rpnx::compression::xz_codec Namespace Reference

Internal implementation of xz and its LZMA2 payload format. More...

Classes
struct	block_record
	Size metadata collected while decoding one xz block. More...
class	crc64_accumulator
	Incremental CRC-64/XZ accumulator for streamed block data. More...
struct	lzma_model
	Adaptive probability tables and state used by LZMA. More...
class	range_decoder
	One-shot LZMA range decoder. More...
class	range_encoder
	One-shot LZMA range encoder used by the deterministic literal encoder. More...

Functions
std::uint32_t	crc32 (std::span< std::byte const > input) noexcept
	Computes the reflected CRC-32 used by xz metadata and checks.
std::uint64_t	read_little_endian (std::span< std::byte const > input, std::size_t &position, std::uint8_t byte_count)
	Reads a fixed-width little-endian integer.
std::uint64_t	read_variable_integer (std::span< std::byte const > input, std::size_t &position)
	Decodes one minimal xz variable-length integer.
std::size_t	check_size (std::uint8_t check_identifier)
	Returns the number of bytes in an xz integrity check.
void	update_literal_state (std::uint8_t &state) noexcept
	Updates an LZMA state after decoding a literal.
std::uint32_t	decode_length (range_decoder &decoder, lzma_model::length_model &model, std::size_t position_state)
	Decodes one LZMA match length.
void	copy_match (std::vector< std::byte > &output, std::size_t history_begin, std::uint32_t dictionary_size, std::uint32_t distance, std::uint32_t length, std::size_t output_limit)
	Copies one validated LZMA match into the output dictionary.
void	decode_lzma_chunk (std::span< std::byte const > encoded, std::size_t uncompressed_size, lzma_model &model, std::vector< std::byte > &output, std::size_t history_begin, std::uint32_t dictionary_size, std::size_t output_limit, std::size_t dictionary_position_offset=0U)
	Decodes one LZMA range-coded chunk into an LZMA2 dictionary.
std::vector< std::byte >	compress_lzma_literals (std::span< std::byte const > input, lzma_model &model, std::size_t dictionary_position, std::uint8_t previous_byte)
	Encodes a literal-only LZMA range-coded chunk.
std::uint32_t	decode_dictionary_size (std::uint8_t properties)
	Decodes the one-byte LZMA2 dictionary-size property.
template<std::input_iterator input_iterator, std::sentinel_for< input_iterator > sentinel, typename output_iterator>
output_iterator	compress (input_iterator first, sentinel last, output_iterator output, compression_options const &options)
	Compresses an iterator range as an xz stream containing LZMA2.
template<std::input_iterator input_iterator, std::sentinel_for< input_iterator > sentinel, typename output_iterator>
output_iterator	decompress (input_iterator first, sentinel last, output_iterator output, decompression_options const &options)
	Decompresses one or more xz streams containing LZMA2.

Variables
constexpr std::uint32_t	probability_total = 1U << 11U
	LZMA probability-model total.
constexpr std::uint32_t	probability_move_bits = 5U
	Adaptation shift applied after every probability decision.
constexpr std::uint32_t	range_top = 1U << 24U
	Range threshold below which the arithmetic coder normalizes.
constexpr std::size_t	state_count = 12U
	Number of LZMA state-machine states.
constexpr std::size_t	position_state_count = 16U
	Maximum number of position states.
constexpr std::size_t	literal_coder_count = 16U
	Number of literal contexts retained by the supported properties.
constexpr std::size_t	literal_coder_size = 0x300U
	Probability count in one LZMA literal coder.

Detailed Description

Internal implementation of xz and its LZMA2 payload format.

Function Documentation

check_size()

std::size_t rpnx::compression::xz_codec::check_size ( std::uint8_t check_identifier )

inlinenodiscard

Returns the number of bytes in an xz integrity check.

Parameters

check_identifier

Four-bit xz check identifier.

Returns

Encoded check size.

Definition at line 484 of file xz.hpp.

Referenced by decompress().

compress()

template<std::input_iterator input_iterator, std::sentinel_for< input_iterator > sentinel, typename output_iterator>

output_iterator rpnx::compression::xz_codec::compress	(	input_iterator	first,
		sentinel	last,
		output_iterator	output,
		compression_options const &	options )

Compresses an iterator range as an xz stream containing LZMA2.

