diff options
| author | Lasse Collin <lasse.collin@tukaani.org> | 2007-12-09 00:42:33 +0200 |
|---|---|---|
| committer | Lasse Collin <lasse.collin@tukaani.org> | 2007-12-09 00:42:33 +0200 |
| commit | 5d018dc03549c1ee4958364712fb0c94e1bf2741 (patch) | |
| tree | 1b211911fb33fddb3f04b77f99e81df23623ffc4 /src/liblzma/api/lzma | |
| download | xz-5d018dc03549c1ee4958364712fb0c94e1bf2741.tar.xz | |
Imported to git.
Diffstat (limited to '')
24 files changed, 3640 insertions, 0 deletions
diff --git a/src/liblzma/api/lzma.h b/src/liblzma/api/lzma.h new file mode 100644 index 00000000..186ae12c --- /dev/null +++ b/src/liblzma/api/lzma.h @@ -0,0 +1,122 @@ +/** + * \file lzma.h + * \brief The public API of liblzma + * + * liblzma is a LZMA compression library with a zlib-like API. + * liblzma is based on LZMA SDK found from http://7-zip.org/sdk.html. + * + * \author Copyright (C) 1999-2006 Igor Pavlov + * \author Copyright (C) 2007 Lasse Collin + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + */ + +#ifndef LZMA_H +#define LZMA_H + +/******************** + * External headers * + ********************/ + +/* size_t */ +#include <sys/types.h> + +/* NULL */ +#include <stddef.h> + +/* uint8_t, uint32_t, uint64_t, UINT32_C, UINT64_C, UINT64_MAX. */ +#include <inttypes.h> + + +/****************** + * GCC extensions * + ******************/ + +/* + * GCC extensions are used conditionally in the public API. It doesn't + * break anything if these are sometimes enabled and sometimes not, only + * affects warnings and optimizations. + */ +#if defined(__GNUC__) && __GNUC__ >= 3 +# ifndef lzma_attribute +# define lzma_attribute(attr) __attribute__(attr) +# endif +# ifndef lzma_restrict +# define lzma_restrict __restrict__ +# endif +#else +# ifndef lzma_attribute +# define lzma_attribute(attr) +# endif +# ifndef lzma_restrict +# define lzma_restrict +# endif +#endif + + +/************** + * Subheaders * + **************/ + +#ifdef __cplusplus +extern "C" { +#endif + +/* + * Subheaders check that this is defined. It is to prevent including + * them directly from applications. + */ +#define LZMA_H_INTERNAL 1 + +/* Basic features */ +#include "lzma/init.h" +#include "lzma/base.h" +#include "lzma/vli.h" +#include "lzma/filter.h" +#include "lzma/check.h" + +/* Filters */ +#include "lzma/copy.h" +#include "lzma/subblock.h" +#include "lzma/simple.h" +#include "lzma/delta.h" +#include "lzma/lzma.h" + +/* Container formats and Metadata */ +#include "lzma/block.h" +#include "lzma/index.h" +#include "lzma/extra.h" +#include "lzma/metadata.h" +#include "lzma/stream.h" +#include "lzma/alone.h" +#include "lzma/raw.h" +#include "lzma/auto.h" + +/* Advanced features */ +#include "lzma/info.h" +#include "lzma/alignment.h" +#include "lzma/stream_flags.h" +#include "lzma/memlimit.h" + +/* Version number */ +#include "lzma/version.h" + +/* + * All subheaders included. Undefine LZMA_H_INTERNAL to prevent applications + * re-including the subheaders. + */ +#undef LZMA_H_INTERNAL + +#ifdef __cplusplus +} +#endif + +#endif /* ifndef LZMA_H */ diff --git a/src/liblzma/api/lzma/alignment.h b/src/liblzma/api/lzma/alignment.h new file mode 100644 index 00000000..6672656c --- /dev/null +++ b/src/liblzma/api/lzma/alignment.h @@ -0,0 +1,60 @@ +/** + * \file lzma/alignment.h + * \brief Calculating input and output alignment of filter chains + * + * \author Copyright (C) 1999-2006 Igor Pavlov + * \author Copyright (C) 2007 Lasse Collin + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + */ + +#ifndef LZMA_H_INTERNAL +# error Never include this file directly. Use <lzma.h> instead. +#endif + + +/** + * \brief Calculates the preferred alignment of the input data + * + * FIXME desc + */ +extern uint32_t lzma_alignment_input( + const lzma_options_filter *filters, uint32_t guess); + + +/** + * \brief Calculates the alignment of the encoded output + * + * Knowing the alignment of the output data is useful e.g. in the Block + * encoder which tries to align the Compressed Data field optimally. + * + * \param filters Pointer to lzma_options_filter array, whose last + * member must have .id = LZMA_VLI_VALUE_UNKNOWN. + * \param guess The value to return if the alignment of the output + * is the same as the alignment of the input data. + * If you want to always detect this special case, + * this guess to zero; this function never returns + * zero unless guess is zero. + * + * \return In most cases, a small positive integer is returned; + * for optimal use, the encoded output of this filter + * chain should start at on offset that is a multiple of + * the returned integer value. + * + * If the alignment of the output is the same as the input + * data (which this function cannot know), \a guess is + * returned. + * + * If an error occurs (that is, unknown Filter IDs or filter + * options), UINT32_MAX is returned. + */ +extern uint32_t lzma_alignment_output( + const lzma_options_filter *filters, uint32_t guess); diff --git a/src/liblzma/api/lzma/alone.h b/src/liblzma/api/lzma/alone.h new file mode 100644 index 00000000..1a6b8e27 --- /dev/null +++ b/src/liblzma/api/lzma/alone.h @@ -0,0 +1,82 @@ +/** + * \file lzma/alone.h + * \brief Handling of the legacy LZMA_Alone format + * + * \author Copyright (C) 1999-2006 Igor Pavlov + * \author Copyright (C) 2007 Lasse Collin + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + */ + +#ifndef LZMA_H_INTERNAL +# error Never include this file directly. Use <lzma.h> instead. +#endif + + +/** + * \brief Options for files in the LZMA_Alone format + */ +typedef struct { + /** + * \brief Uncompressed Size and usage of End of Payload Marker + * + * In contrast to .lzma Blocks, LZMA_Alone format cannot have both + * uncompressed size field in the header and end of payload marker. + * If you don't know the uncompressed size beforehand, set it to + * LZMA_VLI_VALUE_UNKNOWN and liblzma will embed end of payload + * marker. + */ + lzma_vli uncompressed_size; + + /** + * \brief LZMA options + * + * The LZMA_Alone format supports only one filter: the LZMA filter. + * + * \note There exists also an undocumented variant of the + * LZMA_Alone format, which uses the x86 filter in + * addition to LZMA. This format was never supported + * by LZMA Utils and is not supported by liblzma either. + */ + lzma_options_lzma lzma; + +} lzma_options_alone; + + +/** + * \brief Initializes LZMA_Alone encoder + * + * LZMA_Alone files have the suffix .lzma like the .lzma Stream files. + * LZMA_Alone format supports only one filter, the LZMA filter. There is + * no support for integrity checks like CRC32. + * + * Use this format if and only if you need to create files readable by + * legacy LZMA tools. + * + * LZMA_Alone encoder doesn't support LZMA_SYNC_FLUSH or LZMA_FULL_FLUSH. + * + * \return - LZMA_OK + * - LZMA_MEM_ERROR + * - LZMA_PROG_ERROR + */ +extern lzma_ret lzma_alone_encoder( + lzma_stream *strm, const lzma_options_alone *options); + + +/** + * \brief Initializes decoder for LZMA_Alone file + * + * The LZMA_Alone decoder supports LZMA_SYNC_FLUSH. + * + * \return - LZMA_OK + * - LZMA_MEM_ERROR + */ +extern lzma_ret lzma_alone_decoder(lzma_stream *strm); diff --git a/src/liblzma/api/lzma/auto.h b/src/liblzma/api/lzma/auto.h new file mode 100644 index 00000000..327e726f --- /dev/null +++ b/src/liblzma/api/lzma/auto.h @@ -0,0 +1,41 @@ +/** + * \file lzma/auto.h + * \brief Decoder with automatic file format detection + * + * \author Copyright (C) 1999-2006 Igor Pavlov + * \author Copyright (C) 2007 Lasse Collin + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + */ + +#ifndef LZMA_H_INTERNAL +# error Never include this file directly. Use <lzma.h> instead. +#endif + + +/** + * \brief Decode .lzma Streams and LZMA_Alone files with autodetection + * + * Autodetects between the .lzma Stream and LZMA_Alone formats, and + * calls lzma_stream_decoder_init() or lzma_alone_decoder_init() once + * the type of the file has been detected. + * + * \param strm Pointer to propertily prepared lzma_stream + * \param header Pointer to hold a pointer to Extra Records read + * from the Header Metadata Block. Use NULL if + * you don't care about Extra Records. + * \param footer Same as header, but for Footer Metadata Block. + * + * \return - LZMA_OK: Initialization was successful. + * - LZMA_MEM_ERROR: Cannot allocate memory. + */ +extern lzma_ret lzma_auto_decoder(lzma_stream *strm, + lzma_extra **header, lzma_extra **footer); diff --git a/src/liblzma/api/lzma/base.h b/src/liblzma/api/lzma/base.h new file mode 100644 index 00000000..53cf89f5 --- /dev/null +++ b/src/liblzma/api/lzma/base.h @@ -0,0 +1,410 @@ +/** + * \file lzma/base.h + * \brief Data types and functions used in many places of the public API + * + * \author Copyright (C) 1999-2006 Igor Pavlov + * \author Copyright (C) 2007 Lasse Collin + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + */ + +#ifndef LZMA_H_INTERNAL +# error Never include this file directly. Use <lzma.h> instead. +#endif + + +/** + * \brief Boolean + * + * This is here because C89 doesn't have stdbool.h. To set a value for + * variables having type lzma_bool, you can use + * - C99's `true' and `false' from stdbool.h; + * - C++'s internal `true' and `false'; or + * - integers one (true) and zero (false). + */ +typedef unsigned char lzma_bool; + + +/** + * \brief Return values used by several functions in liblzma + * + * Check the descriptions of specific functions to find out which return + * values they can return and the exact meanings of the values in every + * situation. The descriptions given here are only suggestive. + */ +typedef enum { + LZMA_OK = 0, + /**< + * \brief Operation completed successfully + */ + + LZMA_STREAM_END = 1, + /**< + * \brief End of stream was reached + * + * The application should pick the last remaining output + * bytes from strm->next_out. + */ + + LZMA_PROG_ERROR = -2, + /**< + * \brief Programming error + * + * This indicates that the arguments given to the function are + * invalid or the internal state of the decoder is corrupt. + * - Function arguments are invalid or the structures + * pointed by the argument pointers are invalid + * e.g. if strm->next_out has been set to NULL and + * strm->avail_out > 0 when calling lzma_code(). + * - lzma_* functions have been called in wrong order + * e.g. lzma_code() was called right after lzma_end(). + * - If errors occur randomly, the reason might be flaky + * hardware. + * + * If you think that your code is correct, this error code + * can be a sign of a bug in liblzma. See the documentation + * how to report bugs. + */ + + LZMA_DATA_ERROR = -3, + /**< + * \brief Data is corrupt + * + * - Encoder: The input size doesn't match the uncompressed + * size given to lzma_*_encoder_init(). + * - Decoder: The input is corrupt. This includes corrupted + * header, corrupted compressed data, and unmatching + * integrity Check. + * + * \todo What can be done if encoder returns this? + * Probably can continue by fixing the input + * amount, but make sure. + */ + + LZMA_MEM_ERROR = -4, + /**< + * \brief Cannot allocate memory + * + * Memory allocation failed. + */ + + LZMA_BUF_ERROR = -5, + /**< + * \brief No progress is possible + * + * This may happen when avail_in or avail_out is zero. + * + * \note This error is not fatal. Coding can continue + * normally once the reason for this error has + * been fixed. + */ + + LZMA_HEADER_ERROR = -6, + /**< + * \brief Invalid or unsupported header + * + * Invalid or unsupported options, for example + * - unsupported filter(s) or filter options; or + * - reserved bits set in headers (decoder only). + * + * Rebuilding liblzma with more features enabled, or + * upgrading to a newer version of liblzma may help. + */ + + LZMA_UNSUPPORTED_CHECK = -7, + /**< + * \brief Check type is unknown + * + * The type of Check is not supported, and thus the Check + * cannot be calculated. In the encoder, this is an error. + * In the decoder, this is only a warning and decoding can + * still proceed normally (but the Check is ignored). + */ +} lzma_ret; + + +/** + * \brief The `action' argument for lzma_code() + */ +typedef enum { + LZMA_RUN = 0, + /**< + * Encoder: Encode as much input as possible. Some internal + * buffering will probably be done (depends on the filter + * chain in use), which causes latency: the input used won't + * usually be decodeable from the output of the same + * lzma_code() call. + * + * Decoder: Decode as much input as possible and produce as + * much output as possible. This action provides best + * throughput, but may introduce latency, because the + * decoder may decode more data into its internal buffers + * than that fits into next_out. + */ + + LZMA_SYNC_FLUSH = 1, + /**< + * Encoder: Makes all the data given to liblzma via next_in + * available in next_out without resetting the filters. Call + * lzma_code() with LZMA_SYNC_FLUSH until it returns + * LZMA_STREAM_END. Then continue encoding normally. + * + * \note Synchronous flushing is supported only by + * some filters. Some filters support it only + * partially. + * + * Decoder: Asks the decoder to decode only as much as is + * needed to fill next_out. This decreases latency with some + * filters, but is likely to