Merge pull request #1019 from TwanVanDongen/master

ARJ's methods 1, 2 and 3 implemented for streaming
This commit is contained in:
Adam Hathcock
2025-11-18 13:41:49 +00:00
committed by GitHub
21 changed files with 10898 additions and 30 deletions

View File

@@ -1,6 +1,6 @@
# SharpCompress
SharpCompress is a compression library in pure C# for .NET Framework 4.62, .NET Standard 2.1, .NET 6.0 and NET 8.0 that can unrar, un7zip, unzip, untar unbzip2, ungzip, unlzip, unzstd with forward-only reading and file random access APIs. Write support for zip/tar/bzip2/gzip/lzip are implemented.
SharpCompress is a compression library in pure C# for .NET Framework 4.62, .NET Standard 2.1, .NET 6.0 and NET 8.0 that can unrar, un7zip, unzip, untar unbzip2, ungzip, unlzip, unzstd, unarc and unarj with forward-only reading and file random access APIs. Write support for zip/tar/bzip2/gzip/lzip are implemented.
The major feature is support for non-seekable streams so large files can be processed on the fly (i.e. download stream).

View File

@@ -38,6 +38,22 @@ namespace SharpCompress.Common.Arj
Header.CompressedSize
);
break;
case CompressionMethod.CompressedMost:
case CompressionMethod.Compressed:
case CompressionMethod.CompressedFaster:
if (Header.CompressedSize > 128 * 1024)
{
throw new NotSupportedException(
"CompressionMethod: "
+ Header.CompressionMethod
+ " with size > 128KB"
);
}
compressedStream = new LhaStream<Lh7DecoderCfg>(
_stream,
(int)Header.OriginalSize
);
break;
case CompressionMethod.CompressedFastest:
compressedStream = new LHDecoderStream(_stream, (int)Header.OriginalSize);
break;

View File

@@ -4,56 +4,68 @@ using System.IO;
namespace SharpCompress.Compressors.Arj
{
[CLSCompliant(true)]
public sealed class BitReader
public class BitReader
{
private readonly Stream _stream;
private int _bitBuffer;
private int _bitsRemaining;
private bool _disposed;
private readonly Stream _input;
private int _bitBuffer; // currently buffered bits
private int _bitCount; // number of bits in buffer
public BitReader(Stream input)
{
_stream = input ?? throw new ArgumentNullException(nameof(input));
if (!input.CanRead)
throw new ArgumentException("Stream must be readable.", nameof(input));
_input = input ?? throw new ArgumentNullException(nameof(input));
_bitBuffer = 0;
_bitCount = 0;
}
/// <summary>
/// Reads a single bit from the stream. Returns 0 or 1.
/// </summary>
public int ReadBit()
{
if (_bitCount == 0)
{
int nextByte = _input.ReadByte();
if (nextByte < 0)
{
throw new EndOfStreamException("No more data available in BitReader.");
}
_bitBuffer = nextByte;
_bitCount = 8;
}
int bit = (_bitBuffer >> (_bitCount - 1)) & 1;
_bitCount--;
return bit;
}
/// <summary>
/// Reads n bits (up to 32) from the stream.
/// </summary>
public int ReadBits(int count)
{
if (_disposed)
throw new ObjectDisposedException(nameof(BitReader));
if (count <= 0 || count > 32)
if (count < 0 || count > 32)
{
throw new ArgumentOutOfRangeException(
nameof(count),
"Bit count must be between 1 and 32."
"Count must be between 0 and 32."
);
}
int result = 0;
for (int i = 0; i < count; i++)
{
if (_bitsRemaining == 0)
{
int nextByte = _stream.ReadByte();
if (nextByte == -1)
throw new EndOfStreamException();
_bitBuffer = nextByte;
_bitsRemaining = 8;
}
// hoogste bit eerst
result = (result << 1) | ((_bitBuffer >> 7) & 1);
_bitBuffer <<= 1;
_bitsRemaining--;
result = (result << 1) | ReadBit();
}
return result;
}
/// <summary>
/// Resets any buffered bits.
/// </summary>
public void AlignToByte()
{
_bitsRemaining = 0;
_bitCount = 0;
_bitBuffer = 0;
}
}

View File

@@ -0,0 +1,43 @@
using System;
using System.Collections;
using System.Collections.Generic;
namespace SharpCompress.Compressors.Arj
{
/// <summary>
/// Iterator that reads & pushes values back into the ring buffer.
/// </summary>
public class HistoryIterator : IEnumerator<byte>
{
private int _index;
private readonly IRingBuffer _ring;
public HistoryIterator(IRingBuffer ring, int startIndex)
{
_ring = ring;
_index = startIndex;
}
public bool MoveNext()
{
Current = _ring[_index];
_index = unchecked(_index + 1);
// Push value back into the ring buffer
_ring.Push(Current);
return true; // iterator is infinite
}
public void Reset()
{
throw new NotSupportedException();
}
public byte Current { get; private set; }
object IEnumerator.Current => Current;
public void Dispose() { }
}
}