Template Parameters

input_iterator	Single-pass byte iterator.
sentinel	Sentinel for first.
output_iterator	Destination byte iterator.

Parameters

first	First source byte.
last	Sentinel past the source.
output	Destination iterator.
options	Compression level from 0 through 9.

Returns

Destination advanced past the xz footer.

Definition at line 757 of file xz.hpp.

References compress_lzma_literals(), crc32(), rpnx::compression::invalid_option, rpnx::compression::compression_options::level, rpnx::compression::xz_codec::lzma_model::set_properties(), rpnx::compression::implementation::to_byte(), rpnx::compression::implementation::crc32_accumulator::update(), rpnx::compression::xz_codec::crc64_accumulator::update(), rpnx::compression::implementation::crc32_accumulator::value(), rpnx::compression::xz_codec::crc64_accumulator::value(), rpnx::compression::implementation::write_byte(), and rpnx::compression::xz.

compress_lzma_literals()

std::vector< std::byte > rpnx::compression::xz_codec::compress_lzma_literals ( std::span< std::byte const > input, lzma_model & model, std::size_t dictionary_position, std::uint8_t previous_byte )

inlinenodiscard

Encodes a literal-only LZMA range-coded chunk.

Parameters

input	Source literals.
model	Adaptive model updated in place.
dictionary_position	Absolute dictionary position of the first literal.
previous_byte	Byte preceding the chunk, or zero at dictionary start.

Returns

Range-coded chunk payload.

Definition at line 701 of file xz.hpp.

References rpnx::compression::xz_codec::range_encoder::encode_bit(), rpnx::compression::xz_codec::range_encoder::finish(), rpnx::compression::xz_codec::lzma_model::is_match, rpnx::compression::xz_codec::lzma_model::literal, rpnx::compression::xz_codec::lzma_model::literal_context_bits, rpnx::compression::xz_codec::lzma_model::literal_position_bits, rpnx::compression::xz_codec::lzma_model::position_bits, rpnx::compression::xz_codec::lzma_model::state, and update_literal_state().

Referenced by compress().

copy_match()

void rpnx::compression::xz_codec::copy_match ( std::vector< std::byte > & output, std::size_t history_begin, std::uint32_t dictionary_size, std::uint32_t distance, std::uint32_t length, std::size_t output_limit )

inline

Copies one validated LZMA match into the output dictionary.

Parameters

output	Stream output and dictionary storage.
history_begin	Start of the current LZMA2 dictionary history.
dictionary_size	Declared maximum dictionary size.
distance	Zero-based match distance.
length	Number of bytes to reproduce.
output_limit	Absolute output-size limit.

Definition at line 539 of file xz.hpp.

References rpnx::compression::invalid_data, rpnx::compression::output_limit_exceeded, and rpnx::compression::xz.

Referenced by decode_lzma_chunk().

crc32()

std::uint32_t rpnx::compression::xz_codec::crc32 ( std::span< std::byte const > input )

inlinenodiscardnoexcept

Computes the reflected CRC-32 used by xz metadata and checks.

Parameters

input

Bytes to checksum.

Returns

Finalized CRC-32.

Definition at line 384 of file xz.hpp.

Referenced by compress(), and decompress().

decode_dictionary_size()

std::uint32_t rpnx::compression::xz_codec::decode_dictionary_size ( std::uint8_t properties )

inlinenodiscard

Decodes the one-byte LZMA2 dictionary-size property.

Parameters

properties

Encoded property from the xz block filter flags.

Returns

Dictionary size in bytes.

Definition at line 730 of file xz.hpp.

References rpnx::compression::unsupported_feature, and rpnx::compression::xz.

Referenced by decompress().

decode_length()

std::uint32_t rpnx::compression::xz_codec::decode_length ( range_decoder & decoder, lzma_model::length_model & model, std::size_t position_state )

inlinenodiscard

Decodes one LZMA match length.

Parameters

decoder	Range decoder.
model	Match or repeated-match length model updated in place.
position_state	Current low dictionary-position state.

Returns

Match length of at least two bytes.