decrease also throughput. It is + * a good idea to use this flag only when it is likely that + * you don't need more output soon. + * + * \note With decoder, this is not comparable to + * zlib's Z_SYNC_FLUSH. + */ + + LZMA_FULL_FLUSH = 2, + /**< + * Finishes encoding of the current Data Block. All the input + * data going to the current Data Block must have been given + * to the encoder (the last bytes can still be pending in + * next_in). Call lzma_code() with LZMA_FULL_FLUSH until + * it returns LZMA_STREAM_END. Then continue normally with + * LZMA_RUN or finish the Stream with LZMA_FINISH. + * + * This action is supported only by Multi-Block Stream + * encoder. If there is no unfinished Data Block, no empty + * Data Block is created. + */ + + LZMA_FINISH = 3 + /**< + * Finishes the encoding operation. All the input data must + * have been given to the encoder (the last bytes can still + * be pending in next_in). Call lzma_code() with LZMA_FINISH + * until it returns LZMA_STREAM_END. + * + * This action is not supported by decoders. + */ +} lzma_action; + + +/** + * \brief Custom functions for memory handling + * + * A pointer to lzma_allocator may be passed via lzma_stream structure + * to liblzma. The library will use these functions for memory handling + * instead of the default malloc() and free(). + * + * liblzma doesn't make an internal copy of lzma_allocator. Thus, it is + * OK to change these function pointers in the middle of the coding + * process, but obviously it must be done carefully to make sure that the + * replacement `free' can deallocate memory allocated by the earlier + * `alloc' function(s). + */ +typedef struct { + /** + * \brief Pointer to custom memory allocation function + * + * Set this to point to your custom memory allocation function. + * It can be useful for example if you want to limit how much + * memory liblzma is allowed to use: for this, you may use + * a pointer to lzma_memory_alloc(). + * + * If you don't want a custom allocator, but still want + * custom free(), set this to NULL and liblzma will use + * the standard malloc(). + * + * \param opaque lzma_allocator.opaque (see below) + * \param nmemb Number of elements like in calloc(). + * liblzma will always set nmemb to 1. + * This argument exists only for + * compatibility with zlib and libbzip2. + * \param size Size of an element in bytes. + * liblzma never sets this to zero. + * + * \return Pointer to the beginning of a memory block of + * size nmemb * size, or NULL if allocation fails + * for some reason. When allocation fails, functions + * of liblzma return LZMA_MEM_ERROR. + */ + void *(*alloc)(void *opaque, size_t nmemb, size_t size); + + /** + * \brief Pointer to custom memory freeing function + * + * Set this to point to your custom memory freeing function. + * If lzma_memory_alloc() is used as allocator, this should + * be set to lzma_memory_free(). + * + * If you don't want a custom freeing function, but still + * want a custom allocator, set this to NULL and liblzma + * will use the standard free(). + * + * \param opaque lzma_allocator.opaque (see below) + * \param ptr Pointer returned by + * lzma_allocator.alloc(), or when it + * is set to NULL, a pointer returned + * by the standard malloc(). + */ + void (*free)(void *opaque, void *ptr); + + /** + * \brief Pointer passed to .alloc() and .free() + * + * opaque is passed as the first argument to lzma_allocator.alloc() + * and lzma_allocator.free(). This intended to ease implementing + * custom memory allocation functions for use with liblzma. + * + * When using lzma_memory_alloc() and lzma_memory_free(), opaque + * must point to lzma_memory_limitter structure allocated and + * initialized with lzma_memory_limitter_create(). + * + * If you don't need this, you should set it to NULL. + */ + void *opaque; + +} lzma_allocator; + + +/** + * \brief Internal data structure + * + * The contents of this structure is not visible outside the library. + */ +typedef struct lzma_internal_s lzma_internal; + + +/** + * \brief Passing data to and from liblzma + * + * The lzma_stream structure is used for + * - passing pointers to input and output buffers to liblzma; + * - defining custom memory hander functions; and + * - holding a pointer to coder-specific internal data structures. + * + * Before calling any of the lzma_*_init() functions the first time, + * the application must reset lzma_stream to LZMA_STREAM_INIT. The + * lzma_*_init() function will verify the options, allocate internal + * data structures and store pointer to them into `internal'. Finally + * total_in and total_out are reset to zero. In contrast to zlib, + * next_in and avail_in are ignored by the initialization functions. + * + * The actual coding is done with the lzma_code() function. Application + * must update next_in, avail_in, next_out, and avail_out between + * calls to lzma_decode() just like with zlib. + * + * In contrast to zlib, even the decoder requires that there always + * is at least one byte space in next_out; if avail_out == 0, + * LZMA_BUF_ERROR is returned immediatelly. This shouldn't be a problem + * for most applications that already use zlib, but it's still worth + * checking your application. + * + * Application may modify values of total_in and total_out as it wants. + * They are updated by liblzma to match the amount of data read and + * written, but liblzma doesn't use the values internally. + * + * Application must not touch the `internal' pointer. + */ +typedef struct { + uint8_t *next_in; /**< Pointer to the next input byte. */ + size_t avail_in; /**< Number of available input bytes in next_in. */ + uint64_t total_in; /**< Total number of bytes read by liblzma. */ + + uint8_t *next_out; /**< Pointer to the next output position. */ + size_t avail_out; /**< Amount of free space in next_out. */ + uint64_t total_out; /**< Total number of bytes written by liblzma. */ + + /** + * Custom memory allocation functions. Set to NULL to use + * the standard malloc() and free(). + */ + lzma_allocator *allocator; + + /** Internal state is not visible to outsiders. */ + lzma_internal *internal; + +} lzma_stream; + + +/** + * \brief Initialization for lzma_stream + * + * When you declare an instance of lzma_stream, you can immediatelly + * initialize it so that initialization functions know that no memory + * has been allocated yet: + * + * lzma_stream strm = LZMA_STREAM_INIT; + */ +#define LZMA_STREAM_INIT { NULL, 0, 0, NULL, 0, 0, NULL, NULL } + + +/** + * \brief Initialization for lzma_stream + * + * This is like LZMA_STREAM_INIT, but this can be used when the lzma_stream + * has already been allocated: + * + * lzma_stream *strm = malloc(sizeof(lzma_stream)); + * if (strm == NULL) + * return LZMA_MEM_ERROR; + * *strm = LZMA_STREAM_INIT_VAR; + */ +extern const lzma_stream LZMA_STREAM_INIT_VAR; + + +/** + * \brief Encodes or decodes data + * + * Once the lzma_stream has been successfully initialized (e.g. with + * lzma_stream_encoder_single()), the actual encoding or decoding is + * done using this function. + * + * \return Some coders may have more exact meaning for different return + * values, which are mentioned separately in the description of + * the initialization functions. Here are the typical meanings: + * - LZMA_OK: So far all good. + * - LZMA_STREAM_END: + * - Encoder: LZMA_SYNC_FLUSH, LZMA_FULL_FLUSH, or + * LZMA_FINISH completed. + * - Decoder: End of uncompressed data was reached. + * - LZMA_BUF_ERROR: Unable to progress. Provide more input or + * output space, and call this function again. This cannot + * occur if both avail_in and avail_out were non-zero (or + * there's a bug in liblzma). + * - LZMA_MEM_ERROR: Unable to allocate memory. Due to lazy + * programming, the coding cannot continue even if the + * application could free more memory. The next call must + * be lzma_end() or some initialization function. + * - LZMA_DATA_ERROR: + * - Encoder: Filter(s) cannot process the given data. + * - Decoder: Compressed data is corrupt. + * - LZMA_HEADER_ERROR: Unsupported options. Rebuilding liblzma + * with more features enabled or upgrading to a newer version + * may help, although usually this is a sign of invalid options + * (encoder) or corrupted input data (decoder). + * - LZMA_PROG_ERROR: Invalid arguments or the internal state + * of the coder is corrupt. + */ +extern lzma_ret lzma_code(lzma_stream *strm, lzma_action action); + + +/** + * \brief Frees memory allocated for the coder data structures + * + * \param strm Pointer to lzma_stream that is at least initialized + * with LZMA_STREAM_INIT. + * + * \note zlib indicates an error if application end()s unfinished + * stream. liblzma doesn't do this, and assumes that + * application knows what it is doing. + */ +extern void lzma_end(lzma_stream *strm); diff --git a/src/liblzma/api/lzma/block.h b/src/liblzma/api/lzma/block.h new file mode 100644 index 00000000..210c1d87 --- /dev/null +++ b/src/liblzma/api/lzma/block.h @@ -0,0 +1,409 @@ +/** + * \file lzma/block.h + * \brief .lzma Block handling + * + * \author Copyright (C) 1999-2006 Igor Pavlov + * \author Copyright (C) 2007 Lasse Collin + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + */ + +#ifndef LZMA_H_INTERNAL +# error Never include this file directly. Use <lzma.h> instead. +#endif + + +/** + * \brief Options for the Block Header encoder and decoder + * + * Different things use different parts of this structure. Some read + * some members, other functions write, and some do both. Only the + * members listed for reading need to be initialized when the specified + * functions are called. The members marked for writing will be assigned + * new values at some point either by calling the given function or by + * later calls to lzma_code(). + */ +typedef struct { + /** + * \brief Type of integrity Check + * + * The type of the integrity Check is not stored into the Block + * Header, thus its value must be provided also when decoding. + * + * Read by: + * - lzma_block_encoder() + * - lzma_block_decoder() + */ + lzma_check_type check; + + /** + * \brief Precense of CRC32 of the Block Header + * + * Set this to true if CRC32 of the Block Header should be + * calculated and stored in the Block Header. + * + * There is no way to autodetect if CRC32 is present in the Block + * Header, thus this information must be provided also when decoding. + * + * Read by: + * - lzma_block_header_size() + * - lzma_block_header_encoder() + * - lzma_block_header_decoder() + */ + lzma_bool has_crc32; + + /** + * \brief Usage of End of Payload Marker + * + * If this is true, End of Payload Marker is used even if + * Uncompressed Size is known. + * + * Read by: + * - lzma_block_header_encoder() + * - lzma_block_encoder() + * - lzma_block_decoder() + * + * Written by: + * - lzma_block_header_decoder() + */ + lzma_bool has_eopm; + + /** + * \brief True if the Block is a Metadata Block + * + * If this is true, the Metadata bit will be set in the Block Header. + * It is up to the application to store correctly formatted data + * into Metadata Block. + * + * Read by: + * - lzma_block_header_encoder() + * + * Written by: + * - lzma_block_header_decoder() + */ + lzma_bool is_metadata; + + /** + * \brief True if Uncompressed Size is in Block Footer + * + * Read by: + * - lzma_block_encoder() + * - lzma_block_decoder() + */ + lzma_bool has_uncompressed_size_in_footer; + + /** + * \brief True if Backward Size is in Block Footer + * + * Read by: + * - lzma_block_encoder() + * - lzma_block_decoder() + */ + lzma_bool has_backward_size; + + /** + * \brief True if Block coder should take care of Padding + * + * In liblzma, Stream decoder sets this to true when decoding + * Header Metadata Block or Data Blocks from Multi-Block Stream, + * and to false when decoding Single-Block Stream or Footer + * Metadata Block from a Multi-Block Stream. + * + * Read by: + * - lzma_block_encoder() + * - lzma_block_decoder() + */ + lzma_bool handle_padding; + + /** + * \brief Size of the Compressed Data in bytes + * + * Usually you don't know this value when encoding in streamed mode. + * In non-streamed mode you can reserve space for this field when + * encoding the Block Header the first time, and then re-encode the + * Block Header and copy it over the original one after the encoding + * of the Block has been finished. + * + * Read by: + * - lzma_block_header_size() + * - lzma_block_header_encoder() + * - lzma_block_encoder() + * - lzma_block_decoder() + * + * Written by: + * - lzma_block_header_decoder() + * - lzma_block_encoder() + * - lzma_block_decoder() + */ + lzma_vli compressed_size; + + /** + * \brief Uncompressed Size in bytes + * + * Encoder: If this value is not LZMA_VLI_VALUE_UNKNOWN, it is stored + * to the Uncompressed Size field in the Block Header. The real + * uncompressed size of the data being compressed must match + * the Uncompressed Size or LZMA_HEADER_ERROR is returned. + * + * If Uncompressed Size is unknown, End of Payload Marker must + * be used. If uncompressed_size == LZMA_VLI_VALUE_UNKNOWN and + * has_eopm == 0, LZMA_HEADER_ERROR will be returned. + * + * Decoder: If this value is not LZMA_VLI_VALUE_UNKNOWN, it is + * compared to the real