View File

@@ -0,0 +1,218 @@
using System;
using System.Collections.Generic;
using System.IO;
using System.Text;
namespace SharpCompress.Compressors.Arj
{
[CLSCompliant(true)]
public enum NodeType
{
Leaf,
Branch,
}
[CLSCompliant(true)]
public sealed class TreeEntry
{
public readonly NodeType Type;
public readonly int LeafValue;
public readonly int BranchIndex;
public const int MAX_INDEX = 4096;
private TreeEntry(NodeType type, int leafValue, int branchIndex)
{
Type = type;
LeafValue = leafValue;
BranchIndex = branchIndex;
}
public static TreeEntry Leaf(int value)
{
return new TreeEntry(NodeType.Leaf, value, -1);
}
public static TreeEntry Branch(int index)
{
if (index >= MAX_INDEX)
{
throw new ArgumentOutOfRangeException(
nameof(index),
"Branch index exceeds MAX_INDEX"
);
}
return new TreeEntry(NodeType.Branch, 0, index);
}
}
[CLSCompliant(true)]
public sealed class HuffTree
{
private readonly List<TreeEntry> _tree;
public HuffTree(int capacity = 0)
{
_tree = new List<TreeEntry>(capacity);
}
public void SetSingle(int value)
{
_tree.Clear();
_tree.Add(TreeEntry.Leaf(value));
}
public void BuildTree(byte[] lengths, int count)
{
if (lengths == null)
{
throw new ArgumentNullException(nameof(lengths));
}
if (count < 0 || count > lengths.Length)
{
throw new ArgumentOutOfRangeException(nameof(count));
}
if (count > TreeEntry.MAX_INDEX / 2)
{
throw new ArgumentException(
$"Count exceeds maximum allowed: {TreeEntry.MAX_INDEX / 2}"
);
}
byte[] slice = new byte[count];
Array.Copy(lengths, slice, count);
BuildTree(slice);
}
public void BuildTree(byte[] valueLengths)
{
if (valueLengths == null)
{
throw new ArgumentNullException(nameof(valueLengths));
}
if (valueLengths.Length > TreeEntry.MAX_INDEX / 2)
{
throw new InvalidOperationException("Too many code lengths");
}
_tree.Clear();
int maxAllocated = 1; // start with a single (root) node
for (byte currentLen = 1; ; currentLen++)
{
// add missing branches up to current limit
int maxLimit = maxAllocated;
for (int i = _tree.Count; i < maxLimit; i++)
{
// TreeEntry.Branch may throw if index too large
try
{
_tree.Add(TreeEntry.Branch(maxAllocated));
}
catch (ArgumentOutOfRangeException e)
{
_tree.Clear();
throw new InvalidOperationException("Branch index exceeds limit", e);
}
// each branch node allocates two children
maxAllocated += 2;
}
// fill tree with leaves found in the lengths table at the current length
bool moreLeaves = false;
for (int value = 0; value < valueLengths.Length; value++)
{
byte len = valueLengths[value];
if (len == currentLen)
{
_tree.Add(TreeEntry.Leaf(value));
}
else if (len > currentLen)
{
moreLeaves = true; // there are more leaves to process
}
}
// sanity check (too many leaves)
if (_tree.Count > maxAllocated)
{
throw new InvalidOperationException("Too many leaves");
}
// stop when no longer finding longer codes
if (!moreLeaves)
{
break;
}
}
// ensure tree is complete
if (_tree.Count != maxAllocated)
{
throw new InvalidOperationException(
$"Missing some leaves: tree count = {_tree.Count}, expected = {maxAllocated}"
);
}
}
public int ReadEntry(BitReader reader)
{
if (_tree.Count == 0)
{
throw new InvalidOperationException("Tree not initialized");
}
TreeEntry node = _tree[0];
while (true)
{
if (node.Type == NodeType.Leaf)
{
return node.LeafValue;
}
int bit = reader.ReadBit();
int index = node.BranchIndex + bit;
if (index >= _tree.Count)
{
throw new InvalidOperationException("Invalid branch index during read");
}
node = _tree[index];
}
}
public override string ToString()
{
var result = new StringBuilder();
void FormatStep(int index, string prefix)
{
var node = _tree[index];
if (node.Type == NodeType.Leaf)
{
result.AppendLine($"{prefix} -> {node.LeafValue}");
}
else
{
FormatStep(node.BranchIndex, prefix + "0");
FormatStep(node.BranchIndex + 1, prefix + "1");
}
}
if (_tree.Count > 0)
{
FormatStep(0, "");
}
return result.ToString();
}
}
}