Definition at line 517 of file xz.hpp.

References rpnx::compression::xz_codec::lzma_model::length_model::choice, rpnx::compression::xz_codec::lzma_model::length_model::choice2, rpnx::compression::xz_codec::range_decoder::decode_bit(), rpnx::compression::xz_codec::range_decoder::decode_tree(), rpnx::compression::xz_codec::lzma_model::length_model::high, rpnx::compression::xz_codec::lzma_model::length_model::low, and rpnx::compression::xz_codec::lzma_model::length_model::mid.

Referenced by decode_lzma_chunk().

decode_lzma_chunk()

void rpnx::compression::xz_codec::decode_lzma_chunk ( std::span< std::byte const > encoded, std::size_t uncompressed_size, lzma_model & model, std::vector< std::byte > & output, std::size_t history_begin, std::uint32_t dictionary_size, std::size_t output_limit, std::size_t dictionary_position_offset = 0U )

inline

Decodes one LZMA range-coded chunk into an LZMA2 dictionary.

Parameters

encoded	Complete range-coded chunk.
uncompressed_size	Exact number of bytes the chunk must produce.
model	Adaptive model retained across eligible chunks.
output	Stream output and dictionary storage.
history_begin	Start of the current LZMA2 dictionary history.
dictionary_size	Declared maximum dictionary size.
output_limit	Absolute output-size limit.
dictionary_position_offset	Position adjustment for chunk-local buffers.

Definition at line 567 of file xz.hpp.

References copy_match(), rpnx::compression::xz_codec::range_decoder::decode_bit(), rpnx::compression::xz_codec::range_decoder::decode_direct(), decode_length(), rpnx::compression::xz_codec::range_decoder::decode_reverse_tree(), rpnx::compression::xz_codec::range_decoder::decode_tree(), rpnx::compression::xz_codec::lzma_model::distance_align, rpnx::compression::xz_codec::lzma_model::distance_slot, rpnx::compression::xz_codec::lzma_model::distance_special, rpnx::compression::xz_codec::range_decoder::finish(), rpnx::compression::invalid_data, rpnx::compression::xz_codec::lzma_model::is_match, rpnx::compression::xz_codec::lzma_model::is_rep, rpnx::compression::xz_codec::lzma_model::is_rep0, rpnx::compression::xz_codec::lzma_model::is_rep0_long, rpnx::compression::xz_codec::lzma_model::is_rep1, rpnx::compression::xz_codec::lzma_model::is_rep2, rpnx::compression::xz_codec::lzma_model::literal, rpnx::compression::xz_codec::lzma_model::literal_context_bits, rpnx::compression::xz_codec::lzma_model::literal_position_bits, rpnx::compression::xz_codec::lzma_model::match_length, rpnx::compression::output_limit_exceeded, rpnx::compression::xz_codec::lzma_model::position_bits, rpnx::compression::xz_codec::lzma_model::rep0, rpnx::compression::xz_codec::lzma_model::rep1, rpnx::compression::xz_codec::lzma_model::rep2, rpnx::compression::xz_codec::lzma_model::rep3, rpnx::compression::xz_codec::lzma_model::repeated_length, rpnx::compression::xz_codec::lzma_model::state, update_literal_state(), and rpnx::compression::xz.

Referenced by decompress().

decompress()

template<std::input_iterator input_iterator, std::sentinel_for< input_iterator > sentinel, typename output_iterator>

output_iterator rpnx::compression::xz_codec::decompress	(	input_iterator	first,
		sentinel	last,
		output_iterator	output,
		decompression_options const &	options )

Decompresses one or more xz streams containing LZMA2.

Template Parameters

input_iterator	Single-pass byte iterator.
sentinel	Sentinel for first.
output_iterator	Destination byte iterator.

Parameters

first	First compressed byte.
last	Sentinel past the compressed input.
output	Destination iterator.
options	Output limit and concatenated-stream policy.

Returns

Destination advanced past the uncompressed data.

Definition at line 925 of file xz.hpp.