Uncompressed Size. If they do not match, + * LZMA_HEADER_ERROR is returned. + * + * Read by: + * - lzma_block_header_size() + * - lzma_block_header_encoder() + * - lzma_block_encoder() + * - lzma_block_decoder() + * + * Written by: + * - lzma_block_header_decoder() + * - lzma_block_encoder() + * - lzma_block_decoder() + */ + lzma_vli uncompressed_size; + + /** + * \brief Number of bytes to reserve for Compressed Size + * + * This is useful if you want to be able to store the Compressed Size + * to the Block Header, but you don't know it when starting to encode. + * Setting this to non-zero value at maximum of LZMA_VLI_BYTES_MAX, + * the Block Header encoder will force the Compressed Size field to + * occupy specified number of bytes. You can later rewrite the Block + * Header to contain correct information by using otherwise identical + * lzma_options_block structure except the correct compressed_size. + * + * Read by: + * - lzma_block_header_size() + * - lzma_block_header_encoder() + * + * Written by: + * - lzma_block_header_decoder() + */ + uint32_t compressed_reserve; + + /** + * \brief Number of bytes to reserve for Uncompressed Size + * + * See the description of compressed_size above. + * + * Read by: + * - lzma_block_header_size() + * - lzma_block_header_encoder() + * + * Written by: + * - lzma_block_header_decoder() + */ + uint32_t uncompressed_reserve; + + /** + * \brief Total Size of the Block in bytes + * + * This is useful in the decoder, which can verify the Total Size + * if it is known from Index. + * + * Read by: + * - lzma_block_encoder() + * - lzma_block_decoder() + * + * Written by: + * - lzma_block_encoder() + * - lzma_block_decoder() + */ + lzma_vli total_size; + + /** + * \brief Upper limit of Total Size + * + * Read by: + * - lzma_block_encoder() + * - lzma_block_decoder() + */ + lzma_vli total_limit; + + /** + * \brief Upper limit of Uncompressed Size + * + * Read by: + * - lzma_block_encoder() + * - lzma_block_decoder() + */ + lzma_vli uncompressed_limit; + + /** + * \brief Array of filters + * + * There can be at maximum of seven filters. The end of the array + * is marked with .id = LZMA_VLI_VALUE_UNKNOWN. Minimum number of + * filters is zero; in that case, an implicit Copy filter is used. + * + * Read by: + * - lzma_block_header_size() + * - lzma_block_header_encoder() + * - lzma_block_encoder() + * - lzma_block_decoder() + * + * Written by: + * - lzma_block_header_decoder(): Note that this does NOT free() + * the old filter options structures. If decoding fails, the + * caller must take care of freeing the options structures + * that may have been allocated and decoded before the error + * occurred. + */ + lzma_options_filter filters[8]; + + /** + * \brief Size of the Padding field + * + * The Padding field exist to allow aligning the Compressed Data field + * optimally in the Block. See lzma_options_stream.alignment in + * stream.h for more information. + * + * If you want the Block Header encoder to automatically calculate + * optimal size for the Padding field by looking at the information + * in filters[], set this to LZMA_BLOCK_HEADER_PADDING_AUTO. In that + * case, you must also set the aligmnet variable to tell the the + * encoder the aligmnet of the beginning of the Block Header. + * + * The decoder never sets this to LZMA_BLOCK_HEADER_PADDING_AUTO. + * + * Read by: + * - lzma_block_header_size() + * - lzma_block_header_encoder(): Note that this doesn't + * accept LZMA_BLOCK_HEADER_PADDING_AUTO. + * + * Written by (these never set padding to + * LZMA_BLOCK_HEADER_PADDING_AUTO): + * - lzma_block_header_size() + * - lzma_block_header_decoder() + */ + int32_t padding; +# define LZMA_BLOCK_HEADER_PADDING_AUTO (-1) +# define LZMA_BLOCK_HEADER_PADDING_MIN 0 +# define LZMA_BLOCK_HEADER_PADDING_MAX 31 + + /** + * \brief Alignment of the beginning of the Block Header + * + * This variable is read only if padding has been set to + * LZMA_BLOCK_HEADER_PADDING_AUTO. + * + * Read by: + * - lzma_block_header_size() + * - lzma_block_header_encoder() + */ + uint32_t alignment; + + /** + * \brief Size of the Block Header + * + * Read by: + * - lzma_block_encoder() + * - lzma_block_decoder() + * + * Written by: + * - lzma_block_header_size() + * - lzma_block_header_decoder() + */ + uint32_t header_size; + +} lzma_options_block; + + +/** + * \brief Calculates the size of Header Padding and Block Header + * + * \return - LZMA_OK: Size calculated successfully and stored to + * options->header_size. + * - LZMA_HEADER_ERROR: Unsupported filters or filter options. + * - LZMA_PROG_ERROR: Invalid options + * + * \note This doesn't check that all the options are valid i.e. this + * may return LZMA_OK even if lzma_block_header_encode() or + * lzma_block_encoder() would fail. + */ +extern lzma_ret lzma_block_header_size(lzma_options_block *options); + + +/** + * \brief Encodes Block Header + * + * Encoding of the Block options is done with a single call instead of + * first initializing and then doing the actual work with lzma_code(). + * + * \param out Beginning of the output buffer. This must be + * at least options->header_size bytes. + * \param options Block options to be encoded. + * + * \return - LZMA_OK: Encoding was successful. options->header_size + * bytes were written to output buffer. + * - LZMA_HEADER_ERROR: Invalid or unsupported options. + * - LZMA_PROG_ERROR + */ +extern lzma_ret lzma_block_header_encode( + uint8_t *out, const lzma_options_block *options); + + +/** + * \brief Initializes Block Header decoder + * + * Because the results of this decoder are placed into *options, + * strm->next_in, strm->avail_in, and strm->total_in are not used. + * + * The only valid `action' with lzma_code() is LZMA_RUN. + * + * \return - LZMA_OK: Encoding was successful. options->header_size + * bytes were written to output buffer. + * - LZMA_HEADER_ERROR: Invalid or unsupported options. + * - LZMA_PROG_ERROR + */ +extern lzma_ret lzma_block_header_decoder( + lzma_stream *strm, lzma_options_block *options); + + +/** + * \brief Initializes .lzma Block encoder + * + * This function is required for multi-thread encoding. It may also be + * useful when implementing custom file formats. + * + * \return - LZMA_OK: All good, continue with lzma_code(). + * - LZMA_MEM_ERROR + * - LZMA_HEADER_ERROR + * - LZMA_DATA_ERROR: Limits (total_limit and uncompressed_limit) + * have been reached already. + * - LZMA_UNSUPPORTED_CHECK: options->check specfies a Check + * that is not supported by this buid of liblzma. Initializing + * the encoder failed. + * - LZMA_PROG_ERROR + * + * lzma_code() can return FIXME + */ +extern lzma_ret lzma_block_encoder( + lzma_stream *strm, lzma_options_block *options); + + +/** + * \brief Initializes decoder for .lzma Block + * + * \return - LZMA_OK: All good, continue with lzma_code(). + * - LZMA_UNSUPPORTED_CHECK: Initialization was successful, but + * the given Check type is not supported, thus Check will be + * ignored. + * - LZMA_PROG_ERROR + * - LZMA_MEM_ERROR + */ +extern lzma_ret lzma_block_decoder( + lzma_stream *strm, lzma_options_block *options); diff --git a/src/liblzma/api/lzma/check.h b/src/liblzma/api/lzma/check.h new file mode 100644 index 00000000..4a2a453b --- /dev/null +++ b/src/liblzma/api/lzma/check.h @@ -0,0 +1,128 @@ +/** + * \file lzma/check.h + * \brief Integrity checks + * + * \author Copyright (C) 1999-2006 Igor Pavlov + * \author Copyright (C) 2007 Lasse Collin + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + */ + +#ifndef LZMA_H_INTERNAL +# error Never include this file directly. Use <lzma.h> instead. +#endif + + +/** + * \brief Type of the Check + * + * The .lzma format supports multiple types of Checks that are calculated + * from the uncompressed data (unless it is empty; then it's calculated + * from Block Header). + */ +typedef enum { + LZMA_CHECK_NONE = 0, + /**< + * No Check is calculated. + * + * Size of the Check field: 0 bytes + */ + + LZMA_CHECK_CRC32 = 1, + /**< + * CRC32 using the polynomial from the IEEE 802.3 standard + * + * Size of the Check field: 4 bytes + */ + + LZMA_CHECK_CRC64 = 3, + /**< + * CRC64 using the polynomial from the ECMA-182 standard + * + * Size of the Check field: 8 bytes + */ + + LZMA_CHECK_SHA256 = 5 + /**< + * SHA-256 + * + * Size of the Check field: 32 bytes + */ +} lzma_check_type; + + +/** + * \brief Maximum valid Check ID + * + * The .lzma file format specification specifies eight Check IDs (0-7). Some + * of them are only reserved i.e. no actual Check algorithm has been assigned. + * Still liblzma accepts any of these eight IDs for future compatibility + * when decoding files. If a valid but unsupported Check ID is detected, + * liblzma indicates a warning with LZMA_UNSUPPORTED_CHECK. + * + * FIXME bad desc + */ +#define LZMA_CHECK_ID_MAX 7 + + +/** + * \brief Check IDs supported by this liblzma build + * + * If lzma_available_checks[n] is true, the Check ID n is supported by this + * liblzma build. You can assume that LZMA_CHECK_NONE and LZMA_CHECK_CRC32 + * are always available. + */ +extern const lzma_bool lzma_available_checks[LZMA_CHECK_ID_MAX + 1]; + + +/** + * \brief Size of the Check field with different Check IDs + * + * Although not all Check IDs have a check algorithm associated, the size of + * every Check is already frozen. This array contains the size (in bytes) of + * the Check field with specified Check ID. The values are taken from the + * section 2.2.2 of the .lzma file format specification: + * { 0, 4, 4, 8, 16, 32, 32, 64 } + */ +extern const uint32_t lzma_check_sizes[LZMA_CHECK_ID_MAX + 1]; + + +/** + * \brief Calculate CRC32 + * + * Calculates CRC32 using the polynomial from the IEEE 802.3 standard. + * + * \param buf Pointer to the input buffer + * \param size Size of the input buffer + * \param crc Previously returned CRC value. This is used to + * calculate the CRC of a big buffer in smaller chunks. + * Set to zero when there is no previous value. + * + * \return Updated CRC value, which can be passed to this function + * again to continue CRC calculation. + */ +extern uint32_t lzma_crc32(const uint8_t *buf, size_t size, uint32_t crc); + + +/** + * \brief Calculate CRC64 + * + * Calculates CRC64 using the polynomial from the ECMA-182 standard. + * + * This function is used similarly to lzma_crc32(). See its documentation. + */ +extern uint64_t lzma_crc64(const uint8_t *buf, size_t size, uint64_t crc); + + +/* + * SHA256 functions are currently not exported to public API. + * Contact the author if you think it should be. + */ diff --git a/src/liblzma/api/lzma/copy.h b/src/liblzma/api/lzma/copy.h new file mode 100644 index 00000000..f5617462 --- /dev/null +++ b/src/liblzma/api/lzma/copy.h @@ -0,0 +1,29 @@ +/** + * \file lzma/copy.h + * \brief Copy filter + * + * \author Copyright (C) 1999-2006 Igor Pavlov + * \author Copyright (C) 2007 Lasse Collin + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + */ + +#ifndef LZMA_H_INTERNAL +# error Never include this file directly. Use <lzma.h> instead. +#endif + + +/** + * \brief Filter ID + * + * Filter ID of the Copy filter. This is used as lzma_options_filter.id. + */ +#define LZMA_FILTER_COPY LZMA_VLI_C(0x00) diff --git a/src/liblzma/api/lzma/delta.h b/src/liblzma/api/lzma/delta.h new file mode 100644 index 00000000..58afec18 --- /dev/null +++ b/src/liblzma/api/lzma/delta.h @@ -0,0 +1,49 @@ +/** + * \file lzma/delta.h + * \brief Delta filter + * + * \author Copyright (C) 1999-2006 Igor Pavlov + * \author Copyright (C) 2007 Lasse Collin + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + */ + +#ifndef LZMA_H_INTERNAL +# error Never include this file directly. Use <lzma.h> instead. +#endif + + +/** + * \brief Filter ID + * + * Filter ID of the Delta filter. This is used as lzma_options_filter.id. + */ +#define LZMA_FILTER_DELTA LZMA_VLI_C(0x20) + + +/** + * \brief Options for the Delta filter + * + * These options are needed by both encoder and decoder. + */ +typedef struct { + /** + * \brief Delta distance as bytes + * + * Examples: + * - 16-bit stereo audio: distance = 4 bytes + * - 24-bit RGB image data: distance = 3 bytes + */ + uint32_t distance; +# define LZMA_DELTA_DISTANCE_MIN 1 +# define LZMA_DELTA_DISTANCE_MAX 256 + +} lzma_options_delta; diff --git a/src/liblzma/api/lzma/extra.h b/src/liblzma/api/lzma/extra.h new file mode 100644 index 00000000..29426a74 --- /dev/null +++ b/src/liblzma/api/lzma/extra.h @@ -0,0 +1,114 @@ +/** + * \file lzma/extra.h + * \brief Handling of Extra Records in Metadata + * + * \author Copyright (C) 1999-2006 Igor Pavlov + * \author Copyright (C) 2007 Lasse Collin + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + */ + +#ifndef LZMA_H_INTERNAL +# error Never include this file directly. Use <lzma.h> instead. +#endif + + +/* + * Extra Record IDs + * + * See the .lzma file format specification for description what each + * Extra Record type exactly means. + * + * If you ever need to update .lzma files with Extra Records, note that + * the Record IDs are divided in two categories: + * - Safe-to-Copy Records may be preserved as is when the + * Stream is modified in ways that don't change the actual + * uncompressed data. Examples of such operatings