ (revision ) @@ -1,14 +1,0 @@ -/* - * IndexIndicatorException - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -class IndexIndicatorException extends Exception { - private static final long serialVersionUID = 1L; -} Index: trunk/src/org/tukaani/xz/LZMA2Coder.java =================================================================== --- trunk/src/org/tukaani/xz/LZMA2Coder.java (revision 16021) +++ (revision ) @@ -1,26 +1,0 @@ -/* - * LZMA2Coder - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -abstract class LZMA2Coder implements FilterCoder { - public static final long FILTER_ID = 0x21; - - public boolean changesSize() { - return true; - } - - public boolean nonLastOK() { - return false; - } - - public boolean lastOK() { - return true; - } -} Index: trunk/src/org/tukaani/xz/LZMA2Decoder.java =================================================================== --- trunk/src/org/tukaani/xz/LZMA2Decoder.java (revision 16021) +++ (revision ) @@ -1,35 +1,0 @@ -/* - * LZMA2Decoder - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import java.io.InputStream; - -class LZMA2Decoder extends LZMA2Coder implements FilterDecoder { - private int dictSize; - - LZMA2Decoder(byte[] props) throws UnsupportedOptionsException { - // Up to 1.5 GiB dictionary is supported. The bigger ones - // are too big for int. - if (props.length != 1 || (props[0] & 0xFF) > 37) - throw new UnsupportedOptionsException( - "Unsupported LZMA2 properties"); - - dictSize = 2 | (props[0] & 1); - dictSize <<= (props[0] >>> 1) + 11; - } - - public int getMemoryUsage() { - return LZMA2InputStream.getMemoryUsage(dictSize); - } - - public InputStream getInputStream(InputStream in, ArrayCache arrayCache) { - return new LZMA2InputStream(in, dictSize, null, arrayCache); - } -} Index: trunk/src/org/tukaani/xz/LZMA2Encoder.java =================================================================== --- trunk/src/org/tukaani/xz/LZMA2Encoder.java (revision 16021) +++ (revision ) @@ -1,51 +1,0 @@ -/* - * LZMA2Encoder - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import org.tukaani.xz.lzma.LZMAEncoder; - -class LZMA2Encoder extends LZMA2Coder implements FilterEncoder { - private final LZMA2Options options; - private final byte[] props = new byte[1]; - - LZMA2Encoder(LZMA2Options options) { - if (options.getPresetDict() != null) - throw new IllegalArgumentException( - "XZ doesn't support a preset dictionary for now"); - - if (options.getMode() == LZMA2Options.MODE_UNCOMPRESSED) { - props[0] = (byte)0; - } else { - int d = Math.max(options.getDictSize(), - LZMA2Options.DICT_SIZE_MIN); - props[0] = (byte)(LZMAEncoder.getDistSlot(d - 1) - 23); - } - - // Make a private copy so that the caller is free to change its copy. - this.options = (LZMA2Options)options.clone(); - } - - public long getFilterID() { - return FILTER_ID; - } - - public byte[] getFilterProps() { - return props; - } - - public boolean supportsFlushing() { - return true; - } - - public FinishableOutputStream getOutputStream(FinishableOutputStream out, - ArrayCache arrayCache) { - return options.getOutputStream(out, arrayCache); - } -} Index: trunk/src/org/tukaani/xz/LZMA2InputStream.java =================================================================== --- trunk/src/org/tukaani/xz/LZMA2InputStream.java (revision 16021) +++ (revision ) @@ -1,400 +1,0 @@ -/* - * LZMA2InputStream - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import java.io.InputStream; -import java.io.DataInputStream; -import java.io.IOException; -import org.tukaani.xz.lz.LZDecoder; -import org.tukaani.xz.rangecoder.RangeDecoderFromBuffer; -import org.tukaani.xz.lzma.LZMADecoder; - -/** - * Decompresses a raw LZMA2 stream (no XZ headers). - */ -public class LZMA2InputStream extends InputStream { - /** - * Smallest valid LZMA2 dictionary size. - *

- * Very tiny dictionaries would be a performance problem, so - * the minimum is 4 KiB. - */ - public static final int DICT_SIZE_MIN = 4096; - - /** - * Largest dictionary size supported by this implementation. - *

- * The LZMA2 algorithm allows dictionaries up to one byte less than 4 GiB. - * This implementation supports only 16 bytes less than 2 GiB for raw - * LZMA2 streams, and for .xz files the maximum is 1.5 GiB. This - * limitation is due to Java using signed 32-bit integers for array - * indexing. The limitation shouldn't matter much in practice since so - * huge dictionaries are not normally used. - */ - public static final int DICT_SIZE_MAX = Integer.MAX_VALUE & ~15; - - private static final int COMPRESSED_SIZE_MAX = 1 << 16; - - private final ArrayCache arrayCache; - private DataInputStream in; - - private LZDecoder lz; - private RangeDecoderFromBuffer rc; - private LZMADecoder lzma; - - private int uncompressedSize = 0; - private boolean isLZMAChunk = false; - - private boolean needDictReset = true; - private boolean needProps = true; - private boolean endReached = false; - - private IOException exception = null; - - private final byte[] tempBuf = new byte[1]; - - /** - * Gets approximate decompressor memory requirements as kibibytes for - * the given dictionary size. - * - * @param dictSize LZMA2 dictionary size as bytes, must be - * in the range [DICT_SIZE_MIN, - * DICT_SIZE_MAX] - * - * @return approximate memory requirements as kibibytes (KiB) - */ - public static int getMemoryUsage(int dictSize) { - // The base state is around 30-40 KiB (probabilities etc.), - // range decoder needs COMPRESSED_SIZE_MAX bytes for buffering, - // and LZ decoder needs a dictionary buffer. - return 40 + COMPRESSED_SIZE_MAX / 1024 + getDictSize(dictSize) / 1024; - } - - private static int getDictSize(int dictSize) { - if (dictSize < DICT_SIZE_MIN || dictSize > DICT_SIZE_MAX) - throw new IllegalArgumentException( - "Unsupported dictionary size " + dictSize); - - // Round dictionary size upward to a multiple of 16. This way LZMA - // can use LZDecoder.getPos() for calculating LZMA's posMask. - // Note that this check is needed only for raw LZMA2 streams; it is - // redundant with .xz. - return (dictSize + 15) & ~15; - } - - /** - * Creates a new input stream that decompresses raw LZMA2 data - * from in. - *

- * The caller needs to know the dictionary size used when compressing; - * the dictionary size isn't stored as part of a raw LZMA2 stream. - *

- * Specifying a too small dictionary size will prevent decompressing - * the stream. Specifying a too big dictionary is waste of memory but - * decompression will work. - *

- * There is no need to specify a dictionary bigger than - * the uncompressed size of the data even if a bigger dictionary - * was used when compressing. If you know the uncompressed size - * of the data, this might allow saving some memory. - * - * @param in input stream from which LZMA2-compressed - * data is read - * - * @param dictSize LZMA2 dictionary size as bytes, must be - * in the range [DICT_SIZE_MIN, - * DICT_SIZE_MAX] - */ - public LZMA2InputStream(InputStream in, int dictSize) { - this(in, dictSize, null); - } - - /** - * Creates a new LZMA2 decompressor using a preset dictionary. - *

- * This is like LZMA2InputStream(InputStream, int) except - * that the dictionary may be initialized using a preset dictionary. - * If a preset dictionary was used when compressing the data, the - * same preset dictionary must be provided when decompressing. - * - * @param in input stream from which LZMA2-compressed - * data is read - * - * @param dictSize LZMA2 dictionary size as bytes, must be - * in the range [DICT_SIZE_MIN, - * DICT_SIZE_MAX] - * - * @param presetDict preset dictionary or null - * to use no preset dictionary - */ - public LZMA2InputStream(InputStream in, int dictSize, byte[] presetDict) { - this(in, dictSize, presetDict, ArrayCache.getDefaultCache()); - } - - /** - * Creates a new LZMA2 decompressor using a preset dictionary - * and array cache. - *

- * This is like LZMA2InputStream(InputStream, int, byte[]) - * except that this also takes the arrayCache argument. - * - * @param in input stream from which LZMA2-compressed - * data is read - * - * @param dictSize LZMA2 dictionary size as bytes, must be - * in the range [DICT_SIZE_MIN, - * DICT_SIZE_MAX] - * - * @param presetDict preset dictionary or null - * to use no preset dictionary - * - * @param arrayCache cache to be used for allocating large arrays - * - * @since 1.7 - */ - LZMA2InputStream(InputStream in, int dictSize, byte[] presetDict, - ArrayCache arrayCache) { - // Check for null because otherwise null isn't detect - // in this constructor. - if (in == null) - throw new NullPointerException(); - - this.arrayCache = arrayCache; - this.in = new DataInputStream(in); - this.rc = new RangeDecoderFromBuffer(COMPRESSED_SIZE_MAX, arrayCache); - this.lz = new LZDecoder(getDictSize(dictSize), presetDict, arrayCache); - - if (presetDict != null && presetDict.length > 0) - needDictReset = false; - } - - /** - * Decompresses the next byte from this input stream. - *

- * Reading lots of data with read() from this input stream - * may be inefficient. Wrap it in java.io.BufferedInputStream - * if you need to read lots of data one byte at a time. - * - * @return the next decompressed byte, or -1 - * to indicate the end of the compressed stream - * - * @throws CorruptedInputException - * - * @throws XZIOException if the stream has been closed - * - * @throws EOFException - * compressed input is truncated or corrupt - * - * @throws IOException may be thrown by in - */ - public int read() throws IOException { - return read(tempBuf, 0, 1) == -1 ? -1 : (tempBuf[0] & 0xFF); - } - - /** - * Decompresses into an array of bytes. - *

- * If len is zero, no bytes are read and 0 - * is returned. Otherwise this will block until len - * bytes have been decompressed, the end of the LZMA2 stream is reached, - * or an exception is thrown. - * - * @param buf target buffer for uncompressed data - * @param off start offset in buf - * @param len maximum number of uncompressed bytes to read - * - * @return number of bytes read, or -1 to indicate - * the end of the compressed stream - * - * @throws CorruptedInputException - * - * @throws XZIOException if the stream has been closed - * - * @throws EOFException - * compressed input is truncated or corrupt - * - * @throws IOException may be thrown by in - */ - public int read(byte[] buf, int off, int len) throws IOException { - if (off < 0 || len < 0 || off + len < 0 || off + len > buf.length) - throw new IndexOutOfBoundsException(); - - if (len == 0) - return 0; - - if (in == null) - throw new XZIOException("Stream closed"); - - if (exception != null) - throw exception; - - if (endReached) - return -1; - - try { - int size = 0; - - while (len > 0) { - if (uncompressedSize == 0) { - decodeChunkHeader(); - if (endReached) - return size == 0 ? -1 : size; - } - - int copySizeMax = Math.min(uncompressedSize, len); - - if (!isLZMAChunk) { - lz.copyUncompressed(in, copySizeMax); - } else { - lz.setLimit(copySizeMax); - lzma.decode(); - } - - int copiedSize = lz.flush(buf, off); - off += copiedSize; - len -= copiedSize; - size += copiedSize; - uncompressedSize -= copiedSize; - - if (uncompressedSize == 0) - if (!rc.isFinished() || lz.hasPending()) - throw new CorruptedInputException(); - } - - return size; - - } catch (IOException e) { - exception = e; - throw e; - } - } - - private void decodeChunkHeader() throws IOException { - int control = in.readUnsignedByte(); - - if (control == 0x00) { - endReached = true; - putArraysToCache(); - return; - } - - if (control >= 0xE0 || control == 0x01) { - needProps = true; - needDictReset = false; - lz.reset(); - } else if (needDictReset) { - throw new CorruptedInputException(); - } - - if (control >= 0x80) { - isLZMAChunk = true; - - uncompressedSize = (control & 0x1F) << 16; - uncompressedSize += in.readUnsignedShort() + 1; - - int compressedSize = in.readUnsignedShort() + 1; - - if (control >= 0xC0) { - needProps = false; - decodeProps(); - - } else if (needProps) { - throw new CorruptedInputException(); - - } else if (control >= 0xA0) { - lzma.reset(); - } - - rc.prepareInputBuffer(in, compressedSize); - - } else if (control > 0x02) { - throw new CorruptedInputException(); - - } else { - isLZMAChunk = false; - uncompressedSize = in.readUnsignedShort() + 1; - } - } - - private void decodeProps() throws IOException { - int props = in.readUnsignedByte(); - - if (props > (4 * 5 + 4) * 9 + 8) - throw new CorruptedInputException(); - - int pb = props / (9 * 5); - props -= pb * 9 * 5; - int lp = props / 9; - int lc = props - lp * 9; - - if (lc + lp > 4) - throw new CorruptedInputException(); - - lzma = new LZMADecoder(lz, rc, lc, lp, pb); - } - - /** - * Returns the number of uncompressed bytes that can be read - * without blocking. The value is returned with an assumption - * that the compressed input data will be valid. If the compressed - * data is corrupt, CorruptedInputException may get - * thrown before the number of bytes claimed to be available have - * been read from this input stream. - *

- * In LZMA2InputStream, the return value will be non-zero when the - * decompressor is in the middle of an LZMA2 chunk. The return value - * will then be the number of uncompressed bytes remaining from that - * chunk. The return value can also be non-zero in the middle of - * an uncompressed chunk, but then the return value depends also on - * the available() method of the underlying InputStream. - * - * @return the number of uncompressed bytes that can be read - * without blocking - */ - public int available() throws IOException { - if (in == null) - throw new XZIOException("Stream closed"); - - if (exception != null) - throw exception; - - return isLZMAChunk ? uncompressedSize - : Math.min(uncompressedSize, in.available()); - } - - private void putArraysToCache() { - if (lz != null) { - lz.putArraysToCache(arrayCache); - lz = null; - - rc.putArraysToCache(arrayCache); - rc = null; - } - } - - /** - * Closes the stream and calls in.close(). - * If the stream was already closed, this does nothing. - * - * @throws IOException if thrown by in.close() - */ - public void close() throws IOException { - if (in != null) { - putArraysToCache(); - - try { - in.close(); - } finally { - in = null; - } - } - } -} Index: trunk/src/org/tukaani/xz/LZMA2Options.java =================================================================== --- trunk/src/org/tukaani/xz/LZMA2Options.java (revision 16021) +++ (revision ) @@ -1,583 +1,0 @@ -/* - * LZMA2Options - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import java.io.InputStream; -import java.io.IOException; -import org.tukaani.xz.lz.LZEncoder; -import org.tukaani.xz.lzma.LZMAEncoder; - -/** - * LZMA2 compression options. - *

- * While this allows setting the LZMA2 compression options in detail, - * often you only need LZMA2Options() or - * LZMA2Options(int). - */ -public class LZMA2Options extends FilterOptions { - /** - * Minimum valid compression preset level is 0. - */ - public static final int PRESET_MIN = 0; - - /** - * Maximum valid compression preset level is 9. - */ - public static final int PRESET_MAX = 9; - - /** - * Default compression preset level is 6. - */ - public static final int PRESET_DEFAULT = 6; - - /** - * Minimum dictionary size is 4 KiB. - */ - public static final int DICT_SIZE_MIN = 4096; - - /** - * Maximum dictionary size for compression is 768 MiB. - *

- * The decompressor supports bigger dictionaries, up to almost 2 GiB. - * With HC4 the encoder would support dictionaries bigger than 768 MiB. - * The 768 MiB limit comes from the current implementation of BT4 where - * we would otherwise hit the limits of signed ints in array indexing. - *

- * If you really need bigger dictionary for decompression, - * use {@link LZMA2InputStream} directly. - */ - public static final int DICT_SIZE_MAX = 768 << 20; - - /** - * The default dictionary size is 8 MiB. - */ - public static final int DICT_SIZE_DEFAULT = 8 << 20; - - /** - * Maximum value for lc + lp is 4. - */ - public static final int LC_LP_MAX = 4; - - /** - * The default number of literal context bits is 3. - */ - public static final int LC_DEFAULT = 3; - - /** - * The default number of literal position bits is 0. - */ - public static final int LP_DEFAULT = 0; - - /** - * Maximum value for pb is 4. - */ - public static final int PB_MAX = 4; - - /** - * The default number of position bits is 2. - */ - public static final int PB_DEFAULT = 2; - - /** - * Compression mode: uncompressed. - * The data is wrapped into a LZMA2 stream without compression. - */ - public static final int MODE_UNCOMPRESSED = 0; - - /** - * Compression mode: fast. - * This is usually combined with a hash chain match finder. - */ - public static final int MODE_FAST = LZMAEncoder.MODE_FAST; - - /** - * Compression mode: normal. - * This is usually combined with a binary tree match finder. - */ - public static final int MODE_NORMAL = LZMAEncoder.MODE_NORMAL; - - /** - * Minimum value for niceLen is 8. - */ - public static final int NICE_LEN_MIN = 8; - - /** - * Maximum value for niceLen is 273. - */ - public static final int NICE_LEN_MAX = 273; - - /** - * Match finder: Hash Chain 2-3-4 - */ - public static final int MF_HC4 = LZEncoder.MF_HC4; - - /** - * Match finder: Binary tree 2-3-4 - */ - public static final int MF_BT4 = LZEncoder.MF_BT4; - - private static final int[] presetToDictSize = { - 1 << 18, 1 << 20, 1 << 21, 1 << 22, 1 << 22, - 1 << 23, 1 << 23, 1 << 24, 1 << 25, 1 << 26 }; - - private static final int[] presetToDepthLimit = { 4, 8, 24, 48 }; - - private int dictSize; - private byte[] presetDict = null; - private int lc; - private int lp; - private int pb; - private int mode; - private int niceLen; - private int mf; - private int depthLimit; - - /** - * Creates new LZMA2 options and sets them to the default values. - * This is equivalent to LZMA2Options(PRESET_DEFAULT). - */ - public LZMA2Options() { - try { - setPreset(PRESET_DEFAULT); - } catch (UnsupportedOptionsException e) { - assert false; - throw new RuntimeException(); - } - } - - /** - * Creates new LZMA2 options and sets them to the given preset. - * - * @throws UnsupportedOptionsException - * preset is not supported - */ - public LZMA2Options(int preset) throws UnsupportedOptionsException { - setPreset(preset); - } - - /** - * Creates new LZMA2 options and sets them to the given custom values. - * - * @throws UnsupportedOptionsException - * unsupported options were specified - */ - public LZMA2Options(int dictSize, int lc, int lp, int pb, int mode, - int niceLen, int mf, int depthLimit) - throws UnsupportedOptionsException { - setDictSize(dictSize); - setLcLp(lc, lp); - setPb(pb); - setMode(mode); - setNiceLen(niceLen); - setMatchFinder(mf); - setDepthLimit(depthLimit); - } - - /** - * Sets the compression options to the given preset. - *

- * The presets 0-3 are fast presets with medium compression. - * The presets 4-6 are fairly slow presets with high compression. - * The default preset (PRESET_DEFAULT) is 6. - *

- * The presets 7-9 are like the preset 6 but use bigger dictionaries - * and have higher compressor and decompressor memory requirements. - * Unless the uncompressed size of the file exceeds 8 MiB, - * 16 MiB, or 32 MiB, it is waste of memory to use the - * presets 7, 8, or 9, respectively. - * - * @throws UnsupportedOptionsException - * preset is not supported - */ - public void setPreset(int preset) throws UnsupportedOptionsException { - if (preset < 0 || preset > 9) - throw new UnsupportedOptionsException( - "Unsupported preset: " + preset); - - lc = LC_DEFAULT; - lp = LP_DEFAULT; - pb = PB_DEFAULT; - dictSize = presetToDictSize[preset]; - - if (preset <= 3) { - mode = MODE_FAST; - mf = MF_HC4; - niceLen = preset <= 1 ? 128 : NICE_LEN_MAX; - depthLimit = presetToDepthLimit[preset]; - } else { - mode = MODE_NORMAL; - mf = MF_BT4; - niceLen = (preset == 4) ? 16 : (preset == 5) ? 32 : 64; - depthLimit = 0; - } - } - - /** - * Sets the dictionary size in bytes. - *

- * The dictionary (or history buffer) holds the most recently seen - * uncompressed data. Bigger dictionary usually means better compression. - * However, using a dictioanary bigger than the size of the uncompressed - * data is waste of memory. - *

- * Any value in the range [DICT_SIZE_MIN, DICT_SIZE_MAX] is valid, - * but sizes of 2^n and 2^n + 2^(n-1) bytes are somewhat - * recommended. - * - * @throws UnsupportedOptionsException - * dictSize is not supported - */ - public void setDictSize(int dictSize) throws UnsupportedOptionsException { - if (dictSize < DICT_SIZE_MIN) - throw new UnsupportedOptionsException( - "LZMA2 dictionary size must be at least 4 KiB: " - + dictSize + " B"); - - if (dictSize > DICT_SIZE_MAX) - throw new UnsupportedOptionsException( - "LZMA2 dictionary size must not exceed " - + (DICT_SIZE_MAX >> 20) + " MiB: " + dictSize + " B"); - - this.dictSize = dictSize; - } - - /** - * Gets the dictionary size in bytes. - */ - public int getDictSize() { - return dictSize; - } - - /** - * Sets a preset dictionary. Use null to disable the use of - * a preset dictionary. By default there is no preset dictionary. - *

- * The .xz format doesn't support a preset dictionary for now. - * Do not set a preset dictionary unless you use raw LZMA2. - *

- * Preset dictionary can be useful when compressing many similar, - * relatively small chunks of data independently from each other. - * A preset dictionary should contain typical strings that occur in - * the files being compressed. The most probable strings should be - * near the end of the preset dictionary. The preset dictionary used - * for compression is also needed for decompression. - */ - public void setPresetDict(byte[] presetDict) { - this.presetDict = presetDict; - } - - /** - * Gets the preset dictionary. - */ - public byte[] getPresetDict() { - return presetDict; - } - - /** - * Sets the number of literal context bits and literal position bits. - *

- * The sum of lc and lp is limited to 4. - * Trying to exceed it will throw an exception. This function lets - * you change both at the same time. - * - * @throws UnsupportedOptionsException - * lc and lp - * are invalid - */ - public void setLcLp(int lc, int lp) throws UnsupportedOptionsException { - if (lc < 0 || lp < 0 || lc > LC_LP_MAX || lp > LC_LP_MAX - || lc + lp > LC_LP_MAX) - throw new UnsupportedOptionsException( - "lc + lp must not exceed " + LC_LP_MAX + ": " - + lc + " + " + lp); - - this.lc = lc; - this.lp = lp; - } - - /** - * Sets the number of literal context bits. - *

- * All bytes that cannot be encoded as matches are encoded as literals. - * That is, literals are simply 8-bit bytes that are encoded one at - * a time. - *

- * The literal coding makes an assumption that the highest lc - * bits of the previous uncompressed byte correlate with the next byte. - * For example, in typical English text, an upper-case letter is often - * followed by a lower-case letter, and a lower-case letter is usually - * followed by another lower-case letter. In the US-ASCII character set, - * the highest three bits are 010 for upper-case letters and 011 for - * lower-case letters. When lc is at least 3, the literal - * coding can take advantage of this property in the uncompressed data. - *

- * The default value (3) is usually good. If you want maximum compression, - * try setLc(4). Sometimes it helps a little, and sometimes it - * makes compression worse. If it makes it worse, test for example - * setLc(2) too. - * - * @throws UnsupportedOptionsException - * lc is invalid, or the sum - * of lc and lp - * exceed LC_LP_MAX - */ - public void setLc(int lc) throws UnsupportedOptionsException { - setLcLp(lc, lp); - } - - /** - * Sets the number of literal position bits. - *

- * This affets what kind of alignment in the uncompressed data is - * assumed when encoding literals. See {@link #setPb(int) setPb} for - * more information about alignment. - * - * @throws UnsupportedOptionsException - * lp is invalid, or the sum - * of lc and lp - * exceed LC_LP_MAX - */ - public void setLp(int lp) throws UnsupportedOptionsException { - setLcLp(lc, lp); - } - - /** - * Gets the number of literal context bits. - */ - public int getLc() { - return lc; - } - - /** - * Gets the number of literal position bits. - */ - public int getLp() { - return lp; - } - - /** - * Sets the number of position bits. - *

- * This affects what kind of alignment in the uncompressed data is - * assumed in general. The default (2) means four-byte alignment - * (2^pb = 2^2 = 4), which is often a good choice when - * there's no better guess. - *

- * When the alignment is known, setting the number of position bits - * accordingly may reduce the file size a little. For example with text - * files having one-byte alignment (US-ASCII, ISO-8859-*, UTF-8), using - * setPb(0) can improve compression slightly. For UTF-16 - * text, setPb(1) is a good choice. If the alignment is - * an odd number like 3 bytes, setPb(0) might be the best - * choice. - *

- * Even though the assumed alignment can be adjusted with - * setPb and setLp, LZMA2 still slightly favors - * 16-byte alignment. It might be worth taking into account when designing - * file formats that are likely to be often compressed with LZMA2. - * - * @throws UnsupportedOptionsException - * pb is invalid - */ - public void setPb(int pb) throws UnsupportedOptionsException { - if (pb < 0 || pb > PB_MAX) - throw new UnsupportedOptionsException( - "pb must not exceed " + PB_MAX + ": " + pb); - - this.pb = pb; - } - - /** - * Gets the number of position bits. - */ - public int getPb() { - return pb; - } - - /** - * Sets the compression mode. - *

- * This specifies the method to analyze the data produced by - * a match finder. The default is MODE_FAST for presets - * 0-3 and MODE_NORMAL for presets 4-9. - *

- * Usually MODE_FAST is used with Hash Chain match finders - * and MODE_NORMAL with Binary Tree match finders. This is - * also what the presets do. - *

- * The special mode MODE_UNCOMPRESSED doesn't try to - * compress the data at all (and doesn't use a match finder) and will - * simply wrap it in uncompressed LZMA2 chunks. - * - * @throws UnsupportedOptionsException - * mode is not supported - */ - public void setMode(int mode) throws UnsupportedOptionsException { - if (mode < MODE_UNCOMPRESSED || mode > MODE_NORMAL) - throw new UnsupportedOptionsException( - "Unsupported compression mode: " + mode); - - this.mode = mode; - } - - /** - * Gets the compression mode. - */ - public int getMode() { - return mode; - } - - /** - * Sets the nice length of matches. - * Once a match of at least niceLen bytes is found, - * the algorithm stops looking for better matches. Higher values tend - * to give better compression at the expense of speed. The default - * depends on the preset. - * - * @throws UnsupportedOptionsException - * niceLen is invalid - */ - public void setNiceLen(int niceLen) throws UnsupportedOptionsException { - if (niceLen < NICE_LEN_MIN) - throw new UnsupportedOptionsException( - "Minimum nice length of matches is " - + NICE_LEN_MIN + " bytes: " + niceLen); - - if (niceLen > NICE_LEN_MAX) - throw new UnsupportedOptionsException( - "Maximum nice length of matches is " + NICE_LEN_MAX - + ": " + niceLen); - - this.niceLen = niceLen; - } - - /** - * Gets the nice length of matches. - */ - public int getNiceLen() { - return niceLen; - } - - /** - * Sets the match finder type. - *

- * Match finder has a major effect on compression speed, memory usage, - * and compression ratio. Usually Hash Chain match finders are faster - * than Binary Tree match finders. The default depends on the preset: - * 0-3 use MF_HC4 and 4-9 use MF_BT4. - * - * @throws UnsupportedOptionsException - * mf is not supported - */ - public void setMatchFinder(int mf) throws UnsupportedOptionsException { - if (mf != MF_HC4 && mf != MF_BT4) - throw new UnsupportedOptionsException( - "Unsupported match finder: " + mf); - - this.mf = mf; - } - - /** - * Gets the match finder type. - */ - public int getMatchFinder() { - return mf; - } - - /** - * Sets the match finder search depth limit. - *

- * The default is a special value of 0 which indicates that - * the depth limit should be automatically calculated by the selected - * match finder from the nice length of matches. - *

- * Reasonable depth limit for Hash Chain match finders is 4-100 and - * 16-1000 for Binary Tree match finders. Using very high values can - * make the compressor extremely slow with some files. Avoid settings - * higher than 1000 unless you are prepared to interrupt the compression - * in case it is taking far too long. - * - * @throws UnsupportedOptionsException - * depthLimit is invalid - */ - public void setDepthLimit(int depthLimit) - throws UnsupportedOptionsException { - if (depthLimit < 0) - throw new UnsupportedOptionsException( - "Depth limit cannot be negative: " + depthLimit); - - this.depthLimit = depthLimit; - } - - /** - * Gets the match finder search depth limit. - */ - public int getDepthLimit() { - return depthLimit; - } - - public int getEncoderMemoryUsage() { - return (mode == MODE_UNCOMPRESSED) - ? UncompressedLZMA2OutputStream.getMemoryUsage() - : LZMA2OutputStream.getMemoryUsage(this); - } - - public FinishableOutputStream getOutputStream(FinishableOutputStream out, - ArrayCache arrayCache) { - if (mode == MODE_UNCOMPRESSED) - return new UncompressedLZMA2OutputStream(out, arrayCache); - - return new LZMA2OutputStream(out, this, arrayCache); - } - - /** - * Gets how much memory the LZMA2 decoder will need to decompress the data - * that was encoded with these options and stored in a .xz file. - *