References rpnx::compression::decompression_options::allow_concatenated_streams, check_size(), crc32(), decode_dictionary_size(), decode_lzma_chunk(), rpnx::compression::implementation::byte_reader< input_iterator, sentinel >::empty(), rpnx::compression::invalid_data, rpnx::compression::decompression_options::maximum_output_size, rpnx::compression::output_limit_exceeded, rpnx::compression::implementation::byte_reader< input_iterator, sentinel >::peek(), rpnx::compression::implementation::byte_reader< input_iterator, sentinel >::read(), read_little_endian(), read_variable_integer(), rpnx::compression::xz_codec::lzma_model::reset(), rpnx::compression::xz_codec::lzma_model::set_properties(), rpnx::compression::trailing_data, rpnx::compression::unsupported_feature, rpnx::compression::implementation::crc32_accumulator::update(), rpnx::compression::xz_codec::crc64_accumulator::update(), rpnx::compression::implementation::crc32_accumulator::value(), rpnx::compression::xz_codec::crc64_accumulator::value(), rpnx::compression::implementation::write_byte(), and rpnx::compression::xz.

read_little_endian()

std::uint64_t rpnx::compression::xz_codec::read_little_endian ( std::span< std::byte const > input, std::size_t & position, std::uint8_t byte_count )

inlinenodiscard

Reads a fixed-width little-endian integer.

Parameters

input	Source bytes.
position	Current position, advanced by byte_count.
byte_count	Width from zero through eight bytes.

Returns

Decoded value.

Definition at line 435 of file xz.hpp.

References rpnx::compression::invalid_data, and rpnx::compression::xz.

Referenced by decompress().

read_variable_integer()

std::uint64_t rpnx::compression::xz_codec::read_variable_integer ( std::span< std::byte const > input, std::size_t & position )

inlinenodiscard

Decodes one minimal xz variable-length integer.

Parameters

input	Source bytes.
position	Current position, advanced past the encoded integer.

Returns

Decoded unsigned value.

Exceptions

compression_error

If the integer is truncated, non-minimal, or oversized.

Definition at line 456 of file xz.hpp.

References rpnx::compression::invalid_data, and rpnx::compression::xz.

Referenced by decompress().

update_literal_state()

void rpnx::compression::xz_codec::update_literal_state ( std::uint8_t & state )

inlinenoexcept

Updates an LZMA state after decoding a literal.

Parameters

state

State-machine value updated in place.

Definition at line 494 of file xz.hpp.

Referenced by compress_lzma_literals(), and decode_lzma_chunk().

Variable Documentation

literal_coder_count

std::size_t rpnx::compression::xz_codec::literal_coder_count = 16U

inlineconstexpr

Number of literal contexts retained by the supported properties.

Definition at line 39 of file xz.hpp.

literal_coder_size

std::size_t rpnx::compression::xz_codec::literal_coder_size = 0x300U

inlineconstexpr

Probability count in one LZMA literal coder.

Definition at line 41 of file xz.hpp.

position_state_count

std::size_t rpnx::compression::xz_codec::position_state_count = 16U

inlineconstexpr

Maximum number of position states.

Definition at line 37 of file xz.hpp.

probability_move_bits

std::uint32_t rpnx::compression::xz_codec::probability_move_bits = 5U

inlineconstexpr

Adaptation shift applied after every probability decision.

Definition at line 31 of file xz.hpp.

Referenced by rpnx::compression::xz_codec::range_decoder::decode_bit(), and rpnx::compression::xz_codec::range_encoder::encode_bit().

probability_total

std::uint32_t rpnx::compression::xz_codec::probability_total = 1U << 11U

inlineconstexpr

LZMA probability-model total.

Definition at line 29 of file xz.hpp.

Referenced by rpnx::compression::xz_codec::range_decoder::decode_bit(), and rpnx::compression::xz_codec::range_encoder::encode_bit().

range_top

std::uint32_t rpnx::compression::xz_codec::range_top = 1U << 24U

inlineconstexpr

Range threshold below which the arithmetic coder normalizes.

Definition at line 33 of file xz.hpp.

Referenced by rpnx::compression::xz_codec::range_encoder::encode_bit().

state_count

std::size_t rpnx::compression::xz_codec::state_count = 12U

inlineconstexpr

Number of LZMA state-machine states.

Definition at line 35 of file xz.hpp.