include + * recompressing and adding, modifying, or deleting unrelated + * Extra Records. + * - Unsafe-to-Copy Records should be removed (and possibly + * recreated) when any kind of changes are made to the Stream. + */ + +#define LZMA_EXTRA_PADDING 0x00 +#define LZMA_EXTRA_OPENPGP 0x01 +#define LZMA_EXTRA_FILTERS 0x02 +#define LZMA_EXTRA_COMMENT 0x03 +#define LZMA_EXTRA_CHECKS 0x04 +#define LZMA_EXTRA_FILENAME 0x05 +#define LZMA_EXTRA_MTIME 0x07 +#define LZMA_EXTRA_MTIME_HR 0x09 +#define LZMA_EXTRA_MIME_TYPE 0x0B +#define LZMA_EXTRA_HOMEPAGE 0x0D + + +/** + * \brief Extra Records + * + * The .lzma format provides a way to store custom information along + * the actual compressed content. Information about these Records + * are passed to and from liblzma via this linked list. + */ +typedef struct lzma_extra_s lzma_extra; +struct lzma_extra_s { + /** + * \brief Pointer to the next Extra Record + * + * This is NULL on the last Extra Record. + */ + lzma_extra *next; + + /** + * \brief Record ID + * + * Extra Record IDs are divided in three categories: + * - Zero is a special case used for padding. It doesn't have + * Size of Data fields. + * - Odd IDs (1, 3, 5, ...) are Safe-to-Copy IDs. + * These can be preserved as is if the Stream is + * modified in a way that doesn't alter the actual + * uncompressed content. + * - Even IDs (2, 4, 6, ...) are Unsafe-to-Copy IDs. + * If the .lzma Stream is modified in any way, + * the Extra Records having a sensitive ID should + * be removed or updated accordingly. + * + * Refer to the .lzma file format specification for + * the up to date list of Extra Record IDs. + */ + lzma_vli id; + + /** + * \brief Size of the Record data + * + * In case of strings, this should not include the + * trailing '\0'. + */ + size_t size; + + /** + * \brief Record data + * + * Record data is often a string in UTF-8 encoding, + * but it can be arbitrary binary data. In case of + * strings, the trailing '\0' is usually not stored + * in the .lzma file. + * + * To ease working with Extra Records containing strings, + * liblzma always adds '\0' to the end of data even when + * it wasn't present in the .lzma file. This '\0' is not + * counted in the size of the data. + */ + uint8_t *data; +}; + + +extern void lzma_extra_free(lzma_extra *extra, lzma_allocator *allocator); diff --git a/src/liblzma/api/lzma/filter.h b/src/liblzma/api/lzma/filter.h new file mode 100644 index 00000000..a8bdd4bd --- /dev/null +++ b/src/liblzma/api/lzma/filter.h @@ -0,0 +1,166 @@ +/** + * \file lzma/filter.h + * \brief Common filter related types + * + * \author Copyright (C) 1999-2006 Igor Pavlov + * \author Copyright (C) 2007 Lasse Collin + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + */ + +#ifndef LZMA_H_INTERNAL +# error Never include this file directly. Use <lzma.h> instead. +#endif + + +/** + * \brief Filter options + * + * This structure is used to pass Filter ID and a pointer filter's options + * to liblzma. + */ +typedef struct { + /** + * \brief Filter ID + * + * Use constants whose name begin with `LZMA_FILTER_' to specify + * different filters. In an array of lzma_option_filter structures, + * use LZMA_VLI_VALUE_UNKNOWN to indicate end of filters. + */ + lzma_vli id; + + /** + * \brief Pointer to filter-specific options structure + * + * If the filter doesn't need options, set this to NULL. If id is + * set to LZMA_VLI_VALUE_UNKNOWN, options is ignored, and thus + * doesn't need be initialized. + * + * Some filters support changing the options in the middle of + * the encoding process. These filters store the pointer of the + * options structure and communicate with the application via + * modifications of the options structure. + */ + void *options; + +} lzma_options_filter; + + +/** + * \brief Filters available for encoding + * + * Pointer to an array containing the list of available Filter IDs that + * can be used for encoding. The last element is LZMA_VLI_VALUE_UNKNOWN. + * + * If lzma_available_filter_encoders[0] == LZMA_VLI_VALUE_UNKNOWN, the + * encoder components haven't been built at all. This means that the + * encoding-specific functions are probably missing from the library + * API/ABI completely. + */ +extern const lzma_vli *const lzma_available_filter_encoders; + + +/** + * \brief Filters available for decoding + * + * Pointer to an array containing the list of available Filter IDs that + * can be used for decoding. The last element is LZMA_VLI_VALUE_UNKNOWN. + * + * If lzma_available_filter_decoders[0] == LZMA_VLI_VALUE_UNKNOWN, the + * decoder components haven't been built at all. This means that the + * decoding-specific functions are probably missing from the library + * API/ABI completely. + */ +extern const lzma_vli *const lzma_available_filter_decoders; + + +/** + * \brief Calculate rough memory requirements for given filter chain + * + * \param filters Array of filters terminated with + * .id == LZMA_VLI_VALUE_UNKNOWN. + * \param is_encoder Set to true when calculating memory requirements + * of an encoder; false for decoder. + * + * \return Number of mebibytes (MiB i.e. 2^20) required for the given + * encoder or decoder filter chain. + * + * \note If calculating memory requirements of encoder, lzma_init() or + * lzma_init_encoder() must have been called earlier. Similarly, + * if calculating memory requirements of decoder, lzma_init() or + * lzma_init_decoder() must have been called earlier. + */ +extern uint32_t lzma_memory_usage( + const lzma_options_filter *filters, lzma_bool is_encoder); + + +/** + * \brief Calculates encoded size of a Filter Flags field + * + * Knowing the size of Filter Flags is useful to know when allocating + * memory to hold the encoded Filter Flags. + * + * \param size Pointer to integer to hold the calculated size + * \param options Filter ID and associated options whose encoded + * size is to be calculted + * + * \return - LZMA_OK: *size set successfully. Note that this doesn't + * guarantee that options->options is valid, thus + * lzma_filter_flags_encode() may still fail. + * - LZMA_HEADER_ERROR: Unknown Filter ID or unsupported options. + * - LZMA_PROG_ERROR: Invalid options + * + * \note If you need to calculate size of List of Filter Flags, + * you need to loop over every lzma_options_filter entry. + */ +extern lzma_ret lzma_filter_flags_size( + uint32_t *size, const lzma_options_filter *options); + + +/** + * \brief Encodes Filter Flags into given buffer + * + * In contrast to some functions, this doesn't allocate the needed buffer. + * This is due to how this function is used internally by liblzma. + * + * \param out Beginning of the output buffer + * \param out_pos out[*out_pos] is the next write position. This + * is updated by the encoder. + * \param out_size out[out_size] is the first byte to not write. + * \param options Filter options to be encoded + * + * \return - LZMA_OK: Encoding was successful. + * - LZMA_HEADER_ERROR: Invalid or unsupported options. + * - LZMA_PROG_ERROR: Invalid options or not enough output + * buffer space (you should have checked it with + * lzma_filter_flags_size()). + */ +extern lzma_ret lzma_filter_flags_encode(uint8_t *out, size_t *out_pos, + size_t out_size, const lzma_options_filter *options); + + +/** + * \brief Initializes Filter Flags decoder + * + * The decoded result is stored into *options. options->options is + * initialized but the old value is NOT free()d. + * + * Because the results of this decoder are placed into *options, + * strm->next_in, strm->avail_in, and strm->total_in are not used + * when calling lzma_code(). The only valid action for lzma_code() + * is LZMA_RUN + * + * \return - LZMA_OK + * - LZMA_MEM_ERROR + * - LZMA_PROG_ERROR + */ +extern lzma_ret lzma_filter_flags_decoder( + lzma_stream *strm, lzma_options_filter *options); diff --git a/src/liblzma/api/lzma/index.h b/src/liblzma/api/lzma/index.h new file mode 100644 index 00000000..7e59c4b3 --- /dev/null +++ b/src/liblzma/api/lzma/index.h @@ -0,0 +1,84 @@ +/** + * \file lzma/index.h + * \brief Handling of Index lists + * + * \author Copyright (C) 1999-2006 Igor Pavlov + * \author Copyright (C) 2007 Lasse Collin + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + */ + +#ifndef LZMA_H_INTERNAL +# error Never include this file directly. Use <lzma.h> instead. +#endif + + +/** + * \brief + * + * FIXME desc + */ +typedef struct lzma_index_s lzma_index; +struct lzma_index_s { + /** + * \brief Total Size of the Block + * + * This includes Block Header, Compressed Data, and Block Footer. + */ + lzma_vli total_size; + + /** + * \brief Uncompressed Size of the Block + */ + lzma_vli uncompressed_size; + + /** + * \brief Pointer to the next Index Record + * + * This is NULL on the last Index Record. + */ + lzma_index *next; +}; + + +/** + * \brief Duplicates an Index list + * + * \return A copy of the Index list, or NULL if memory allocation + * failed or the original Index was empty. + */ +extern lzma_index *lzma_index_dup( + const lzma_index *index, lzma_allocator *allocator); + + +/** + * \brief Frees an Index list + * + * All Index Recors in the list are freed. This function is convenient when + * getting rid of lzma_metadata structures containing an Index. + */ +extern void lzma_index_free(lzma_index *index, lzma_allocator *allocator); + + +/** + * \brief Calculates information about the Index + * + * \return LZMA_OK on success, LZMA_PROG_ERROR on error. FIXME + */ +extern lzma_ret lzma_index_count(const lzma_index *index, size_t *count, + lzma_vli *lzma_restrict total_size, + lzma_vli *lzma_restrict uncompressed_size); + + +/** + * \brief Compares if two Index lists are identical + */ +extern lzma_bool lzma_index_is_equal(const lzma_index *a, const lzma_index *b); diff --git a/src/liblzma/api/lzma/info.h b/src/liblzma/api/lzma/info.h new file mode 100644 index 00000000..3a91850f --- /dev/null +++ b/src/liblzma/api/lzma/info.h @@ -0,0 +1,315 @@ +/** + * \file lzma/info.h + * \brief Handling of Stream size information + * + * \author Copyright (C) 1999-2006 Igor Pavlov + * \author Copyright (C) 2007 Lasse Collin + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + */ + +#ifndef LZMA_H_INTERNAL +# error Never include this file directly. Use <lzma.h> instead. +#endif + + +/********** + * Basics * + **********/ + +/** + * \brief Opaque data type to hold the size information + */ +typedef struct lzma_info_s lzma_info; + + +typedef struct { + /** + * \brief Total Size of this Block + * + * This can be LZMA_VLI_VALUE_UNKNOWN. + */ + lzma_vli total_size; + + /** + * \brief Uncompressed Size of this Block + * + * This can be LZMA_VLI_VALUE_UNKNOWN. + */ + lzma_vli uncompressed_size; + + /** + * \brief Offset of the first byte of the Block + * + * In encoder, this is useful to find out the alignment of the Block. + * + * In decoder, this is useful when doing random-access reading + * with help from lzma_info_data_locate(). + */ + lzma_vli stream_offset; + + /** + * \brief Uncompressed offset of the Block + * + * Offset of the first uncompressed byte of the Block relative to + * all uncompressed data in the Block. + * FIXME desc + */ + lzma_vli uncompressed_offset; + + /** + * \brief Pointers to internal data structures + * + * Applications must not touch these. + */ + void *internal[4]; + +} lzma_info_iter; + + +typedef enum { + LZMA_INFO_STREAM_START, + LZMA_INFO_HEADER_METADATA, + LZMA_INFO_TOTAL, + LZMA_INFO_UNCOMPRESSED, + LZMA_INFO_FOOTER_METADATA +} lzma_info_size; + + +/** + * \brief Allocates and initializes a new lzma_info structure + * + * If info is NULL, a new lzma_info structure is allocated, initialized, and + * a pointer to it returned. If allocation fails, NULL is returned. + * + * If info is non-NULL, it is reinitialized and the same pointer returned. + * (In this case, return value cannot be NULL or a different pointer than + * the info given as argument.) + */ +extern lzma_info *lzma_info_init(lzma_info *info, lzma_allocator *allocator); + + +/** + * \brief Resets lzma_info + * + * This is like calling lzma_info_end() and lzma_info_create(), but + * re-uses the existing base structure. + */ +extern void lzma_info_reset( + lzma_info *info, lzma_allocator *allocator); + + +/** + * \brief Frees memory allocated for a lzma_info structure + */ +extern void lzma_info_free(lzma_info *info, lzma_allocator *allocator); + + +/************************ + * Setting known values * + ************************/ + +/** + * \brief Set a known size value + * + * \param info Pointer returned by lzma_info_create() + * \param type Any value from lzma_info_size + * \param size Value to set or verify + * + * \return LZMA_OK on success, LZMA_DATA_ERROR if the size doesn't + * match the existing information, or LZMA_PROG_ERROR + * if type is invalid or size is not a valid VLI. + */ +extern lzma_ret lzma_info_size_set( + lzma_info *info, lzma_info_size type, lzma_vli size); + + +/** + * \brief Sets the Index + * + * The given lzma_index list is "absorbed" by this function. The application + * must not access it after this function call, even if this function returns + * an error. + * + * \note The given lzma_index will at some point get freed by the + * lzma_info_* functions. If you use a custom