View File

@@ -0,0 +1,9 @@
namespace SharpCompress.Compressors.Arj
{
public interface ILhaDecoderConfig
{
int HistoryBits { get; }
int OffsetBits { get; }
RingBuffer RingBuffer { get; }
}
}

View File

@@ -0,0 +1,17 @@
namespace SharpCompress.Compressors.Arj
{
public interface IRingBuffer
{
int BufferSize { get; }
int Cursor { get; }
void SetCursor(int pos);
void Push(byte value);
HistoryIterator IterFromOffset(int offset);
HistoryIterator IterFromPos(int pos);
byte this[int index] { get; }
}
}

View File

@@ -0,0 +1,9 @@
namespace SharpCompress.Compressors.Arj
{
public class Lh5DecoderCfg : ILhaDecoderConfig
{
public int HistoryBits => 14;
public int OffsetBits => 4;
public RingBuffer RingBuffer { get; } = new RingBuffer(1 << 14);
}
}

View File

@@ -0,0 +1,9 @@
namespace SharpCompress.Compressors.Arj
{
public class Lh7DecoderCfg : ILhaDecoderConfig
{
public int HistoryBits => 17;
public int OffsetBits => 5;
public RingBuffer RingBuffer { get; } = new RingBuffer(1 << 17);
}
}