- * The returned value may bigger than the value returned by a direct call - * to {@link LZMA2InputStream#getMemoryUsage(int)} if the dictionary size - * is not 2^n or 2^n + 2^(n-1) bytes. This is because the .xz - * headers store the dictionary size in such a format and other values - * are rounded up to the next such value. Such rounding is harmess except - * it might waste some memory if an unsual dictionary size is used. - *

- * If you use raw LZMA2 streams and unusual dictioanary size, call - * {@link LZMA2InputStream#getMemoryUsage} directly to get raw decoder - * memory requirements. - */ - public int getDecoderMemoryUsage() { - // Round the dictionary size up to the next 2^n or 2^n + 2^(n-1). - int d = dictSize - 1; - d |= d >>> 2; - d |= d >>> 3; - d |= d >>> 4; - d |= d >>> 8; - d |= d >>> 16; - return LZMA2InputStream.getMemoryUsage(d + 1); - } - - public InputStream getInputStream(InputStream in, ArrayCache arrayCache) - throws IOException { - return new LZMA2InputStream(in, dictSize, presetDict, arrayCache); - } - - FilterEncoder getFilterEncoder() { - return new LZMA2Encoder(this); - } - - public Object clone() { - try { - return super.clone(); - } catch (CloneNotSupportedException e) { - assert false; - throw new RuntimeException(); - } - } -} Index: trunk/src/org/tukaani/xz/LZMA2OutputStream.java =================================================================== --- trunk/src/org/tukaani/xz/LZMA2OutputStream.java (revision 16021) +++ (revision ) @@ -1,270 +1,0 @@ -/* - * LZMA2OutputStream - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import java.io.DataOutputStream; -import java.io.IOException; -import org.tukaani.xz.lz.LZEncoder; -import org.tukaani.xz.rangecoder.RangeEncoderToBuffer; -import org.tukaani.xz.lzma.LZMAEncoder; - -class LZMA2OutputStream extends FinishableOutputStream { - static final int COMPRESSED_SIZE_MAX = 64 << 10; - - private final ArrayCache arrayCache; - - private FinishableOutputStream out; - private final DataOutputStream outData; - - private LZEncoder lz; - private RangeEncoderToBuffer rc; - private LZMAEncoder lzma; - - private final int props; // Cannot change props on the fly for now. - private boolean dictResetNeeded = true; - private boolean stateResetNeeded = true; - private boolean propsNeeded = true; - - private int pendingSize = 0; - private boolean finished = false; - private IOException exception = null; - - private final byte[] tempBuf = new byte[1]; - - private static int getExtraSizeBefore(int dictSize) { - return COMPRESSED_SIZE_MAX > dictSize - ? COMPRESSED_SIZE_MAX - dictSize : 0; - } - - static int getMemoryUsage(LZMA2Options options) { - // 64 KiB buffer for the range encoder + a little extra + LZMAEncoder - int dictSize = options.getDictSize(); - int extraSizeBefore = getExtraSizeBefore(dictSize); - return 70 + LZMAEncoder.getMemoryUsage(options.getMode(), - dictSize, extraSizeBefore, - options.getMatchFinder()); - } - - LZMA2OutputStream(FinishableOutputStream out, LZMA2Options options, - ArrayCache arrayCache) { - if (out == null) - throw new NullPointerException(); - - this.arrayCache = arrayCache; - this.out = out; - outData = new DataOutputStream(out); - rc = new RangeEncoderToBuffer(COMPRESSED_SIZE_MAX, arrayCache); - - int dictSize = options.getDictSize(); - int extraSizeBefore = getExtraSizeBefore(dictSize); - lzma = LZMAEncoder.getInstance(rc, - options.getLc(), options.getLp(), options.getPb(), - options.getMode(), - dictSize, extraSizeBefore, options.getNiceLen(), - options.getMatchFinder(), options.getDepthLimit(), - this.arrayCache); - - lz = lzma.getLZEncoder(); - - byte[] presetDict = options.getPresetDict(); - if (presetDict != null && presetDict.length > 0) { - lz.setPresetDict(dictSize, presetDict); - dictResetNeeded = false; - } - - props = (options.getPb() * 5 + options.getLp()) * 9 + options.getLc(); - } - - public void write(int b) throws IOException { - tempBuf[0] = (byte)b; - write(tempBuf, 0, 1); - } - - public void write(byte[] buf, int off, int len) throws IOException { - if (off < 0 || len < 0 || off + len < 0 || off + len > buf.length) - throw new IndexOutOfBoundsException(); - - if (exception != null) - throw exception; - - if (finished) - throw new XZIOException("Stream finished or closed"); - - try { - while (len > 0) { - int used = lz.fillWindow(buf, off, len); - off += used; - len -= used; - pendingSize += used; - - if (lzma.encodeForLZMA2()) - writeChunk(); - } - } catch (IOException e) { - exception = e; - throw e; - } - } - - private void writeChunk() throws IOException { - int compressedSize = rc.finish(); - int uncompressedSize = lzma.getUncompressedSize(); - - assert compressedSize > 0 : compressedSize; - assert uncompressedSize > 0 : uncompressedSize; - - // +2 because the header of a compressed chunk is 2 bytes - // bigger than the header of an uncompressed chunk. - if (compressedSize + 2 < uncompressedSize) { - writeLZMA(uncompressedSize, compressedSize); - } else { - lzma.reset(); - uncompressedSize = lzma.getUncompressedSize(); - assert uncompressedSize > 0 : uncompressedSize; - writeUncompressed(uncompressedSize); - } - - pendingSize -= uncompressedSize; - lzma.resetUncompressedSize(); - rc.reset(); - } - - private void writeLZMA(int uncompressedSize, int compressedSize) - throws IOException { - int control; - - if (propsNeeded) { - if (dictResetNeeded) - control = 0x80 + (3 << 5); - else - control = 0x80 + (2 << 5); - } else { - if (stateResetNeeded) - control = 0x80 + (1 << 5); - else - control = 0x80; - } - - control |= (uncompressedSize - 1) >>> 16; - outData.writeByte(control); - - outData.writeShort(uncompressedSize - 1); - outData.writeShort(compressedSize - 1); - - if (propsNeeded) - outData.writeByte(props); - - rc.write(out); - - propsNeeded = false; - stateResetNeeded = false; - dictResetNeeded = false; - } - - private void writeUncompressed(int uncompressedSize) throws IOException { - while (uncompressedSize > 0) { - int chunkSize = Math.min(uncompressedSize, COMPRESSED_SIZE_MAX); - outData.writeByte(dictResetNeeded ? 0x01 : 0x02); - outData.writeShort(chunkSize - 1); - lz.copyUncompressed(out, uncompressedSize, chunkSize); - uncompressedSize -= chunkSize; - dictResetNeeded = false; - } - - stateResetNeeded = true; - } - - private void writeEndMarker() throws IOException { - assert !finished; - - if (exception != null) - throw exception; - - lz.setFinishing(); - - try { - while (pendingSize > 0) { - lzma.encodeForLZMA2(); - writeChunk(); - } - - out.write(0x00); - } catch (IOException e) { - exception = e; - throw e; - } - - finished = true; - - lzma.putArraysToCache(arrayCache); - lzma = null; - lz = null; - rc.putArraysToCache(arrayCache); - rc = null; - } - - public void flush() throws IOException { - if (exception != null) - throw exception; - - if (finished) - throw new XZIOException("Stream finished or closed"); - - try { - lz.setFlushing(); - - while (pendingSize > 0) { - lzma.encodeForLZMA2(); - writeChunk(); - } - - out.flush(); - } catch (IOException e) { - exception = e; - throw e; - } - } - - public void finish() throws IOException { - if (!finished) { - writeEndMarker(); - - try { - out.finish(); - } catch (IOException e) { - exception = e; - throw e; - } - } - } - - public void close() throws IOException { - if (out != null) { - if (!finished) { - try { - writeEndMarker(); - } catch (IOException e) {} - } - - try { - out.close(); - } catch (IOException e) { - if (exception == null) - exception = e; - } - - out = null; - } - - if (exception != null) - throw exception; - } -} Index: trunk/src/org/tukaani/xz/MemoryLimitException.java =================================================================== --- trunk/src/org/tukaani/xz/MemoryLimitException.java (revision 16021) +++ (revision ) @@ -1,60 +1,0 @@ -/* - * MemoryLimitException - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -/** - * Thrown when the memory usage limit given to the XZ decompressor - * would be exceeded. - *

- * The amount of memory required and the memory usage limit are - * included in the error detail message in human readable format. - */ -public class MemoryLimitException extends XZIOException { - private static final long serialVersionUID = 3L; - - private final int memoryNeeded; - private final int memoryLimit; - - /** - * Creates a new MemoryLimitException. - *

- * The amount of memory needed and the memory usage limit are - * included in the error detail message. - * - * @param memoryNeeded amount of memory needed as kibibytes (KiB) - * @param memoryLimit specified memory usage limit as kibibytes (KiB) - */ - public MemoryLimitException(int memoryNeeded, int memoryLimit) { - super("" + memoryNeeded + " KiB of memory would be needed; limit was " - + memoryLimit + " KiB"); - - this.memoryNeeded = memoryNeeded; - this.memoryLimit = memoryLimit; - } - - /** - * Gets how much memory is required to decompress the data. - * - * @return amount of memory needed as kibibytes (KiB) - */ - public int getMemoryNeeded() { - return memoryNeeded; - } - - /** - * Gets what the memory usage limit was at the time the exception - * was created. - * - * @return memory usage limit as kibibytes (KiB) - */ - public int getMemoryLimit() { - return memoryLimit; - } -} Index: trunk/src/org/tukaani/xz/RawCoder.java =================================================================== --- trunk/src/org/tukaani/xz/RawCoder.java (revision 16021) +++ (revision ) @@ -1,33 +1,0 @@ -/* - * RawCoder - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -class RawCoder { - static void validate(FilterCoder[] filters) - throws UnsupportedOptionsException { - for (int i = 0; i < filters.length - 1; ++i) - if (!filters[i].nonLastOK()) - throw new UnsupportedOptionsException( - "Unsupported XZ filter chain"); - - if (!filters[filters.length - 1].lastOK()) - throw new UnsupportedOptionsException( - "Unsupported XZ filter chain"); - - int changesSizeCount = 0; - for (int i = 0; i < filters.length; ++i) - if (filters[i].changesSize()) - ++changesSizeCount; - - if (changesSizeCount > 3) - throw new UnsupportedOptionsException( - "Unsupported XZ filter chain"); - } -} Index: trunk/src/org/tukaani/xz/SeekableInputStream.java =================================================================== --- trunk/src/org/tukaani/xz/SeekableInputStream.java (revision 16021) +++ (revision ) @@ -1,81 +1,0 @@ -/* - * SeekableInputStream - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import java.io.InputStream; -import java.io.IOException; - -/** - * Input stream with random access support. - */ -public abstract class SeekableInputStream extends InputStream { - /** - * Seeks n bytes forward in this stream. - *

- * This will not seek past the end of the file. If the current position - * is already at or past the end of the file, this doesn't seek at all - * and returns 0. Otherwise, if skipping n bytes - * would cause the position to exceed the stream size, this will do - * equivalent of seek(length()) and the return value will - * be adjusted accordingly. - *

- * If n is negative, the position isn't changed and - * the return value is 0. It doesn't seek backward - * because it would conflict with the specification of - * {@link java.io.InputStream#skip(long) InputStream.skip}. - * - * @return 0 if n is negative, - * less than n if skipping n - * bytes would seek past the end of the file, - * n otherwise - * - * @throws IOException might be thrown by {@link #seek(long)} - */ - public long skip(long n) throws IOException { - if (n <= 0) - return 0; - - long size = length(); - long pos = position(); - if (pos >= size) - return 0; - - if (size - pos < n) - n = size - pos; - - seek(pos + n); - return n; - } - - /** - * Gets the size of the stream. - */ - public abstract long length() throws IOException; - - /** - * Gets the current position in the stream. - */ - public abstract long position() throws IOException; - - /** - * Seeks to the specified absolute position in the stream. - *

- * Seeking past the end of the file should be supported by the subclasses - * unless there is a good reason to do otherwise. If one has seeked - * past the end of the stream, read will return - * -1 to indicate end of stream. - * - * @param pos new read position in the stream - * - * @throws IOException if pos is negative or if - * a stream-specific I/O error occurs - */ - public abstract void seek(long pos) throws IOException; -} Index: trunk/src/org/tukaani/xz/SimpleInputStream.java =================================================================== --- trunk/src/org/tukaani/xz/SimpleInputStream.java (revision 16021) +++ (revision ) @@ -1,138 +1,0 @@ -/* - * SimpleInputStream - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import java.io.InputStream; -import java.io.IOException; -import org.tukaani.xz.simple.SimpleFilter; - -class SimpleInputStream extends InputStream { - private static final int FILTER_BUF_SIZE = 4096; - - private InputStream in; - private final SimpleFilter simpleFilter; - - private final byte[] filterBuf = new byte[FILTER_BUF_SIZE]; - private int pos = 0; - private int filtered = 0; - private int unfiltered = 0; - - private boolean endReached = false; - private IOException exception = null; - - private final byte[] tempBuf = new byte[1]; - - static int getMemoryUsage() { - return 1 + FILTER_BUF_SIZE / 1024; - } - - SimpleInputStream(InputStream in, SimpleFilter simpleFilter) { - // Check for null because otherwise null isn't detect - // in this constructor. - if (in == null) - throw new NullPointerException(); - - // The simpleFilter argument comes from this package - // so it is known to be non-null already. - assert simpleFilter != null; - - this.in = in; - this.simpleFilter = simpleFilter; - } - - public int read() throws IOException { - return read(tempBuf, 0, 1) == -1 ? -1 : (tempBuf[0] & 0xFF); - } - - public int read(byte[] buf, int off, int len) throws IOException { - if (off < 0 || len < 0 || off + len < 0 || off + len > buf.length) - throw new IndexOutOfBoundsException(); - - if (len == 0) - return 0; - - if (in == null) - throw new XZIOException("Stream closed"); - - if (exception != null) - throw exception; - - try { - int size = 0; - - while (true) { - // Copy filtered data into the caller-provided buffer. - int copySize = Math.min(filtered, len); - System.arraycopy(filterBuf, pos, buf, off, copySize); - pos += copySize; - filtered -= copySize; - off += copySize; - len -= copySize; - size += copySize; - - // If end of filterBuf was reached, move the pending data to - // the beginning of the buffer so that more data can be - // copied into filterBuf on the next loop iteration. - if (pos + filtered + unfiltered == FILTER_BUF_SIZE) { - System.arraycopy(filterBuf, pos, filterBuf, 0, - filtered + unfiltered); - pos = 0; - } - - if (len == 0 || endReached) - return size > 0 ? size : -1; - - assert filtered == 0; - - // Get more data into the temporary buffer. - int inSize = FILTER_BUF_SIZE - (pos + filtered + unfiltered); - inSize = in.read(filterBuf, pos + filtered + unfiltered, - inSize); - - if (inSize == -1) { - // Mark the remaining unfiltered bytes to be ready - // to be copied out. - endReached = true; - filtered = unfiltered; - unfiltered = 0; - } else { - // Filter the data in filterBuf. - unfiltered += inSize; - filtered = simpleFilter.code(filterBuf, pos, unfiltered); - assert filtered <= unfiltered; - unfiltered -= filtered; - } - } - } catch (IOException e) { - exception = e; - throw e; - } - } - - public int available() throws IOException { - if (in == null) - throw new XZIOException("Stream closed"); - - if (exception != null) - throw exception; - - return filtered; - } - - public void close() throws IOException { - if (in != null) { - try { - in.close(); - } finally { - in = null; - } - } - } -} Index: trunk/src/org/tukaani/xz/SingleXZInputStream.java =================================================================== --- trunk/src/org/tukaani/xz/SingleXZInputStream.java (revision 16021) +++ (revision ) @@ -1,535 +1,0 @@ -/* - * SingleXZInputStream - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import java.io.InputStream; -import java.io.DataInputStream; -import java.io.IOException; -import java.io.EOFException; -import org.tukaani.xz.common.DecoderUtil; -import org.tukaani.xz.common.StreamFlags; -import org.tukaani.xz.index.IndexHash; -import org.tukaani.xz.check.Check; - -/** - * Decompresses exactly one XZ Stream in streamed mode (no seeking). - * The decompression stops after the first XZ Stream has been decompressed, - * and the read position in the input stream is left at the first byte - * after the end of the XZ Stream. This can be useful when XZ data has - * been stored inside some other file format or protocol. - *

- * Unless you know what you are doing, don't use this class to decompress - * standalone .xz files. For that purpose, use XZInputStream. - * - *

When uncompressed size is known beforehand

- *

- * If you are decompressing complete XZ streams and your application knows - * exactly how much uncompressed data there should be, it is good to try - * reading one more byte by calling read() and checking - * that it returns -1. This way the decompressor will parse the - * file footers and verify the integrity checks, giving the caller more - * confidence that the uncompressed data is valid. - * - * @see XZInputStream - */ -public class SingleXZInputStream extends InputStream { - private InputStream in; - private final ArrayCache arrayCache; - private final int memoryLimit; - private final StreamFlags streamHeaderFlags; - private final Check check; - private final boolean verifyCheck; - private BlockInputStream blockDecoder = null; - private final IndexHash indexHash = new IndexHash(); - private boolean endReached = false; - private IOException exception = null; - - private final byte[] tempBuf = new byte[1]; - - /** - * Reads the Stream Header into a buffer. - * This is a helper function for the constructors. - */ - private static byte[] readStreamHeader(InputStream in) throws IOException { - byte[] streamHeader = new byte[DecoderUtil.STREAM_HEADER_SIZE]; - new DataInputStream(in).readFully(streamHeader); - return streamHeader; - } - - /** - * Creates a new XZ decompressor that decompresses exactly one - * XZ Stream from in without a memory usage limit. - *

- * This constructor reads and parses the XZ Stream Header (12 bytes) - * from in. The header of the first Block is not read - * until read is called. - * - * @param in input stream from which XZ-compressed - * data is read - * - * @throws XZFormatException - * input is not in the XZ format - * - * @throws CorruptedInputException - * XZ header CRC32 doesn't match - * - * @throws UnsupportedOptionsException - * XZ header is valid but specifies options - * not supported by this implementation - * - * @throws EOFException - * less than 12 bytes of input was available - * from in - * - * @throws IOException may be thrown by in - */ - public SingleXZInputStream(InputStream in) throws IOException { - this(in, -1); - } - - /** - * Creates a new XZ decompressor that decompresses exactly one - * XZ Stream from in without a memory usage limit. - *

- * This is identical to SingleXZInputStream(InputStream) - * except that this also takes the arrayCache argument. - * - * @param in input stream from which XZ-compressed - * data is read - * - * @param arrayCache cache to be used for allocating large arrays - * - * @throws XZFormatException - * input is not in the XZ format - * - * @throws CorruptedInputException - * XZ header CRC32 doesn't match - * - * @throws UnsupportedOptionsException - * XZ header is valid but specifies options - * not supported by this implementation - * - * @throws EOFException - * less than 12 bytes of input was available - * from in - * - * @throws IOException may be thrown by in - * - * @since 1.7 - */ - public SingleXZInputStream(InputStream in, ArrayCache arrayCache) - throws IOException { - this(in, -1, arrayCache); - } - - /** - * Creates a new XZ decompressor that decompresses exactly one - * XZ Stream from in with an optional memory usage limit. - *

- * This is identical to SingleXZInputStream(InputStream) - * except that this also takes the memoryLimit argument. - * - * @param in input stream from which XZ-compressed - * data is read - * - * @param memoryLimit memory usage limit in kibibytes (KiB) - * or -1 to impose no - * memory usage limit - * - * @throws XZFormatException - * input is not in the XZ format - * - * @throws CorruptedInputException - * XZ header CRC32 doesn't match - * - * @throws UnsupportedOptionsException - * XZ header is valid but specifies options - * not supported by this implementation - * - * @throws EOFException - * less than 12 bytes of input was available - * from in - * - * @throws IOException may be thrown by in - */ - public SingleXZInputStream(InputStream in, int memoryLimit) - throws IOException { - this(in, memoryLimit, true); - } - - /** - * Creates a new XZ decompressor that decompresses exactly one - * XZ Stream from in with an optional memory usage limit. - *

- * This is identical to SingleXZInputStream(InputStream) - * except that this also takes the memoryLimit and - * arrayCache arguments. - * - * @param in input stream from which XZ-compressed - * data is read - * - * @param memoryLimit memory usage limit in kibibytes (KiB) - * or -1 to impose no - * memory usage limit - * - * @param arrayCache cache to be used for allocating large arrays - * - * @throws XZFormatException - * input is not in the XZ format - * - * @throws CorruptedInputException - * XZ header CRC32 doesn't match - * - * @throws UnsupportedOptionsException - * XZ header is valid but specifies options - * not supported by this implementation - * - * @throws EOFException - * less than 12 bytes of input was available - * from in - * - * @throws IOException may be thrown by in - * - * @since 1.7 - */ - public SingleXZInputStream(InputStream in, int memoryLimit, - ArrayCache arrayCache) throws IOException { - this(in, memoryLimit, true, arrayCache); - } - - /** - * Creates a new XZ decompressor that decompresses exactly one - * XZ Stream from in with an optional memory usage limit - * and ability to disable verification of integrity checks. - *

- * This is identical to SingleXZInputStream(InputStream,int) - * except that this also takes the verifyCheck argument. - *

- * Note that integrity check verification should almost never be disabled. - * Possible reasons to disable integrity check verification: - *

- *

- * verifyCheck only affects the integrity check of - * the actual compressed data. The CRC32 fields in the headers - * are always verified. - * - * @param in input stream from which XZ-compressed - * data is read - * - * @param memoryLimit memory usage limit in kibibytes (KiB) - * or -1 to impose no - * memory usage limit - * - * @param verifyCheck if true, the integrity checks - * will be verified; this should almost never - * be set to false - * - * @throws XZFormatException - * input is not in the XZ format - * - * @throws CorruptedInputException - * XZ header CRC32 doesn't match - * - * @throws UnsupportedOptionsException - * XZ header is valid but specifies options - * not supported by this implementation - * - * @throws EOFException - * less than 12 bytes of input was available - * from in - * - * @throws IOException may be thrown by in - * - * @since 1.6 - */ - public SingleXZInputStream(InputStream in, int memoryLimit, - boolean verifyCheck) throws IOException { - this(in, memoryLimit, verifyCheck, ArrayCache.getDefaultCache()); - } - - /** - * Creates a new XZ decompressor that decompresses exactly one - * XZ Stream from in with an optional memory usage limit - * and ability to disable verification of integrity checks. - *

- * This is identical to - * SingleXZInputStream(InputStream,int,boolean) - * except that this also takes the arrayCache argument. - * - * @param in input stream from which XZ-compressed - * data is read - * - * @param memoryLimit memory usage limit in kibibytes (KiB) - * or -1 to impose no - * memory usage limit - * - * @param verifyCheck if true, the integrity checks - * will be verified; this should almost never - * be set to false - * - * @param arrayCache cache to be used for allocating large arrays - * - * @throws XZFormatException - * input is not in the XZ format - * - * @throws CorruptedInputException - * XZ header CRC32 doesn't match - * - * @throws UnsupportedOptionsException - * XZ header is valid but specifies options - * not supported by this implementation - * - * @throws EOFException - * less than 12 bytes of input was available - * from in - * - * @throws IOException may be thrown by in - * - * @since 1.7 - */ - public SingleXZInputStream(InputStream in, int memoryLimit, - boolean verifyCheck, ArrayCache arrayCache) - throws IOException { - this(in, memoryLimit, verifyCheck, readStreamHeader(in), arrayCache); - } - - SingleXZInputStream(InputStream in, int memoryLimit, boolean verifyCheck, - byte[] streamHeader, ArrayCache arrayCache) - throws IOException { - this.arrayCache = arrayCache; - this.in = in; - this.memoryLimit = memoryLimit; - this.verifyCheck = verifyCheck; - streamHeaderFlags = DecoderUtil.decodeStreamHeader(streamHeader); - check = Check.getInstance(streamHeaderFlags.checkType); - } - - /** - * Gets the ID of the integrity check used in this XZ Stream. - * - * @return the Check ID specified in the XZ Stream Header - */ - public int getCheckType() { - return streamHeaderFlags.checkType; - } - - /** - * Gets the name of the integrity check used in this XZ Stream. - * - * @return the name of the check specified in the XZ Stream Header - */ - public String getCheckName() { - return check.getName(); - } - - /** - * Decompresses the next byte from this input stream. - *

- * Reading lots of data with read() from this input stream - * may be inefficient. Wrap it in {@link java.io.BufferedInputStream} - * if you need to read lots of data one byte at a time. - * - * @return the next decompressed byte, or -1 - * to indicate the end of the compressed stream - * - * @throws CorruptedInputException - * @throws UnsupportedOptionsException - * @throws MemoryLimitException - * - * @throws XZIOException if the stream has been closed - * - * @throws EOFException - * compressed input is truncated or corrupt - * - * @throws IOException may be thrown by in - */ - public int read() throws IOException { - return read(tempBuf, 0, 1) == -1 ? -1 : (tempBuf[0] & 0xFF); - } - - /** - * Decompresses into an array of bytes. - *

- * If len is zero, no bytes are read and 0 - * is returned. Otherwise this will try to decompress len - * bytes of uncompressed data. Less than len bytes may - * be read only in the following situations: - *

- * - * @param buf target buffer for uncompressed data - * @param off start offset in buf - * @param len maximum number of uncompressed bytes to read - * - * @return number of bytes read, or -1 to indicate - * the end of the compressed stream - * - * @throws CorruptedInputException - * @throws UnsupportedOptionsException - * @throws MemoryLimitException - * - * @throws XZIOException if the stream has been closed - * - * @throws EOFException - * compressed input is truncated or corrupt - * - * @throws IOException may be thrown by in - */ - public int read(byte[] buf, int off, int len) throws IOException { - if (off < 0 || len < 0 || off + len < 0 || off + len > buf.length) - throw new IndexOutOfBoundsException(); - - if (len == 0) - return 0; - - if (in == null) - throw new XZIOException("Stream closed"); - - if (exception != null) - throw exception; - - if (endReached) - return -1; - - int size = 0; - - try { - while (len > 0) { - if (blockDecoder == null) { - try { - blockDecoder = new BlockInputStream( - in, check, verifyCheck, memoryLimit, -1, -1, - arrayCache); - } catch (IndexIndicatorException e) { - indexHash.validate(in); - validateStreamFooter(); - endReached = true; - return size > 0 ? size : -1; - } - } - - int ret = blockDecoder.read(buf, off, len); - - if (ret > 0) { - size += ret; - off += ret; - len -= ret; - } else if (ret == -1) { - indexHash.add(blockDecoder.getUnpaddedSize(), - blockDecoder.getUncompressedSize()); - blockDecoder = null; - } - } - } catch (IOException e) { - exception = e; - if (size == 0) - throw e; - } - - return size; - } - - private void validateStreamFooter() throws IOException { - byte[] buf = new byte[DecoderUtil.STREAM_HEADER_SIZE]; - new DataInputStream(in).readFully(buf); - StreamFlags streamFooterFlags = DecoderUtil.decodeStreamFooter(buf); - - if (!DecoderUtil.areStreamFlagsEqual(streamHeaderFlags, - streamFooterFlags) - || indexHash.getIndexSize() != streamFooterFlags.backwardSize) - throw new CorruptedInputException( - "XZ Stream Footer does not match Stream Header"); - } - - /** - * Returns the number of uncompressed bytes that can be read - * without blocking. The value is returned with an assumption - * that the compressed input data will be valid. If the compressed - * data is corrupt, CorruptedInputException may get - * thrown before the number of bytes claimed to be available have - * been read from this input stream. - * - * @return the number of uncompressed bytes that can be read - * without blocking - */ - public int available() throws IOException { - if (in == null) - throw new XZIOException("Stream closed"); - - if (exception != null) - throw exception; - - return blockDecoder == null ? 0 : blockDecoder.available(); - } - - /** - * Closes the stream and calls in.close(). - * If the stream was already closed, this does nothing. - *

- * This is equivalent to close(true). - * - * @throws IOException if thrown by in.close() - */ - public void close() throws IOException { - close(true); - } - - /** - * Closes the stream and optionally calls in.close(). - * If the stream was already closed, this does nothing. - * If close(false) has been called, a further - * call of close(true) does nothing (it doesn't call - * in.close()). - *

- * If you don't want to close the underlying InputStream, - * there is usually no need to worry about closing this stream either; - * it's fine to do nothing and let the garbage collector handle it. - * However, if you are using {@link ArrayCache}, close(false) - * can be useful to put the allocated arrays back to the cache without - * closing the underlying InputStream. - *