lzma_allocator, + * make sure that it can free the lzma_index. + */ +extern lzma_ret lzma_info_index_set( + lzma_info *info, lzma_allocator *allocator, + lzma_index *index, lzma_bool eat_index); + + +/** + * \brief Sets information from a known Metadata Block + * + * This is a shortcut for calling lzma_info_size_set() with different type + * arguments, lzma_info_index_set() with metadata->index. + */ +extern lzma_ret lzma_info_metadata_set(lzma_info *info, + lzma_allocator *allocator, lzma_metadata *metadata, + lzma_bool is_header_metadata, lzma_bool eat_index); + + +/*************** + * Incremental * + ***************/ + +/** + * \brief Prepares an iterator to be used with given lzma_info structure + * + * + */ +extern void lzma_info_iter_begin(lzma_info *info, lzma_info_iter *iter); + + +/** + * \brief Moves to the next Index Record + * + * + */ +extern lzma_ret lzma_info_iter_next( + lzma_info_iter *iter, lzma_allocator *allocator); + + +/** + * \brief Sets or verifies the sizes in the Index Record + * + * \param iter Pointer to iterator to be set or verified + * \param total_size + * Total Size in bytes or LZMA_VLI_VALUE_UNKNOWN + * \param uncompressed_size + * Uncompressed Size or LZMA_VLI_VALUE_UNKNOWN + * + * \return - LZMA_OK: All OK. + * - LZMA_DATA_ERROR: Given sizes don't match with the already + * known sizes. + * - LZMA_PROG_ERROR: Internal error, possibly integer + * overflow (e.g. the sum of all the known sizes is too big) + */ +extern lzma_ret lzma_info_iter_set(lzma_info_iter *iter, + lzma_vli total_size, lzma_vli uncompressed_size); + + +/** + * \brief Locates a Data Block + * + * \param iter Properly initialized iterator + * \param allocator Pointer to lzma_allocator or NULL + * \param uncompressed_offset + * Target offset to locate. The final offset + * will be equal or smaller than this. + * \param allow_alloc True if this function is allowed to call + * lzma_info_iter_next() to allocate a new Record + * if the requested offset reached end of Index + * Record list. Note that if Index has been marked + * final, lzma_info_iter_next() is never called. + * + * \return - LZMA_OK: All OK, *iter updated accordingly. + * - LZMA_DATA_ERROR: Trying to search past the end of the Index + * Record list, and allocating a new Record was not allowed + * either because allow_alloc was false or Index was final. + * - LZMA_PROG_ERROR: Internal error (probably integer + * overflow causing some lzma_vli getting too big). + */ +extern lzma_ret lzma_info_iter_locate(lzma_info_iter *iter, + lzma_allocator *allocator, lzma_vli uncompressed_offset, + lzma_bool allow_alloc); + + +/** + * \brief Finishes incrementally constructed Index + * + * This sets the known Total Size and Uncompressed of the Data Blocks + * based on the information collected from the Index Records, and marks + * the Index as final. + */ +extern lzma_ret lzma_info_index_finish(lzma_info *info); + + +/*************************** + * Reading the information * + ***************************/ + +/** + * \brief Gets a known size + * + * + */ +extern lzma_vli lzma_info_size_get( + const lzma_info *info, lzma_info_size type); + +extern lzma_vli lzma_info_metadata_locate( + const lzma_info *info, lzma_bool is_header_metadata); + +/** + * \brief Gets a pointer to the beginning of the Index list + * + * If detach is true, the Index will be detached from the lzma_info + * structure, and thus not be modified or freed by lzma_info_end(). + * + * If detach is false, the application must not modify the Index in any way. + * Also, the Index list is guaranteed to be valid only till the next call + * to any lzma_info_* function. + */ +extern lzma_index *lzma_info_index_get(lzma_info *info, lzma_bool detach); + + +extern size_t lzma_info_index_count_get(const lzma_info *info); + + +extern uint32_t lzma_info_metadata_alignment_get( + const lzma_info *info, lzma_bool is_header_metadata); + + + +/** + * \brief Locate a Block containing the given uncompressed offset + * + * This function is useful when you need to do random-access reading in + * a Multi-Block Stream. + * + * \param info Pointer to lzma_info that has at least one + * Index Record. The Index doesn't need to be finished. + * \param uncompressed_target + * Uncompressed target offset which the caller would + * like to locate from the Stream. + * \param stream_offset + * Starting offset (relative to the beginning the Stream) + * of the Block containing the requested location. + * \param uncompressed_offset + * The actual uncompressed offset of the beginning of + * the Block. uncompressed_offset <= uncompressed_target + * is always true; the application needs to uncompress + * uncompressed_target - uncompressed_offset bytes to + * reach the requested target offset. + * \param total_size + * Total Size of the Block. If the Index is incomplete, + * this may be LZMA_VLI_VALUE_UNKNOWN indicating unknown + * size. + * \param uncompressed_size + * Uncompressed Size of the Block. If the Index is + * incomplete, this may be LZMA_VLI_VALUE_UNKNOWN + * indicating unknown size. The application must pass + * this value to the Block decoder to verify FIXME + * + * \return + * + * \note This function is currently implemented as a linear search. + * If there are many Index Records, this can be really slow. + * This can be improved in newer liblzma versions if needed. + */ +extern lzma_bool lzma_info_data_locate(const lzma_info *info, + lzma_vli uncompressed_target, + lzma_vli *lzma_restrict stream_offset, + lzma_vli *lzma_restrict uncompressed_offset, + lzma_vli *lzma_restrict total_size, + lzma_vli *lzma_restrict uncompressed_size); diff --git a/src/liblzma/api/lzma/init.h b/src/liblzma/api/lzma/init.h new file mode 100644 index 00000000..f7b79246 --- /dev/null +++ b/src/liblzma/api/lzma/init.h @@ -0,0 +1,85 @@ +/** + * \file lzma/init.h + * \brief Initializations + * + * \author Copyright (C) 1999-2006 Igor Pavlov + * \author Copyright (C) 2007 Lasse Collin + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + */ + +#ifndef LZMA_H_INTERNAL +# error Never include this file directly. Use <lzma.h> instead. +#endif + + +/** + * \brief Initialize all internal static variables + * + * Depending on the build options, liblzma may have some internal static + * variables, that must be initialized before using any other part of + * the library (*). It is recommended to do these initializations in the very + * beginning of the application by calling appropriate initialization function. + * + * (*) There are some exceptions to this rule. FIXME + * + * The initialization functions are not necessarily thread-safe, thus the + * required initializations must be done before creating any threads. (The + * rest of the functions of liblzma are thread-safe.) Calling the + * initialization functions multiple times does no harm, although it + * still shouldn't be done when there are multiple threads running. + * + * lzma_init() initializes all internal static variables by calling + * lzma_lzma_init_encoder() and lzma_init_decoder(). + * + * If you need only encoder, decoder, or neither-encoder-nor-decoder + * functions, you may use other initialization functions, which initialize + * only a subset of liblzma's internal static variables. Using those + * functions have the following advantages: + * - When linking statically against liblzma, less useless functions will + * get linked into the binary. E.g. if you need only the decoder functions, + * using lzma_init_decoder() avoids linking bunch of encoder related code. + * - There is less things to initialize, making the initialization + * process slightly faster. + */ +extern void lzma_init(void); + + +/** + * \brief Initialize internal static variables needed by encoders + * + * If you need only the encoder functions, you may use this function to + * initialize only the things required by encoders. + * + * This function also calls lzma_init_check(). + */ +extern void lzma_init_encoder(void); + + +/** + * \brief Initialize internal static variables needed by decoders + * + * If you need only the decoder functions, you may use this function to + * initialize only the things required by decoders. + * + * This function also calls lzma_init_check(). + */ +extern void lzma_init_decoder(void); + + +/** + * \brief Initialize internal static variables needed by integrity checks + * + * Currently this initializes CRC32 and CRC64 lookup tables if precalculated + * tables haven't been built into the library. This function can be useful + * if the only thing you need from liblzma is the integrity check functions. + */ +extern void lzma_init_check(void); diff --git a/src/liblzma/api/lzma/lzma.h b/src/liblzma/api/lzma/lzma.h new file mode 100644 index 00000000..0fe74854 --- /dev/null +++ b/src/liblzma/api/lzma/lzma.h @@ -0,0 +1,312 @@ +/** + * \file lzma/lzma.h + * \brief LZMA filter + * + * \author Copyright (C) 1999-2006 Igor Pavlov + * \author Copyright (C) 2007 Lasse Collin + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + */ + +#ifndef LZMA_H_INTERNAL +# error Never include this file directly. Use <lzma.h> instead. +#endif + + +/** + * \brief Filter ID + * + * Filter ID of the LZMA filter. This is used as lzma_options_filter.id. + */ +#define LZMA_FILTER_LZMA LZMA_VLI_C(0x40) + + +/** + * \brief LZMA compression modes + * + * Currently there are only two modes. Earlier LZMA SDKs had also third + * mode between fast and best. + */ +typedef enum { + LZMA_MODE_INVALID = -1, + /**< + * \brief Invalid mode + * + * Used as array terminator in lzma_available_modes. + */ + + + LZMA_MODE_FAST = 0, + /**< + * \brief Fast compression + * + * Fast mode is usually at its best when combined with + * a hash chain match finder. + */ + + LZMA_MODE_BEST = 2 + /**< + * \brief Best compression ratio + * + * This is usually notably slower than fast mode. Use this + * together with binary tree match finders to expose the + * full potential of the LZMA encoder. + */ +} lzma_mode; + + +/** + * \brief Match finders + * + * Match finder has major effect on both speed and compression ratio. + * Usually hash chains are faster than binary trees. + */ +typedef enum { + LZMA_MF_INVALID = -1, + /**< + * \brief Invalid match finder ID + * + * Used as array terminator in lzma_available_match_finders. + */ + + LZMA_MF_HC3 = 0x03, + /**< + * \brief Hash Chain with 3 bytes hashing + * + * \todo Memory requirements + * + * \note It's possible that this match finder gets + * removed in future. The definition will stay + * in this header, but liblzma may return + * LZMA_HEADER_ERROR if it is specified (just + * like it would if the match finder had been + * disabled at compile time). + */ + + LZMA_MF_HC4 = 0x04, + /**< + * \brief Hash Chain with 4 bytes hashing + * + * Memory requirements: 7.5 * dictionary_size + 4 MiB + * + * \note It's possible that this match finder gets + * removed in future. The definition will stay + * in this header, but liblzma may return + * LZMA_HEADER_ERROR if it is specified (just + * like it would if the match finder had been + * disabled at compile time). + */ + + LZMA_MF_BT2 = 0x12, + /**< + * \brief Binary Tree with 2 bytes hashing + * + * Memory requirements: 9.5 * dictionary_size + 4 MiB + */ + + LZMA_MF_BT3 = 0x13, + /**< + * \brief Binary Tree with 3 bytes hashing + * + * Memory requirements: 11.5 * dictionary_size + 4 MiB + */ + + LZMA_MF_BT4 = 0x14 + /**< + * \brief Binary Tree with 4 bytes hashing + * + * Memory requirements: 11.5 * dictionary_size + 4 MiB + */ +} lzma_match_finder; + + +/** + * \brief Options specific to the LZMA method handler + */ +typedef struct { + /********************************** + * LZMA encoding/decoding options * + **********************************/ + + /* These options are required in encoder and also with raw decoding. */ + + /** + * \brief Dictionary size in bytes + * + * Dictionary size indicates how many bytes of the recently processed + * uncompressed data is kept in memory. One method to reduce size of + * the uncompressed data is to store distance-length pairs, which + * indicate what data to repeat from the dictionary buffer. Thus, + * the bigger the dictionary, the better compression ratio usually is. + * + * Raw decoding: Too big dictionary does no other harm than + * wasting memory. This value is ignored by lzma_raw_decode_buffer(), + * because it uses the target buffer as the dictionary. + */ + uint32_t dictionary_size; +# define LZMA_DICTIONARY_SIZE_MIN 1 +# define LZMA_DICTIONARY_SIZE_MAX (UINT32_C(1) << 30) +# define LZMA_DICTIONARY_SIZE_DEFAULT (UINT32_C(1) << 23) + + /** + * \brief Number of literal context bits + * + * How many of the highest bits of the previous uncompressed + * eight-bit byte (also known as `literal') are taken into + * account when predicting the bits of the next literal. + * + * \todo Example + */ + uint32_t literal_context_bits; +# define LZMA_LITERAL_CONTEXT_BITS_MIN 0 +# define LZMA_LITERAL_CONTEXT_BITS_MAX 8 +# define LZMA_LITERAL_CONTEXT_BITS_DEFAULT 3 + + /** + * \brief Number of literal position bits + * + * How many of the lowest bits of the current position (number + * of bytes from the beginning of the uncompressed data) in the + * uncompressed data is taken into account when