LibCompress: Implement GZip compression

This commit implements a stream compressor for the gzip
specification (RFC 1952), which is essentially a thin
wrapper around the DEFLATE compression format.
This commit is contained in:
Idan Horowitz 2021-03-13 01:26:44 +02:00 committed by Andreas Kling
parent 02569bec11
commit 135751c3a2
Notes: sideshowbarker 2024-07-18 21:24:23 +09:00
2 changed files with 94 additions and 25 deletions

View File

@ -1,5 +1,6 @@
/*
* Copyright (c) 2020, the SerenityOS developers.
* Copyright (c) 2021, Idan Horowitz <idan.horowitz@gmail.com>
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
@ -36,12 +37,12 @@ bool GzipDecompressor::is_likely_compressed(ReadonlyBytes bytes)
return bytes.size() >= 2 && bytes[0] == gzip_magic_1 && bytes[1] == gzip_magic_2;
}
bool GzipDecompressor::BlockHeader::valid_magic_number() const
bool BlockHeader::valid_magic_number() const
{
return identification_1 == gzip_magic_1 && identification_2 == gzip_magic_2;
}
bool GzipDecompressor::BlockHeader::supported_by_implementation() const
bool BlockHeader::supported_by_implementation() const
{
if (compression_method != 0x08) {
// RFC 1952 does not define any compression methods other than deflate.
@ -187,4 +188,58 @@ Optional<ByteBuffer> GzipDecompressor::decompress_all(ReadonlyBytes bytes)
bool GzipDecompressor::unreliable_eof() const { return m_eof; }
GzipCompressor::GzipCompressor(OutputStream& stream)
: m_output_stream(stream)
{
}
GzipCompressor::~GzipCompressor()
{
}
size_t GzipCompressor::write(ReadonlyBytes bytes)
{
BlockHeader header;
header.identification_1 = 0x1f;
header.identification_2 = 0x8b;
header.compression_method = 0x08;
header.flags = 0;
header.modification_time = 0;
header.extra_flags = 3; // DEFLATE sets 2 for maximum compression and 4 for minimum compression
header.operating_system = 3; // unix
m_output_stream << Bytes { &header, sizeof(header) };
DeflateCompressor compressed_stream { m_output_stream };
VERIFY(compressed_stream.write_or_error(bytes));
compressed_stream.final_flush();
Crypto::Checksum::CRC32 crc32;
crc32.update(bytes);
LittleEndian<u32> digest = crc32.digest();
LittleEndian<u32> size = bytes.size();
m_output_stream << digest << size;
return bytes.size();
}
bool GzipCompressor::write_or_error(ReadonlyBytes bytes)
{
if (write(bytes) < bytes.size()) {
set_fatal_error();
return false;
}
return true;
}
Optional<ByteBuffer> GzipCompressor::compress_all(const ReadonlyBytes& bytes)
{
DuplexMemoryStream output_stream;
GzipCompressor gzip_stream { output_stream };
gzip_stream.write_or_error(bytes);
if (gzip_stream.handle_any_error())
return {};
return output_stream.copy_into_contiguous_buffer();
}
}

View File

@ -1,5 +1,6 @@
/*
* Copyright (c) 2020, the SerenityOS developers.
* Copyright (c) 2021, Idan Horowitz <idan.horowitz@gmail.com>
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
@ -33,6 +34,28 @@ namespace Compress {
constexpr u8 gzip_magic_1 = 0x1f;
constexpr u8 gzip_magic_2 = 0x8b;
struct [[gnu::packed]] BlockHeader {
u8 identification_1;
u8 identification_2;
u8 compression_method;
u8 flags;
LittleEndian<u32> modification_time;
u8 extra_flags;
u8 operating_system;
bool valid_magic_number() const;
bool supported_by_implementation() const;
};
struct Flags {
static constexpr u8 FTEXT = 1 << 0;
static constexpr u8 FHCRC = 1 << 1;
static constexpr u8 FEXTRA = 1 << 2;
static constexpr u8 FNAME = 1 << 3;
static constexpr u8 FCOMMENT = 1 << 4;
static constexpr u8 MAX = FTEXT | FHCRC | FEXTRA | FNAME | FCOMMENT;
};
class GzipDecompressor final : public InputStream {
public:
@ -49,29 +72,6 @@ public:
static bool is_likely_compressed(ReadonlyBytes bytes);
private:
struct [[gnu::packed]] BlockHeader {
u8 identification_1;
u8 identification_2;
u8 compression_method;
u8 flags;
LittleEndian<u32> modification_time;
u8 extra_flags;
u8 operating_system;
bool valid_magic_number() const;
bool supported_by_implementation() const;
};
struct Flags {
static constexpr u8 FTEXT = 1 << 0;
static constexpr u8 FHCRC = 1 << 1;
static constexpr u8 FEXTRA = 1 << 2;
static constexpr u8 FNAME = 1 << 3;
static constexpr u8 FCOMMENT = 1 << 4;
static constexpr u8 MAX = FTEXT | FHCRC | FEXTRA | FNAME | FCOMMENT;
};
class Member {
public:
Member(BlockHeader header, InputStream& stream)
@ -95,4 +95,18 @@ private:
bool m_eof { false };
};
class GzipCompressor final : public OutputStream {
public:
GzipCompressor(OutputStream&);
~GzipCompressor();
size_t write(ReadonlyBytes) override;
bool write_or_error(ReadonlyBytes) override;
static Optional<ByteBuffer> compress_all(const ReadonlyBytes& bytes);
private:
OutputStream& m_output_stream;
};
}