- * Note that if you successfully reach the end of the stream - * (read returns -1), the arrays are - * automatically put back to the cache by that read call. In - * this situation close(false) is redundant (but harmless). - * - * @throws IOException if thrown by in.close() - * - * @since 1.7 - */ - public void close(boolean closeInput) throws IOException { - if (in != null) { - if (blockDecoder != null) { - blockDecoder.close(); - blockDecoder = null; - } - - try { - if (closeInput) - in.close(); - } finally { - in = null; - } - } - } -} Index: trunk/src/org/tukaani/xz/UncompressedLZMA2OutputStream.java =================================================================== --- trunk/src/org/tukaani/xz/UncompressedLZMA2OutputStream.java (revision 16021) +++ (revision ) @@ -1,164 +1,0 @@ -/* - * UncompressedLZMA2OutputStream - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import java.io.DataOutputStream; -import java.io.IOException; - -class UncompressedLZMA2OutputStream extends FinishableOutputStream { - private final ArrayCache arrayCache; - - private FinishableOutputStream out; - private final DataOutputStream outData; - - private final byte[] uncompBuf; - private int uncompPos = 0; - private boolean dictResetNeeded = true; - - private boolean finished = false; - private IOException exception = null; - - private final byte[] tempBuf = new byte[1]; - - static int getMemoryUsage() { - // uncompBuf + a little extra - return 70; - } - - UncompressedLZMA2OutputStream(FinishableOutputStream out, - ArrayCache arrayCache) { - if (out == null) - throw new NullPointerException(); - - this.out = out; - outData = new DataOutputStream(out); - - // We only allocate one array from the cache. We will call - // putArray directly in writeEndMarker and thus we don't use - // ResettableArrayCache here. - this.arrayCache = arrayCache; - uncompBuf = arrayCache.getByteArray( - LZMA2OutputStream.COMPRESSED_SIZE_MAX, false); - } - - public void write(int b) throws IOException { - tempBuf[0] = (byte)b; - write(tempBuf, 0, 1); - } - - public void write(byte[] buf, int off, int len) throws IOException { - if (off < 0 || len < 0 || off + len < 0 || off + len > buf.length) - throw new IndexOutOfBoundsException(); - - if (exception != null) - throw exception; - - if (finished) - throw new XZIOException("Stream finished or closed"); - - try { - while (len > 0) { - int copySize = Math.min(LZMA2OutputStream.COMPRESSED_SIZE_MAX - - uncompPos, len); - System.arraycopy(buf, off, uncompBuf, uncompPos, copySize); - len -= copySize; - uncompPos += copySize; - - if (uncompPos == LZMA2OutputStream.COMPRESSED_SIZE_MAX) - writeChunk(); - } - } catch (IOException e) { - exception = e; - throw e; - } - } - - private void writeChunk() throws IOException { - outData.writeByte(dictResetNeeded ? 0x01 : 0x02); - outData.writeShort(uncompPos - 1); - outData.write(uncompBuf, 0, uncompPos); - uncompPos = 0; - dictResetNeeded = false; - } - - private void writeEndMarker() throws IOException { - if (exception != null) - throw exception; - - if (finished) - throw new XZIOException("Stream finished or closed"); - - try { - if (uncompPos > 0) - writeChunk(); - - out.write(0x00); - } catch (IOException e) { - exception = e; - throw e; - } - - finished = true; - arrayCache.putArray(uncompBuf); - } - - public void flush() throws IOException { - if (exception != null) - throw exception; - - if (finished) - throw new XZIOException("Stream finished or closed"); - - try { - if (uncompPos > 0) - writeChunk(); - - out.flush(); - } catch (IOException e) { - exception = e; - throw e; - } - } - - public void finish() throws IOException { - if (!finished) { - writeEndMarker(); - - try { - out.finish(); - } catch (IOException e) { - exception = e; - throw e; - } - } - } - - public void close() throws IOException { - if (out != null) { - if (!finished) { - try { - writeEndMarker(); - } catch (IOException e) {} - } - - try { - out.close(); - } catch (IOException e) { - if (exception == null) - exception = e; - } - - out = null; - } - - if (exception != null) - throw exception; - } -} Index: trunk/src/org/tukaani/xz/UnsupportedOptionsException.java =================================================================== --- trunk/src/org/tukaani/xz/UnsupportedOptionsException.java (revision 16021) +++ (revision ) @@ -1,34 +1,0 @@ -/* - * UnsupportedOptionsException - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -/** - * Thrown when compression options not supported by this implementation - * are detected. Some other implementation might support those options. - */ -public class UnsupportedOptionsException extends XZIOException { - private static final long serialVersionUID = 3L; - - /** - * Creates a new UnsupportedOptionsException with null - * as its error detail message. - */ - public UnsupportedOptionsException() {} - - /** - * Creates a new UnsupportedOptionsException with the given - * error detail message. - * - * @param s error detail message - */ - public UnsupportedOptionsException(String s) { - super(s); - } -} Index: trunk/src/org/tukaani/xz/XZ.java =================================================================== --- trunk/src/org/tukaani/xz/XZ.java (revision 16021) +++ (revision ) @@ -1,53 +1,0 @@ -/* - * XZ - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -/** - * XZ constants. - */ -public class XZ { - /** - * XZ Header Magic Bytes begin a XZ file. - * This can be useful to detect XZ compressed data. - */ - public static final byte[] HEADER_MAGIC = { - (byte)0xFD, '7', 'z', 'X', 'Z', '\0' }; - - /** - * XZ Footer Magic Bytes are the last bytes of a XZ Stream. - */ - public static final byte[] FOOTER_MAGIC = { 'Y', 'Z' }; - - /** - * Integrity check ID indicating that no integrity check is calculated. - *

- * Omitting the integrity check is strongly discouraged except when - * the integrity of the data will be verified by other means anyway, - * and calculating the check twice would be useless. - */ - public static final int CHECK_NONE = 0; - - /** - * Integrity check ID for CRC32. - */ - public static final int CHECK_CRC32 = 1; - - /** - * Integrity check ID for CRC64. - */ - public static final int CHECK_CRC64 = 4; - - /** - * Integrity check ID for SHA-256. - */ - public static final int CHECK_SHA256 = 10; - - private XZ() {} -} Index: trunk/src/org/tukaani/xz/XZFormatException.java =================================================================== --- trunk/src/org/tukaani/xz/XZFormatException.java (revision 16021) +++ (revision ) @@ -1,24 +1,0 @@ -/* - * XZFormatException - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -/** - * Thrown when the input data is not in the XZ format. - */ -public class XZFormatException extends XZIOException { - private static final long serialVersionUID = 3L; - - /** - * Creates a new exception with the default error detail message. - */ - public XZFormatException() { - super("Input is not in the XZ format"); - } -} Index: trunk/src/org/tukaani/xz/XZIOException.java =================================================================== --- trunk/src/org/tukaani/xz/XZIOException.java (revision 16021) +++ (revision ) @@ -1,27 +1,0 @@ -/* - * XZIOException - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -/** - * Generic {@link java.io.IOException IOException} specific to this package. - * The other IOExceptions in this package extend - * from XZIOException. - */ -public class XZIOException extends java.io.IOException { - private static final long serialVersionUID = 3L; - - public XZIOException() { - super(); - } - - public XZIOException(String s) { - super(s); - } -} Index: trunk/src/org/tukaani/xz/XZInputStream.java =================================================================== --- trunk/src/org/tukaani/xz/XZInputStream.java (revision 16021) +++ (revision ) @@ -1,527 +1,0 @@ -/* - * XZInputStream - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import java.io.InputStream; -import java.io.DataInputStream; -import java.io.IOException; -import java.io.EOFException; -import org.tukaani.xz.common.DecoderUtil; - -/** - * Decompresses a .xz file in streamed mode (no seeking). - *

- * Use this to decompress regular standalone .xz files. This reads from - * its input stream until the end of the input or until an error occurs. - * This supports decompressing concatenated .xz files. - * - *

Typical use cases

- *

- * Getting an input stream to decompress a .xz file: - * 

- * InputStream infile = new FileInputStream("foo.xz");
- * XZInputStream inxz = new XZInputStream(infile);
- *
- *

- * It's important to keep in mind that decompressor memory usage depends - * on the settings used to compress the file. The worst-case memory usage - * of XZInputStream is currently 1.5 GiB. Still, very few files will - * require more than about 65 MiB because that's how much decompressing - * a file created with the highest preset level will need, and only a few - * people use settings other than the predefined presets. - *

- * It is possible to specify a memory usage limit for - * XZInputStream. If decompression requires more memory than - * the specified limit, MemoryLimitException will be thrown when reading - * from the stream. For example, the following sets the memory usage limit - * to 100 MiB: - * 

- * InputStream infile = new FileInputStream("foo.xz");
- * XZInputStream inxz = new XZInputStream(infile, 100 * 1024);
- *
- * - *

When uncompressed size is known beforehand

- *

- * If you are decompressing complete files and your application knows - * exactly how much uncompressed data there should be, it is good to try - * reading one more byte by calling read() and checking - * that it returns -1. This way the decompressor will parse the - * file footers and verify the integrity checks, giving the caller more - * confidence that the uncompressed data is valid. (This advice seems to - * apply to - * {@link java.util.zip.GZIPInputStream java.util.zip.GZIPInputStream} too.) - * - * @see SingleXZInputStream - */ -public class XZInputStream extends InputStream { - private final ArrayCache arrayCache; - - private final int memoryLimit; - private InputStream in; - private SingleXZInputStream xzIn; - private final boolean verifyCheck; - private boolean endReached = false; - private IOException exception = null; - - private final byte[] tempBuf = new byte[1]; - - /** - * Creates a new XZ decompressor without a memory usage limit. - *

- * This constructor reads and parses the XZ Stream Header (12 bytes) - * from in. The header of the first Block is not read - * until read is called. - * - * @param in input stream from which XZ-compressed - * data is read - * - * @throws XZFormatException - * input is not in the XZ format - * - * @throws CorruptedInputException - * XZ header CRC32 doesn't match - * - * @throws UnsupportedOptionsException - * XZ header is valid but specifies options - * not supported by this implementation - * - * @throws EOFException - * less than 12 bytes of input was available - * from in - * - * @throws IOException may be thrown by in - */ - public XZInputStream(InputStream in) throws IOException { - this(in, -1); - } - - /** - * Creates a new XZ decompressor without a memory usage limit. - *

- * This is identical to XZInputStream(InputStream) - * except that this takes also the arrayCache argument. - * - * @param in input stream from which XZ-compressed - * data is read - * - * @param arrayCache cache to be used for allocating large arrays - * - * @throws XZFormatException - * input is not in the XZ format - * - * @throws CorruptedInputException - * XZ header CRC32 doesn't match - * - * @throws UnsupportedOptionsException - * XZ header is valid but specifies options - * not supported by this implementation - * - * @throws EOFException - * less than 12 bytes of input was available - * from in - * - * @throws IOException may be thrown by in - * - * @since 1.7 - */ - public XZInputStream(InputStream in, ArrayCache arrayCache) - throws IOException { - this(in, -1, arrayCache); - } - - /** - * Creates a new XZ decompressor with an optional memory usage limit. - *

- * This is identical to XZInputStream(InputStream) except - * that this takes also the memoryLimit argument. - * - * @param in input stream from which XZ-compressed - * data is read - * - * @param memoryLimit memory usage limit in kibibytes (KiB) - * or -1 to impose no - * memory usage limit - * - * @throws XZFormatException - * input is not in the XZ format - * - * @throws CorruptedInputException - * XZ header CRC32 doesn't match - * - * @throws UnsupportedOptionsException - * XZ header is valid but specifies options - * not supported by this implementation - * - * @throws EOFException - * less than 12 bytes of input was available - * from in - * - * @throws IOException may be thrown by in - */ - public XZInputStream(InputStream in, int memoryLimit) throws IOException { - this(in, memoryLimit, true); - } - - /** - * Creates a new XZ decompressor with an optional memory usage limit. - *

- * This is identical to XZInputStream(InputStream) except - * that this takes also the memoryLimit and - * arrayCache arguments. - * - * @param in input stream from which XZ-compressed - * data is read - * - * @param memoryLimit memory usage limit in kibibytes (KiB) - * or -1 to impose no - * memory usage limit - * - * @param arrayCache cache to be used for allocating large arrays - * - * @throws XZFormatException - * input is not in the XZ format - * - * @throws CorruptedInputException - * XZ header CRC32 doesn't match - * - * @throws UnsupportedOptionsException - * XZ header is valid but specifies options - * not supported by this implementation - * - * @throws EOFException - * less than 12 bytes of input was available - * from in - * - * @throws IOException may be thrown by in - * - * @since 1.7 - */ - public XZInputStream(InputStream in, int memoryLimit, - ArrayCache arrayCache) throws IOException { - this(in, memoryLimit, true, arrayCache); - } - - /** - * Creates a new XZ decompressor with an optional memory usage limit - * and ability to disable verification of integrity checks. - *

- * This is identical to XZInputStream(InputStream,int) except - * that this takes also the verifyCheck argument. - *

- * Note that integrity check verification should almost never be disabled. - * Possible reasons to disable integrity check verification: - *

- *

- * verifyCheck only affects the integrity check of - * the actual compressed data. The CRC32 fields in the headers - * are always verified. - * - * @param in input stream from which XZ-compressed - * data is read - * - * @param memoryLimit memory usage limit in kibibytes (KiB) - * or -1 to impose no - * memory usage limit - * - * @param verifyCheck if true, the integrity checks - * will be verified; this should almost never - * be set to false - * - * @throws XZFormatException - * input is not in the XZ format - * - * @throws CorruptedInputException - * XZ header CRC32 doesn't match - * - * @throws UnsupportedOptionsException - * XZ header is valid but specifies options - * not supported by this implementation - * - * @throws EOFException - * less than 12 bytes of input was available - * from in - * - * @throws IOException may be thrown by in - * - * @since 1.6 - */ - public XZInputStream(InputStream in, int memoryLimit, boolean verifyCheck) - throws IOException { - this(in, memoryLimit, verifyCheck, ArrayCache.getDefaultCache()); - } - - /** - * Creates a new XZ decompressor with an optional memory usage limit - * and ability to disable verification of integrity checks. - *

- * This is identical to XZInputStream(InputStream,int,boolean) - * except that this takes also the arrayCache argument. - * - * @param in input stream from which XZ-compressed - * data is read - * - * @param memoryLimit memory usage limit in kibibytes (KiB) - * or -1 to impose no - * memory usage limit - * - * @param verifyCheck if true, the integrity checks - * will be verified; this should almost never - * be set to false - * - * @param arrayCache cache to be used for allocating large arrays - * - * @throws XZFormatException - * input is not in the XZ format - * - * @throws CorruptedInputException - * XZ header CRC32 doesn't match - * - * @throws UnsupportedOptionsException - * XZ header is valid but specifies options - * not supported by this implementation - * - * @throws EOFException - * less than 12 bytes of input was available - * from in - * - * @throws IOException may be thrown by in - * - * @since 1.7 - */ - public XZInputStream(InputStream in, int memoryLimit, boolean verifyCheck, - ArrayCache arrayCache) throws IOException { - this.arrayCache = arrayCache; - this.in = in; - this.memoryLimit = memoryLimit; - this.verifyCheck = verifyCheck; - this.xzIn = new SingleXZInputStream(in, memoryLimit, verifyCheck, - arrayCache); - } - - /** - * Decompresses the next byte from this input stream. - *

- * Reading lots of data with read() from this input stream - * may be inefficient. Wrap it in {@link java.io.BufferedInputStream} - * if you need to read lots of data one byte at a time. - * - * @return the next decompressed byte, or -1 - * to indicate the end of the compressed stream - * - * @throws CorruptedInputException - * @throws UnsupportedOptionsException - * @throws MemoryLimitException - * - * @throws XZIOException if the stream has been closed - * - * @throws EOFException - * compressed input is truncated or corrupt - * - * @throws IOException may be thrown by in - */ - public int read() throws IOException { - return read(tempBuf, 0, 1) == -1 ? -1 : (tempBuf[0] & 0xFF); - } - - /** - * Decompresses into an array of bytes. - *

- * If len is zero, no bytes are read and 0 - * is returned. Otherwise this will try to decompress len - * bytes of uncompressed data. Less than len bytes may - * be read only in the following situations: - *

- * - * @param buf target buffer for uncompressed data - * @param off start offset in buf - * @param len maximum number of uncompressed bytes to read - * - * @return number of bytes read, or -1 to indicate - * the end of the compressed stream - * - * @throws CorruptedInputException - * @throws UnsupportedOptionsException - * @throws MemoryLimitException - * - * @throws XZIOException if the stream has been closed - * - * @throws EOFException - * compressed input is truncated or corrupt - * - * @throws IOException may be thrown by in - */ - public int read(byte[] buf, int off, int len) throws IOException { - if (off < 0 || len < 0 || off + len < 0 || off + len > buf.length) - throw new IndexOutOfBoundsException(); - - if (len == 0) - return 0; - - if (in == null) - throw new XZIOException("Stream closed"); - - if (exception != null) - throw exception; - - if (endReached) - return -1; - - int size = 0; - - try { - while (len > 0) { - if (xzIn == null) { - prepareNextStream(); - if (endReached) - return size == 0 ? -1 : size; - } - - int ret = xzIn.read(buf, off, len); - - if (ret > 0) { - size += ret; - off += ret; - len -= ret; - } else if (ret == -1) { - xzIn = null; - } - } - } catch (IOException e) { - exception = e; - if (size == 0) - throw e; - } - - return size; - } - - private void prepareNextStream() throws IOException { - DataInputStream inData = new DataInputStream(in); - byte[] buf = new byte[DecoderUtil.STREAM_HEADER_SIZE]; - - // The size of Stream Padding must be a multiple of four bytes, - // all bytes zero. - do { - // First try to read one byte to see if we have reached the end - // of the file. - int ret = inData.read(buf, 0, 1); - if (ret == -1) { - endReached = true; - return; - } - - // Since we got one byte of input, there must be at least - // three more available in a valid file. - inData.readFully(buf, 1, 3); - - } while (buf[0] == 0 && buf[1] == 0 && buf[2] == 0 && buf[3] == 0); - - // Not all bytes are zero. In a valid Stream it indicates the - // beginning of the next Stream. Read the rest of the Stream Header - // and initialize the XZ decoder. - inData.readFully(buf, 4, DecoderUtil.STREAM_HEADER_SIZE - 4); - - try { - xzIn = new SingleXZInputStream(in, memoryLimit, verifyCheck, buf, - arrayCache); - } catch (XZFormatException e) { - // Since this isn't the first .xz Stream, it is more - // logical to tell that the data is corrupt. - throw new CorruptedInputException( - "Garbage after a valid XZ Stream"); - } - } - - /** - * Returns the number of uncompressed bytes that can be read - * without blocking. The value is returned with an assumption - * that the compressed input data will be valid. If the compressed - * data is corrupt, CorruptedInputException may get - * thrown before the number of bytes claimed to be available have - * been read from this input stream. - * - * @return the number of uncompressed bytes that can be read - * without blocking - */ - public int available() throws IOException { - if (in == null) - throw new XZIOException("Stream closed"); - - if (exception != null) - throw exception; - - return xzIn == null ? 0 : xzIn.available(); - } - - /** - * Closes the stream and calls in.close(). - * If the stream was already closed, this does nothing. - *

- * This is equivalent to close(true). - * - * @throws IOException if thrown by in.close() - */ - public void close() throws IOException { - close(true); - } - - /** - * Closes the stream and optionally calls in.close(). - * If the stream was already closed, this does nothing. - * If close(false) has been called, a further - * call of close(true) does nothing (it doesn't call - * in.close()). - *

- * If you don't want to close the underlying InputStream, - * there is usually no need to worry about closing this stream either; - * it's fine to do nothing and let the garbage collector handle it. - * However, if you are using {@link ArrayCache}, close(false) - * can be useful to put the allocated arrays back to the cache without - * closing the underlying InputStream. - *

- * Note that if you successfully reach the end of the stream - * (read returns -1), the arrays are - * automatically put back to the cache by that read call. In - * this situation close(false) is redundant (but harmless). - * - * @throws IOException if thrown by in.close() - * - * @since 1.7 - */ - public void close(boolean closeInput) throws IOException { - if (in != null) { - if (xzIn != null) { - xzIn.close(false); - xzIn = null; - } - - try { - if (closeInput) - in.close(); - } finally { - in = null; - } - } - } -} Index: trunk/src/org/tukaani/xz/XZOutputStream.java =================================================================== --- trunk/src/org/tukaani/xz/XZOutputStream.java (revision 16021) +++ (revision ) @@ -1,606 +1,0 @@ -/* - * XZOutputStream - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz; - -import java.io.OutputStream; -import java.io.IOException; -import org.tukaani.xz.common.EncoderUtil; -import org.tukaani.xz.common.StreamFlags; -import org.tukaani.xz.check.Check; -import org.tukaani.xz.index.IndexEncoder; - -/** - * Compresses into the .xz file format. - * - *

Examples

- *

- * Getting an output stream to compress with LZMA2 using the default - * settings and the default integrity check type (CRC64): - * 

- * FileOutputStream outfile = new FileOutputStream("foo.xz");
- * XZOutputStream outxz = new XZOutputStream(outfile, new LZMA2Options());
- *
- *

- * Using the preset level 8 for LZMA2 (the default - * is 6) and SHA-256 instead of CRC64 for integrity checking: - * 

- * XZOutputStream outxz = new XZOutputStream(outfile, new LZMA2Options(8),
- *                                           XZ.CHECK_SHA256);
- *
- *

- * Using the x86 BCJ filter together with LZMA2 to compress x86 executables - * and printing the memory usage information before creating the - * XZOutputStream: - * 

- * X86Options x86 = new X86Options();
- * LZMA2Options lzma2 = new LZMA2Options();
- * FilterOptions[] options = { x86, lzma2 };
- * System.out.println("Encoder memory usage: "
- *                    + FilterOptions.getEncoderMemoryUsage(options)
- *                    + " KiB");
- * System.out.println("Decoder memory usage: "
- *                    + FilterOptions.getDecoderMemoryUsage(options)
- *                    + " KiB");
- * XZOutputStream outxz = new XZOutputStream(outfile, options);
- *
- */ -public class XZOutputStream extends FinishableOutputStream { - private final ArrayCache arrayCache; - - private OutputStream out; - private final StreamFlags streamFlags = new StreamFlags(); - private final Check check; - private final IndexEncoder index = new IndexEncoder(); - - private BlockOutputStream blockEncoder = null; - private FilterEncoder[] filters; - - /** - * True if the current filter chain supports flushing. - * If it doesn't support flushing, flush() - * will use endBlock() as a fallback. - */ - private boolean filtersSupportFlushing; - - private IOException exception = null; - private boolean finished = false; - - private final byte[] tempBuf = new byte[1]; - - /** - * Creates a new XZ compressor using one filter and CRC64 as - * the integrity check. This constructor is equivalent to passing - * a single-member FilterOptions array to - * XZOutputStream(OutputStream, FilterOptions[]). - * - * @param out output stream to which the compressed data - * will be written - * - * @param filterOptions - * filter options to use - * - * @throws UnsupportedOptionsException - * invalid filter chain - * - * @throws IOException may be thrown from out - */ - public XZOutputStream(OutputStream out, FilterOptions filterOptions) - throws IOException { - this(out, filterOptions, XZ.CHECK_CRC64); - } - - /** - * Creates a new XZ compressor using one filter and CRC64 as - * the integrity check. This constructor is equivalent to passing - * a single-member FilterOptions array to - * XZOutputStream(OutputStream, FilterOptions[], ArrayCache). - * - * @param out output stream to which the compressed data - * will be written - * - * @param filterOptions - * filter options to use - * - * @param arrayCache cache to be used for allocating large arrays - * - * @throws UnsupportedOptionsException - * invalid filter chain - * - * @throws IOException may be thrown from out - * - * @since 1.7 - */ - public XZOutputStream(OutputStream out, FilterOptions filterOptions, - ArrayCache arrayCache) - throws IOException { - this(out, filterOptions, XZ.CHECK_CRC64, arrayCache); - } - - /** - * Creates a new XZ compressor using one filter and the specified - * integrity check type. This constructor is equivalent to - * passing a single-member FilterOptions array to - * XZOutputStream(OutputStream, FilterOptions[], int). - * - * @param out output stream to which the compressed data - * will be written - * - * @param filterOptions - * filter options to use - * - * @param checkType type of the integrity check, - * for example XZ.CHECK_CRC32 - * - * @throws UnsupportedOptionsException - * invalid filter chain - * - * @throws IOException may be thrown from out - */ - public XZOutputStream(OutputStream out, FilterOptions filterOptions, - int checkType) throws IOException { - this(out, new FilterOptions[] { filterOptions }, checkType); - } - - /** - * Creates a new XZ compressor using one filter and the specified - * integrity check type. This constructor is equivalent to - * passing a single-member FilterOptions array to - * XZOutputStream(OutputStream, FilterOptions[], int, - * ArrayCache). - * - * @param out output stream to which the compressed data - * will be written - * - * @param filterOptions - * filter options to use - * - * @param checkType type of the integrity check, - * for example XZ.CHECK_CRC32 - * - * @param arrayCache cache to be used for allocating large arrays - * - * @throws UnsupportedOptionsException - * invalid filter chain - * - * @throws IOException may be thrown from out - * - * @since 1.7 - */ - public XZOutputStream(OutputStream out, FilterOptions filterOptions, - int checkType, ArrayCache arrayCache) - throws IOException { - this(out, new FilterOptions[] { filterOptions }, checkType, - arrayCache); - } - - /** - * Creates a new XZ compressor using 1-4 filters and CRC64 as - * the integrity check. This constructor is equivalent - * XZOutputStream(out, filterOptions, XZ.CHECK_CRC64). - * - * @param out output stream to which the compressed data - * will be written - * - * @param filterOptions - * array of filter options to use - * - * @throws UnsupportedOptionsException - * invalid filter chain - * - * @throws IOException may be thrown from out - */ - public XZOutputStream(OutputStream out, FilterOptions[] filterOptions) - throws IOException { - this(out, filterOptions, XZ.CHECK_CRC64); - } - - /** - * Creates a new XZ compressor using 1-4 filters and CRC64 as - * the integrity check. This constructor is equivalent - * XZOutputStream(out, filterOptions, XZ.CHECK_CRC64, - * arrayCache). - * - * @param out output stream to which the compressed data - * will be written - * - * @param filterOptions - * array of filter options to use - * - * @param arrayCache cache to be used for allocating large arrays - * - * @throws UnsupportedOptionsException - * invalid filter chain - * - * @throws IOException may be thrown from out - * - * @since 1.7 - */ - public XZOutputStream(OutputStream out, FilterOptions[] filterOptions, - ArrayCache arrayCache) - throws IOException { - this(out, filterOptions, XZ.CHECK_CRC64, arrayCache); - } - - /** - * Creates a new XZ compressor using 1-4 filters and the specified - * integrity check type. - * - * @param out output stream to which the compressed data - * will be written - * - * @param filterOptions - * array of filter options to use - * - * @param checkType type of the integrity check, - * for example XZ.CHECK_CRC32 - * - * @throws UnsupportedOptionsException - * invalid filter chain - * - * @throws IOException may be thrown from out - */ - public XZOutputStream(OutputStream out, FilterOptions[] filterOptions, - int checkType) throws IOException { - this(out, filterOptions, checkType, ArrayCache.getDefaultCache()); - } - - /** - * Creates a new XZ compressor using 1-4 filters and the specified - * integrity check type. - * - * @param out output stream to which the compressed data - * will be written - * - * @param filterOptions - * array of filter options to use - * - * @param checkType type of the integrity check, - * for example XZ.CHECK_CRC32 - * - * @param arrayCache cache to be used for allocating large arrays - * - * @throws UnsupportedOptionsException - * invalid filter chain - * - * @throws IOException may be thrown from out - * - * @since 1.7 - */ - public XZOutputStream(OutputStream out, FilterOptions[] filterOptions, - int checkType, ArrayCache arrayCache) - throws IOException { - this.arrayCache = arrayCache; - this.out = out; - updateFilters(filterOptions); - - streamFlags.checkType = checkType; - check = Check.getInstance(checkType); - - encodeStreamHeader(); - } - - /** - * Updates the filter chain with a single filter. - * This is equivalent to passing a single-member FilterOptions array - * to updateFilters(FilterOptions[]). - * - * @param filterOptions - * new filter to use - * - * @throws UnsupportedOptionsException - * unsupported filter chain, or trying to change - * the filter chain in the middle of a Block - */ - public void updateFilters(FilterOptions filterOptions) - throws XZIOException { - FilterOptions[] opts = new FilterOptions[1]; - opts[0] = filterOptions; - updateFilters(opts); - } - - /** - * Updates the filter chain with 1-4 filters. - *

