+/**************************************************************************
+ *
+ * Copyright 2011 Zack Rusin
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ **************************************************************************/
+
+
#include "trace_snappyfile.hpp"
#include <snappy.h>
+#include <iostream>
+
#include <assert.h>
#include <string.h>
using namespace Trace;
+/*
+ * Snappy file format.
+ * -------------------
+ *
+ * Snappy at its core is just a compressoin algorithm so we're
+ * creating a new file format which uses snappy compression
+ * to hold the trace data.
+ *
+ * The file is composed of a number of chunks, they are:
+ * chunk {
+ * uint32 - specifying the length of the compressed data
+ * compressed data
+ * }
+ * File can contain any number of such chunks.
+ * The default size of an uncompressed chunk is specified in
+ * SNAPPY_CHUNK_SIZE.
+ *
+ * Note:
+ * Currently the default size for a a to-be-compressed data is
+ * 1mb, meaning that the compressed data will be <= 1mb.
+ * The reason it's 1mb is because it seems
+ * to offer a pretty good compression/disk io speed ratio
+ * but that might change.
+ *
+ */
+
SnappyFile::SnappyFile(const std::string &filename,
File::Mode mode)
: File(),
m_cachePtr(0),
m_cacheSize(0)
{
- m_compressedCache = new char[SNAPPY_CHUNK_SIZE];
+ size_t maxCompressedLength =
+ snappy::MaxCompressedLength(SNAPPY_CHUNK_SIZE);
+ m_compressedCache = new char[maxCompressedLength];
}
SnappyFile::~SnappyFile()
bool SnappyFile::rawRead(void *buffer, int length)
{
- if (m_stream.eof()) {
+ if (endOfData()) {
return false;
}
- if (freeCacheSize() > length) {
- memcpy(buffer, m_cachePtr, length);
- m_cachePtr += length;
- } else if (freeCacheSize() == length) {
+
+ if (freeCacheSize() >= length) {
memcpy(buffer, m_cachePtr, length);
m_cachePtr += length;
- flushCache();
} else {
int sizeToRead = length;
int offset = 0;
m_cachePtr = NULL;
}
-void SnappyFile::rawFlush(FlushType type)
+void SnappyFile::rawFlush()
{
- if (type == FlushDeep) {
- flushCache();
- }
+ flushCache();
m_stream.flush();
}
if (m_stream.eof())
return;
//assert(m_cachePtr == m_cache + m_cacheSize);
+ m_currentOffset.chunk = m_stream.tellg();
size_t compressedLength;
compressedLength = readCompressedLength();
m_stream.read((char*)m_compressedCache, compressedLength);
+ /*
+ * The reason we peek here is because the last read will
+ * read all the way until the last character, but that will not
+ * trigger m_stream.eof() to be set, so by calling peek
+ * we assure that if we in fact have read the entire stream
+ * then the m_stream.eof() is always set.
+ */
+ m_stream.peek();
::snappy::GetUncompressedLength(m_compressedCache, compressedLength,
&m_cacheSize);
if (m_cache)
m_cacheSize = size;
}
-void SnappyFile::writeCompressedLength(size_t value)
+void SnappyFile::writeCompressedLength(uint32_t value)
{
- m_stream.write((char*)&value, sizeof value);
+ m_stream.write((const char*)&value, sizeof value);
}
-size_t SnappyFile::readCompressedLength()
+uint32_t SnappyFile::readCompressedLength()
{
- size_t len;
+ uint32_t len;
m_stream.read((char*)&len, sizeof len);
return len;
}
+
+bool SnappyFile::supportsOffsets() const
+{
+ return true;
+}
+
+File::Offset SnappyFile::currentOffset()
+{
+ m_currentOffset.offsetInChunk = m_cachePtr - m_cache;
+ return m_currentOffset;
+}
+
+void SnappyFile::setCurrentOffset(const File::Offset &offset)
+{
+ // to remove eof bit
+ m_stream.clear();
+ // seek to the start of a chunk
+ m_stream.seekg(offset.chunk, std::ios::beg);
+ // load the chunk
+ flushCache();
+ assert(m_cacheSize >= offset.offsetInChunk);
+ // seek within our cache to the correct location within the chunk
+ m_cachePtr = m_cache + offset.offsetInChunk;
+
+}