predicting the + * bits of the next literal (a single eight-bit byte). + * + * \todo Example + */ + uint32_t literal_pos_bits; +# define LZMA_LITERAL_POS_BITS_MIN 0 +# define LZMA_LITERAL_POS_BITS_MAX 4 +# define LZMA_LITERAL_POS_BITS_DEFAULT 0 + + /** + * \brief Number of position bits + * + * How many of the lowest bits of the current position in the + * uncompressed data is taken into account when estimating + * probabilities of matches. A match is a sequence of bytes for + * which a matching sequence is found from the dictionary and + * thus can be stored as distance-length pair. + * + * Example: If most of the matches occur at byte positions + * of 8 * n + 3, that is, 3, 11, 19, ... set pos_bits to 3, + * because 2**3 == 8. + */ + uint32_t pos_bits; +# define LZMA_POS_BITS_MIN 0 +# define LZMA_POS_BITS_MAX 4 +# define LZMA_POS_BITS_DEFAULT 2 + + /** + * \brief Pointer to an initial dictionary + * + * It is possible to initialize the LZ77 history window using + * a preset dictionary. Here is a good quote from zlib's + * documentation; this applies to LZMA as is: + * + * "The dictionary should consist of strings (byte sequences) that + * are likely to be encountered later in the data to be compressed, + * with the most commonly used strings preferably put towards the + * end of the dictionary. Using a dictionary is most useful when + * the data to be compressed is short and can be predicted with + * good accuracy; the data can then be compressed better than + * with the default empty dictionary." + * (From deflateSetDictionary() in zlib.h of zlib version 1.2.3) + * + * This feature should be used only in special situations. + * It works correctly only with raw encoding and decoding. + * Currently none of the container formats supported by + * liblzma allow preset dictionary when decoding, thus if + * you create a .lzma file with preset dictionary, it cannot + * be decoded with the regular .lzma decoder functions. + * + * \todo This feature is not implemented yet. + */ + const uint8_t *preset_dictionary; + + /** + * \brief Size of the preset dictionary + * + * Specifies the size of the preset dictionary. If the size is + * bigger than dictionary_size, only the last dictionary_size + * bytes are processed. + * + * This variable is read only when preset_dictionary is not NULL. + */ + uint32_t preset_dictionary_size; + + /****************************************** + * LZMA options needed only when encoding * + ******************************************/ + + /** LZMA compression mode */ + lzma_mode mode; + + /** + * \brief Number of fast bytes + * + * Number of fast bytes determines how many bytes the encoder + * compares from the match candidates when looking for the best + * match. Bigger fast bytes value usually increase both compression + * ratio and time. + */ + uint32_t fast_bytes; +# define LZMA_FAST_BYTES_MIN 5 +# define LZMA_FAST_BYTES_MAX 273 +# define LZMA_FAST_BYTES_DEFAULT 128 + + /** Match finder ID */ + lzma_match_finder match_finder; + + /** + * \brief Match finder cycles + * + * Higher values give slightly better compression ratio but + * decrease speed. Use special value 0 to let liblzma use + * match-finder-dependent default value. + * + * \todo Write much better description. + */ + uint32_t match_finder_cycles; + +} lzma_options_lzma; + + +/** + * \brief Available LZMA encoding modes + * + * Pointer to an array containing the list of available encoding modes. + * + * This variable is available only if LZMA encoder has been enabled. + */ +extern const lzma_mode *const lzma_available_modes; + + +/** + * \brief Available match finders + * + * Pointer to an array containing the list of available match finders. + * The last element is LZMA_MF_INVALID. + * + * This variable is available only if LZMA encoder has been enabled. + */ +extern const lzma_match_finder *const lzma_available_match_finders; + + +/** + * \brief Table of presets for the LZMA filter + * + * lzma_presets[0] is the fastest and lzma_preset_lzma[8] is the slowest. + * These presets match the switches -1 .. -9 of the lzma command line tool + * + * The preset values are subject to changes between liblzma versions. + * + * This variable is available only if LZMA encoder has been enabled. + */ +extern const lzma_options_lzma lzma_preset_lzma[9]; diff --git a/src/liblzma/api/lzma/memlimit.h b/src/liblzma/api/lzma/memlimit.h new file mode 100644 index 00000000..26ec50fe --- /dev/null +++ b/src/liblzma/api/lzma/memlimit.h @@ -0,0 +1,157 @@ +/** + * \file lzma/memlimit.h + * \brief Memory usage limitter + * + * \author Copyright (C) 1999-2006 Igor Pavlov + * \author Copyright (C) 2007 Lasse Collin + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + */ + +#ifndef LZMA_H_INTERNAL +# error Never include this file directly. Use <lzma.h> instead. +#endif + + +/** + * \brief Opaque data type used with the memory usage limitting functions + */ +typedef struct lzma_memlimit_s lzma_memlimit; + + +/** + * \brief Allocates and initializes a new lzma_memlimit structure + * + * It is easy to make liblzma to use huge amounts of memory. This can + * be a problem especially with the decoder, since it a file requiring + * huge amounts of memory to uncompress could allow even a denial of + * service attack if the memory usage wasn't limited. + * + * liblzma provides a set of functions to control memory usage. Pointers + * to these functions can be used in lzma_allocator structure, which makes + * it easy to limit memory usage with liblzma. + * + * The memory limitter functions are not tied to limitting memory usage + * with liblzma itself. You can use them with anything you like. + * + * In multi-threaded applications, only one thread at once may use the same + * lzma_memlimit structure. If there is a need, this limitation may + * be removed in future versions without breaking the libary API/ABI. + * + * \param limit Initial memory usage limit in bytes + * + * \return Pointer to allocated and initialized lzma_memlimit + * structure. On error, NULL is returned. The reason behind + * an error is either that malloc() failed or that the given + * limit was so small that it didn't allow allocating even + * the lzma_memlimit structure itself. + * + * \note Excluding lzma_memlimit_usage(), the functions whose name begin + * lzma_memlimit_ can be used even if lzma_init() hasn't been + * called. + */ +extern lzma_memlimit *lzma_memlimit_create(size_t limit); + + +/** + * \brief Sets a new memory usage limit + * + * \param mem Pointer to a lzma_memlimit structure returned + * earlier by lzma_memry_limit_create(). + * \param limit New memory usage limit + * + * The new usage limit may be smaller than the amount of memory currently + * allocated via *mem: New allocations will fail until enough memory has + * been freed or a new limit is set, but the existing allocatations will + * stay untouched. + */ +extern void lzma_memlimit_set(lzma_memlimit *mem, size_t limit); + + +/** + * \brief Gets the current memory usage limit + */ +extern size_t lzma_memlimit_get(const lzma_memlimit *mem); + + +/** + * \brief Gets the amount of currently allocated memory + * + * \note This value includes the sizes of some helper structures, + * thus it will always be larger than the total number of + * bytes allocated via lzma_memlimit_alloc(). + */ +extern size_t lzma_memlimit_used(const lzma_memlimit *mem); + + +/** + * \brief Allocates memory with malloc() if memory limit allows + * + * \param mem Pointer to a lzma_memlimit structure returned + * earlier by lzma_memry_limit_create(). + * \param nmemb Number of elements to allocate. While liblzma always + * sets this to one, this function still takes the + * value of nmemb into account to keep the function + * usable with zlib and libbzip2. + * \param size Size of an element. + * + * \return Pointer to memory allocated with malloc(nmemb * size), + * except if nmemb * size == 0 which returns malloc(1). + * On error, NULL is returned. + * + * \note This function assumes that nmemb * size is at maximum of + * SIZE_MAX. If it isn't, an overflow will occur resulting + * invalid amount of memory being allocated. + */ +extern void *lzma_memlimit_alloc( + lzma_memlimit *mem, size_t nmemb, size_t size); + + +/** + * \brief Removes the pointer from memory limitting list + * + * \param mem Pointer to a lzma_memlimit structure returned + * earlier by lzma_memry_limit_create(). + * \param ptr Pointer returned earlier by lzma_memlimit_alloc(). + * + * This function removes ptr from the internal list and decreases the + * counter of used memory accordingly. The ptr itself isn't freed. This is + * useful when Extra Records allocated by liblzma using lzma_memlimit + * are needed by the application and must not be freed when the + * lzma_memlimit structure is destroyed. + * + * It is OK to call this function with ptr that hasn't been allocated with + * lzma_memlimit_alloc(). In that case, this has no effect other than wasting + * a few CPU cycles. + */ +extern void lzma_memlimit_detach(lzma_memlimit *mem, void *ptr); + + +/** + * \brief Frees memory and updates the memory limit list + * + * This is like lzma_memlimit_detach() but also frees the given pointer. + */ +extern void lzma_memlimit_free(lzma_memlimit *mem, void *ptr); + + +/** + * \brief Frees the memory allocated for and by the memory usage limitter + * + * \param mem Pointer to memory limitter + * \param free_allocated If this is non-zero, all the memory allocated + * by lzma_memlimit_alloc() using *mem is also + * freed if it hasn't already been freed with + * lzma_memlimit_free(). Usually this should be + * set to true. + */ +extern void lzma_memlimit_end( + lzma_memlimit *mem, lzma_bool free_allocated); diff --git a/src/liblzma/api/lzma/metadata.h b/src/liblzma/api/lzma/metadata.h new file mode 100644 index 00000000..69592a3a --- /dev/null +++ b/src/liblzma/api/lzma/metadata.h @@ -0,0 +1,100 @@ +/** + * \file lzma/metadata.h + * \brief Metadata handling + * + * \author Copyright (C) 1999-2006 Igor Pavlov + * \author Copyright (C) 2007 Lasse Collin + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + */ + +#ifndef LZMA_H_INTERNAL +# error Never include this file directly. Use <lzma.h> instead. +#endif + + +/** + * \brief Information stored into a Metadata Block + * + * This structure holds all the information that can be stored to + * a Metadata Block. + */ +typedef struct { + /** + * \brief Size of Header Metadata Block + */ + lzma_vli header_metadata_size; + + /** + * \brief Total Size of the Stream + */ + lzma_vli total_size; + + /** + * \brief Uncompressed Size of the Stream + */ + lzma_vli uncompressed_size; + + /** + * \brief Index of the Blocks stored in the Stream + */ + lzma_index *index; + + /** + * \brief Extra information + */ + lzma_extra *extra; + +} lzma_metadata; + + +/** + * \brief Calculate the encoded size of Metadata + * + * \return Uncompressed size of the Metadata in encoded form. This value + * may be passed to Block encoder as Uncompressed Size when using + * Metadata filter. On error, zero is returned. + */ +extern lzma_vli lzma_metadata_size(const lzma_metadata *metadata); + + +/** + * \brief Initializes Metadata encoder + * + * \param coder Pointer to a pointer to hold Metadata encoder's + * internal state. Original value is ignored, thus + * you don't need to initialize the pointer. + * \param allocator Custom memory allocator; usually NULL. + * \param metadata Pointer to Metadata to encoded + * + * \return - LZMA_OK: Initialization succeeded. + * - LZMA_MEM_ERROR: Cannot allocate memory for *coder. + * + * The initialization function makes internal copy of the *metadata structure. + * However, the linked lists metadata->index and metadata->extra are NOT + * copied. Thus, the application may destroy *metadata after initialization + * if it likes, but not Index or Extra. + */ +extern lzma_ret lzma_metadata_encoder(lzma_stream *strm, + lzma_options_block *options, const lzma_metadata *metadata); + + +/** + * \brief Initializes Metadata decoder + * + * \param want_extra If this is true, Extra Records will be stored + * to metadata->extra. If this is false, Extra + * Records will be parsed but not stored anywhere, + * metadata->extra will be set to NULL. + */ +extern lzma_ret lzma_metadata_decoder( + lzma_stream *strm, lzma_options_block *options, + lzma_metadata *metadata, lzma_bool want_extra); diff --git a/src/liblzma/api/lzma/raw.h b/src/liblzma/api/lzma/raw.h new file mode 100644 index 00000000..c1ee41d8 --- /dev/null +++ b/src/liblzma/api/lzma/raw.h @@ -0,0 +1,72 @@ +/** + * \file lzma/raw.h + * \brief Raw encoder and decoder + * + * \author Copyright (C) 1999-2006 Igor Pavlov + * \author Copyright (C) 2007 Lasse Collin + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + */ + +#ifndef LZMA_H_INTERNAL +# error Never include this file directly. Use <lzma.h> instead. +#endif + + +/** + * \brief Initializes raw encoder + * + * This function may be useful when implementing custom file formats. + * + * \param strm Pointer to properly prepared lzma_stream + * \param options Array of lzma_options_filter structures. + * The end of the array must be marked with + * .id = LZMA_VLI_VALUE_UNKNOWN. The minimum + * number of filters is zero; the maximum is + * determined by available memory. + * \param uncompressed_size + * Size of the uncompressed data. If it is unknown, + * use LZMA_VLI_VALUE_UNKNOWN. You need to give the + * same value to the raw decoder to decode the data. + * \param allow_implicit + * If true, an implicit Copy or Subblock filter should be + * automatically added when needed. If this is false and + * an implicit filter would be needed, LZMA_PROG_ERROR is + * returned. + * + * The `action' with lzma_code() can be LZMA_RUN, LZMA_SYNC_FLUSH (if the + * filter chain support it), or LZMA_FINISH. + * + * \return - LZMA_OK + * - LZMA_MEM_ERROR + * - LZMA_HEADER_ERROR + * - LZMA_PROG_ERROR + */ +extern lzma_ret lzma_raw_encoder( + lzma_stream *strm, const lzma_options_filter *options, + lzma_vli uncompressed_size, lzma_bool allow_implicit); + + +/** + * \brief Initializes raw decoder + * + * The initialization of raw decoder goes similarly to raw encoder. + * + * The `action' with lzma_code() can be LZMA_RUN or LZMA_SYNC_FLUSH. + * + * \return - LZMA_OK + * - LZMA_MEM_ERROR + * - LZMA_HEADER_ERROR + * - LZMA_PROG_ERROR + */ +extern lzma_ret lzma_raw_decoder( + lzma_stream *strm, const lzma_options_filter *options, + lzma_vli uncompressed_size, lzma_bool allow_implicit); diff --git a/src/liblzma/api/lzma/simple.h b/src/liblzma/api/lzma/simple.h new file mode 100644 index 00000000..fb78d01f --- /dev/null +++ b/src/liblzma/api/lzma/simple.h @@ -0,0 +1,85 @@ +/** + * \file lzma/simple.h + * \brief So called "simple" filters + * + * \author Copyright (C) 1999-2006 Igor Pavlov + * \author Copyright (C) 2007 Lasse Collin + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + */ + +#ifndef LZMA_H_INTERNAL +# error Never include this file directly. Use <lzma.h> instead. +#endif + + +/* Filter IDs for lzma_options_filter.id */ + +#define LZMA_FILTER_X86 LZMA_VLI_C(0x04) + /**< + * BCJ (Branch, Call, Jump) filter for x86 binaries + */ + +#define LZMA_FILTER_POWERPC LZMA_VLI_C(0x05) + /**< + * Filter for Big endian PowerPC binaries + */ + +#define LZMA_FILTER_IA64 LZMA_VLI_C(0x06) + /**< + * Filter for IA64 (Itanium) binaries. + */ + +#define LZMA_FILTER_ARM LZMA_VLI_C(0x07) + /**< + * Filter for ARM binaries. + */ + +#define LZMA_FILTER_ARMTHUMB LZMA_VLI_C(0x08) + /**< + * Filter for ARMThumb binaries. + */ + +#define LZMA_FILTER_SPARC LZMA_VLI_C(0x09) + /**< + * Filter for SPARC binaries. + */ + + +/** + * \brief Options for so called "simple" filters + * + * The simple filters never change the size of the data. Specifying options + * for them is optional: if pointer to options is NULL, default values are + * used. You probably never need to specify these options, so just set the + * options pointer to NULL and be happy. + * + * If options with non-default values have been specified when encoding, + * the same options must also be specified when decoding. + */ +typedef struct { + /** + * \brief Start offset for branch conversions + * + * This setting is useful only when the same filter is used + * _separately_ for multiple sections of the same executable file, + * and the sections contain cross-section branch/call/jump + * instructions. In that case it is benefical to set the start + * offset of the non-first sections so that the relative addresses + * of the cross-section branch/call/jump instructions will use the + * same absolute addresses as in the first section. + * + * When the pointer to options is NULL, the default value is used. + * The default value is zero. + */ + uint32_t start_offset; + +} lzma_options_simple; diff --git a/src/liblzma/api/lzma/stream.h b/src/liblzma/api/lzma/stream.h new file mode 100644 index 00000000..be86075f --- /dev/null +++ b/src/liblzma/api/lzma/stream.h @@ -0,0 +1,178 @@ +/** + * \file lzma/stream.h + * \brief .lzma Stream handling + * + * \author Copyright (C) 1999-2006 Igor Pavlov + * \author Copyright (C) 2007 Lasse Collin + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + */ + +#ifndef LZMA_H_INTERNAL +# error Never include this file directly. Use <lzma.h> instead. +#endif + + +/** + * \brief Options for .lzma Stream encoder + */ +typedef struct { + /** + * \brief Type of integrity Check + * + * The type of the integrity Check is stored into Stream Header + * and Stream Footer. The same Check is used for all Blocks in + * the Stream. + */ + lzma_check_type check; + + /** + * \brief Precense of CRC32 of the Block Header + * + * Set this to true if CRC32 of every Block Header should be + * calculated and stored in the Block Header. This is recommended. + * + * This setting is stored into Stream Header and Stream Footer. + */ + lzma_bool has_crc32; + + /** + * \brief Uncompressed Size in bytes + * + * This is somewhat advanced feature. Most users want to set this to + * LZMA_VLI_VALUE_UNKNOWN to indicate unknown Uncompressed Size. + * + * If the Uncompressed Size of the Stream being encoded is known, + * it can be stored to the beginning of the Stream. The details + * differ for Single-Block and Multi-Block Streams: + * - With Single-Block Streams, the Uncompressed Size is stored to + * the Block Header and End of Payload Marker is omitted. + * - With Multi-Block Streams, the Uncompressed Size is stored to + * the Header Metadata Block. The Uncompressed Size of the Blocks + * will be unknown, because liblzma cannot predict how the + * application is going to split the data in Blocks. + */ + lzma_vli uncompressed_size; + + /** + * \brief Alignment of the beginning of the Stream + * + * Certain filters handle data in bigger chunks than single bytes. + * This affects two things: + * - Performance: aligned memory access is usually faster. + * - Further compression ratio in custom file formats: if you + * encode multiple Blocks with some non-compression filter + * such as LZMA_FILTER_POWERPC, it is a good idea to keep + * the inter-Block alignment correct to maximize compression + * ratio when all these Blocks are finally compressed as a + * single step. + * + * Usually the Stream is stored into its own file, thus + * the alignment is usually zero. + */ + uint32_t alignment; + + /** + * \brief Array of filters used to encode Data Blocks + * + * There can be at maximum of seven filters. The end of the array is + * marked with .id = LZMA_VLI_VALUE_UNKNOWN. (That's why the array + * has eight members.) Minimum number of filters is zero; in that + * case, an implicit Copy filter is used. + */ + lzma_options_filter filters[8]; + + /** + * \brief Array of filters used to encode Metadata Blocks + * + * This is like filters[] but for Metadata Blocks. If Metadata + * Blocks are compressed, they usually are compressed with + * settings that require only little memory to uncompress e.g. + * LZMA with 64 KiB dictionary. + * + * \todo Recommend a preset. + * + * When liblzma sees that the Metadata Block would be very small + * even in uncompressed form, it is not compressed no matter + * what filter have been set here. This is to avoid possibly + * increasing the size of the Metadata Block with bad compression, + * and to avoid useless overhead of filters in uncompression phase. + */ + lzma_options_filter metadata_filters[8]; + + /** + * \brief Extra information in the Header Metadata Block + */ + lzma_extra *header; + + /** + * \brief Extra information in the Footer Metadata Block + * + * It is enough to set this pointer any time before calling + * lzma_code() with LZMA_FINISH as the second argument. + */ + lzma_extra *footer; + +} lzma_options_stream; + + +/** + * \brief Initializes Single-Block .lzma Stream encoder + * + * This is the function that most developers are looking for. :-) It + * compresses using the specified options without storing any extra + * information. + * + * \todo Describe that is_metadata is ignored, maybe something else. + */ +extern lzma_ret lzma_stream_encoder_single( + lzma_stream *strm, const lzma_options_stream *options); + + +/** + * \brief Initializes Multi-Block .lzma Stream encoder + * + */ +extern lzma_ret lzma_stream_encoder_multi( + lzma_stream *strm, const lzma_options_stream *options); + + +/** + * \brief Initializes decoder for .lzma Stream + * + * \param strm Pointer to propertily prepared lzma_stream + * \param header Pointer to hold a pointer to Extra Records read + * from the Header Metadata Block. Use NULL if + * you don't care about Extra Records. + * \param footer Same as header, but for Footer Metadata Block. + * + * \return - LZMA_OK: Initialization was successful. + * - LZMA_MEM_ERROR: Cannot allocate memory. + * + * If header and/or footer are not NULL, *header and/or *footer will be + * initially set to NULL. + * + * The application can detect that Header Metadata Block has been completely + * parsed when the decoder procudes some output the first time. If *header + * is still NULL, there was no Extra field in the Header Metadata Block (or + * the whole Header Metadata Block wasn't present at all). + * + * The application can detect that Footer Metadata Block has been parsed + * completely when lzma_code() returns LZMA_STREAM_END. If *footer is still + * NULL, there was no Extra field in the Footer Metadata Block. + * + * \note If you use lzma_memory_limitter, the Extra Records will be + * allocated with it, and thus remain in the lzma_memory_limitter + * even after they get exported to the application via *header + * and *footer pointers. + */ +extern lzma_ret lzma_stream_decoder(lzma_stream *strm, + lzma_extra **header, lzma_extra **footer); diff --git a/src/liblzma/api/lzma/stream_flags.h b/src/liblzma/api/lzma/stream_flags.h new file mode 100644 index 00000000..070c91c9 --- /dev/null +++ b/src/liblzma/api/lzma/stream_flags.h @@ -0,0 +1,142 @@ +/** + * \file lzma/stream_flags.h + * \brief .lzma Stream Header and Stream tail encoder and decoder + * + * \author Copyright (C) 1999-2006 Igor Pavlov + * \author Copyright (C) 2007 Lasse Collin + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + */ + +#ifndef LZMA_H_INTERNAL +# error Never include this file directly. Use <lzma.h> instead. +#endif + + +/** + * \brief Size of Stream Header + * + * Magic Bytes (6) + Stream Flags (1) + CRC32 (4) + */ +#define LZMA_STREAM_HEADER_SIZE (6 + 1 + 4) + + +/** + * \brief Size of Stream tail + * + * Because Stream Footer already has a defined meaning in the file format + * specification, we use Stream tail to denote these two fields: + * Stream Flags (1) + Magic Bytes (2) + */ +#define LZMA_STREAM_TAIL_SIZE (1 + 2) + + +/** + * Options for encoding and decoding Stream Header and Stream tail + */ +typedef struct { + /** + * Type of the Check calculated from uncompressed data + */ + lzma_check_type check; + + /** + * True if Block Headers have the CRC32 field. Note that the CRC32 + * field is always present in the Stream Header. + */ + lzma_bool has_crc32; + + /** + * True if the Stream is a Multi-Block Stream. + */ + lzma_bool is_multi; + +} lzma_stream_flags; + + +#define lzma_stream_flags_is_equal(a, b) \ + ((a).check == (b).check \ + && (a).has_crc32 == (b).has_crc32 \ + && (a).is_multi == (b).is_multi) + + +/** + * \brief Encodes Stream Header + * + * Encoding of the Stream Header is done with a single call instead of + * first initializing and then doing the actual work with lzma_code(). + * + * \param out Beginning of the output buffer + * \param out_pos out[*out_pos] is the next write position. This + * is updated by the encoder. + * \param out_size out[out_size] is the first byte to not write. + * \param options Stream Header options to be encoded. + * + * \return - LZMA_OK: Encoding was successful. + * - LZMA_PROG_ERROR: Invalid options. + * - LZMA_BUF_ERROR: Not enough output buffer space. + */ +extern lzma_ret lzma_stream_header_encode( + uint8_t *out, const lzma_stream_flags *options); + + +/** + * \brief Encodes Stream tail + * + * \param footer Pointer to a pointer that will hold the + * allocated buffer. Caller must free it once + * it isn't needed anymore. + * \param footer_size Pointer to a variable that will the final size + * of the footer buffer. + * \param allocator lzma_allocator for custom allocator functions. + * Set to NULL to use malloc(). + * \param options Stream Header options to be encoded. + * + * \return - LZMA_OK: Success; *header and *header_size set. + * - LZMA_PROG_ERROR: *options is invalid. + * - LZMA_MEM_ERROR: Cannot allocate memory. + */ +extern lzma_ret lzma_stream_tail_encode( + uint8_t *out, const lzma_stream_flags *options); + + +/** + * \brief Initializes Stream Header decoder + * + * \param strm Pointer to lzma_stream used to pass input data + * \param options Target structure for parsed results + * + * \return - LZMA_OK: Successfully initialized + * - LZMA_MEM_ERROR: Cannot allocate memory + * + * The actual decoding is done with lzma_code() and freed with lzma_end(). + */ +extern lzma_ret lzma_stream_header_decoder( + lzma_stream *strm, lzma_stream_flags *options); + + +/** + * \brief Initializes Stream tail decoder + * + * \param strm Pointer to lzma_stream used to pass input data + * \param options Target structure for parsed results. + * \param decode_uncompressed_size + * Set to true if the first field to decode is + * Uncompressed Size. Set to