- * Currently this cannot be used to update e.g. LZMA2 options in the - * middle of a XZ Block. Use endBlock() to finish the - * current XZ Block before calling this function. The new filter chain - * will then be used for the next XZ Block. - * - * @param filterOptions - * new filter chain to use - * - * @throws UnsupportedOptionsException - * unsupported filter chain, or trying to change - * the filter chain in the middle of a Block - */ - public void updateFilters(FilterOptions[] filterOptions) - throws XZIOException { - if (blockEncoder != null) - throw new UnsupportedOptionsException("Changing filter options " - + "in the middle of a XZ Block not implemented"); - - if (filterOptions.length < 1 || filterOptions.length > 4) - throw new UnsupportedOptionsException( - "XZ filter chain must be 1-4 filters"); - - filtersSupportFlushing = true; - FilterEncoder[] newFilters = new FilterEncoder[filterOptions.length]; - for (int i = 0; i < filterOptions.length; ++i) { - newFilters[i] = filterOptions[i].getFilterEncoder(); - filtersSupportFlushing &= newFilters[i].supportsFlushing(); - } - - RawCoder.validate(newFilters); - filters = newFilters; - } - - /** - * Writes one byte to be compressed. - * - * @throws XZIOException - * XZ Stream has grown too big - * - * @throws XZIOException - * finish() or close() - * was already called - * - * @throws IOException may be thrown by the underlying output stream - */ - public void write(int b) throws IOException { - tempBuf[0] = (byte)b; - write(tempBuf, 0, 1); - } - - /** - * Writes an array of bytes to be compressed. - * The compressors tend to do internal buffering and thus the written - * data won't be readable from the compressed output immediately. - * Use flush() to force everything written so far to - * be written to the underlaying output stream, but be aware that - * flushing reduces compression ratio. - * - * @param buf buffer of bytes to be written - * @param off start offset in buf - * @param len number of bytes to write - * - * @throws XZIOException - * XZ Stream has grown too big: total file size - * about 8 EiB or the Index field exceeds - * 16 GiB; you shouldn't reach these sizes - * in practice - * - * @throws XZIOException - * finish() or close() - * was already called and len > 0 - * - * @throws IOException may be thrown by the underlying output stream - */ - public void write(byte[] buf, int off, int len) throws IOException { - if (off < 0 || len < 0 || off + len < 0 || off + len > buf.length) - throw new IndexOutOfBoundsException(); - - if (exception != null) - throw exception; - - if (finished) - throw new XZIOException("Stream finished or closed"); - - try { - if (blockEncoder == null) - blockEncoder = new BlockOutputStream(out, filters, check, - arrayCache); - - blockEncoder.write(buf, off, len); - } catch (IOException e) { - exception = e; - throw e; - } - } - - /** - * Finishes the current XZ Block (but not the whole XZ Stream). - * This doesn't flush the stream so it's possible that not all data will - * be decompressible from the output stream when this function returns. - * Call also flush() if flushing is wanted in addition to - * finishing the current XZ Block. - *

- * If there is no unfinished Block open, this function will do nothing. - * (No empty XZ Block will be created.) - *

- * This function can be useful, for example, to create - * random-accessible .xz files. - *

- * Starting a new XZ Block means that the encoder state is reset. - * Doing this very often will increase the size of the compressed - * file a lot (more than plain flush() would do). - * - * @throws XZIOException - * XZ Stream has grown too big - * - * @throws XZIOException - * stream finished or closed - * - * @throws IOException may be thrown by the underlying output stream - */ - public void endBlock() throws IOException { - if (exception != null) - throw exception; - - if (finished) - throw new XZIOException("Stream finished or closed"); - - // NOTE: Once there is threading with multiple Blocks, it's possible - // that this function will be more like a barrier that returns - // before the last Block has been finished. - if (blockEncoder != null) { - try { - blockEncoder.finish(); - index.add(blockEncoder.getUnpaddedSize(), - blockEncoder.getUncompressedSize()); - blockEncoder = null; - } catch (IOException e) { - exception = e; - throw e; - } - } - } - - /** - * Flushes the encoder and calls out.flush(). - * All buffered pending data will then be decompressible from - * the output stream. - *

- * Calling this function very often may increase the compressed - * file size a lot. The filter chain options may affect the size - * increase too. For example, with LZMA2 the HC4 match finder has - * smaller penalty with flushing than BT4. - *

- * Some filters don't support flushing. If the filter chain has - * such a filter, flush() will call endBlock() - * before flushing. - * - * @throws XZIOException - * XZ Stream has grown too big - * - * @throws XZIOException - * stream finished or closed - * - * @throws IOException may be thrown by the underlying output stream - */ - public void flush() throws IOException { - if (exception != null) - throw exception; - - if (finished) - throw new XZIOException("Stream finished or closed"); - - try { - if (blockEncoder != null) { - if (filtersSupportFlushing) { - // This will eventually call out.flush() so - // no need to do it here again. - blockEncoder.flush(); - } else { - endBlock(); - out.flush(); - } - } else { - out.flush(); - } - } catch (IOException e) { - exception = e; - throw e; - } - } - - /** - * Finishes compression without closing the underlying stream. - * No more data can be written to this stream after finishing - * (calling write with an empty buffer is OK). - *

- * Repeated calls to finish() do nothing unless - * an exception was thrown by this stream earlier. In that case - * the same exception is thrown again. - *