View File

@@ -0,0 +1,363 @@
using System;
using System.Data;
using System.IO;
using System.Linq;
using SharpCompress.IO;
namespace SharpCompress.Compressors.Arj
{
[CLSCompliant(true)]
public sealed class LhaStream<C> : Stream, IStreamStack
where C : ILhaDecoderConfig, new()
{
private readonly BitReader _bitReader;
private readonly Stream _stream;
private readonly HuffTree _commandTree;
private readonly HuffTree _offsetTree;
private int _remainingCommands;
private (int offset, int count)? _copyProgress;
private readonly RingBuffer _ringBuffer;
private readonly C _config = new C();
private const int NUM_COMMANDS = 510;
private const int NUM_TEMP_CODELEN = 20;
private readonly int _originalSize;
private int _producedBytes = 0;
#if DEBUG_STREAMS
long IStreamStack.InstanceId { get; set; }
#endif
int IStreamStack.DefaultBufferSize { get; set; }
Stream IStreamStack.BaseStream() => _stream;
int IStreamStack.BufferSize
{
get => 0;
set { }
}
int IStreamStack.BufferPosition
{
get => 0;
set { }
}
void IStreamStack.SetPosition(long position) { }
public LhaStream(Stream compressedStream, int originalSize)
{
_stream = compressedStream ?? throw new ArgumentNullException(nameof(compressedStream));
_bitReader = new BitReader(compressedStream);
_ringBuffer = _config.RingBuffer;
_commandTree = new HuffTree(NUM_COMMANDS * 2);
_offsetTree = new HuffTree(NUM_TEMP_CODELEN * 2);
_remainingCommands = 0;
_copyProgress = null;
_originalSize = originalSize;
}
public override bool CanRead => true;
public override bool CanSeek => false;
public override bool CanWrite => false;
public override long Length => throw new NotSupportedException();
public override long Position
{
get => throw new NotSupportedException();
set => throw new NotSupportedException();
}
public override void Flush() { }
public override long Seek(long offset, SeekOrigin origin) =>
throw new NotSupportedException();
public override void SetLength(long value) => throw new NotSupportedException();
public override void Write(byte[] buffer, int offset, int count) =>
throw new NotSupportedException();
public override int Read(byte[] buffer, int offset, int count)
{
if (buffer == null)
{
throw new ArgumentNullException(nameof(buffer));
}
if (offset < 0 || count < 0 || (offset + count) > buffer.Length)
{
throw new ArgumentOutOfRangeException();
}
if (_producedBytes >= _originalSize)
{
return 0; // EOF
}
if (count == 0)
{
return 0;
}
int bytesRead = FillBuffer(buffer);
return bytesRead;
}
private byte ReadCodeLength()
{
byte len = (byte)_bitReader.ReadBits(3);
if (len == 7)
{
while (_bitReader.ReadBit() != 0)
{
len++;
if (len > 255)
{
throw new InvalidOperationException("Code length overflow");
}
}
}
return len;
}
private int ReadCodeSkip(int skipRange)
{
int bits;
int increment;
switch (skipRange)
{
case 0:
return 1;
case 1:
bits = 4;
increment = 3; // 3..=18
break;
default:
bits = 9;
increment = 20; // 20..=531
break;
}
int skip = _bitReader.ReadBits(bits);
return skip + increment;
}
private void ReadTempTree()
{
byte[] codeLengths = new byte[NUM_TEMP_CODELEN];
// number of codes to read (5 bits)
int numCodes = _bitReader.ReadBits(5);
// single code only
if (numCodes == 0)
{
int code = _bitReader.ReadBits(5);
_offsetTree.SetSingle((byte)code);
return;
}
if (numCodes > NUM_TEMP_CODELEN)
{
throw new Exception("temporary codelen table has invalid size");
}
// read actual lengths
int count = Math.Min(3, numCodes);
for (int i = 0; i < count; i++)
{
codeLengths[i] = (byte)ReadCodeLength();
}
// 2-bit skip value follows
int skip = _bitReader.ReadBits(2);
if (3 + skip > numCodes)
{
throw new Exception("temporary codelen table has invalid size");
}
for (int i = 3 + skip; i < numCodes; i++)
{
codeLengths[i] = (byte)ReadCodeLength();
}
_offsetTree.BuildTree(codeLengths, numCodes);
}
private void ReadCommandTree()
{
byte[] codeLengths = new byte[NUM_COMMANDS];
// number of codes to read (9 bits)
int numCodes = _bitReader.ReadBits(9);
// single code only
if (numCodes == 0)
{
int code = _bitReader.ReadBits(9);
_commandTree.SetSingle((ushort)code);
return;
}
if (numCodes > NUM_COMMANDS)
{
throw new Exception("commands codelen table has invalid size");
}
int index = 0;
while (index < numCodes)
{
for (int n = 0; n < numCodes - index; n++)
{
int code = _offsetTree.ReadEntry(_bitReader);
if (code >= 0 && code <= 2) // skip range
{
int skipCount = ReadCodeSkip(code);
index += n + skipCount;
goto outerLoop;
}
else
{
codeLengths[index + n] = (byte)(code - 2);
}
}
break;
outerLoop:
;
}
_commandTree.BuildTree(codeLengths, numCodes);
}
private void ReadOffsetTree()
{
int numCodes = _bitReader.ReadBits(_config.OffsetBits);
if (numCodes == 0)
{
int code = _bitReader.ReadBits(_config.OffsetBits);
_offsetTree.SetSingle(code);
return;
}
if (numCodes > _config.HistoryBits)
{
throw new InvalidDataException("Offset code table too large");
}
byte[] codeLengths = new byte[NUM_TEMP_CODELEN];
for (int i = 0; i < numCodes; i++)
{
codeLengths[i] = (byte)ReadCodeLength();
}
_offsetTree.BuildTree(codeLengths, numCodes);
}
private void BeginNewBlock()
{
ReadTempTree();
ReadCommandTree();
ReadOffsetTree();
}
private int ReadCommand() => _commandTree.ReadEntry(_bitReader);
private int ReadOffset()
{
int bits = _offsetTree.ReadEntry(_bitReader);
if (bits <= 1)
{
return bits;
}
int res = _bitReader.ReadBits(bits - 1);
return res | (1 << (bits - 1));
}
private int CopyFromHistory(byte[] target, int targetIndex, int offset, int count)
{
var historyIter = _ringBuffer.IterFromOffset(offset);
int copied = 0;
while (
copied < count && historyIter.MoveNext() && (targetIndex + copied) < target.Length
)
{
target[targetIndex + copied] = historyIter.Current;
copied++;
}
if (copied < count)
{
_copyProgress = (offset, count - copied);
}
return copied;
}
public int FillBuffer(byte[] buffer)
{
int bufLen = buffer.Length;
int bufIndex = 0;
// stop when we reached original size
if (_producedBytes >= _originalSize)
{
return 0;
}
// calculate limit, so that we don't go over the original size
int remaining = (int)Math.Min(bufLen, _originalSize - _producedBytes);
while (bufIndex < remaining)
{
if (_copyProgress.HasValue)
{
var (offset, count) = _copyProgress.Value;
int copied = CopyFromHistory(
buffer,
bufIndex,
offset,
(int)Math.Min(count, remaining - bufIndex)
);
bufIndex += copied;
_copyProgress = null;
}
if (_remainingCommands == 0)
{
_remainingCommands = _bitReader.ReadBits(16);
if (bufIndex + _remainingCommands > remaining)
{
break;
}
BeginNewBlock();
}
_remainingCommands--;
int command = ReadCommand();
if (command >= 0 && command <= 0xFF)
{
byte value = (byte)command;
buffer[bufIndex++] = value;
_ringBuffer.Push(value);
}
else
{
int count = command - 0x100 + 3;
int offset = ReadOffset();
int copyCount = (int)Math.Min(count, remaining - bufIndex);
bufIndex += CopyFromHistory(buffer, bufIndex, offset, copyCount);
}
}
_producedBytes += bufIndex;
return bufIndex;
}
}
}