- * After finishing, the stream may be closed normally with - * close(). If the stream will be closed anyway, there - * usually is no need to call finish() separately. - * - * @throws XZIOException - * XZ Stream has grown too big - * - * @throws IOException may be thrown by the underlying output stream - */ - public void finish() throws IOException { - if (!finished) { - // This checks for pending exceptions so we don't need to - // worry about it here. - endBlock(); - - try { - index.encode(out); - encodeStreamFooter(); - } catch (IOException e) { - exception = e; - throw e; - } - - // Set it to true only if everything goes fine. Setting it earlier - // would cause repeated calls to finish() do nothing instead of - // throwing an exception to indicate an earlier error. - finished = true; - } - } - - /** - * Finishes compression and closes the underlying stream. - * The underlying stream out is closed even if finishing - * fails. If both finishing and closing fail, the exception thrown - * by finish() is thrown and the exception from the failed - * out.close() is lost. - * - * @throws XZIOException - * XZ Stream has grown too big - * - * @throws IOException may be thrown by the underlying output stream - */ - public void close() throws IOException { - if (out != null) { - // If finish() throws an exception, it stores the exception to - // the variable "exception". So we can ignore the possible - // exception here. - try { - finish(); - } catch (IOException e) {} - - try { - out.close(); - } catch (IOException e) { - // Remember the exception but only if there is no previous - // pending exception. - if (exception == null) - exception = e; - } - - out = null; - } - - if (exception != null) - throw exception; - } - - private void encodeStreamFlags(byte[] buf, int off) { - buf[off] = 0x00; - buf[off + 1] = (byte)streamFlags.checkType; - } - - private void encodeStreamHeader() throws IOException { - out.write(XZ.HEADER_MAGIC); - - byte[] buf = new byte[2]; - encodeStreamFlags(buf, 0); - out.write(buf); - - EncoderUtil.writeCRC32(out, buf); - } - - private void encodeStreamFooter() throws IOException { - byte[] buf = new byte[6]; - long backwardSize = index.getIndexSize() / 4 - 1; - for (int i = 0; i < 4; ++i) - buf[i] = (byte)(backwardSize >>> (i * 8)); - - encodeStreamFlags(buf, 4); - - EncoderUtil.writeCRC32(out, buf); - out.write(buf); - out.write(XZ.FOOTER_MAGIC); - } -} Index: trunk/src/org/tukaani/xz/check/CRC32.java =================================================================== --- trunk/src/org/tukaani/xz/check/CRC32.java (revision 16021) +++ (revision ) @@ -1,33 +1,0 @@ -/* - * CRC32 - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.check; - -public class CRC32 extends Check { - private final java.util.zip.CRC32 state = new java.util.zip.CRC32(); - - public CRC32() { - size = 4; - name = "CRC32"; - } - - public void update(byte[] buf, int off, int len) { - state.update(buf, off, len); - } - - public byte[] finish() { - long value = state.getValue(); - byte[] buf = { (byte)(value), - (byte)(value >>> 8), - (byte)(value >>> 16), - (byte)(value >>> 24) }; - state.reset(); - return buf; - } -} Index: trunk/src/org/tukaani/xz/check/CRC64.java =================================================================== --- trunk/src/org/tukaani/xz/check/CRC64.java (revision 16021) +++ (revision ) @@ -1,54 +1,0 @@ -/* - * CRC64 - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.check; - -public class CRC64 extends Check { - private static final long poly = 0xC96C5795D7870F42L; - private static final long[] crcTable = new long[256]; - - private long crc = -1; - - static { - for (int b = 0; b < crcTable.length; ++b) { - long r = b; - for (int i = 0; i < 8; ++i) { - if ((r & 1) == 1) - r = (r >>> 1) ^ poly; - else - r >>>= 1; - } - - crcTable[b] = r; - } - } - - public CRC64() { - size = 8; - name = "CRC64"; - } - - public void update(byte[] buf, int off, int len) { - int end = off + len; - - while (off < end) - crc = crcTable[(buf[off++] ^ (int)crc) & 0xFF] ^ (crc >>> 8); - } - - public byte[] finish() { - long value = ~crc; - crc = -1; - - byte[] buf = new byte[8]; - for (int i = 0; i < buf.length; ++i) - buf[i] = (byte)(value >> (i * 8)); - - return buf; - } -} Index: trunk/src/org/tukaani/xz/check/Check.java =================================================================== --- trunk/src/org/tukaani/xz/check/Check.java (revision 16021) +++ (revision ) @@ -1,57 +1,0 @@ -/* - * Check - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.check; - -import org.tukaani.xz.XZ; -import org.tukaani.xz.UnsupportedOptionsException; - -public abstract class Check { - int size; - String name; - - public abstract void update(byte[] buf, int off, int len); - public abstract byte[] finish(); - - public void update(byte[] buf) { - update(buf, 0, buf.length); - } - - public int getSize() { - return size; - } - - public String getName() { - return name; - } - - public static Check getInstance(int checkType) - throws UnsupportedOptionsException { - switch (checkType) { - case XZ.CHECK_NONE: - return new None(); - - case XZ.CHECK_CRC32: - return new CRC32(); - - case XZ.CHECK_CRC64: - return new CRC64(); - - case XZ.CHECK_SHA256: - try { - return new SHA256(); - } catch (java.security.NoSuchAlgorithmException e) {} - - break; - } - - throw new UnsupportedOptionsException( - "Unsupported Check ID " + checkType); - } -} Index: trunk/src/org/tukaani/xz/check/None.java =================================================================== --- trunk/src/org/tukaani/xz/check/None.java (revision 16021) +++ (revision ) @@ -1,24 +1,0 @@ -/* - * None - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.check; - -public class None extends Check { - public None() { - size = 0; - name = "None"; - } - - public void update(byte[] buf, int off, int len) {} - - public byte[] finish() { - byte[] empty = new byte[0]; - return empty; - } -} Index: trunk/src/org/tukaani/xz/check/SHA256.java =================================================================== --- trunk/src/org/tukaani/xz/check/SHA256.java (revision 16021) +++ (revision ) @@ -1,30 +1,0 @@ -/* - * SHA256 - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.check; - -public class SHA256 extends Check { - private final java.security.MessageDigest sha256; - - public SHA256() throws java.security.NoSuchAlgorithmException { - size = 32; - name = "SHA-256"; - sha256 = java.security.MessageDigest.getInstance("SHA-256"); - } - - public void update(byte[] buf, int off, int len) { - sha256.update(buf, off, len); - } - - public byte[] finish() { - byte[] buf = sha256.digest(); - sha256.reset(); - return buf; - } -} Index: trunk/src/org/tukaani/xz/common/DecoderUtil.java =================================================================== --- trunk/src/org/tukaani/xz/common/DecoderUtil.java (revision 16021) +++ (revision ) @@ -1,121 +1,0 @@ -/* - * DecoderUtil - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.common; - -import java.io.InputStream; -import java.io.IOException; -import java.io.EOFException; -import java.util.zip.CRC32; -import org.tukaani.xz.XZ; -import org.tukaani.xz.XZFormatException; -import org.tukaani.xz.CorruptedInputException; -import org.tukaani.xz.UnsupportedOptionsException; - -public class DecoderUtil extends Util { - public static boolean isCRC32Valid(byte[] buf, int off, int len, - int ref_off) { - CRC32 crc32 = new CRC32(); - crc32.update(buf, off, len); - long value = crc32.getValue(); - - for (int i = 0; i < 4; ++i) - if ((byte)(value >>> (i * 8)) != buf[ref_off + i]) - return false; - - return true; - } - - public static StreamFlags decodeStreamHeader(byte[] buf) - throws IOException { - for (int i = 0; i < XZ.HEADER_MAGIC.length; ++i) - if (buf[i] != XZ.HEADER_MAGIC[i]) - throw new XZFormatException(); - - if (!isCRC32Valid(buf, XZ.HEADER_MAGIC.length, 2, - XZ.HEADER_MAGIC.length + 2)) - throw new CorruptedInputException("XZ Stream Header is corrupt"); - - try { - return decodeStreamFlags(buf, XZ.HEADER_MAGIC.length); - } catch (UnsupportedOptionsException e) { - throw new UnsupportedOptionsException( - "Unsupported options in XZ Stream Header"); - } - } - - public static StreamFlags decodeStreamFooter(byte[] buf) - throws IOException { - if (buf[10] != XZ.FOOTER_MAGIC[0] || buf[11] != XZ.FOOTER_MAGIC[1]) { - // NOTE: The exception could be XZFormatException too. - // It depends on the situation which one is better. - throw new CorruptedInputException("XZ Stream Footer is corrupt"); - } - - if (!isCRC32Valid(buf, 4, 6, 0)) - throw new CorruptedInputException("XZ Stream Footer is corrupt"); - - StreamFlags streamFlags; - try { - streamFlags = decodeStreamFlags(buf, 8); - } catch (UnsupportedOptionsException e) { - throw new UnsupportedOptionsException( - "Unsupported options in XZ Stream Footer"); - } - - streamFlags.backwardSize = 0; - for (int i = 0; i < 4; ++i) - streamFlags.backwardSize |= (buf[i + 4] & 0xFF) << (i * 8); - - streamFlags.backwardSize = (streamFlags.backwardSize + 1) * 4; - - return streamFlags; - } - - private static StreamFlags decodeStreamFlags(byte[] buf, int off) - throws UnsupportedOptionsException { - if (buf[off] != 0x00 || (buf[off + 1] & 0xFF) >= 0x10) - throw new UnsupportedOptionsException(); - - StreamFlags streamFlags = new StreamFlags(); - streamFlags.checkType = buf[off + 1]; - - return streamFlags; - } - - public static boolean areStreamFlagsEqual(StreamFlags a, StreamFlags b) { - // backwardSize is intentionally not compared. - return a.checkType == b.checkType; - } - - public static long decodeVLI(InputStream in) throws IOException { - int b = in.read(); - if (b == -1) - throw new EOFException(); - - long num = b & 0x7F; - int i = 0; - - while ((b & 0x80) != 0x00) { - if (++i >= VLI_SIZE_MAX) - throw new CorruptedInputException(); - - b = in.read(); - if (b == -1) - throw new EOFException(); - - if (b == 0x00) - throw new CorruptedInputException(); - - num |= (long)(b & 0x7F) << (i * 7); - } - - return num; - } -} Index: trunk/src/org/tukaani/xz/common/EncoderUtil.java =================================================================== --- trunk/src/org/tukaani/xz/common/EncoderUtil.java (revision 16021) +++ (revision ) @@ -1,36 +1,0 @@ -/* - * EncoderUtil - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.common; - -import java.io.OutputStream; -import java.io.IOException; -import java.util.zip.CRC32; - -public class EncoderUtil extends Util { - public static void writeCRC32(OutputStream out, byte[] buf) - throws IOException { - CRC32 crc32 = new CRC32(); - crc32.update(buf); - long value = crc32.getValue(); - - for (int i = 0; i < 4; ++i) - out.write((byte)(value >>> (i * 8))); - } - - public static void encodeVLI(OutputStream out, long num) - throws IOException { - while (num >= 0x80) { - out.write((byte)(num | 0x80)); - num >>>= 7; - } - - out.write((byte)num); - } -} Index: trunk/src/org/tukaani/xz/common/StreamFlags.java =================================================================== --- trunk/src/org/tukaani/xz/common/StreamFlags.java (revision 16021) +++ (revision ) @@ -1,15 +1,0 @@ -/* - * StreamFlags - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.common; - -public class StreamFlags { - public int checkType = -1; - public long backwardSize = -1; -} Index: trunk/src/org/tukaani/xz/common/Util.java =================================================================== --- trunk/src/org/tukaani/xz/common/Util.java (revision 16021) +++ (revision ) @@ -1,28 +1,0 @@ -/* - * Util - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.common; - -public class Util { - public static final int STREAM_HEADER_SIZE = 12; - public static final long BACKWARD_SIZE_MAX = 1L << 34; - public static final int BLOCK_HEADER_SIZE_MAX = 1024; - public static final long VLI_MAX = Long.MAX_VALUE; - public static final int VLI_SIZE_MAX = 9; - - public static int getVLISize(long num) { - int size = 0; - do { - ++size; - num >>= 7; - } while (num != 0); - - return size; - } -} Index: trunk/src/org/tukaani/xz/delta/DeltaCoder.java =================================================================== --- trunk/src/org/tukaani/xz/delta/DeltaCoder.java (revision 16021) +++ (revision ) @@ -1,27 +1,0 @@ -/* - * DeltaCoder - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.delta; - -abstract class DeltaCoder { - static final int DISTANCE_MIN = 1; - static final int DISTANCE_MAX = 256; - static final int DISTANCE_MASK = DISTANCE_MAX - 1; - - final int distance; - final byte[] history = new byte[DISTANCE_MAX]; - int pos = 0; - - DeltaCoder(int distance) { - if (distance < DISTANCE_MIN || distance > DISTANCE_MAX) - throw new IllegalArgumentException(); - - this.distance = distance; - } -} Index: trunk/src/org/tukaani/xz/delta/DeltaDecoder.java =================================================================== --- trunk/src/org/tukaani/xz/delta/DeltaDecoder.java (revision 16021) +++ (revision ) @@ -1,24 +1,0 @@ -/* - * DeltaDecoder - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.delta; - -public class DeltaDecoder extends DeltaCoder { - public DeltaDecoder(int distance) { - super(distance); - } - - public void decode(byte[] buf, int off, int len) { - int end = off + len; - for (int i = off; i < end; ++i) { - buf[i] += history[(distance + pos) & DISTANCE_MASK]; - history[pos-- & DISTANCE_MASK] = buf[i]; - } - } -} Index: trunk/src/org/tukaani/xz/delta/DeltaEncoder.java =================================================================== --- trunk/src/org/tukaani/xz/delta/DeltaEncoder.java (revision 16021) +++ (revision ) @@ -1,24 +1,0 @@ -/* - * DeltaEncoder - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.delta; - -public class DeltaEncoder extends DeltaCoder { - public DeltaEncoder(int distance) { - super(distance); - } - - public void encode(byte[] in, int in_off, int len, byte[] out) { - for (int i = 0; i < len; ++i) { - byte tmp = history[(distance + pos) & DISTANCE_MASK]; - history[pos-- & DISTANCE_MASK] = in[in_off + i]; - out[i] = (byte)(in[in_off + i] - tmp); - } - } -} Index: trunk/src/org/tukaani/xz/index/BlockInfo.java =================================================================== --- trunk/src/org/tukaani/xz/index/BlockInfo.java (revision 16021) +++ (revision ) @@ -1,38 +1,0 @@ -/* - * BlockInfo - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.index; - -import org.tukaani.xz.common.StreamFlags; - -public class BlockInfo { - public int blockNumber = -1; - public long compressedOffset = -1; - public long uncompressedOffset = -1; - public long unpaddedSize = -1; - public long uncompressedSize = -1; - - IndexDecoder index; - - public BlockInfo(IndexDecoder indexOfFirstStream) { - index = indexOfFirstStream; - } - - public int getCheckType() { - return index.getStreamFlags().checkType; - } - - public boolean hasNext() { - return index.hasRecord(blockNumber + 1); - } - - public void setNext() { - index.setBlockInfo(this, blockNumber + 1); - } -} Index: trunk/src/org/tukaani/xz/index/IndexBase.java =================================================================== --- trunk/src/org/tukaani/xz/index/IndexBase.java (revision 16021) +++ (revision ) @@ -1,56 +1,0 @@ -/* - * IndexBase - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.index; - -import org.tukaani.xz.common.Util; -import org.tukaani.xz.XZIOException; - -abstract class IndexBase { - private final XZIOException invalidIndexException; - long blocksSum = 0; - long uncompressedSum = 0; - long indexListSize = 0; - long recordCount = 0; - - IndexBase(XZIOException invalidIndexException) { - this.invalidIndexException = invalidIndexException; - } - - private long getUnpaddedIndexSize() { - // Index Indicator + Number of Records + List of Records + CRC32 - return 1 + Util.getVLISize(recordCount) + indexListSize + 4; - } - - public long getIndexSize() { - return (getUnpaddedIndexSize() + 3) & ~3; - } - - public long getStreamSize() { - return Util.STREAM_HEADER_SIZE + blocksSum + getIndexSize() - + Util.STREAM_HEADER_SIZE; - } - - int getIndexPaddingSize() { - return (int)((4 - getUnpaddedIndexSize()) & 3); - } - - void add(long unpaddedSize, long uncompressedSize) throws XZIOException { - blocksSum += (unpaddedSize + 3) & ~3; - uncompressedSum += uncompressedSize; - indexListSize += Util.getVLISize(unpaddedSize) - + Util.getVLISize(uncompressedSize); - ++recordCount; - - if (blocksSum < 0 || uncompressedSum < 0 - || getIndexSize() > Util.BACKWARD_SIZE_MAX - || getStreamSize() < 0) - throw invalidIndexException; - } -} Index: trunk/src/org/tukaani/xz/index/IndexDecoder.java =================================================================== --- trunk/src/org/tukaani/xz/index/IndexDecoder.java (revision 16021) +++ (revision ) @@ -1,223 +1,0 @@ -/* - * IndexDecoder - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.index; - -import java.io.IOException; -import java.io.EOFException; -import java.util.zip.CheckedInputStream; -import org.tukaani.xz.common.DecoderUtil; -import org.tukaani.xz.common.StreamFlags; -import org.tukaani.xz.SeekableInputStream; -import org.tukaani.xz.CorruptedInputException; -import org.tukaani.xz.MemoryLimitException; -import org.tukaani.xz.UnsupportedOptionsException; - -public class IndexDecoder extends IndexBase { - private final StreamFlags streamFlags; - private final long streamPadding; - private final int memoryUsage; - - // Unpadded Size and Uncompressed Size fields - private final long[] unpadded; - private final long[] uncompressed; - - // Uncompressed size of the largest Block. It is used by - // SeekableXZInputStream to find out the largest Block of the .xz file. - private long largestBlockSize = 0; - - // Offsets relative to the beginning of the .xz file. These are all zero - // for the first Stream in the file. - private int recordOffset = 0; - private long compressedOffset = 0; - private long uncompressedOffset = 0; - - public IndexDecoder(SeekableInputStream in, StreamFlags streamFooterFlags, - long streamPadding, int memoryLimit) - throws IOException { - super(new CorruptedInputException("XZ Index is corrupt")); - this.streamFlags = streamFooterFlags; - this.streamPadding = streamPadding; - - // If endPos is exceeded before the CRC32 field has been decoded, - // the Index is corrupt. - long endPos = in.position() + streamFooterFlags.backwardSize - 4; - - java.util.zip.CRC32 crc32 = new java.util.zip.CRC32(); - CheckedInputStream inChecked = new CheckedInputStream(in, crc32); - - // Index Indicator - if (inChecked.read() != 0x00) - throw new CorruptedInputException("XZ Index is corrupt"); - - try { - // Number of Records - long count = DecoderUtil.decodeVLI(inChecked); - - // Catch Record counts that obviously too high to be valid. - // This test isn't exact because it ignores Index Indicator, - // Number of Records, and CRC32 fields, but this is good enough - // to catch the most obvious problems. - if (count >= streamFooterFlags.backwardSize / 2) - throw new CorruptedInputException("XZ Index is corrupt"); - - // If the Record count doesn't fit into an int, we cannot - // allocate the arrays to hold the Records. - if (count > Integer.MAX_VALUE) - throw new UnsupportedOptionsException("XZ Index has over " - + Integer.MAX_VALUE + " Records"); - - // Calculate approximate memory requirements and check the - // memory usage limit. - memoryUsage = 1 + (int)((16L * count + 1023) / 1024); - if (memoryLimit >= 0 && memoryUsage > memoryLimit) - throw new MemoryLimitException(memoryUsage, memoryLimit); - - // Allocate the arrays for the Records. - unpadded = new long[(int)count]; - uncompressed = new long[(int)count]; - int record = 0; - - // Decode the Records. - for (int i = (int)count; i > 0; --i) { - // Get the next Record. - long unpaddedSize = DecoderUtil.decodeVLI(inChecked); - long uncompressedSize = DecoderUtil.decodeVLI(inChecked); - - // Check that the input position stays sane. Since this is - // checked only once per loop iteration instead of for - // every input byte read, it's still possible that - // EOFException gets thrown with corrupt input. - if (in.position() > endPos) - throw new CorruptedInputException("XZ Index is corrupt"); - - // Add the new Record. - unpadded[record] = blocksSum + unpaddedSize; - uncompressed[record] = uncompressedSum + uncompressedSize; - ++record; - super.add(unpaddedSize, uncompressedSize); - assert record == recordCount; - - // Remember the uncompressed size of the largest Block. - if (largestBlockSize < uncompressedSize) - largestBlockSize = uncompressedSize; - } - } catch (EOFException e) { - // EOFException is caught just in case a corrupt input causes - // DecoderUtil.decodeVLI to read too much at once. - throw new CorruptedInputException("XZ Index is corrupt"); - } - - // Validate that the size of the Index field matches - // Backward Size. - int indexPaddingSize = getIndexPaddingSize(); - if (in.position() + indexPaddingSize != endPos) - throw new CorruptedInputException("XZ Index is corrupt"); - - // Index Padding - while (indexPaddingSize-- > 0) - if (inChecked.read() != 0x00) - throw new CorruptedInputException("XZ Index is corrupt"); - - // CRC32 - long value = crc32.getValue(); - for (int i = 0; i < 4; ++i) - if (((value >>> (i * 8)) & 0xFF) != in.read()) - throw new CorruptedInputException("XZ Index is corrupt"); - } - - public void setOffsets(IndexDecoder prev) { - // NOTE: SeekableXZInputStream checks that the total number of Blocks - // in concatenated Streams fits into an int. - recordOffset = prev.recordOffset + (int)prev.recordCount; - compressedOffset = prev.compressedOffset - + prev.getStreamSize() + prev.streamPadding; - assert (compressedOffset & 3) == 0; - uncompressedOffset = prev.uncompressedOffset + prev.uncompressedSum; - } - - public int getMemoryUsage() { - return memoryUsage; - } - - public StreamFlags getStreamFlags() { - return streamFlags; - } - - public int getRecordCount() { - // It was already checked in the constructor that it fits into an int. - // Otherwise we couldn't have allocated the arrays. - return (int)recordCount; - } - - public long getUncompressedSize() { - return uncompressedSum; - } - - public long getLargestBlockSize() { - return largestBlockSize; - } - - public boolean hasUncompressedOffset(long pos) { - return pos >= uncompressedOffset - && pos < uncompressedOffset + uncompressedSum; - } - - public boolean hasRecord(int blockNumber) { - return blockNumber >= recordOffset - && blockNumber < recordOffset + recordCount; - } - - public void locateBlock(BlockInfo info, long target) { - assert target >= uncompressedOffset; - target -= uncompressedOffset; - assert target < uncompressedSum; - - int left = 0; - int right = unpadded.length - 1; - - while (left < right) { - int i = left + (right - left) / 2; - - if (uncompressed[i] <= target) - left = i + 1; - else - right = i; - } - - setBlockInfo(info, recordOffset + left); - } - - public void setBlockInfo(BlockInfo info, int blockNumber) { - // The caller has checked that the given Block number is inside - // this Index. - assert blockNumber >= recordOffset; - assert blockNumber - recordOffset < recordCount; - - info.index = this; - info.blockNumber = blockNumber; - - int pos = blockNumber - recordOffset; - - if (pos == 0) { - info.compressedOffset = 0; - info.uncompressedOffset = 0; - } else { - info.compressedOffset = (unpadded[pos - 1] + 3) & ~3; - info.uncompressedOffset = uncompressed[pos - 1]; - } - - info.unpaddedSize = unpadded[pos] - info.compressedOffset; - info.uncompressedSize = uncompressed[pos] - info.uncompressedOffset; - - info.compressedOffset += compressedOffset - + DecoderUtil.STREAM_HEADER_SIZE; - info.uncompressedOffset += uncompressedOffset; - } -} Index: trunk/src/org/tukaani/xz/index/IndexEncoder.java =================================================================== --- trunk/src/org/tukaani/xz/index/IndexEncoder.java (revision 16021) +++ (revision ) @@ -1,59 +1,0 @@ -/* - * IndexEncoder - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.index; - -import java.io.OutputStream; -import java.io.IOException; -import java.util.ArrayList; -import java.util.Iterator; -import java.util.zip.CheckedOutputStream; -import org.tukaani.xz.common.EncoderUtil; -import org.tukaani.xz.XZIOException; - -public class IndexEncoder extends IndexBase { - private final ArrayList records - = new ArrayList(); - - public IndexEncoder() { - super(new XZIOException("XZ Stream or its Index has grown too big")); - } - - public void add(long unpaddedSize, long uncompressedSize) - throws XZIOException { - super.add(unpaddedSize, uncompressedSize); - records.add(new IndexRecord(unpaddedSize, uncompressedSize)); - } - - public void encode(OutputStream out) throws IOException { - java.util.zip.CRC32 crc32 = new java.util.zip.CRC32(); - CheckedOutputStream outChecked = new CheckedOutputStream(out, crc32); - - // Index Indicator - outChecked.write(0x00); - - // Number of Records - EncoderUtil.encodeVLI(outChecked, recordCount); - - // List of Records - for (IndexRecord record : records) { - EncoderUtil.encodeVLI(outChecked, record.unpadded); - EncoderUtil.encodeVLI(outChecked, record.uncompressed); - } - - // Index Padding - for (int i = getIndexPaddingSize(); i > 0; --i) - outChecked.write(0x00); - - // CRC32 - long value = crc32.getValue(); - for (int i = 0; i < 4; ++i) - out.write((byte)(value >>> (i * 8))); - } -} Index: trunk/src/org/tukaani/xz/index/IndexHash.java =================================================================== --- trunk/src/org/tukaani/xz/index/IndexHash.java (revision 16021) +++ (revision ) @@ -1,98 +1,0 @@ -/* - * IndexHash - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.index; - -import java.io.InputStream; -import java.io.DataInputStream; -import java.io.IOException; -import java.nio.ByteBuffer; -import java.util.Arrays; -import java.util.zip.CheckedInputStream; -import org.tukaani.xz.common.DecoderUtil; -import org.tukaani.xz.XZIOException; -import org.tukaani.xz.CorruptedInputException; - -public class IndexHash extends IndexBase { - private org.tukaani.xz.check.Check hash; - - public IndexHash() { - super(new CorruptedInputException()); - - try { - hash = new org.tukaani.xz.check.SHA256(); - } catch (java.security.NoSuchAlgorithmException e) { - hash = new org.tukaani.xz.check.CRC32(); - } - } - - public void add(long unpaddedSize, long uncompressedSize) - throws XZIOException { - super.add(unpaddedSize, uncompressedSize); - - ByteBuffer buf = ByteBuffer.allocate(2 * 8); - buf.putLong(unpaddedSize); - buf.putLong(uncompressedSize); - hash.update(buf.array()); - } - - public void validate(InputStream in) throws IOException { - // Index Indicator (0x00) has already been read by BlockInputStream - // so add 0x00 to the CRC32 here. - java.util.zip.CRC32 crc32 = new java.util.zip.CRC32(); - crc32.update('\0'); - CheckedInputStream inChecked = new CheckedInputStream(in, crc32); - - // Get and validate the Number of Records field. - // If Block Header Size was corrupt and became Index Indicator, - // this error would actually be about corrupt Block Header. - // This is why the error message mentions both possibilities. - long storedRecordCount = DecoderUtil.decodeVLI(inChecked); - if (storedRecordCount != recordCount) - throw new CorruptedInputException( - "XZ Block Header or the start of XZ Index is corrupt"); - - // Decode and hash the Index field and compare it to - // the hash value calculated from the decoded Blocks. - IndexHash stored = new IndexHash(); - for (long i = 0; i < recordCount; ++i) { - long unpaddedSize = DecoderUtil.decodeVLI(inChecked); - long uncompressedSize = DecoderUtil.decodeVLI(inChecked); - - try { - stored.add(unpaddedSize, uncompressedSize); - } catch (XZIOException e) { - throw new CorruptedInputException("XZ Index is corrupt"); - } - - if (stored.blocksSum > blocksSum - || stored.uncompressedSum > uncompressedSum - || stored.indexListSize > indexListSize) - throw new CorruptedInputException("XZ Index is corrupt"); - } - - if (stored.blocksSum != blocksSum - || stored.uncompressedSum != uncompressedSum - || stored.indexListSize != indexListSize - || !Arrays.equals(stored.hash.finish(), hash.finish())) - throw new CorruptedInputException("XZ Index is corrupt"); - - // Index Padding - DataInputStream inData = new DataInputStream(inChecked); - for (int i = getIndexPaddingSize(); i > 0; --i) - if (inData.readUnsignedByte() != 0x00) - throw new CorruptedInputException("XZ Index is corrupt"); - - // CRC32 - long value = crc32.getValue(); - for (int i = 0; i < 4; ++i) - if (((value >>> (i * 8)) & 0xFF) != inData.readUnsignedByte()) - throw new CorruptedInputException("XZ Index is corrupt"); - } -} Index: trunk/src/org/tukaani/xz/index/IndexRecord.java =================================================================== --- trunk/src/org/tukaani/xz/index/IndexRecord.java (revision 16021) +++ (revision ) @@ -1,20 +1,0 @@ -/* - * IndexRecord - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.index; - -class IndexRecord { - final long unpadded; - final long uncompressed; - - IndexRecord(long unpadded, long uncompressed) { - this.unpadded = unpadded; - this.uncompressed = uncompressed; - } -} Index: trunk/src/org/tukaani/xz/lz/BT4.java =================================================================== --- trunk/src/org/tukaani/xz/lz/BT4.java (revision 16021) +++ (revision ) @@ -1,265 +1,0 @@ -/* - * Binary Tree match finder with 2-, 3-, and 4-byte hashing - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.lz; - -import org.tukaani.xz.ArrayCache; - -final class BT4 extends LZEncoder { - private final Hash234 hash; - private final int[] tree; - private final Matches matches; - private final int depthLimit; - - private final int cyclicSize; - private int cyclicPos = -1; - private int lzPos; - - static int getMemoryUsage(int dictSize) { - return Hash234.getMemoryUsage(dictSize) + dictSize / (1024 / 8) + 10; - } - - BT4(int dictSize, int beforeSizeMin, int readAheadMax, - int niceLen, int matchLenMax, int depthLimit, - ArrayCache arrayCache) { - super(dictSize, beforeSizeMin, readAheadMax, niceLen, matchLenMax, - arrayCache); - - cyclicSize = dictSize + 1; - lzPos = cyclicSize; - - hash = new Hash234(dictSize, arrayCache); - tree = arrayCache.getIntArray(cyclicSize * 2, false); - - // Substracting 1 because the shortest match that this match - // finder can find is 2 bytes, so there's no need to reserve - // space for one-byte matches. - matches = new Matches(niceLen - 1); - - this.depthLimit = depthLimit > 0 ? depthLimit : 16 + niceLen / 2; - } - - public void putArraysToCache(ArrayCache arrayCache) { - arrayCache.putArray(tree); - hash.putArraysToCache(arrayCache); - super.putArraysToCache(arrayCache); - } - - private int movePos() { - int avail = movePos(niceLen, 4); - - if (avail != 0) { - if (++lzPos == Integer.MAX_VALUE) { - int normalizationOffset = Integer.MAX_VALUE - cyclicSize; - hash.normalize(normalizationOffset); - normalize(tree, cyclicSize * 2, normalizationOffset); - lzPos -= normalizationOffset; - } - - if (++cyclicPos == cyclicSize) - cyclicPos = 0; - } - - return avail; - } - - public Matches getMatches() { - matches.count = 0; - - int matchLenLimit = matchLenMax; - int niceLenLimit = niceLen; - int avail = movePos(); - - if (avail < matchLenLimit) { - if (avail == 0) - return matches; - - matchLenLimit = avail; - if (niceLenLimit > avail) - niceLenLimit = avail; - } - - hash.calcHashes(buf, readPos); - int delta2 = lzPos - hash.getHash2Pos(); - int delta3 = lzPos - hash.getHash3Pos(); - int currentMatch = hash.getHash4Pos(); - hash.updateTables(lzPos); - - int lenBest = 0; - - // See if the hash from the first two bytes found a match. - // The hashing algorithm guarantees that if the first byte - // matches, also the second byte does, so there's no need to - // test the second byte. - if (delta2 < cyclicSize && buf[readPos - delta2] == buf[readPos]) { - lenBest = 2; - matches.len[0] = 2; - matches.dist[0] = delta2 - 1; - matches.count = 1; - } - - // See if the hash from the first three bytes found a match that - // is different from the match possibly found by the two-byte hash. - // Also here the hashing algorithm guarantees that if the first byte - // matches, also the next two bytes do. - if (delta2 != delta3 && delta3 < cyclicSize - && buf[readPos - delta3] == buf[readPos]) { - lenBest = 3; - matches.dist[matches.count++] = delta3 - 1; - delta2 = delta3; - } - - // If a match was found, see how long it is. - if (matches.count > 0) { - while (lenBest < matchLenLimit && buf[readPos + lenBest - delta2] - == buf[readPos + lenBest]) - ++lenBest; - - matches.len[matches.count - 1] = lenBest; - - // Return if it is long enough (niceLen or reached the end of - // the dictionary). - if (lenBest >= niceLenLimit) { - skip(niceLenLimit, currentMatch); - return matches; - } - } - - // Long enough match wasn't found so easily. Look for better matches - // from the binary tree. - if (lenBest < 3) - lenBest = 3; - - int depth = depthLimit; - - int ptr0 = (cyclicPos << 1) + 1; - int ptr1 = cyclicPos << 1; - int len0 = 0; - int len1 = 0; - - while (true) { - int delta = lzPos - currentMatch; - - // Return if the search depth limit has been reached or - // if the distance of the potential match exceeds the - // dictionary size. - if (depth-- == 0 || delta >= cyclicSize) { - tree[ptr0] = 0; - tree[ptr1] = 0; - return matches; - } - - int pair = (cyclicPos - delta - + (delta > cyclicPos ? cyclicSize : 0)) << 1; - int len = Math.min(len0, len1); - - if (buf[readPos + len - delta] == buf[readPos + len]) { - while (++len < matchLenLimit) - if (buf[readPos + len - delta] != buf[readPos + len]) - break; - - if (len > lenBest) { - lenBest = len; - matches.len[matches.count] = len; - matches.dist[matches.count] = delta - 1; - ++matches.count; - - if (len >= niceLenLimit) { - tree[ptr1] = tree[pair]; - tree[ptr0] = tree[pair + 1]; - return matches; - } - } - } - - if ((buf[readPos + len - delta] & 0xFF) - < (buf[readPos + len] & 0xFF)) { - tree[ptr1] = currentMatch; - ptr1 = pair + 1; - currentMatch = tree[ptr1]; - len1 = len; - } else { - tree[ptr0] = currentMatch; - ptr0 = pair; - currentMatch = tree[ptr0]; - len0 = len; - } - } - } - - private void skip(int niceLenLimit, int currentMatch) { - int depth = depthLimit; - - int ptr0 = (cyclicPos << 1) + 1; - int ptr1 = cyclicPos << 1; - int len0 = 0; - int len1 = 0; - - while (true) { - int delta = lzPos - currentMatch; - - if (depth-- == 0 || delta >= cyclicSize) { - tree[ptr0] = 0; - tree[ptr1] = 0; - return; - } - - int pair = (cyclicPos - delta - + (delta > cyclicPos ? cyclicSize : 0)) << 1; - int len = Math.min(len0, len1); - - if (buf[readPos + len - delta] == buf[readPos + len]) { - // No need to look for longer matches than niceLenLimit - // because we only are updating the tree, not returning - // matches found to the caller. - do { - if (++len == niceLenLimit) { - tree[ptr1] = tree[pair]; - tree[ptr0] = tree[pair + 1]; - return; - } - } while (buf[readPos + len - delta] == buf[readPos + len]); - } - - if ((buf[readPos + len - delta] & 0xFF) - < (buf[readPos + len] & 0xFF)) { - tree[ptr1] = currentMatch; - ptr1 = pair + 1; - currentMatch = tree[ptr1]; - len1 = len; - } else { - tree[ptr0] = currentMatch; - ptr0 = pair; - currentMatch = tree[ptr0]; - len0 = len; - } - } - } - - public void skip(int len) { - while (len-- > 0) { - int niceLenLimit = niceLen; - int avail = movePos(); - - if (avail < niceLenLimit) { - if (avail == 0) - continue; - - niceLenLimit = avail; - } - - hash.calcHashes(buf, readPos); - int currentMatch = hash.getHash4Pos(); - hash.updateTables(lzPos); - - skip(niceLenLimit, currentMatch); - } - } -} Index: trunk/src/org/tukaani/xz/lz/CRC32Hash.java =================================================================== --- trunk/src/org/tukaani/xz/lz/CRC32Hash.java (revision 16021) +++ (revision ) @@ -1,35 +1,0 @@ -/* - * CRC32Hash - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.lz; - -/** - * Provides a CRC32 table using the polynomial from IEEE 802.3. - */ -class CRC32Hash { - private static final int CRC32_POLY = 0xEDB88320; - - static final int[] crcTable = new int[256]; - - static { - for (int i = 0; i < 256; ++i) { - int r = i; - - for (int j = 0; j < 8; ++j) { - if ((r & 1) != 0) - r = (r >>> 1) ^ CRC32_POLY; - else - r >>>= 1; - } - - crcTable[i] = r; - } - } -} Index: trunk/src/org/tukaani/xz/lz/HC4.java =================================================================== --- trunk/src/org/tukaani/xz/lz/HC4.java (revision 16021) +++ (revision ) @@ -1,210 +1,0 @@ -/* - * Hash Chain match finder with 2-, 3-, and 4-byte hashing - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.lz; - -import org.tukaani.xz.ArrayCache; - -final class HC4 extends LZEncoder { - private final Hash234 hash; - private final int[] chain; - private final Matches matches; - private final int depthLimit; - - private final int cyclicSize; - private int cyclicPos = -1; - private int lzPos; - - /** - * Gets approximate memory usage of the match finder as kibibytes. - */ - static int getMemoryUsage(int dictSize) { - return Hash234.getMemoryUsage(dictSize) + dictSize / (1024 / 4) + 10; - } - - /** - * Creates a new LZEncoder with the HC4 match finder. - * See LZEncoder.getInstance for parameter descriptions. - */ - HC4(int dictSize, int beforeSizeMin, int readAheadMax, - int niceLen, int matchLenMax, int depthLimit, - ArrayCache arrayCache) { - super(dictSize, beforeSizeMin, readAheadMax, niceLen, matchLenMax, - arrayCache); - - hash = new Hash234(dictSize, arrayCache); - - // +1 because we need dictSize bytes of history + the current byte. - cyclicSize = dictSize + 1; - chain = arrayCache.getIntArray(cyclicSize, false); - lzPos = cyclicSize; - - // Substracting 1 because the shortest match that this match - // finder can find is 2 bytes, so there's no need to reserve - // space for one-byte matches. - matches = new Matches(niceLen - 1); - - // Use a default depth limit if no other value was specified. - // The default is just something based on experimentation; - // it's nothing magic. - this.depthLimit = (depthLimit > 0) ? depthLimit : 4 + niceLen / 4; - } - - public void putArraysToCache(ArrayCache arrayCache) { - arrayCache.putArray(chain); - hash.putArraysToCache(arrayCache); - super.putArraysToCache(arrayCache); - } - - /** - * Moves to the next byte, checks that there is enough available space, - * and possibly normalizes the hash tables and the hash chain. - * - * @return number of bytes available, including the current byte - */ - private int movePos() { - int avail = movePos(4, 4); - - if (avail != 0) { - if (++lzPos == Integer.MAX_VALUE) { - int normalizationOffset = Integer.MAX_VALUE - cyclicSize; - hash.normalize(normalizationOffset); - normalize(chain, cyclicSize, normalizationOffset); - lzPos -= normalizationOffset; - } - - if (++cyclicPos == cyclicSize) - cyclicPos = 0; - } - - return avail; - } - - public Matches getMatches() { - matches.count = 0; - int matchLenLimit = matchLenMax; - int niceLenLimit = niceLen; - int avail = movePos(); - - if (avail < matchLenLimit) { - if (avail == 0) - return matches; - - matchLenLimit = avail; - if (niceLenLimit > avail) - niceLenLimit = avail; - } - - hash.calcHashes(buf, readPos); - int delta2 = lzPos - hash.getHash2Pos(); - int delta3 = lzPos - hash.getHash3Pos(); - int currentMatch = hash.getHash4Pos(); - hash.updateTables(lzPos); - - chain[cyclicPos] = currentMatch; - - int lenBest = 0; - - // See if the hash from the first two bytes found a match. - // The hashing algorithm guarantees that if the first byte - // matches, also the second byte does, so there's no need to - // test the second byte. - if (delta2 < cyclicSize && buf[readPos - delta2] == buf[readPos]) { - lenBest = 2; - matches.len[0] = 2; - matches.dist[0] = delta2 - 1; - matches.count = 1; - } - - // See if the hash from the first three bytes found a match that - // is different from the match possibly found by the two-byte hash. - // Also here the hashing algorithm guarantees that if the first byte - // matches, also the next two bytes do. - if (delta2 != delta3 && delta3 < cyclicSize - && buf[readPos - delta3] == buf[readPos]) { - lenBest = 3; - matches.dist[matches.count++] = delta3 - 1; - delta2 = delta3; - } - - // If a match was found, see how long it is. - if (matches.count > 0) { - while (lenBest < matchLenLimit && buf[readPos + lenBest - delta2] - == buf[readPos + lenBest]) - ++lenBest; - - matches.len[matches.count - 1] = lenBest; - - // Return if it is long enough (niceLen or reached the end of - // the dictionary). - if (lenBest >= niceLenLimit) - return matches; - } - - // Long enough match wasn't found so easily. Look for better matches - // from the hash chain. - if (lenBest < 3) - lenBest = 3; - - int depth = depthLimit; - - while (true) { - int delta = lzPos - currentMatch; - - // Return if the search depth limit has been reached or - // if the distance of the potential match exceeds the - // dictionary size. - if (depth-- == 0 || delta >= cyclicSize) - return matches; - - currentMatch = chain[cyclicPos - delta - + (delta > cyclicPos ? cyclicSize : 0)]; - - // Test the first byte and the first new byte that would give us - // a match that is at least one byte longer than lenBest. This - // too short matches get quickly skipped. - if (buf[readPos + lenBest - delta] == buf[readPos + lenBest] - && buf[readPos - delta] == buf[readPos]) { - // Calculate the length of the match. - int len = 0; - while (++len < matchLenLimit) - if (buf[readPos + len - delta] != buf[readPos + len]) - break; - - // Use the match if and only if it is better than the longest - // match found so far. - if (len > lenBest) { - lenBest = len; - matches.len[matches.count] = len; - matches.dist[matches.count] = delta - 1; - ++matches.count; - - // Return if it is long enough (niceLen or reached the - // end of the dictionary). - if (len >= niceLenLimit) - return matches; - } - } - } - } - - public void skip(int len) { - assert len >= 0; - - while (len-- > 0) { - if (movePos() != 0) { - // Update the hash chain and hash tables. - hash.calcHashes(buf, readPos); - chain[cyclicPos] = hash.getHash4Pos(); - hash.updateTables(lzPos); - } - } - } -} Index: trunk/src/org/tukaani/xz/lz/Hash234.java =================================================================== --- trunk/src/org/tukaani/xz/lz/Hash234.java (revision 16021) +++ (revision ) @@ -1,102 +1,0 @@ -/* - * 2-, 3-, and 4-byte hashing - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.lz; - -import org.tukaani.xz.ArrayCache; - -final class Hash234 extends CRC32Hash { - private static final int HASH_2_SIZE = 1 << 10; - private static final int HASH_2_MASK = HASH_2_SIZE - 1; - - private static final int HASH_3_SIZE = 1 << 16; - private static final int HASH_3_MASK = HASH_3_SIZE - 1; - - private final int hash4Mask; - - private final int[] hash2Table; - private final int[] hash3Table; - private final int[] hash4Table; - private final int hash4Size; - - private int hash2Value = 0; - private int hash3Value = 0; - private int hash4Value = 0; - - static int getHash4Size(int dictSize) { - int h = dictSize - 1; - h |= h >>> 1; - h |= h >>> 2; - h |= h >>> 4; - h |= h >>> 8; - h >>>= 1; - h |= 0xFFFF; - if (h > (1 << 24)) - h >>>= 1; - - return h + 1; - } - - static int getMemoryUsage(int dictSize) { - // Sizes of the hash arrays + a little extra - return (HASH_2_SIZE + HASH_3_SIZE + getHash4Size(dictSize)) - / (1024 / 4) + 4; - } - - Hash234(int dictSize, ArrayCache arrayCache) { - hash2Table = arrayCache.getIntArray(HASH_2_SIZE, true); - hash3Table = arrayCache.getIntArray(HASH_3_SIZE, true); - - hash4Size = getHash4Size(dictSize); - hash4Table = arrayCache.getIntArray(hash4Size, true); - hash4Mask = hash4Size - 1; - } - - void putArraysToCache(ArrayCache arrayCache) { - arrayCache.putArray(hash4Table); - arrayCache.putArray(hash3Table); - arrayCache.putArray(hash2Table); - } - - void calcHashes(byte[] buf, int off) { - int temp = crcTable[buf[off] & 0xFF] ^ (buf[off + 1] & 0xFF); - hash2Value = temp & HASH_2_MASK; - - temp ^= (buf[off + 2] & 0xFF) << 8; - hash3Value = temp & HASH_3_MASK; - - temp ^= crcTable[buf[off + 3] & 0xFF] << 5; - hash4Value = temp & hash4Mask; - } - - int getHash2Pos() { - return hash2Table[hash2Value]; - } - - int getHash3Pos() { - return hash3Table[hash3Value]; - } - - int getHash4Pos() { - return hash4Table[hash4Value]; - } - - void updateTables(int pos) { - hash2Table[hash2Value] = pos; - hash3Table[hash3Value] = pos; - hash4Table[hash4Value] = pos; - } - - void normalize(int normalizeOffset) { - LZEncoder.normalize(hash2Table, HASH_2_SIZE, normalizeOffset); - LZEncoder.normalize(hash3Table, HASH_3_SIZE, normalizeOffset); - LZEncoder.normalize(hash4Table, hash4Size, normalizeOffset); - } -} Index: trunk/src/org/tukaani/xz/lz/LZDecoder.java =================================================================== --- trunk/src/org/tukaani/xz/lz/LZDecoder.java (revision 16021) +++ (revision ) @@ -1,133 +1,0 @@ -/* - * LZDecoder - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.lz; - -import java.io.DataInputStream; -import java.io.IOException; -import org.tukaani.xz.ArrayCache; -import org.tukaani.xz.CorruptedInputException; - -public final class LZDecoder { - private final byte[] buf; - private final int bufSize; // To avoid buf.length with an array-cached buf. - private int start = 0; - private int pos = 0; - private int full = 0; - private int limit = 0; - private int pendingLen = 0; - private int pendingDist = 0; - - public LZDecoder(int dictSize, byte[] presetDict, ArrayCache arrayCache) { - bufSize = dictSize; - buf = arrayCache.getByteArray(bufSize, false); - - if (presetDict != null) { - pos = Math.min(presetDict.length, dictSize); - full = pos; - start = pos; - System.arraycopy(presetDict, presetDict.length - pos, buf, 0, pos); - } - } - - public void putArraysToCache(ArrayCache arrayCache) { - arrayCache.putArray(buf); - } - - public void reset() { - start = 0; - pos = 0; - full = 0; - limit = 0; - buf[bufSize - 1] = 0x00; - } - - public void setLimit(int outMax) { - if (bufSize - pos <= outMax) - limit = bufSize; - else - limit = pos + outMax; - } - - public boolean hasSpace() { - return pos < limit; - } - - public boolean hasPending() { - return pendingLen > 0; - } - - public int getPos() { - return pos; - } - - public int getByte(int dist) { - int offset = pos - dist - 1; - if (dist >= pos) - offset += bufSize; - - return buf[offset] & 0xFF; - } - - public void putByte(byte b) { - buf[pos++] = b; - - if (full < pos) - full = pos; - } - - public void repeat(int dist, int len) throws IOException { - if (dist < 0 || dist >= full) - throw new CorruptedInputException(); - - int left = Math.min(limit - pos, len); - pendingLen = len - left; - pendingDist = dist; - - int back = pos - dist - 1; - if (dist >= pos) - back += bufSize; - - do { - buf[pos++] = buf[back++]; - if (back == bufSize) - back = 0; - } while (--left > 0); - - if (full < pos) - full = pos; - } - - public void repeatPending() throws IOException { - if (pendingLen > 0) - repeat(pendingDist, pendingLen); - } - - public void copyUncompressed(DataInputStream inData, int len) - throws IOException { - int copySize = Math.min(bufSize - pos, len); - inData.readFully(buf, pos, copySize); - pos += copySize; - - if (full < pos) - full = pos; - } - - public int flush(byte[] out, int outOff) { - int copySize = pos - start; - if (pos == bufSize) - pos = 0; - - System.arraycopy(buf, start, out, outOff, copySize); - start = pos; - - return copySize; - } -} Index: trunk/src/org/tukaani/xz/lz/LZEncoder.java =================================================================== --- trunk/src/org/tukaani/xz/lz/LZEncoder.java (revision 16021) +++ (revision ) @@ -1,428 +1,0 @@ -/* - * LZEncoder - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.lz; - -import java.io.OutputStream; -import java.io.IOException; -import org.tukaani.xz.ArrayCache; - -public abstract class LZEncoder { - public static final int MF_HC4 = 0x04; - public static final int MF_BT4 = 0x14; - - /** - * Number of bytes to keep available before the current byte - * when moving the LZ window. - */ - private final int keepSizeBefore; - - /** - * Number of bytes that must be available, the current byte included, - * to make hasEnoughData return true. Flushing and finishing are - * naturally exceptions to this since there cannot be any data after - * the end of the uncompressed input. - */ - private final int keepSizeAfter; - - final int matchLenMax; - final int niceLen; - - final byte[] buf; - final int bufSize; // To avoid buf.length with an array-cached buf. - - int readPos = -1; - private int readLimit = -1; - private boolean finishing = false; - private int writePos = 0; - private int pendingSize = 0; - - static void normalize(int[] positions, int positionsCount, - int normalizationOffset) { - for (int i = 0; i < positionsCount; ++i) { - if (positions[i] <= normalizationOffset) - positions[i] = 0; - else - positions[i] -= normalizationOffset; - } - } - - /** - * Gets the size of the LZ window buffer that needs to be allocated. - */ - private static int getBufSize( - int dictSize, int extraSizeBefore, int extraSizeAfter, - int matchLenMax) { - int keepSizeBefore = extraSizeBefore + dictSize; - int keepSizeAfter = extraSizeAfter + matchLenMax; - int reserveSize = Math.min(dictSize / 2 + (256 << 10), 512 << 20); - return keepSizeBefore + keepSizeAfter + reserveSize; - } - - /** - * Gets approximate memory usage of the LZEncoder base structure and - * the match finder as kibibytes. - */ - public static int getMemoryUsage( - int dictSize, int extraSizeBefore, int extraSizeAfter, - int matchLenMax, int mf) { - // Buffer size + a little extra - int m = getBufSize(dictSize, extraSizeBefore, extraSizeAfter, - matchLenMax) / 1024 + 10; - - switch (mf) { - case MF_HC4: - m += HC4.getMemoryUsage(dictSize); - break; - - case MF_BT4: - m += BT4.getMemoryUsage(dictSize); - break; - - default: - throw new IllegalArgumentException(); - } - - return m; - } - - /** - * Creates a new LZEncoder. - *

- * @param dictSize dictionary size - * - * @param extraSizeBefore - * number of bytes to keep available in the - * history in addition to dictSize - * - * @param extraSizeAfter - * number of bytes that must be available - * after current position + matchLenMax - * - * @param niceLen if a match of at least niceLen - * bytes is found, be happy with it and don't - * stop looking for longer matches - * - * @param matchLenMax don't test for matches longer than - * matchLenMax bytes - * - * @param mf match finder ID - * - * @param depthLimit match finder search depth limit - */ - public static LZEncoder getInstance( - int dictSize, int extraSizeBefore, int extraSizeAfter, - int niceLen, int matchLenMax, int mf, int depthLimit, - ArrayCache arrayCache) { - switch (mf) { - case MF_HC4: - return new HC4(dictSize, extraSizeBefore, extraSizeAfter, - niceLen, matchLenMax, depthLimit, arrayCache); - - case MF_BT4: - return new BT4(dictSize, extraSizeBefore, extraSizeAfter, - niceLen, matchLenMax, depthLimit, arrayCache); - } - - throw new IllegalArgumentException(); - } - - /** - * Creates a new LZEncoder. See getInstance. - */ - LZEncoder(int dictSize, int extraSizeBefore, int extraSizeAfter, - int niceLen, int matchLenMax, ArrayCache arrayCache) { - bufSize = getBufSize(dictSize, extraSizeBefore, extraSizeAfter, - matchLenMax); - buf = arrayCache.getByteArray(bufSize, false); - - keepSizeBefore = extraSizeBefore + dictSize; - keepSizeAfter = extraSizeAfter + matchLenMax; - - this.matchLenMax = matchLenMax; - this.niceLen = niceLen; - } - - public void putArraysToCache(ArrayCache arrayCache) { - arrayCache.putArray(buf); - } - - /** - * Sets a preset dictionary. If a preset dictionary is wanted, this - * function must be called immediately after creating the LZEncoder - * before any data has been encoded. - */ - public void setPresetDict(int dictSize, byte[] presetDict) { - assert !isStarted(); - assert writePos == 0; - - if (presetDict != null) { - // If the preset dictionary buffer is bigger than the dictionary - // size, copy only the tail of the preset dictionary. - int copySize = Math.min(presetDict.length, dictSize); - int offset = presetDict.length - copySize; - System.arraycopy(presetDict, offset, buf, 0, copySize); - writePos += copySize; - skip(copySize); - } - } - - /** - * Moves data from the end of the buffer to the beginning, discarding - * old data and making space for new input. - */ - private void moveWindow() { - // Align the move to a multiple of 16 bytes. LZMA2 needs this - // because it uses the lowest bits from readPos to get the - // alignment of the uncompressed data. - int moveOffset = (readPos + 1 - keepSizeBefore) & ~15; - int moveSize = writePos - moveOffset; - System.arraycopy(buf, moveOffset, buf, 0, moveSize); - - readPos -= moveOffset; - readLimit -= moveOffset; - writePos -= moveOffset; - } - - /** - * Copies new data into the LZEncoder's buffer. - */ - public int fillWindow(byte[] in, int off, int len) { - assert !finishing; - - // Move the sliding window if needed. - if (readPos >= bufSize - keepSizeAfter) - moveWindow(); - - // Try to fill the dictionary buffer. If it becomes full, - // some of the input bytes may be left unused. - if (len > bufSize - writePos) - len = bufSize - writePos; - - System.arraycopy(in, off, buf, writePos, len); - writePos += len; - - // Set the new readLimit but only if there's enough data to allow - // encoding of at least one more byte. - if (writePos >= keepSizeAfter) - readLimit = writePos - keepSizeAfter; - - processPendingBytes(); - - // Tell the caller how much input we actually copied into - // the dictionary. - return len; - } - - /** - * Process pending bytes remaining from preset dictionary initialization - * or encoder flush operation. - */ - private void processPendingBytes() { - // After flushing or setting a preset dictionary there will be - // pending data that hasn't been ran through the match finder yet. - // Run it through the match finder now if there is enough new data - // available (readPos < readLimit) that the encoder may encode at - // least one more input byte. This way we don't waste any time - // looping in the match finder (and marking the same bytes as - // pending again) if the application provides very little new data - // per write call. - if (pendingSize > 0 && readPos < readLimit) { - readPos -= pendingSize; - int oldPendingSize = pendingSize; - pendingSize = 0; - skip(oldPendingSize); - assert pendingSize < oldPendingSize; - } - } - - /** - * Returns true if at least one byte has already been run through - * the match finder. - */ - public boolean isStarted() { - return readPos != -1; - } - - /** - * Marks that all the input needs to be made available in - * the encoded output. - */ - public void setFlushing() { - readLimit = writePos - 1; - processPendingBytes(); - } - - /** - * Marks that there is no more input remaining. The read position - * can be advanced until the end of the data. - */ - public void setFinishing() { - readLimit = writePos - 1; - finishing = true; - processPendingBytes(); - } - - /** - * Tests if there is enough input available to let the caller encode - * at least one more byte. - */ - public boolean hasEnoughData(int alreadyReadLen) { - return readPos - alreadyReadLen < readLimit; - } - - public void copyUncompressed(OutputStream out, int backward, int len) - throws IOException { - out.write(buf, readPos + 1 - backward, len); - } - - /** - * Get the number of bytes available, including the current byte. - *

- * Note that the result is undefined if getMatches or - * skip hasn't been called yet and no preset dictionary - * is being used. - */ - public int getAvail() { - assert isStarted(); - return writePos - readPos; - } - - /** - * Gets the lowest four bits of the absolute offset of the current byte. - * Bits other than the lowest four are undefined. - */ - public int getPos() { - return readPos; - } - - /** - * Gets the byte from the given backward offset. - *

- * The current byte is at 0, the previous byte - * at 1 etc. To get a byte at zero-based distance, - * use getByte(dist + 1). - *

- * This function is equivalent to getByte(0, backward). - */ - public int getByte(int backward) { - return buf[readPos - backward] & 0xFF; - } - - /** - * Gets the byte from the given forward minus backward offset. - * The forward offset is added to the current position. This lets - * one read bytes ahead of the current byte. - */ - public int getByte(int forward, int backward) { - return buf[readPos + forward - backward] & 0xFF; - } - - /** - * Get the length of a match at the given distance. - * - * @param dist zero-based distance of the match to test - * @param lenLimit don't test for a match longer than this - * - * @return length of the match; it is in the range [0, lenLimit] - */ - public int getMatchLen(int dist, int lenLimit) { - int backPos = readPos - dist - 1; - int len = 0; - - while (len < lenLimit && buf[readPos + len] == buf[backPos + len]) - ++len; - - return len; - } - - /** - * Get the length of a match at the given distance and forward offset. - * - * @param forward forward offset - * @param dist zero-based distance of the match to test - * @param lenLimit don't test for a match longer than this - * - * @return length of the match; it is in the range [0, lenLimit] - */ - public int getMatchLen(int forward, int dist, int lenLimit) { - int curPos = readPos + forward; - int backPos = curPos - dist - 1; - int len = 0; - - while (len < lenLimit && buf[curPos + len] == buf[backPos + len]) - ++len; - - return len; - } - - /** - * Verifies that the matches returned by the match finder are valid. - * This is meant to be used in an assert statement. This is totally - * useless for actual encoding since match finder's results should - * naturally always be valid if it isn't broken. - * - * @param matches return value from getMatches - * - * @return true if matches are valid, false if match finder is broken - */ - public boolean verifyMatches(Matches matches) { - int lenLimit = Math.min(getAvail(), matchLenMax); - - for (int i = 0; i < matches.count; ++i) - if (getMatchLen(matches.dist[i], lenLimit) != matches.len[i]) - return false; - - return true; - } - - /** - * Moves to the next byte, checks if there is enough input available, - * and returns the amount of input available. - * - * @param requiredForFlushing - * minimum number of available bytes when - * flushing; encoding may be continued with - * new input after flushing - * @param requiredForFinishing - * minimum number of available bytes when - * finishing; encoding must not be continued - * after finishing or the match finder state - * may be corrupt - * - * @return the number of bytes available or zero if there - * is not enough input available - */ - int movePos(int requiredForFlushing, int requiredForFinishing) { - assert requiredForFlushing >= requiredForFinishing; - - ++readPos; - int avail = writePos - readPos; - - if (avail < requiredForFlushing) { - if (avail < requiredForFinishing || !finishing) { - ++pendingSize; - avail = 0; - } - } - - return avail; - } - - /** - * Runs match finder for the next byte and returns the matches found. - */ - public abstract Matches getMatches(); - - /** - * Skips the given number of bytes in the match finder. - */ - public abstract void skip(int len); -} Index: trunk/src/org/tukaani/xz/lz/Matches.java =================================================================== --- trunk/src/org/tukaani/xz/lz/Matches.java (revision 16021) +++ (revision ) @@ -1,22 +1,0 @@ -/* - * Matches - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.lz; - -public final class Matches { - public final int[] len; - public final int[] dist; - public int count = 0; - - Matches(int countMax) { - len = new int[countMax]; - dist = new int[countMax]; - } -} Index: trunk/src/org/tukaani/xz/lzma/LZMACoder.java =================================================================== --- trunk/src/org/tukaani/xz/lzma/LZMACoder.java (revision 16021) +++ (revision ) @@ -1,140 +1,0 @@ -/* - * LZMACoder - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.lzma; - -import org.tukaani.xz.rangecoder.RangeCoder; - -abstract class LZMACoder { - static final int POS_STATES_MAX = 1 << 4; - - static final int MATCH_LEN_MIN = 2; - static final int MATCH_LEN_MAX = MATCH_LEN_MIN + LengthCoder.LOW_SYMBOLS - + LengthCoder.MID_SYMBOLS - + LengthCoder.HIGH_SYMBOLS - 1; - - static final int DIST_STATES = 4; - static final int DIST_SLOTS = 1 << 6; - static final int DIST_MODEL_START = 4; - static final int DIST_MODEL_END = 14; - static final int FULL_DISTANCES = 1 << (DIST_MODEL_END / 2); - - static final int ALIGN_BITS = 4; - static final int ALIGN_SIZE = 1 << ALIGN_BITS; - static final int ALIGN_MASK = ALIGN_SIZE - 1; - - static final int REPS = 4; - - final int posMask; - - final int[] reps = new int[REPS]; - final State state = new State(); - - final short[][] isMatch = new short[State.STATES][POS_STATES_MAX]; - final short[] isRep = new short[State.STATES]; - final short[] isRep0 = new short[State.STATES]; - final short[] isRep1 = new short[State.STATES]; - final short[] isRep2 = new short[State.STATES]; - final short[][] isRep0Long = new short[State.STATES][POS_STATES_MAX]; - final short[][] distSlots = new short[DIST_STATES][DIST_SLOTS]; - final short[][] distSpecial = { new short[2], new short[2], - new short[4], new short[4], - new short[8], new short[8], - new short[16], new short[16], - new short[32], new short[32] }; - final short[] distAlign = new short[ALIGN_SIZE]; - - static final int getDistState(int len) { - return len < DIST_STATES + MATCH_LEN_MIN - ? len - MATCH_LEN_MIN - : DIST_STATES - 1; - } - - LZMACoder(int pb) { - posMask = (1 << pb) - 1; - } - - void reset() { - reps[0] = 0; - reps[1] = 0; - reps[2] = 0; - reps[3] = 0; - state.reset(); - - for (int i = 0; i < isMatch.length; ++i) - RangeCoder.initProbs(isMatch[i]); - - RangeCoder.initProbs(isRep); - RangeCoder.initProbs(isRep0); - RangeCoder.initProbs(isRep1); - RangeCoder.initProbs(isRep2); - - for (int i = 0; i < isRep0Long.length; ++i) - RangeCoder.initProbs(isRep0Long[i]); - - for (int i = 0; i < distSlots.length; ++i) - RangeCoder.initProbs(distSlots[i]); - - for (int i = 0; i < distSpecial.length; ++i) - RangeCoder.initProbs(distSpecial[i]); - - RangeCoder.initProbs(distAlign); - } - - - abstract class LiteralCoder { - private final int lc; - private final int literalPosMask; - - LiteralCoder(int lc, int lp) { - this.lc = lc; - this.literalPosMask = (1 << lp) - 1; - } - - final int getSubcoderIndex(int prevByte, int pos) { - int low = prevByte >> (8 - lc); - int high = (pos & literalPosMask) << lc; - return low + high; - } - - - abstract class LiteralSubcoder { - final short[] probs = new short[0x300]; - - void reset() { - RangeCoder.initProbs(probs); - } - } - } - - - abstract class LengthCoder { - static final int LOW_SYMBOLS = 1 << 3; - static final int MID_SYMBOLS = 1 << 3; - static final int HIGH_SYMBOLS = 1 << 8; - - final short[] choice = new short[2]; - final short[][] low = new short[POS_STATES_MAX][LOW_SYMBOLS]; - final short[][] mid = new short[POS_STATES_MAX][MID_SYMBOLS]; - final short[] high = new short[HIGH_SYMBOLS]; - - void reset() { - RangeCoder.initProbs(choice); - - for (int i = 0; i < low.length; ++i) - RangeCoder.initProbs(low[i]); - - for (int i = 0; i < low.length; ++i) - RangeCoder.initProbs(mid[i]); - - RangeCoder.initProbs(high); - } - } -} Index: trunk/src/org/tukaani/xz/lzma/LZMADecoder.java =================================================================== --- trunk/src/org/tukaani/xz/lzma/LZMADecoder.java (revision 16021) +++ (revision ) @@ -1,199 +1,0 @@ -/* - * LZMADecoder - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.lzma; - -import java.io.IOException; -import org.tukaani.xz.lz.LZDecoder; -import org.tukaani.xz.rangecoder.RangeDecoder; - -public final class LZMADecoder extends LZMACoder { - private final LZDecoder lz; - private final RangeDecoder rc; - private final LiteralDecoder literalDecoder; - private final LengthDecoder matchLenDecoder = new LengthDecoder(); - private final LengthDecoder repLenDecoder = new LengthDecoder(); - - public LZMADecoder(LZDecoder lz, RangeDecoder rc, int lc, int lp, int pb) { - super(pb); - this.lz = lz; - this.rc = rc; - this.literalDecoder = new LiteralDecoder(lc, lp); - reset(); - } - - public void reset() { - super.reset(); - literalDecoder.reset(); - matchLenDecoder.reset(); - repLenDecoder.reset(); - } - - /** - * Returns true if LZMA end marker was detected. It is encoded as - * the maximum match distance which with signed ints becomes -1. This - * function is needed only for LZMA1. LZMA2 doesn't use the end marker - * in the LZMA layer. - */ - public boolean endMarkerDetected() { - return reps[0] == -1; - } - - public void decode() throws IOException { - lz.repeatPending(); - - while (lz.hasSpace()) { - int posState = lz.getPos() & posMask; - - if (rc.decodeBit(isMatch[state.get()], posState) == 0) { - literalDecoder.decode(); - } else { - int len = rc.decodeBit(isRep, state.get()) == 0 - ? decodeMatch(posState) - : decodeRepMatch(posState); - - // NOTE: With LZMA1 streams that have the end marker, - // this will throw CorruptedInputException. LZMAInputStream - // handles it specially. - lz.repeat(reps[0], len); - } - } - - rc.normalize(); - } - - private int decodeMatch(int posState) throws IOException { - state.updateMatch(); - - reps[3] = reps[2]; - reps[2] = reps[1]; - reps[1] = reps[0]; - - int len = matchLenDecoder.decode(posState); - int distSlot = rc.decodeBitTree(distSlots[getDistState(len)]); - - if (distSlot < DIST_MODEL_START) { - reps[0] = distSlot; - } else { - int limit = (distSlot >> 1) - 1; - reps[0] = (2 | (distSlot & 1)) << limit; - - if (distSlot < DIST_MODEL_END) { - reps[0] |= rc.decodeReverseBitTree( - distSpecial[distSlot - DIST_MODEL_START]); - } else { - reps[0] |= rc.decodeDirectBits(limit - ALIGN_BITS) - << ALIGN_BITS; - reps[0] |= rc.decodeReverseBitTree(distAlign); - } - } - - return len; - } - - private int decodeRepMatch(int posState) throws IOException { - if (rc.decodeBit(isRep0, state.get()) == 0) { - if (rc.decodeBit(isRep0Long[state.get()], posState) == 0) { - state.updateShortRep(); - return 1; - } - } else { - int tmp; - - if (rc.decodeBit(isRep1, state.get()) == 0) { - tmp = reps[1]; - } else { - if (rc.decodeBit(isRep2, state.get()) == 0) { - tmp = reps[2]; - } else { - tmp = reps[3]; - reps[3] = reps[2]; - } - - reps[2] = reps[1]; - } - - reps[1] = reps[0]; - reps[0] = tmp; - } - - state.updateLongRep(); - - return repLenDecoder.decode(posState); - } - - - private class LiteralDecoder extends LiteralCoder { - private final LiteralSubdecoder[] subdecoders; - - LiteralDecoder(int lc, int lp) { - super(lc, lp); - - subdecoders = new LiteralSubdecoder[1 << (lc + lp)]; - for (int i = 0; i < subdecoders.length; ++i) - subdecoders[i] = new LiteralSubdecoder(); - } - - void reset() { - for (int i = 0; i < subdecoders.length; ++i) - subdecoders[i].reset(); - } - - void decode() throws IOException { - int i = getSubcoderIndex(lz.getByte(0), lz.getPos()); - subdecoders[i].decode(); - } - - - private class LiteralSubdecoder extends LiteralSubcoder { - void decode() throws IOException { - int symbol = 1; - - if (state.isLiteral()) { - do { - symbol = (symbol << 1) | rc.decodeBit(probs, symbol); - } while (symbol < 0x100); - - } else { - int matchByte = lz.getByte(reps[0]); - int offset = 0x100; - int matchBit; - int bit; - - do { - matchByte <<= 1; - matchBit = matchByte & offset; - bit = rc.decodeBit(probs, offset + matchBit + symbol); - symbol = (symbol << 1) | bit; - offset &= (0 - bit) ^ ~matchBit; - } while (symbol < 0x100); - } - - lz.putByte((byte)symbol); - state.updateLiteral(); - } - } - } - - - private class LengthDecoder extends LengthCoder { - int decode(int posState) throws IOException { - if (rc.decodeBit(choice, 0) == 0) - return rc.decodeBitTree(low[posState]) + MATCH_LEN_MIN; - - if (rc.decodeBit(choice, 1) == 0) - return rc.decodeBitTree(mid[posState]) - + MATCH_LEN_MIN + LOW_SYMBOLS; - - return rc.decodeBitTree(high) - + MATCH_LEN_MIN + LOW_SYMBOLS + MID_SYMBOLS; - } - } -} Index: trunk/src/org/tukaani/xz/lzma/LZMAEncoder.java =================================================================== --- trunk/src/org/tukaani/xz/lzma/LZMAEncoder.java (revision 16021) +++ (revision ) @@ -1,750 +1,0 @@ -/* - * LZMAEncoder - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.lzma; - -import java.io.IOException; -import org.tukaani.xz.ArrayCache; -import org.tukaani.xz.lz.LZEncoder; -import org.tukaani.xz.lz.Matches; -import org.tukaani.xz.rangecoder.RangeEncoder; - -public abstract class LZMAEncoder extends LZMACoder { - public static final int MODE_FAST = 1; - public static final int MODE_NORMAL = 2; - - /** - * LZMA2 chunk is considered full when its uncompressed size exceeds - * LZMA2_UNCOMPRESSED_LIMIT. - *

- * A compressed LZMA2 chunk can hold 2 MiB of uncompressed data. - * A single LZMA symbol may indicate up to MATCH_LEN_MAX bytes - * of data, so the LZMA2 chunk is considered full when there is - * less space than MATCH_LEN_MAX bytes. - */ - private static final int LZMA2_UNCOMPRESSED_LIMIT - = (2 << 20) - MATCH_LEN_MAX; - - /** - * LZMA2 chunk is considered full when its compressed size exceeds - * LZMA2_COMPRESSED_LIMIT. - *

- * The maximum compressed size of a LZMA2 chunk is 64 KiB. - * A single LZMA symbol might use 20 bytes of space even though - * it usually takes just one byte or so. Two more bytes are needed - * for LZMA2 uncompressed chunks (see LZMA2OutputStream.writeChunk). - * Leave a little safety margin and use 26 bytes. - */ - private static final int LZMA2_COMPRESSED_LIMIT = (64 << 10) - 26; - - private static final int DIST_PRICE_UPDATE_INTERVAL = FULL_DISTANCES; - private static final int ALIGN_PRICE_UPDATE_INTERVAL = ALIGN_SIZE; - - private final RangeEncoder rc; - final LZEncoder lz; - final LiteralEncoder literalEncoder; - final LengthEncoder matchLenEncoder; - final LengthEncoder repLenEncoder; - final int niceLen; - - private int distPriceCount = 0; - private int alignPriceCount = 0; - - private final int distSlotPricesSize; - private final int[][] distSlotPrices; - private final int[][] fullDistPrices - = new int[DIST_STATES][FULL_DISTANCES]; - private final int[] alignPrices = new int[ALIGN_SIZE]; - - int back = 0; - int readAhead = -1; - private int uncompressedSize = 0; - - public static int getMemoryUsage(int mode, int dictSize, - int extraSizeBefore, int mf) { - int m = 80; - - switch (mode) { - case MODE_FAST: - m += LZMAEncoderFast.getMemoryUsage( - dictSize, extraSizeBefore, mf); - break; - - case MODE_NORMAL: - m += LZMAEncoderNormal.getMemoryUsage( - dictSize, extraSizeBefore, mf); - break; - - default: - throw new IllegalArgumentException(); - } - - return m; - } - - public static LZMAEncoder getInstance( - RangeEncoder rc, int lc, int lp, int pb, int mode, - int dictSize, int extraSizeBefore, - int niceLen, int mf, int depthLimit, - ArrayCache arrayCache) { - switch (mode) { - case MODE_FAST: - return new LZMAEncoderFast(rc, lc, lp, pb, - dictSize, extraSizeBefore, - niceLen, mf, depthLimit, - arrayCache); - - case MODE_NORMAL: - return new LZMAEncoderNormal(rc, lc, lp, pb, - dictSize, extraSizeBefore, - niceLen, mf, depthLimit, - arrayCache); - } - - throw new IllegalArgumentException(); - } - - public void putArraysToCache(ArrayCache arrayCache) { - lz.putArraysToCache(arrayCache); - } - - /** - * Gets an integer [0, 63] matching the highest two bits of an integer. - * This is like bit scan reverse (BSR) on x86 except that this also - * cares about the second highest bit. - */ - public static int getDistSlot(int dist) { - if (dist <= DIST_MODEL_START && dist >= 0) - return dist; - - int n = dist; - int i = 31; - - if ((n & 0xFFFF0000) == 0) { - n <<= 16; - i = 15; - } - - if ((n & 0xFF000000) == 0) { - n <<= 8; - i -= 8; - } - - if ((n & 0xF0000000) == 0) { - n <<= 4; - i -= 4; - } - - if ((n & 0xC0000000) == 0) { - n <<= 2; - i -= 2; - } - - if ((n & 0x80000000) == 0) - --i; - - return (i << 1) + ((dist >>> (i - 1)) & 1); - } - - /** - * Gets the next LZMA symbol. - *

- * There are three types of symbols: literal (a single byte), - * repeated match, and normal match. The symbol is indicated - * by the return value and by the variable back. - *

- * Literal: back == -1 and return value is 1. - * The literal itself needs to be read from lz separately. - *

- * Repeated match: back is in the range [0, 3] and - * the return value is the length of the repeated match. - *