View File

@@ -0,0 +1,67 @@
using System;
using System.Collections;
using System.Collections.Generic;
namespace SharpCompress.Compressors.Arj
{
/// <summary>
/// A fixed-size ring buffer where N must be a power of two.
/// </summary>
public class RingBuffer : IRingBuffer
{
private readonly byte[] _buffer;
private int _cursor;
public int BufferSize { get; }
public int Cursor => _cursor;
private readonly int _mask;
public RingBuffer(int size)
{
if ((size & (size - 1)) != 0)
{
throw new ArgumentException("RingArrayBuffer size must be a power of two");
}
BufferSize = size;
_buffer = new byte[size];
_cursor = 0;
_mask = size - 1;
// Fill with spaces
for (int i = 0; i < size; i++)
{
_buffer[i] = (byte)' ';
}
}
public void SetCursor(int pos)
{
_cursor = pos & _mask;
}
public void Push(byte value)
{
int index = _cursor;
_buffer[index & _mask] = value;
_cursor = (index + 1) & _mask;
}
public byte this[int index] => _buffer[index & _mask];
public HistoryIterator IterFromOffset(int offset)
{
int masked = (offset & _mask) + 1;
int startIndex = _cursor + BufferSize - masked;
return new HistoryIterator(this, startIndex);
}
public HistoryIterator IterFromPos(int pos)
{
int startIndex = pos & _mask;
return new HistoryIterator(this, startIndex);
}
}
}

View File

@@ -8,6 +8,7 @@ using SharpCompress.Common;
using SharpCompress.Readers;
using SharpCompress.Readers.Arj;
using Xunit;
using Xunit.Sdk;
namespace SharpCompress.Test.Arj
{
@@ -22,6 +23,15 @@ namespace SharpCompress.Test.Arj
[Fact]
public void Arj_Uncompressed_Read() => Read("Arj.store.arj", CompressionType.None);
[Fact]
public void Arj_Method1_Read() => Read("Arj.method1.arj");
[Fact]
public void Arj_Method2_Read() => Read("Arj.method2.arj");
[Fact]
public void Arj_Method3_Read() => Read("Arj.method3.arj");
[Fact]
public void Arj_Method4_Read() => Read("Arj.method4.arj");
@@ -48,6 +58,42 @@ namespace SharpCompress.Test.Arj
);
}
[Theory]
[InlineData("Arj.method1.largefile.arj", CompressionType.ArjLZ77)]
[InlineData("Arj.method2.largefile.arj", CompressionType.ArjLZ77)]
[InlineData("Arj.method3.largefile.arj", CompressionType.ArjLZ77)]
public void Arj_LargeFile_ShouldThrow(string fileName, CompressionType compressionType)
{
var exception = Assert.Throws<NotSupportedException>(() =>
Arj_LargeFileTest_Read(fileName, compressionType)
);
}
[Theory]
[InlineData("Arj.store.largefile.arj", CompressionType.None)]
[InlineData("Arj.method4.largefile.arj", CompressionType.ArjLZ77)]
public void Arj_LargeFileTest_Read(string fileName, CompressionType compressionType)
{
using (var stream = File.OpenRead(Path.Combine(TEST_ARCHIVES_PATH, fileName)))
using (
var reader = ReaderFactory.Open(stream, new ReaderOptions { LookForHeader = true })
)
{
while (reader.MoveToNextEntry())
{
Assert.Equal(compressionType, reader.Entry.CompressionType);
reader.WriteEntryToDirectory(
SCRATCH_FILES_PATH,
new ExtractionOptions { ExtractFullPath = true, Overwrite = true }
);
}
}
CompareFilesByPath(
Path.Combine(SCRATCH_FILES_PATH, "news.txt"),
Path.Combine(MISC_TEST_FILES_PATH, "news.txt")
);
}
private void DoArj_Multi_Reader(string[] archives)
{
using (

Binary file not shown.

Binary file not shown.

Binary file not shown.

Binary file not shown.

Binary file not shown.

Binary file not shown.

Binary file not shown.

Binary file not shown.

File diff suppressed because it is too large Load Diff