- * Normal match: back - REPS (back - 4) - * is the distance of the match and the return value is the length - * of the match. - */ - abstract int getNextSymbol(); - - LZMAEncoder(RangeEncoder rc, LZEncoder lz, - int lc, int lp, int pb, int dictSize, int niceLen) { - super(pb); - this.rc = rc; - this.lz = lz; - this.niceLen = niceLen; - - literalEncoder = new LiteralEncoder(lc, lp); - matchLenEncoder = new LengthEncoder(pb, niceLen); - repLenEncoder = new LengthEncoder(pb, niceLen); - - distSlotPricesSize = getDistSlot(dictSize - 1) + 1; - distSlotPrices = new int[DIST_STATES][distSlotPricesSize]; - - reset(); - } - - public LZEncoder getLZEncoder() { - return lz; - } - - public void reset() { - super.reset(); - literalEncoder.reset(); - matchLenEncoder.reset(); - repLenEncoder.reset(); - distPriceCount = 0; - alignPriceCount = 0; - - uncompressedSize += readAhead + 1; - readAhead = -1; - } - - public int getUncompressedSize() { - return uncompressedSize; - } - - public void resetUncompressedSize() { - uncompressedSize = 0; - } - - /** - * Compress for LZMA1. - */ - public void encodeForLZMA1() throws IOException { - if (!lz.isStarted() && !encodeInit()) - return; - - while (encodeSymbol()) {} - } - - public void encodeLZMA1EndMarker() throws IOException { - // End of stream marker is encoded as a match with the maximum - // possible distance. The length is ignored by the decoder, - // but the minimum length has been used by the LZMA SDK. - // - // Distance is a 32-bit unsigned integer in LZMA. - // With Java's signed int, UINT32_MAX becomes -1. - int posState = (lz.getPos() - readAhead) & posMask; - rc.encodeBit(isMatch[state.get()], posState, 1); - rc.encodeBit(isRep, state.get(), 0); - encodeMatch(-1, MATCH_LEN_MIN, posState); - } - - /** - * Compresses for LZMA2. - * - * @return true if the LZMA2 chunk became full, false otherwise - */ - public boolean encodeForLZMA2() { - // LZMA2 uses RangeEncoderToBuffer so IOExceptions aren't possible. - try { - if (!lz.isStarted() && !encodeInit()) - return false; - - while (uncompressedSize <= LZMA2_UNCOMPRESSED_LIMIT - && rc.getPendingSize() <= LZMA2_COMPRESSED_LIMIT) - if (!encodeSymbol()) - return false; - } catch (IOException e) { - throw new Error(); - } - - return true; - } - - private boolean encodeInit() throws IOException { - assert readAhead == -1; - if (!lz.hasEnoughData(0)) - return false; - - // The first symbol must be a literal unless using - // a preset dictionary. This code isn't run if using - // a preset dictionary. - skip(1); - rc.encodeBit(isMatch[state.get()], 0, 0); - literalEncoder.encodeInit(); - - --readAhead; - assert readAhead == -1; - - ++uncompressedSize; - assert uncompressedSize == 1; - - return true; - } - - private boolean encodeSymbol() throws IOException { - if (!lz.hasEnoughData(readAhead + 1)) - return false; - - int len = getNextSymbol(); - - assert readAhead >= 0; - int posState = (lz.getPos() - readAhead) & posMask; - - if (back == -1) { - // Literal i.e. eight-bit byte - assert len == 1; - rc.encodeBit(isMatch[state.get()], posState, 0); - literalEncoder.encode(); - } else { - // Some type of match - rc.encodeBit(isMatch[state.get()], posState, 1); - if (back < REPS) { - // Repeated match i.e. the same distance - // has been used earlier. - assert lz.getMatchLen(-readAhead, reps[back], len) == len; - rc.encodeBit(isRep, state.get(), 1); - encodeRepMatch(back, len, posState); - } else { - // Normal match - assert lz.getMatchLen(-readAhead, back - REPS, len) == len; - rc.encodeBit(isRep, state.get(), 0); - encodeMatch(back - REPS, len, posState); - } - } - - readAhead -= len; - uncompressedSize += len; - - return true; - } - - private void encodeMatch(int dist, int len, int posState) - throws IOException { - state.updateMatch(); - matchLenEncoder.encode(len, posState); - - int distSlot = getDistSlot(dist); - rc.encodeBitTree(distSlots[getDistState(len)], distSlot); - - if (distSlot >= DIST_MODEL_START) { - int footerBits = (distSlot >>> 1) - 1; - int base = (2 | (distSlot & 1)) << footerBits; - int distReduced = dist - base; - - if (distSlot < DIST_MODEL_END) { - rc.encodeReverseBitTree( - distSpecial[distSlot - DIST_MODEL_START], - distReduced); - } else { - rc.encodeDirectBits(distReduced >>> ALIGN_BITS, - footerBits - ALIGN_BITS); - rc.encodeReverseBitTree(distAlign, distReduced & ALIGN_MASK); - --alignPriceCount; - } - } - - reps[3] = reps[2]; - reps[2] = reps[1]; - reps[1] = reps[0]; - reps[0] = dist; - - --distPriceCount; - } - - private void encodeRepMatch(int rep, int len, int posState) - throws IOException { - if (rep == 0) { - rc.encodeBit(isRep0, state.get(), 0); - rc.encodeBit(isRep0Long[state.get()], posState, len == 1 ? 0 : 1); - } else { - int dist = reps[rep]; - rc.encodeBit(isRep0, state.get(), 1); - - if (rep == 1) { - rc.encodeBit(isRep1, state.get(), 0); - } else { - rc.encodeBit(isRep1, state.get(), 1); - rc.encodeBit(isRep2, state.get(), rep - 2); - - if (rep == 3) - reps[3] = reps[2]; - - reps[2] = reps[1]; - } - - reps[1] = reps[0]; - reps[0] = dist; - } - - if (len == 1) { - state.updateShortRep(); - } else { - repLenEncoder.encode(len, posState); - state.updateLongRep(); - } - } - - Matches getMatches() { - ++readAhead; - Matches matches = lz.getMatches(); - assert lz.verifyMatches(matches); - return matches; - } - - void skip(int len) { - readAhead += len; - lz.skip(len); - } - - int getAnyMatchPrice(State state, int posState) { - return RangeEncoder.getBitPrice(isMatch[state.get()][posState], 1); - } - - int getNormalMatchPrice(int anyMatchPrice, State state) { - return anyMatchPrice - + RangeEncoder.getBitPrice(isRep[state.get()], 0); - } - - int getAnyRepPrice(int anyMatchPrice, State state) { - return anyMatchPrice - + RangeEncoder.getBitPrice(isRep[state.get()], 1); - } - - int getShortRepPrice(int anyRepPrice, State state, int posState) { - return anyRepPrice - + RangeEncoder.getBitPrice(isRep0[state.get()], 0) - + RangeEncoder.getBitPrice(isRep0Long[state.get()][posState], - 0); - } - - int getLongRepPrice(int anyRepPrice, int rep, State state, int posState) { - int price = anyRepPrice; - - if (rep == 0) { - price += RangeEncoder.getBitPrice(isRep0[state.get()], 0) - + RangeEncoder.getBitPrice( - isRep0Long[state.get()][posState], 1); - } else { - price += RangeEncoder.getBitPrice(isRep0[state.get()], 1); - - if (rep == 1) - price += RangeEncoder.getBitPrice(isRep1[state.get()], 0); - else - price += RangeEncoder.getBitPrice(isRep1[state.get()], 1) - + RangeEncoder.getBitPrice(isRep2[state.get()], - rep - 2); - } - - return price; - } - - int getLongRepAndLenPrice(int rep, int len, State state, int posState) { - int anyMatchPrice = getAnyMatchPrice(state, posState); - int anyRepPrice = getAnyRepPrice(anyMatchPrice, state); - int longRepPrice = getLongRepPrice(anyRepPrice, rep, state, posState); - return longRepPrice + repLenEncoder.getPrice(len, posState); - } - - int getMatchAndLenPrice(int normalMatchPrice, - int dist, int len, int posState) { - int price = normalMatchPrice - + matchLenEncoder.getPrice(len, posState); - int distState = getDistState(len); - - if (dist < FULL_DISTANCES) { - price += fullDistPrices[distState][dist]; - } else { - // Note that distSlotPrices includes also - // the price of direct bits. - int distSlot = getDistSlot(dist); - price += distSlotPrices[distState][distSlot] - + alignPrices[dist & ALIGN_MASK]; - } - - return price; - } - - private void updateDistPrices() { - distPriceCount = DIST_PRICE_UPDATE_INTERVAL; - - for (int distState = 0; distState < DIST_STATES; ++distState) { - for (int distSlot = 0; distSlot < distSlotPricesSize; ++distSlot) - distSlotPrices[distState][distSlot] - = RangeEncoder.getBitTreePrice( - distSlots[distState], distSlot); - - for (int distSlot = DIST_MODEL_END; distSlot < distSlotPricesSize; - ++distSlot) { - int count = (distSlot >>> 1) - 1 - ALIGN_BITS; - distSlotPrices[distState][distSlot] - += RangeEncoder.getDirectBitsPrice(count); - } - - for (int dist = 0; dist < DIST_MODEL_START; ++dist) - fullDistPrices[distState][dist] - = distSlotPrices[distState][dist]; - } - - int dist = DIST_MODEL_START; - for (int distSlot = DIST_MODEL_START; distSlot < DIST_MODEL_END; - ++distSlot) { - int footerBits = (distSlot >>> 1) - 1; - int base = (2 | (distSlot & 1)) << footerBits; - - int limit = distSpecial[distSlot - DIST_MODEL_START].length; - for (int i = 0; i < limit; ++i) { - int distReduced = dist - base; - int price = RangeEncoder.getReverseBitTreePrice( - distSpecial[distSlot - DIST_MODEL_START], - distReduced); - - for (int distState = 0; distState < DIST_STATES; ++distState) - fullDistPrices[distState][dist] - = distSlotPrices[distState][distSlot] + price; - - ++dist; - } - } - - assert dist == FULL_DISTANCES; - } - - private void updateAlignPrices() { - alignPriceCount = ALIGN_PRICE_UPDATE_INTERVAL; - - for (int i = 0; i < ALIGN_SIZE; ++i) - alignPrices[i] = RangeEncoder.getReverseBitTreePrice(distAlign, - i); - } - - /** - * Updates the lookup tables used for calculating match distance - * and length prices. The updating is skipped for performance reasons - * if the tables haven't changed much since the previous update. - */ - void updatePrices() { - if (distPriceCount <= 0) - updateDistPrices(); - - if (alignPriceCount <= 0) - updateAlignPrices(); - - matchLenEncoder.updatePrices(); - repLenEncoder.updatePrices(); - } - - - class LiteralEncoder extends LiteralCoder { - private final LiteralSubencoder[] subencoders; - - LiteralEncoder(int lc, int lp) { - super(lc, lp); - - subencoders = new LiteralSubencoder[1 << (lc + lp)]; - for (int i = 0; i < subencoders.length; ++i) - subencoders[i] = new LiteralSubencoder(); - } - - void reset() { - for (int i = 0; i < subencoders.length; ++i) - subencoders[i].reset(); - } - - void encodeInit() throws IOException { - // When encoding the first byte of the stream, there is - // no previous byte in the dictionary so the encode function - // wouldn't work. - assert readAhead >= 0; - subencoders[0].encode(); - } - - void encode() throws IOException { - assert readAhead >= 0; - int i = getSubcoderIndex(lz.getByte(1 + readAhead), - lz.getPos() - readAhead); - subencoders[i].encode(); - } - - int getPrice(int curByte, int matchByte, - int prevByte, int pos, State state) { - int price = RangeEncoder.getBitPrice( - isMatch[state.get()][pos & posMask], 0); - - int i = getSubcoderIndex(prevByte, pos); - price += state.isLiteral() - ? subencoders[i].getNormalPrice(curByte) - : subencoders[i].getMatchedPrice(curByte, matchByte); - - return price; - } - - private class LiteralSubencoder extends LiteralSubcoder { - void encode() throws IOException { - int symbol = lz.getByte(readAhead) | 0x100; - - if (state.isLiteral()) { - int subencoderIndex; - int bit; - - do { - subencoderIndex = symbol >>> 8; - bit = (symbol >>> 7) & 1; - rc.encodeBit(probs, subencoderIndex, bit); - symbol <<= 1; - } while (symbol < 0x10000); - - } else { - int matchByte = lz.getByte(reps[0] + 1 + readAhead); - int offset = 0x100; - int subencoderIndex; - int matchBit; - int bit; - - do { - matchByte <<= 1; - matchBit = matchByte & offset; - subencoderIndex = offset + matchBit + (symbol >>> 8); - bit = (symbol >>> 7) & 1; - rc.encodeBit(probs, subencoderIndex, bit); - symbol <<= 1; - offset &= ~(matchByte ^ symbol); - } while (symbol < 0x10000); - } - - state.updateLiteral(); - } - - int getNormalPrice(int symbol) { - int price = 0; - int subencoderIndex; - int bit; - - symbol |= 0x100; - - do { - subencoderIndex = symbol >>> 8; - bit = (symbol >>> 7) & 1; - price += RangeEncoder.getBitPrice(probs[subencoderIndex], - bit); - symbol <<= 1; - } while (symbol < (0x100 << 8)); - - return price; - } - - int getMatchedPrice(int symbol, int matchByte) { - int price = 0; - int offset = 0x100; - int subencoderIndex; - int matchBit; - int bit; - - symbol |= 0x100; - - do { - matchByte <<= 1; - matchBit = matchByte & offset; - subencoderIndex = offset + matchBit + (symbol >>> 8); - bit = (symbol >>> 7) & 1; - price += RangeEncoder.getBitPrice(probs[subencoderIndex], - bit); - symbol <<= 1; - offset &= ~(matchByte ^ symbol); - } while (symbol < (0x100 << 8)); - - return price; - } - } - } - - - class LengthEncoder extends LengthCoder { - /** - * The prices are updated after at least - * PRICE_UPDATE_INTERVAL many lengths - * have been encoded with the same posState. - */ - private static final int PRICE_UPDATE_INTERVAL = 32; // FIXME? - - private final int[] counters; - private final int[][] prices; - - LengthEncoder(int pb, int niceLen) { - int posStates = 1 << pb; - counters = new int[posStates]; - - // Always allocate at least LOW_SYMBOLS + MID_SYMBOLS because - // it makes updatePrices slightly simpler. The prices aren't - // usually needed anyway if niceLen < 18. - int lenSymbols = Math.max(niceLen - MATCH_LEN_MIN + 1, - LOW_SYMBOLS + MID_SYMBOLS); - prices = new int[posStates][lenSymbols]; - } - - void reset() { - super.reset(); - - // Reset counters to zero to force price update before - // the prices are needed. - for (int i = 0; i < counters.length; ++i) - counters[i] = 0; - } - - void encode(int len, int posState) throws IOException { - len -= MATCH_LEN_MIN; - - if (len < LOW_SYMBOLS) { - rc.encodeBit(choice, 0, 0); - rc.encodeBitTree(low[posState], len); - } else { - rc.encodeBit(choice, 0, 1); - len -= LOW_SYMBOLS; - - if (len < MID_SYMBOLS) { - rc.encodeBit(choice, 1, 0); - rc.encodeBitTree(mid[posState], len); - } else { - rc.encodeBit(choice, 1, 1); - rc.encodeBitTree(high, len - MID_SYMBOLS); - } - } - - --counters[posState]; - } - - int getPrice(int len, int posState) { - return prices[posState][len - MATCH_LEN_MIN]; - } - - void updatePrices() { - for (int posState = 0; posState < counters.length; ++posState) { - if (counters[posState] <= 0) { - counters[posState] = PRICE_UPDATE_INTERVAL; - updatePrices(posState); - } - } - } - - private void updatePrices(int posState) { - int choice0Price = RangeEncoder.getBitPrice(choice[0], 0); - - int i = 0; - for (; i < LOW_SYMBOLS; ++i) - prices[posState][i] = choice0Price - + RangeEncoder.getBitTreePrice(low[posState], i); - - choice0Price = RangeEncoder.getBitPrice(choice[0], 1); - int choice1Price = RangeEncoder.getBitPrice(choice[1], 0); - - for (; i < LOW_SYMBOLS + MID_SYMBOLS; ++i) - prices[posState][i] = choice0Price + choice1Price - + RangeEncoder.getBitTreePrice(mid[posState], - i - LOW_SYMBOLS); - - choice1Price = RangeEncoder.getBitPrice(choice[1], 1); - - for (; i < prices[posState].length; ++i) - prices[posState][i] = choice0Price + choice1Price - + RangeEncoder.getBitTreePrice(high, i - LOW_SYMBOLS - - MID_SYMBOLS); - } - } -} Index: trunk/src/org/tukaani/xz/lzma/LZMAEncoderFast.java =================================================================== --- trunk/src/org/tukaani/xz/lzma/LZMAEncoderFast.java (revision 16021) +++ (revision ) @@ -1,153 +1,0 @@ -/* - * LZMAEncoderFast - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.lzma; - -import org.tukaani.xz.ArrayCache; -import org.tukaani.xz.lz.LZEncoder; -import org.tukaani.xz.lz.Matches; -import org.tukaani.xz.rangecoder.RangeEncoder; - -final class LZMAEncoderFast extends LZMAEncoder { - private static final int EXTRA_SIZE_BEFORE = 1; - private static final int EXTRA_SIZE_AFTER = MATCH_LEN_MAX - 1; - - private Matches matches = null; - - static int getMemoryUsage(int dictSize, int extraSizeBefore, int mf) { - return LZEncoder.getMemoryUsage( - dictSize, Math.max(extraSizeBefore, EXTRA_SIZE_BEFORE), - EXTRA_SIZE_AFTER, MATCH_LEN_MAX, mf); - } - - LZMAEncoderFast(RangeEncoder rc, int lc, int lp, int pb, - int dictSize, int extraSizeBefore, - int niceLen, int mf, int depthLimit, - ArrayCache arrayCache) { - super(rc, LZEncoder.getInstance(dictSize, - Math.max(extraSizeBefore, - EXTRA_SIZE_BEFORE), - EXTRA_SIZE_AFTER, - niceLen, MATCH_LEN_MAX, - mf, depthLimit, arrayCache), - lc, lp, pb, dictSize, niceLen); - } - - private boolean changePair(int smallDist, int bigDist) { - return smallDist < (bigDist >>> 7); - } - - int getNextSymbol() { - // Get the matches for the next byte unless readAhead indicates - // that we already got the new matches during the previous call - // to this function. - if (readAhead == -1) - matches = getMatches(); - - back = -1; - - // Get the number of bytes available in the dictionary, but - // not more than the maximum match length. If there aren't - // enough bytes remaining to encode a match at all, return - // immediately to encode this byte as a literal. - int avail = Math.min(lz.getAvail(), MATCH_LEN_MAX); - if (avail < MATCH_LEN_MIN) - return 1; - - // Look for a match from the previous four match distances. - int bestRepLen = 0; - int bestRepIndex = 0; - for (int rep = 0; rep < REPS; ++rep) { - int len = lz.getMatchLen(reps[rep], avail); - if (len < MATCH_LEN_MIN) - continue; - - // If it is long enough, return it. - if (len >= niceLen) { - back = rep; - skip(len - 1); - return len; - } - - // Remember the index and length of the best repeated match. - if (len > bestRepLen) { - bestRepIndex = rep; - bestRepLen = len; - } - } - - int mainLen = 0; - int mainDist = 0; - - if (matches.count > 0) { - mainLen = matches.len[matches.count - 1]; - mainDist = matches.dist[matches.count - 1]; - - if (mainLen >= niceLen) { - back = mainDist + REPS; - skip(mainLen - 1); - return mainLen; - } - - while (matches.count > 1 - && mainLen == matches.len[matches.count - 2] + 1) { - if (!changePair(matches.dist[matches.count - 2], mainDist)) - break; - - --matches.count; - mainLen = matches.len[matches.count - 1]; - mainDist = matches.dist[matches.count - 1]; - } - - if (mainLen == MATCH_LEN_MIN && mainDist >= 0x80) - mainLen = 1; - } - - if (bestRepLen >= MATCH_LEN_MIN) { - if (bestRepLen + 1 >= mainLen - || (bestRepLen + 2 >= mainLen && mainDist >= (1 << 9)) - || (bestRepLen + 3 >= mainLen && mainDist >= (1 << 15))) { - back = bestRepIndex; - skip(bestRepLen - 1); - return bestRepLen; - } - } - - if (mainLen < MATCH_LEN_MIN || avail <= MATCH_LEN_MIN) - return 1; - - // Get the next match. Test if it is better than the current match. - // If so, encode the current byte as a literal. - matches = getMatches(); - - if (matches.count > 0) { - int newLen = matches.len[matches.count - 1]; - int newDist = matches.dist[matches.count - 1]; - - if ((newLen >= mainLen && newDist < mainDist) - || (newLen == mainLen + 1 - && !changePair(mainDist, newDist)) - || newLen > mainLen + 1 - || (newLen + 1 >= mainLen - && mainLen >= MATCH_LEN_MIN + 1 - && changePair(newDist, mainDist))) - return 1; - } - - int limit = Math.max(mainLen - 1, MATCH_LEN_MIN); - for (int rep = 0; rep < REPS; ++rep) - if (lz.getMatchLen(reps[rep], limit) == limit) - return 1; - - back = mainDist + REPS; - skip(mainLen - 2); - return mainLen; - } -} Index: trunk/src/org/tukaani/xz/lzma/LZMAEncoderNormal.java =================================================================== --- trunk/src/org/tukaani/xz/lzma/LZMAEncoderNormal.java (revision 16021) +++ (revision ) @@ -1,568 +1,0 @@ -/* - * LZMAEncoderNormal - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.lzma; - -import org.tukaani.xz.ArrayCache; -import org.tukaani.xz.lz.LZEncoder; -import org.tukaani.xz.lz.Matches; -import org.tukaani.xz.rangecoder.RangeEncoder; - -final class LZMAEncoderNormal extends LZMAEncoder { - private static final int OPTS = 4096; - - private static final int EXTRA_SIZE_BEFORE = OPTS; - private static final int EXTRA_SIZE_AFTER = OPTS; - - private final Optimum[] opts = new Optimum[OPTS]; - private int optCur = 0; - private int optEnd = 0; - - private Matches matches; - - // These are fields solely to avoid allocating the objects again and - // again on each function call. - private final int[] repLens = new int[REPS]; - private final State nextState = new State(); - - static int getMemoryUsage(int dictSize, int extraSizeBefore, int mf) { - return LZEncoder.getMemoryUsage(dictSize, - Math.max(extraSizeBefore, EXTRA_SIZE_BEFORE), - EXTRA_SIZE_AFTER, MATCH_LEN_MAX, mf) - + OPTS * 64 / 1024; - } - - LZMAEncoderNormal(RangeEncoder rc, int lc, int lp, int pb, - int dictSize, int extraSizeBefore, - int niceLen, int mf, int depthLimit, - ArrayCache arrayCache) { - super(rc, LZEncoder.getInstance(dictSize, - Math.max(extraSizeBefore, - EXTRA_SIZE_BEFORE), - EXTRA_SIZE_AFTER, - niceLen, MATCH_LEN_MAX, - mf, depthLimit, arrayCache), - lc, lp, pb, dictSize, niceLen); - - for (int i = 0; i < OPTS; ++i) - opts[i] = new Optimum(); - } - - public void reset() { - optCur = 0; - optEnd = 0; - super.reset(); - } - - /** - * Converts the opts array from backward indexes to forward indexes. - * Then it will be simple to get the next symbol from the array - * in later calls to getNextSymbol(). - */ - private int convertOpts() { - optEnd = optCur; - - int optPrev = opts[optCur].optPrev; - - do { - Optimum opt = opts[optCur]; - - if (opt.prev1IsLiteral) { - opts[optPrev].optPrev = optCur; - opts[optPrev].backPrev = -1; - optCur = optPrev--; - - if (opt.hasPrev2) { - opts[optPrev].optPrev = optPrev + 1; - opts[optPrev].backPrev = opt.backPrev2; - optCur = optPrev; - optPrev = opt.optPrev2; - } - } - - int temp = opts[optPrev].optPrev; - opts[optPrev].optPrev = optCur; - optCur = optPrev; - optPrev = temp; - } while (optCur > 0); - - optCur = opts[0].optPrev; - back = opts[optCur].backPrev; - return optCur; - } - - int getNextSymbol() { - // If there are pending symbols from an earlier call to this - // function, return those symbols first. - if (optCur < optEnd) { - int len = opts[optCur].optPrev - optCur; - optCur = opts[optCur].optPrev; - back = opts[optCur].backPrev; - return len; - } - - assert optCur == optEnd; - optCur = 0; - optEnd = 0; - back = -1; - - if (readAhead == -1) - matches = getMatches(); - - // Get the number of bytes available in the dictionary, but - // not more than the maximum match length. If there aren't - // enough bytes remaining to encode a match at all, return - // immediately to encode this byte as a literal. - int avail = Math.min(lz.getAvail(), MATCH_LEN_MAX); - if (avail < MATCH_LEN_MIN) - return 1; - - // Get the lengths of repeated matches. - int repBest = 0; - for (int rep = 0; rep < REPS; ++rep) { - repLens[rep] = lz.getMatchLen(reps[rep], avail); - - if (repLens[rep] < MATCH_LEN_MIN) { - repLens[rep] = 0; - continue; - } - - if (repLens[rep] > repLens[repBest]) - repBest = rep; - } - - // Return if the best repeated match is at least niceLen bytes long. - if (repLens[repBest] >= niceLen) { - back = repBest; - skip(repLens[repBest] - 1); - return repLens[repBest]; - } - - // Initialize mainLen and mainDist to the longest match found - // by the match finder. - int mainLen = 0; - int mainDist = 0; - if (matches.count > 0) { - mainLen = matches.len[matches.count - 1]; - mainDist = matches.dist[matches.count - 1]; - - // Return if it is at least niceLen bytes long. - if (mainLen >= niceLen) { - back = mainDist + REPS; - skip(mainLen - 1); - return mainLen; - } - } - - int curByte = lz.getByte(0); - int matchByte = lz.getByte(reps[0] + 1); - - // If the match finder found no matches and this byte cannot be - // encoded as a repeated match (short or long), we must be return - // to have the byte encoded as a literal. - if (mainLen < MATCH_LEN_MIN && curByte != matchByte - && repLens[repBest] < MATCH_LEN_MIN) - return 1; - - - int pos = lz.getPos(); - int posState = pos & posMask; - - // Calculate the price of encoding the current byte as a literal. - { - int prevByte = lz.getByte(1); - int literalPrice = literalEncoder.getPrice(curByte, matchByte, - prevByte, pos, state); - opts[1].set1(literalPrice, 0, -1); - } - - int anyMatchPrice = getAnyMatchPrice(state, posState); - int anyRepPrice = getAnyRepPrice(anyMatchPrice, state); - - // If it is possible to encode this byte as a short rep, see if - // it is cheaper than encoding it as a literal. - if (matchByte == curByte) { - int shortRepPrice = getShortRepPrice(anyRepPrice, - state, posState); - if (shortRepPrice < opts[1].price) - opts[1].set1(shortRepPrice, 0, 0); - } - - // Return if there is neither normal nor long repeated match. Use - // a short match instead of a literal if is is possible and cheaper. - optEnd = Math.max(mainLen, repLens[repBest]); - if (optEnd < MATCH_LEN_MIN) { - assert optEnd == 0 : optEnd; - back = opts[1].backPrev; - return 1; - } - - - // Update the lookup tables for distances and lengths before using - // those price calculation functions. (The price function above - // don't need these tables.) - updatePrices(); - - // Initialize the state and reps of this position in opts[]. - // updateOptStateAndReps() will need these to get the new - // state and reps for the next byte. - opts[0].state.set(state); - System.arraycopy(reps, 0, opts[0].reps, 0, REPS); - - // Initialize the prices for latter opts that will be used below. - for (int i = optEnd; i >= MATCH_LEN_MIN; --i) - opts[i].reset(); - - // Calculate the prices of repeated matches of all lengths. - for (int rep = 0; rep < REPS; ++rep) { - int repLen = repLens[rep]; - if (repLen < MATCH_LEN_MIN) - continue; - - int longRepPrice = getLongRepPrice(anyRepPrice, rep, - state, posState); - do { - int price = longRepPrice + repLenEncoder.getPrice(repLen, - posState); - if (price < opts[repLen].price) - opts[repLen].set1(price, 0, rep); - } while (--repLen >= MATCH_LEN_MIN); - } - - // Calculate the prices of normal matches that are longer than rep0. - { - int len = Math.max(repLens[0] + 1, MATCH_LEN_MIN); - if (len <= mainLen) { - int normalMatchPrice = getNormalMatchPrice(anyMatchPrice, - state); - - // Set i to the index of the shortest match that is - // at least len bytes long. - int i = 0; - while (len > matches.len[i]) - ++i; - - while (true) { - int dist = matches.dist[i]; - int price = getMatchAndLenPrice(normalMatchPrice, - dist, len, posState); - if (price < opts[len].price) - opts[len].set1(price, 0, dist + REPS); - - if (len == matches.len[i]) - if (++i == matches.count) - break; - - ++len; - } - } - } - - - avail = Math.min(lz.getAvail(), OPTS - 1); - - // Get matches for later bytes and optimize the use of LZMA symbols - // by calculating the prices and picking the cheapest symbol - // combinations. - while (++optCur < optEnd) { - matches = getMatches(); - if (matches.count > 0 - && matches.len[matches.count - 1] >= niceLen) - break; - - --avail; - ++pos; - posState = pos & posMask; - - updateOptStateAndReps(); - anyMatchPrice = opts[optCur].price - + getAnyMatchPrice(opts[optCur].state, posState); - anyRepPrice = getAnyRepPrice(anyMatchPrice, opts[optCur].state); - - calc1BytePrices(pos, posState, avail, anyRepPrice); - - if (avail >= MATCH_LEN_MIN) { - int startLen = calcLongRepPrices(pos, posState, - avail, anyRepPrice); - if (matches.count > 0) - calcNormalMatchPrices(pos, posState, avail, - anyMatchPrice, startLen); - } - } - - return convertOpts(); - } - - /** - * Updates the state and reps for the current byte in the opts array. - */ - private void updateOptStateAndReps() { - int optPrev = opts[optCur].optPrev; - assert optPrev < optCur; - - if (opts[optCur].prev1IsLiteral) { - --optPrev; - - if (opts[optCur].hasPrev2) { - opts[optCur].state.set(opts[opts[optCur].optPrev2].state); - if (opts[optCur].backPrev2 < REPS) - opts[optCur].state.updateLongRep(); - else - opts[optCur].state.updateMatch(); - } else { - opts[optCur].state.set(opts[optPrev].state); - } - - opts[optCur].state.updateLiteral(); - } else { - opts[optCur].state.set(opts[optPrev].state); - } - - if (optPrev == optCur - 1) { - // Must be either a short rep or a literal. - assert opts[optCur].backPrev == 0 || opts[optCur].backPrev == -1; - - if (opts[optCur].backPrev == 0) - opts[optCur].state.updateShortRep(); - else - opts[optCur].state.updateLiteral(); - - System.arraycopy(opts[optPrev].reps, 0, - opts[optCur].reps, 0, REPS); - } else { - int back; - if (opts[optCur].prev1IsLiteral && opts[optCur].hasPrev2) { - optPrev = opts[optCur].optPrev2; - back = opts[optCur].backPrev2; - opts[optCur].state.updateLongRep(); - } else { - back = opts[optCur].backPrev; - if (back < REPS) - opts[optCur].state.updateLongRep(); - else - opts[optCur].state.updateMatch(); - } - - if (back < REPS) { - opts[optCur].reps[0] = opts[optPrev].reps[back]; - - int rep; - for (rep = 1; rep <= back; ++rep) - opts[optCur].reps[rep] = opts[optPrev].reps[rep - 1]; - - for (; rep < REPS; ++rep) - opts[optCur].reps[rep] = opts[optPrev].reps[rep]; - } else { - opts[optCur].reps[0] = back - REPS; - System.arraycopy(opts[optPrev].reps, 0, - opts[optCur].reps, 1, REPS - 1); - } - } - } - - /** - * Calculates prices of a literal, a short rep, and literal + rep0. - */ - private void calc1BytePrices(int pos, int posState, - int avail, int anyRepPrice) { - // This will be set to true if using a literal or a short rep. - boolean nextIsByte = false; - - int curByte = lz.getByte(0); - int matchByte = lz.getByte(opts[optCur].reps[0] + 1); - - // Try a literal. - int literalPrice = opts[optCur].price - + literalEncoder.getPrice(curByte, matchByte, lz.getByte(1), - pos, opts[optCur].state); - if (literalPrice < opts[optCur + 1].price) { - opts[optCur + 1].set1(literalPrice, optCur, -1); - nextIsByte = true; - } - - // Try a short rep. - if (matchByte == curByte && (opts[optCur + 1].optPrev == optCur - || opts[optCur + 1].backPrev != 0)) { - int shortRepPrice = getShortRepPrice(anyRepPrice, - opts[optCur].state, - posState); - if (shortRepPrice <= opts[optCur + 1].price) { - opts[optCur + 1].set1(shortRepPrice, optCur, 0); - nextIsByte = true; - } - } - - // If neither a literal nor a short rep was the cheapest choice, - // try literal + long rep0. - if (!nextIsByte && matchByte != curByte && avail > MATCH_LEN_MIN) { - int lenLimit = Math.min(niceLen, avail - 1); - int len = lz.getMatchLen(1, opts[optCur].reps[0], lenLimit); - - if (len >= MATCH_LEN_MIN) { - nextState.set(opts[optCur].state); - nextState.updateLiteral(); - int nextPosState = (pos + 1) & posMask; - int price = literalPrice - + getLongRepAndLenPrice(0, len, - nextState, nextPosState); - - int i = optCur + 1 + len; - while (optEnd < i) - opts[++optEnd].reset(); - - if (price < opts[i].price) - opts[i].set2(price, optCur, 0); - } - } - } - - /** - * Calculates prices of long rep and long rep + literal + rep0. - */ - private int calcLongRepPrices(int pos, int posState, - int avail, int anyRepPrice) { - int startLen = MATCH_LEN_MIN; - int lenLimit = Math.min(avail, niceLen); - - for (int rep = 0; rep < REPS; ++rep) { - int len = lz.getMatchLen(opts[optCur].reps[rep], lenLimit); - if (len < MATCH_LEN_MIN) - continue; - - while (optEnd < optCur + len) - opts[++optEnd].reset(); - - int longRepPrice = getLongRepPrice(anyRepPrice, rep, - opts[optCur].state, posState); - - for (int i = len; i >= MATCH_LEN_MIN; --i) { - int price = longRepPrice - + repLenEncoder.getPrice(i, posState); - if (price < opts[optCur + i].price) - opts[optCur + i].set1(price, optCur, rep); - } - - if (rep == 0) - startLen = len + 1; - - int len2Limit = Math.min(niceLen, avail - len - 1); - int len2 = lz.getMatchLen(len + 1, opts[optCur].reps[rep], - len2Limit); - - if (len2 >= MATCH_LEN_MIN) { - // Rep - int price = longRepPrice - + repLenEncoder.getPrice(len, posState); - nextState.set(opts[optCur].state); - nextState.updateLongRep(); - - // Literal - int curByte = lz.getByte(len, 0); - int matchByte = lz.getByte(0); // lz.getByte(len, len) - int prevByte = lz.getByte(len, 1); - price += literalEncoder.getPrice(curByte, matchByte, prevByte, - pos + len, nextState); - nextState.updateLiteral(); - - // Rep0 - int nextPosState = (pos + len + 1) & posMask; - price += getLongRepAndLenPrice(0, len2, - nextState, nextPosState); - - int i = optCur + len + 1 + len2; - while (optEnd < i) - opts[++optEnd].reset(); - - if (price < opts[i].price) - opts[i].set3(price, optCur, rep, len, 0); - } - } - - return startLen; - } - - /** - * Calculates prices of a normal match and normal match + literal + rep0. - */ - private void calcNormalMatchPrices(int pos, int posState, int avail, - int anyMatchPrice, int startLen) { - // If the longest match is so long that it would not fit into - // the opts array, shorten the matches. - if (matches.len[matches.count - 1] > avail) { - matches.count = 0; - while (matches.len[matches.count] < avail) - ++matches.count; - - matches.len[matches.count++] = avail; - } - - if (matches.len[matches.count - 1] < startLen) - return; - - while (optEnd < optCur + matches.len[matches.count - 1]) - opts[++optEnd].reset(); - - int normalMatchPrice = getNormalMatchPrice(anyMatchPrice, - opts[optCur].state); - - int match = 0; - while (startLen > matches.len[match]) - ++match; - - for (int len = startLen; ; ++len) { - int dist = matches.dist[match]; - - // Calculate the price of a match of len bytes from the nearest - // possible distance. - int matchAndLenPrice = getMatchAndLenPrice(normalMatchPrice, - dist, len, posState); - if (matchAndLenPrice < opts[optCur + len].price) - opts[optCur + len].set1(matchAndLenPrice, - optCur, dist + REPS); - - if (len != matches.len[match]) - continue; - - // Try match + literal + rep0. First get the length of the rep0. - int len2Limit = Math.min(niceLen, avail - len - 1); - int len2 = lz.getMatchLen(len + 1, dist, len2Limit); - - if (len2 >= MATCH_LEN_MIN) { - nextState.set(opts[optCur].state); - nextState.updateMatch(); - - // Literal - int curByte = lz.getByte(len, 0); - int matchByte = lz.getByte(0); // lz.getByte(len, len) - int prevByte = lz.getByte(len, 1); - int price = matchAndLenPrice - + literalEncoder.getPrice(curByte, matchByte, - prevByte, pos + len, - nextState); - nextState.updateLiteral(); - - // Rep0 - int nextPosState = (pos + len + 1) & posMask; - price += getLongRepAndLenPrice(0, len2, - nextState, nextPosState); - - int i = optCur + len + 1 + len2; - while (optEnd < i) - opts[++optEnd].reset(); - - if (price < opts[i].price) - opts[i].set3(price, optCur, dist + REPS, len, 0); - } - - if (++match == matches.count) - break; - } - } -} Index: trunk/src/org/tukaani/xz/lzma/Optimum.java =================================================================== --- trunk/src/org/tukaani/xz/lzma/Optimum.java (revision 16021) +++ (revision ) @@ -1,73 +1,0 @@ -/* - * Optimum - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.lzma; - -final class Optimum { - private static final int INFINITY_PRICE = 1 << 30; - - final State state = new State(); - final int[] reps = new int[LZMACoder.REPS]; - - /** - * Cumulative price of arriving to this byte. - */ - int price; - - int optPrev; - int backPrev; - boolean prev1IsLiteral; - - boolean hasPrev2; - int optPrev2; - int backPrev2; - - /** - * Resets the price. - */ - void reset() { - price = INFINITY_PRICE; - } - - /** - * Sets to indicate one LZMA symbol (literal, rep, or match). - */ - void set1(int newPrice, int optCur, int back) { - price = newPrice; - optPrev = optCur; - backPrev = back; - prev1IsLiteral = false; - } - - /** - * Sets to indicate two LZMA symbols of which the first one is a literal. - */ - void set2(int newPrice, int optCur, int back) { - price = newPrice; - optPrev = optCur + 1; - backPrev = back; - prev1IsLiteral = true; - hasPrev2 = false; - } - - /** - * Sets to indicate three LZMA symbols of which the second one - * is a literal. - */ - void set3(int newPrice, int optCur, int back2, int len2, int back) { - price = newPrice; - optPrev = optCur + len2 + 1; - backPrev = back; - prev1IsLiteral = true; - hasPrev2 = true; - optPrev2 = optCur; - backPrev2 = back2; - } -} Index: trunk/src/org/tukaani/xz/lzma/State.java =================================================================== --- trunk/src/org/tukaani/xz/lzma/State.java (revision 16021) +++ (revision ) @@ -1,75 +1,0 @@ -/* - * State - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.lzma; - -final class State { - static final int STATES = 12; - - private static final int LIT_STATES = 7; - - private static final int LIT_LIT = 0; - private static final int MATCH_LIT_LIT = 1; - private static final int REP_LIT_LIT = 2; - private static final int SHORTREP_LIT_LIT = 3; - private static final int MATCH_LIT = 4; - private static final int REP_LIT = 5; - private static final int SHORTREP_LIT = 6; - private static final int LIT_MATCH = 7; - private static final int LIT_LONGREP = 8; - private static final int LIT_SHORTREP = 9; - private static final int NONLIT_MATCH = 10; - private static final int NONLIT_REP = 11; - - private int state; - - State() {} - - State(State other) { - state = other.state; - } - - void reset() { - state = LIT_LIT; - } - - int get() { - return state; - } - - void set(State other) { - state = other.state; - } - - void updateLiteral() { - if (state <= SHORTREP_LIT_LIT) - state = LIT_LIT; - else if (state <= LIT_SHORTREP) - state -= 3; - else - state -= 6; - } - - void updateMatch() { - state = state < LIT_STATES ? LIT_MATCH : NONLIT_MATCH; - } - - void updateLongRep() { - state = state < LIT_STATES ? LIT_LONGREP : NONLIT_REP; - } - - void updateShortRep() { - state = state < LIT_STATES ? LIT_SHORTREP : NONLIT_REP; - } - - boolean isLiteral() { - return state < LIT_STATES; - } -} Index: trunk/src/org/tukaani/xz/package-info.java =================================================================== --- trunk/src/org/tukaani/xz/package-info.java (revision 16021) +++ (revision ) @@ -1,45 +1,0 @@ -/* - * package-info - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -/** - * XZ data compression support. - * - *

Introduction

- *

- * This aims to be a complete implementation of XZ data compression - * in pure Java. Features: - *

    - *
  • Full support for the .xz file format specification version 1.0.4
    • - *
  • Single-threaded streamed compression and decompression
    • - *
  • Single-threaded decompression with limited random access support
    • - *
  • Raw streams (no .xz headers) for advanced users, including LZMA2 - * with preset dictionary
    • - *
- *

- * Threading is planned but it is unknown when it will be implemented. - *

- * For the latest source code, see the - * home page of XZ for Java. - * - *

Getting started

- *

- * Start by reading the documentation of {@link org.tukaani.xz.XZOutputStream} - * and {@link org.tukaani.xz.XZInputStream}. - * If you use XZ inside another file format or protocol, - * see also {@link org.tukaani.xz.SingleXZInputStream}. - * - *

Licensing

- *

- * XZ for Java has been put into the public domain, thus you can do - * whatever you want with it. All the files in the package have been - * written by Lasse Collin and/or Igor Pavlov. - *

- * This software is provided "as is", without any warranty. - */ -package org.tukaani.xz; Index: trunk/src/org/tukaani/xz/rangecoder/RangeCoder.java =================================================================== --- trunk/src/org/tukaani/xz/rangecoder/RangeCoder.java (revision 16021) +++ (revision ) @@ -1,26 +1,0 @@ -/* - * RangeCoder - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.rangecoder; - -import java.util.Arrays; - -public abstract class RangeCoder { - static final int SHIFT_BITS = 8; - static final int TOP_MASK = 0xFF000000; - static final int BIT_MODEL_TOTAL_BITS = 11; - static final int BIT_MODEL_TOTAL = 1 << BIT_MODEL_TOTAL_BITS; - static final short PROB_INIT = (short)(BIT_MODEL_TOTAL / 2); - static final int MOVE_BITS = 5; - - public static final void initProbs(short[] probs) { - Arrays.fill(probs, PROB_INIT); - } -} Index: trunk/src/org/tukaani/xz/rangecoder/RangeDecoder.java =================================================================== --- trunk/src/org/tukaani/xz/rangecoder/RangeDecoder.java (revision 16021) +++ (revision ) @@ -1,83 +1,0 @@ -/* - * RangeDecoder - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.rangecoder; - -import java.io.DataInputStream; -import java.io.IOException; - -public abstract class RangeDecoder extends RangeCoder { - int range = 0; - int code = 0; - - public abstract void normalize() throws IOException; - - public int decodeBit(short[] probs, int index) throws IOException { - normalize(); - - int prob = probs[index]; - int bound = (range >>> BIT_MODEL_TOTAL_BITS) * prob; - int bit; - - // Compare code and bound as if they were unsigned 32-bit integers. - if ((code ^ 0x80000000) < (bound ^ 0x80000000)) { - range = bound; - probs[index] = (short)( - prob + ((BIT_MODEL_TOTAL - prob) >>> MOVE_BITS)); - bit = 0; - } else { - range -= bound; - code -= bound; - probs[index] = (short)(prob - (prob >>> MOVE_BITS)); - bit = 1; - } - - return bit; - } - - public int decodeBitTree(short[] probs) throws IOException { - int symbol = 1; - - do { - symbol = (symbol << 1) | decodeBit(probs, symbol); - } while (symbol < probs.length); - - return symbol - probs.length; - } - - public int decodeReverseBitTree(short[] probs) throws IOException { - int symbol = 1; - int i = 0; - int result = 0; - - do { - int bit = decodeBit(probs, symbol); - symbol = (symbol << 1) | bit; - result |= bit << i++; - } while (symbol < probs.length); - - return result; - } - - public int decodeDirectBits(int count) throws IOException { - int result = 0; - - do { - normalize(); - - range >>>= 1; - int t = (code - range) >>> 31; - code -= range & (t - 1); - result = (result << 1) | (1 - t); - } while (--count != 0); - - return result; - } -} Index: trunk/src/org/tukaani/xz/rangecoder/RangeDecoderFromBuffer.java =================================================================== --- trunk/src/org/tukaani/xz/rangecoder/RangeDecoderFromBuffer.java (revision 16021) +++ (revision ) @@ -1,71 +1,0 @@ -/* - * RangeDecoderFromBuffer - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.rangecoder; - -import java.io.DataInputStream; -import java.io.IOException; -import org.tukaani.xz.ArrayCache; -import org.tukaani.xz.CorruptedInputException; - -public final class RangeDecoderFromBuffer extends RangeDecoder { - private static final int INIT_SIZE = 5; - - private final byte[] buf; - private int pos; - - public RangeDecoderFromBuffer(int inputSizeMax, ArrayCache arrayCache) { - // We will use the *end* of the array so if the cache gives us - // a bigger-than-requested array, we still want to use buf.length. - buf = arrayCache.getByteArray(inputSizeMax - INIT_SIZE, false); - pos = buf.length; - } - - public void putArraysToCache(ArrayCache arrayCache) { - arrayCache.putArray(buf); - } - - public void prepareInputBuffer(DataInputStream in, int len) - throws IOException { - if (len < INIT_SIZE) - throw new CorruptedInputException(); - - if (in.readUnsignedByte() != 0x00) - throw new CorruptedInputException(); - - code = in.readInt(); - range = 0xFFFFFFFF; - - // Read the data to the end of the buffer. If the data is corrupt - // and the decoder, reading from buf, tries to read past the end of - // the data, ArrayIndexOutOfBoundsException will be thrown and - // the problem is detected immediately. - len -= INIT_SIZE; - pos = buf.length - len; - in.readFully(buf, pos, len); - } - - public boolean isFinished() { - return pos == buf.length && code == 0; - } - - public void normalize() throws IOException { - if ((range & TOP_MASK) == 0) { - try { - // If the input is corrupt, this might throw - // ArrayIndexOutOfBoundsException. - code = (code << SHIFT_BITS) | (buf[pos++] & 0xFF); - range <<= SHIFT_BITS; - } catch (ArrayIndexOutOfBoundsException e) { - throw new CorruptedInputException(); - } - } - } -} Index: trunk/src/org/tukaani/xz/rangecoder/RangeEncoder.java =================================================================== --- trunk/src/org/tukaani/xz/rangecoder/RangeEncoder.java (revision 16021) +++ (revision ) @@ -1,200 +1,0 @@ -/* - * RangeEncoder - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.rangecoder; - -import java.io.IOException; - -public abstract class RangeEncoder extends RangeCoder { - private static final int MOVE_REDUCING_BITS = 4; - private static final int BIT_PRICE_SHIFT_BITS = 4; - - private static final int[] prices - = new int[BIT_MODEL_TOTAL >>> MOVE_REDUCING_BITS]; - - private long low; - private int range; - - // NOTE: int is OK for LZMA2 because a compressed chunk - // is not more than 64 KiB, but with LZMA1 there is no chunking - // so in theory cacheSize can grow very big. To be very safe, - // use long instead of int since this code is used for LZMA1 too. - long cacheSize; - private byte cache; - - static { - for (int i = (1 << MOVE_REDUCING_BITS) / 2; i < BIT_MODEL_TOTAL; - i += (1 << MOVE_REDUCING_BITS)) { - int w = i; - int bitCount = 0; - - for (int j = 0; j < BIT_PRICE_SHIFT_BITS; ++j) { - w *= w; - bitCount <<= 1; - - while ((w & 0xFFFF0000) != 0) { - w >>>= 1; - ++bitCount; - } - } - - prices[i >> MOVE_REDUCING_BITS] - = (BIT_MODEL_TOTAL_BITS << BIT_PRICE_SHIFT_BITS) - - 15 - bitCount; - } - } - - public void reset() { - low = 0; - range = 0xFFFFFFFF; - cache = 0x00; - cacheSize = 1; - } - - public int getPendingSize() { - // This function is only needed by users of RangeEncoderToBuffer, - // but providing a must-be-never-called version here makes - // LZMAEncoder simpler. - throw new Error(); - } - - public int finish() throws IOException { - for (int i = 0; i < 5; ++i) - shiftLow(); - - // RangeEncoderToBuffer.finish() needs a return value to tell - // how big the finished buffer is. RangeEncoderToStream has no - // buffer and thus no return value is needed. Here we use a dummy - // value which can be overridden in RangeEncoderToBuffer.finish(). - return -1; - } - - abstract void writeByte(int b) throws IOException; - - private void shiftLow() throws IOException { - int lowHi = (int)(low >>> 32); - - if (lowHi != 0 || low < 0xFF000000L) { - int temp = cache; - - do { - writeByte(temp + lowHi); - temp = 0xFF; - } while (--cacheSize != 0); - - cache = (byte)(low >>> 24); - } - - ++cacheSize; - low = (low & 0x00FFFFFF) << 8; - } - - public void encodeBit(short[] probs, int index, int bit) - throws IOException { - int prob = probs[index]; - int bound = (range >>> BIT_MODEL_TOTAL_BITS) * prob; - - // NOTE: Any non-zero value for bit is taken as 1. - if (bit == 0) { - range = bound; - probs[index] = (short)( - prob + ((BIT_MODEL_TOTAL - prob) >>> MOVE_BITS)); - } else { - low += bound & 0xFFFFFFFFL; - range -= bound; - probs[index] = (short)(prob - (prob >>> MOVE_BITS)); - } - - if ((range & TOP_MASK) == 0) { - range <<= SHIFT_BITS; - shiftLow(); - } - } - - public static int getBitPrice(int prob, int bit) { - // NOTE: Unlike in encodeBit(), here bit must be 0 or 1. - assert bit == 0 || bit == 1; - return prices[(prob ^ ((-bit) & (BIT_MODEL_TOTAL - 1))) - >>> MOVE_REDUCING_BITS]; - } - - public void encodeBitTree(short[] probs, int symbol) throws IOException { - int index = 1; - int mask = probs.length; - - do { - mask >>>= 1; - int bit = symbol & mask; - encodeBit(probs, index, bit); - - index <<= 1; - if (bit != 0) - index |= 1; - - } while (mask != 1); - } - - public static int getBitTreePrice(short[] probs, int symbol) { - int price = 0; - symbol |= probs.length; - - do { - int bit = symbol & 1; - symbol >>>= 1; - price += getBitPrice(probs[symbol], bit); - } while (symbol != 1); - - return price; - } - - public void encodeReverseBitTree(short[] probs, int symbol) - throws IOException { - int index = 1; - symbol |= probs.length; - - do { - int bit = symbol & 1; - symbol >>>= 1; - encodeBit(probs, index, bit); - index = (index << 1) | bit; - } while (symbol != 1); - } - - public static int getReverseBitTreePrice(short[] probs, int symbol) { - int price = 0; - int index = 1; - symbol |= probs.length; - - do { - int bit = symbol & 1; - symbol >>>= 1; - price += getBitPrice(probs[index], bit); - index = (index << 1) | bit; - } while (symbol != 1); - - return price; - } - - public void encodeDirectBits(int value, int count) throws IOException { - do { - range >>>= 1; - low += range & (0 - ((value >>> --count) & 1)); - - if ((range & TOP_MASK) == 0) { - range <<= SHIFT_BITS; - shiftLow(); - } - } while (count != 0); - } - - public static int getDirectBitsPrice(int count) { - return count << BIT_PRICE_SHIFT_BITS; - } -} Index: trunk/src/org/tukaani/xz/rangecoder/RangeEncoderToBuffer.java =================================================================== --- trunk/src/org/tukaani/xz/rangecoder/RangeEncoderToBuffer.java (revision 16021) +++ (revision ) @@ -1,59 +1,0 @@ -/* - * RangeEncoderToBuffer - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.rangecoder; - -import java.io.OutputStream; -import java.io.IOException; -import org.tukaani.xz.ArrayCache; - -public final class RangeEncoderToBuffer extends RangeEncoder { - private final byte[] buf; - private int bufPos; - - public RangeEncoderToBuffer(int bufSize, ArrayCache arrayCache) { - buf = arrayCache.getByteArray(bufSize, false); - reset(); - } - - public void putArraysToCache(ArrayCache arrayCache) { - arrayCache.putArray(buf); - } - - public void reset() { - super.reset(); - bufPos = 0; - } - - public int getPendingSize() { - // With LZMA2 it is known that cacheSize fits into an int. - return bufPos + (int)cacheSize + 5 - 1; - } - - public int finish() { - // super.finish() cannot throw an IOException because writeByte() - // provided in this file cannot throw an IOException. - try { - super.finish(); - } catch (IOException e) { - throw new Error(); - } - - return bufPos; - } - - public void write(OutputStream out) throws IOException { - out.write(buf, 0, bufPos); - } - - void writeByte(int b) { - buf[bufPos++] = (byte)b; - } -} Index: trunk/src/org/tukaani/xz/simple/ARM.java =================================================================== --- trunk/src/org/tukaani/xz/simple/ARM.java (revision 16021) +++ (revision ) @@ -1,50 +1,0 @@ -/* - * BCJ filter for little endian ARM instructions - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.simple; - -public final class ARM implements SimpleFilter { - private final boolean isEncoder; - private int pos; - - public ARM(boolean isEncoder, int startPos) { - this.isEncoder = isEncoder; - pos = startPos + 8; - } - - public int code(byte[] buf, int off, int len) { - int end = off + len - 4; - int i; - - for (i = off; i <= end; i += 4) { - if ((buf[i + 3] & 0xFF) == 0xEB) { - int src = ((buf[i + 2] & 0xFF) << 16) - | ((buf[i + 1] & 0xFF) << 8) - | (buf[i] & 0xFF); - src <<= 2; - - int dest; - if (isEncoder) - dest = src + (pos + i - off); - else - dest = src - (pos + i - off); - - dest >>>= 2; - buf[i + 2] = (byte)(dest >>> 16); - buf[i + 1] = (byte)(dest >>> 8); - buf[i] = (byte)dest; - } - } - - i -= off; - pos += i; - return i; - } -} Index: trunk/src/org/tukaani/xz/simple/ARMThumb.java =================================================================== --- trunk/src/org/tukaani/xz/simple/ARMThumb.java (revision 16021) +++ (revision ) @@ -1,53 +1,0 @@ -/* - * BCJ filter for little endian ARM-Thumb instructions - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.simple; - -public final class ARMThumb implements SimpleFilter { - private final boolean isEncoder; - private int pos; - - public ARMThumb(boolean isEncoder, int startPos) { - this.isEncoder = isEncoder; - pos = startPos + 4; - } - - public int code(byte[] buf, int off, int len) { - int end = off + len - 4; - int i; - - for (i = off; i <= end; i += 2) { - if ((buf[i + 1] & 0xF8) == 0xF0 && (buf[i + 3] & 0xF8) == 0xF8) { - int src = ((buf[i + 1] & 0x07) << 19) - | ((buf[i] & 0xFF) << 11) - | ((buf[i + 3] & 0x07) << 8) - | (buf[i + 2] & 0xFF); - src <<= 1; - - int dest; - if (isEncoder) - dest = src + (pos + i - off); - else - dest = src - (pos + i - off); - - dest >>>= 1; - buf[i + 1] = (byte)(0xF0 | ((dest >>> 19) & 0x07)); - buf[i] = (byte)(dest >>> 11); - buf[i + 3] = (byte)(0xF8 | ((dest >>> 8) & 0x07)); - buf[i + 2] = (byte)dest; - i += 2; - } - } - - i -= off; - pos += i; - return i; - } -} Index: trunk/src/org/tukaani/xz/simple/IA64.java =================================================================== --- trunk/src/org/tukaani/xz/simple/IA64.java (revision 16021) +++ (revision ) @@ -1,81 +1,0 @@ -/* - * BCJ filter for Itanium (IA-64) instructions - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.simple; - -public final class IA64 implements SimpleFilter { - private static final int[] BRANCH_TABLE = { - 0, 0, 0, 0, 0, 0, 0, 0, - 0, 0, 0, 0, 0, 0, 0, 0, - 4, 4, 6, 6, 0, 0, 7, 7, - 4, 4, 0, 0, 4, 4, 0, 0 }; - - private final boolean isEncoder; - private int pos; - - public IA64(boolean isEncoder, int startPos) { - this.isEncoder = isEncoder; - pos = startPos; - } - - public int code(byte[] buf, int off, int len) { - int end = off + len - 16; - int i; - - for (i = off; i <= end; i += 16) { - int instrTemplate = buf[i] & 0x1F; - int mask = BRANCH_TABLE[instrTemplate]; - - for (int slot = 0, bitPos = 5; slot < 3; ++slot, bitPos += 41) { - if (((mask >>> slot) & 1) == 0) - continue; - - int bytePos = bitPos >>> 3; - int bitRes = bitPos & 7; - - long instr = 0; - for (int j = 0; j < 6; ++j) - instr |= (buf[i + bytePos + j] & 0xFFL) << (8 * j); - - long instrNorm = instr >>> bitRes; - - if (((instrNorm >>> 37) & 0x0F) != 0x05 - || ((instrNorm >>> 9) & 0x07) != 0x00) - continue; - - int src = (int)((instrNorm >>> 13) & 0x0FFFFF); - src |= ((int)(instrNorm >>> 36) & 1) << 20; - src <<= 4; - - int dest; - if (isEncoder) - dest = src + (pos + i - off); - else - dest = src - (pos + i - off); - - dest >>>= 4; - - instrNorm &= ~(0x8FFFFFL << 13); - instrNorm |= (dest & 0x0FFFFFL) << 13; - instrNorm |= (dest & 0x100000L) << (36 - 20); - - instr &= (1 << bitRes) - 1; - instr |= instrNorm << bitRes; - - for (int j = 0; j < 6; ++j) - buf[i + bytePos + j] = (byte)(instr >>> (8 * j)); - } - } - - i -= off; - pos += i; - return i; - } -} Index: trunk/src/org/tukaani/xz/simple/PowerPC.java =================================================================== --- trunk/src/org/tukaani/xz/simple/PowerPC.java (revision 16021) +++ (revision ) @@ -1,50 +1,0 @@ -/* - * BCJ filter for big endian PowerPC instructions - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.simple; - -public final class PowerPC implements SimpleFilter { - private final boolean isEncoder; - private int pos; - - public PowerPC(boolean isEncoder, int startPos) { - this.isEncoder = isEncoder; - pos = startPos; - } - - public int code(byte[] buf, int off, int len) { - int end = off + len - 4; - int i; - - for (i = off; i <= end; i += 4) { - if ((buf[i] & 0xFC) == 0x48 && (buf[i + 3] & 0x03) == 0x01) { - int src = ((buf[i] & 0x03) << 24) - | ((buf[i + 1] & 0xFF) << 16) - | ((buf[i + 2] & 0xFF) << 8) - | (buf[i + 3] & 0xFC); - - int dest; - if (isEncoder) - dest = src + (pos + i - off); - else - dest = src - (pos + i - off); - - buf[i] = (byte)(0x48 | ((dest >>> 24) & 0x03)); - buf[i + 1] = (byte)(dest >>> 16); - buf[i + 2] = (byte)(dest >>> 8); - buf[i + 3] = (byte)((buf[i + 3] & 0x03) | dest); - } - } - - i -= off; - pos += i; - return i; - } -} Index: trunk/src/org/tukaani/xz/simple/SPARC.java =================================================================== --- trunk/src/org/tukaani/xz/simple/SPARC.java (revision 16021) +++ (revision ) @@ -1,56 +1,0 @@ -/* - * BCJ filter for SPARC instructions - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.simple; - -public final class SPARC implements SimpleFilter { - private final boolean isEncoder; - private int pos; - - public SPARC(boolean isEncoder, int startPos) { - this.isEncoder = isEncoder; - pos = startPos; - } - - public int code(byte[] buf, int off, int len) { - int end = off + len - 4; - int i; - - for (i = off; i <= end; i += 4) { - if ((buf[i] == 0x40 && (buf[i + 1] & 0xC0) == 0x00) - || (buf[i] == 0x7F && (buf[i + 1] & 0xC0) == 0xC0)) { - int src = ((buf[i] & 0xFF) << 24) - | ((buf[i + 1] & 0xFF) << 16) - | ((buf[i + 2] & 0xFF) << 8) - | (buf[i + 3] & 0xFF); - src <<= 2; - - int dest; - if (isEncoder) - dest = src + (pos + i - off); - else - dest = src - (pos + i - off); - - dest >>>= 2; - dest = (((0 - ((dest >>> 22) & 1)) << 22) & 0x3FFFFFFF) - | (dest & 0x3FFFFF) | 0x40000000; - - buf[i] = (byte)(dest >>> 24); - buf[i + 1] = (byte)(dest >>> 16); - buf[i + 2] = (byte)(dest >>> 8); - buf[i + 3] = (byte)dest; - } - } - - i -= off; - pos += i; - return i; - } -} Index: trunk/src/org/tukaani/xz/simple/SimpleFilter.java =================================================================== --- trunk/src/org/tukaani/xz/simple/SimpleFilter.java (revision 16021) +++ (revision ) @@ -1,14 +1,0 @@ -/* - * BCJ filter for little endian ARM instructions - * - * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.simple; - -public interface SimpleFilter { - int code(byte[] buf, int off, int len); -} Index: trunk/src/org/tukaani/xz/simple/X86.java =================================================================== --- trunk/src/org/tukaani/xz/simple/X86.java (revision 16021) +++ (revision ) @@ -1,98 +1,0 @@ -/* - * BCJ filter for x86 instructions - * - * Authors: Lasse Collin - * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. - */ - -package org.tukaani.xz.simple; - -public final class X86 implements SimpleFilter { - private static final boolean[] MASK_TO_ALLOWED_STATUS - = {true, true, true, false, true, false, false, false}; - - private static final int[] MASK_TO_BIT_NUMBER = {0, 1, 2, 2, 3, 3, 3, 3}; - - private final boolean isEncoder; - private int pos; - private int prevMask = 0; - - private static boolean test86MSByte(byte b) { - int i = b & 0xFF; - return i == 0x00 || i == 0xFF; - } - - public X86(boolean isEncoder, int startPos) { - this.isEncoder = isEncoder; - pos = startPos + 5; - } - - public int code(byte[] buf, int off, int len) { - int prevPos = off - 1; - int end = off + len - 5; - int i; - - for (i = off; i <= end; ++i) { - if ((buf[i] & 0xFE) != 0xE8) - continue; - - prevPos = i - prevPos; - if ((prevPos & ~3) != 0) { // (unsigned)prevPos > 3 - prevMask = 0; - } else { - prevMask = (prevMask << (prevPos - 1)) & 7; - if (prevMask != 0) { - if (!MASK_TO_ALLOWED_STATUS[prevMask] || test86MSByte( - buf[i + 4 - MASK_TO_BIT_NUMBER[prevMask]])) { - prevPos = i; - prevMask = (prevMask << 1) | 1; - continue; - } - } - } - - prevPos = i; - - if (test86MSByte(buf[i + 4])) { - int src = (buf[i + 1] & 0xFF) - | ((buf[i + 2] & 0xFF) << 8) - | ((buf[i + 3] & 0xFF) << 16) - | ((buf[i + 4] & 0xFF) << 24); - int dest; - while (true) { - if (isEncoder) - dest = src + (pos + i - off); - else - dest = src - (pos + i - off); - - if (prevMask == 0) - break; - - int index = MASK_TO_BIT_NUMBER[prevMask] * 8; - if (!test86MSByte((byte)(dest >>> (24 - index)))) - break; - - src = dest ^ ((1 << (32 - index)) - 1); - } - - buf[i + 1] = (byte)dest; - buf[i + 2] = (byte)(dest >>> 8); - buf[i + 3] = (byte)(dest >>> 16); - buf[i + 4] = (byte)(~(((dest >>> 24) & 1) - 1)); - i += 4; - } else { - prevMask = (prevMask << 1) | 1; - } - } - - prevPos = i - prevPos; - prevMask = ((prevPos & ~3) != 0) ? 0 : prevMask << (prevPos - 1); - - i -= off; - pos += i; - return i; - } -}