false if the first + * field to decode is Backward Size. + * + * \return - LZMA_OK: Successfully initialized + * - LZMA_MEM_ERROR: Cannot allocate memory + * + * The actual decoding is done with lzma_code() and freed with lzma_end(). + */ +extern lzma_ret lzma_stream_tail_decoder( + lzma_stream *strm, lzma_stream_flags *options); diff --git a/src/liblzma/api/lzma/subblock.h b/src/liblzma/api/lzma/subblock.h new file mode 100644 index 00000000..0474b6af --- /dev/null +++ b/src/liblzma/api/lzma/subblock.h @@ -0,0 +1,197 @@ +/** + * \file lzma/subblock.h + * \brief Subblock filter + * + * \author Copyright (C) 1999-2006 Igor Pavlov + * \author Copyright (C) 2007 Lasse Collin + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + */ + +#ifndef LZMA_H_INTERNAL +# error Never include this file directly. Use <lzma.h> instead. +#endif + + +/** + * \brief Filter ID + * + * Filter ID of the Subblock filter. This is used as lzma_options_filter.id. + */ +#define LZMA_FILTER_SUBBLOCK LZMA_VLI_C(0x01) + + +/** + * \brief Subfilter mode + * + * See lzma_options_subblock.subfilter_mode for details. + */ +typedef enum { + LZMA_SUBFILTER_NONE, + /**< + * No Subfilter is in use. + */ + + LZMA_SUBFILTER_SET, + /**< + * New Subfilter has been requested to be initialized. + */ + + LZMA_SUBFILTER_RUN, + /**< + * Subfilter is active. + */ + + LZMA_SUBFILTER_FINISH + /**< + * Subfilter has been requested to be finished. + */ +} lzma_subfilter_mode; + + +/** + * \brief Options for the Subblock filter + * + * Specifying options for the Subblock filter is optional: if the pointer + * options is NULL, no subfilters are allowed and the default value is used + * for subblock_data_size. + */ +typedef struct { + /* Options for encoder and decoder */ + + /** + * \brief Allowing subfilters + * + * If this true, subfilters are allowed. + * + * In the encoder, if this is set to false, subfilter_mode and + * subfilter_options are completely ignored. + */ + lzma_bool allow_subfilters; + + /* Options for encoder only */ + + /** + * \brief Alignment + * + * The Subblock filter encapsulates the input data into Subblocks. + * Each Subblock has a header which takes a few bytes of space. + * When the output of the Subblock encoder is fed to another filter + * that takes advantage of the alignment of the input data (e.g. LZMA), + * the Subblock filter can add padding to keep the actual data parts + * in the Subblocks aligned correctly. + * + * The alignment should be a positive integer. Subblock filter will + * add enough padding between Subblocks so that this is true for + * every payload byte: + * input_offset % alignment == output_offset % alignment + * + * The Subblock filter assumes that the first output byte will be + * written to a position in the output stream that is properly aligned. + * + * FIXME desc + */ + uint32_t alignment; +# define LZMA_SUBBLOCK_ALIGNMENT_MIN 1 +# define LZMA_SUBBLOCK_ALIGNMENT_MAX 32 +# define LZMA_SUBBLOCK_ALIGNMENT_DEFAULT 4 + + /** + * \brief Size of the Subblock Data part of each Subblock + * + * This value is re-read every time a new Subblock is started. + * + * Bigger values + * - save a few bytes of space; + * - increase latency in the encoder (but no effect for decoding); + * - decrease memory locality (increased cache pollution) in the + * encoder (no effect in decoding). + */ + uint32_t subblock_data_size; +# define LZMA_SUBBLOCK_DATA_SIZE_MIN 1 +# define LZMA_SUBBLOCK_DATA_SIZE_MAX (UINT32_C(1) << 28) +# define LZMA_SUBBLOCK_DATA_SIZE_DEFAULT 4096 + + /** + * \brief Run-length encoder remote control + * + * The Subblock filter has an internal run-length encoder (RLE). It + * can be useful when the data includes byte sequences that repeat + * very many times. The RLE can be used also when a Subfilter is + * in use; the RLE will be applied to the output of the Subfilter. + * + * Note that in contrast to traditional RLE, this RLE is intended to + * be used only when there's a lot of data to be repeated. If the + * input data has e.g. 500 bytes of NULs now and then, this RLE + * is probably useless, because plain LZMA should provide better + * results. + * + * Due to above reasons, it was decided to keep the implementation + * of the RLE very simple. When the rle variable is non-zero, it + * subblock_data_size must be a multiple of rle. Once the Subblock + * encoder has got subblock_data_size bytes of input, it will check + * if the whole buffer of the last subblock_data_size can be + * represented with repeats of chunks having size of rle bytes. + * + * If there are consecutive identical buffers of subblock_data_size + * bytes, they will be encoded using a single repeat entry if + * possible. + * + * If need arises, more advanced RLE can be implemented later + * without breaking API or ABI. + */ + uint32_t rle; +# define LZMA_SUBBLOCK_RLE_OFF 0 +# define LZMA_SUBBLOCK_RLE_MIN 1 +# define LZMA_SUBBLOCK_RLE_MAX 256 + + /** + * \brief Subfilter remote control + * + * When the Subblock filter is initialized, this variable must be + * LZMA_SUBFILTER_NONE or LZMA_SUBFILTER_SET. + * + * When subfilter_mode is LZMA_SUBFILTER_NONE, the application may + * put Subfilter options to subfilter_options structure, and then + * set subfilter_mode to LZMA_SUBFILTER_SET. This implies setting + * flush to true. No new input data will be read until the Subfilter + * has been enabled. Once the Subfilter has been enabled, liblzma + * will set subfilter_mode to LZMA_SUBFILTER_RUN. + * + * When subfilter_mode is LZMA_SUBFILTER_RUN, the application may + * set subfilter_mode to LZMA_SUBFILTER_FINISH. No new input data + * will be read until the Subfilter has been finished. Once the + * Subfilter has been finished, liblzma will set subfilter_mode + * to LZMA_SUBFILTER_NONE. + * + * If the intent is to have Subfilter enabled to the very end of + * the data, it is not needed to separately disable Subfilter with + * LZMA_SUBFILTER_FINISH. Using LZMA_FINISH as the second argument + * of lzma_code() will make the Subblock encoder to disable the + * Subfilter once all the data has been ran through the Subfilter. + * + * \note This variable is ignored if allow_subfilters is false. + */ + lzma_subfilter_mode subfilter_mode; + + /** + * \brief Subfilter and its options + * + * When no Subfilter is used, the data is copied as is into Subblocks. + * Setting a Subfilter allows encoding some parts of the data with + * an additional filter. It is possible to many different Subfilters + * in the same Block, although only one can be used at once. + * + * \note This variable is ignored if allow_subfilters is false. + */ + lzma_options_filter subfilter_options; + +} lzma_options_subblock; diff --git a/src/liblzma/api/lzma/version.h b/src/liblzma/api/lzma/version.h new file mode 100644 index 00000000..ffbf8a81 --- /dev/null +++ b/src/liblzma/api/lzma/version.h @@ -0,0 +1,59 @@ +/** + * \file lzma/version.h + * \brief Version number + * + * \author Copyright (C) 1999-2006 Igor Pavlov + * \author Copyright (C) 2007 Lasse Collin + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + */ + +#ifndef LZMA_H_INTERNAL +# error Never include this file directly. Use <lzma.h> instead. +#endif + + +/** + * \brief Compile-time version number + * + * The version number is of format xyyyuuus where + * - x is the major LZMA SDK version + * - yyy is the minor LZMA SDK version + * - uuu is LZMA Utils version (reset to zero every time SDK version + * is incremented) + * - s indicates stability: 0 = alpha, 1 = beta, 2 = stable + */ +#define LZMA_VERSION UINT32_C(40420010) + + +/** + * \brief liblzma version number as an integer + * + * This is the value of LZMA_VERSION macro at the compile time of liblzma. + * This allows the application to compare if it was built against the same, + * older, or newer version of liblzma that is currently running. + */ +extern const uint32_t lzma_version_number; + + +/** + * \brief Returns versions number of liblzma as a string + * + * This function may be useful if you want to display which version of + * libilzma your application is currently using. + * + * \return Returns a pointer to a statically allocated string constant, + * which contains the version number of liblzma. The format of + * the version string is usually (but not necessarily) x.y.z + * e.g. "4.42.1". Alpha and beta versions contain a suffix + * ("4.42.0alpha"). + */ +extern const char *const lzma_version_string; diff --git a/src/liblzma/api/lzma/vli.h b/src/liblzma/api/lzma/vli.h new file mode 100644 index 00000000..322014e1 --- /dev/null +++ b/src/liblzma/api/lzma/vli.h @@ -0,0 +1,244 @@ +/** + * \file lzma/vli.h + * \brief Variable-length integer handling + * + * \author Copyright (C) 1999-2006 Igor Pavlov + * \author Copyright (C) 2007 Lasse Collin + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + */ + +#ifndef LZMA_H_INTERNAL +# error Never include this file directly. Use <lzma.h> instead. +#endif + + +/** + * \brief Maximum supported value of variable-length integer + */ +#define LZMA_VLI_VALUE_MAX (UINT64_MAX / 2) + +/** + * \brief VLI value to denote that the value is unknown + */ +#define LZMA_VLI_VALUE_UNKNOWN UINT64_MAX + +/** + * \brief Maximum supported length of variable length integers + */ +#define LZMA_VLI_BYTES_MAX 9 + + +/** + * \brief VLI constant suffix + */ +#define LZMA_VLI_C(n) UINT64_C(n) + + +/** + * \brief Variable-length integer type + * + * This will always be unsigned integer. Valid VLI values are in the range + * [0, LZMA_VLI_VALUE_MAX]. Unknown value is indicated with + * LZMA_VLI_VALUE_UNKNOWN, which is the maximum value of the underlaying + * integer type (this feature is useful in several situations). + * + * In future, even if lzma_vli is typdefined to something else than uint64_t, + * it is guaranteed that 2 * LZMA_VLI_VALUE_MAX will not overflow lzma_vli. + * This simplifies integer overflow detection. + */ +typedef uint64_t lzma_vli; + + +/** + * \brief Simple macro to validate variable-length integer + * + * This is useful to test that application has given acceptable values + * for example in the uncompressed_size and compressed_size variables. + * + * \return True if the integer is representable as VLI or if it + * indicates unknown value. + */ +#define lzma_vli_is_valid(vli) \ + ((vli) <= LZMA_VLI_VALUE_MAX || (vli) == LZMA_VLI_VALUE_UNKNOWN) + + +/** + * \brief Sets VLI to given value with error checking + * + * \param dest Target variable which must have type of lzma_vli. + * \param src New value to be stored to dest. + * \param limit Maximum allowed value for src. + * + * \return False on success, true on error. If an error occurred, + * dest is left in undefined state (i.e. it's possible that + * it will be different in newer liblzma versions). + */ +#define lzma_vli_set_lim(dest, src, limit) \ + ((src) > (limit) || ((dest) = (src)) > (limit)) + +/** + * \brief + */ +#define lzma_vli_add_lim(dest, src, limit) \ + ((src) > (limit) || ((dest) += (src)) > (limit)) + +#define lzma_vli_add2_lim(dest, src1, src2, limit) \ + (lzma_vli_add_lim(dest, src1, limit) \ + || lzma_vli_add_lim(dest, src2, limit)) + +#define lzma_vli_add3_lim(dest, src1, src2, src3, limit) \ + (lzma_vli_add_lim(dest, src1, limit) \ + || lzma_vli_add_lim(dest, src2, limit) \ + || lzma_vli_add_lim(dest, src3, limit)) + +#define lzma_vli_add4_lim(dest, src1, src2, src3, src4, limit) \ + (lzma_vli_add_lim(dest, src1, limit) \ + || lzma_vli_add_lim(dest, src2, limit) \ + || lzma_vli_add_lim(dest, src3, limit) \ + || lzma_vli_add_lim(dest, src4, limit)) + +#define lzma_vli_sum_lim(dest, src1, src2, limit) \ + (lzma_vli_set_lim(dest, src1, limit) \ + || lzma_vli_add_lim(dest, src2, limit)) + +#define lzma_vli_sum3_lim(dest, src1, src2, src3, limit) \ + (lzma_vli_set_lim(dest, src1, limit) \ + || lzma_vli_add_lim(dest, src2, limit) \ + || lzma_vli_add_lim(dest, src3, limit)) + +#define lzma_vli_sum4_lim(dest, src1, src2, src3, src4, limit) \ + (lzma_vli_set_lim(dest, src1, limit) \ + || lzma_vli_add_lim(dest, src2, limit) \ + || lzma_vli_add_lim(dest, src3, limit) \ + || lzma_vli_add_lim(dest, src4, limit)) + +#define lzma_vli_set(dest, src) lzma_vli_set_lim(dest, src, LZMA_VLI_VALUE_MAX) + +#define lzma_vli_add(dest, src) lzma_vli_add_lim(dest, src, LZMA_VLI_VALUE_MAX) + +#define lzma_vli_add2(dest, src1, src2) \ + lzma_vli_add2_lim(dest, src1, src2, LZMA_VLI_VALUE_MAX) + +#define lzma_vli_add3(dest, src1, src2, src3) \ + lzma_vli_add3_lim(dest, src1, src2, src3, LZMA_VLI_VALUE_MAX) + +#define lzma_vli_add4(dest, src1, src2, src3, src4) \ + lzma_vli_add4_lim(dest, src1, src2, src3, src4, LZMA_VLI_VALUE_MAX) + +#define lzma_vli_sum(dest, src1, src2) \ + lzma_vli_sum_lim(dest, src1, src2, LZMA_VLI_VALUE_MAX) + +#define lzma_vli_sum3(dest, src1, src2, src3) \ + lzma_vli_sum3_lim(dest, src1, src2, src3, LZMA_VLI_VALUE_MAX) + +#define lzma_vli_sum4(dest, src1, src2, src3, src4) \ + lzma_vli_sum4_lim(dest, src1, src2, src3, src4, LZMA_VLI_VALUE_MAX) + + +/** + * \brief Encodes variable-length integer + * + * In the new .lzma format, most integers are encoded in variable-length + * representation. This saves space when smaller values are more likely + * than bigger values. + * + * The encoding scheme encodes seven bits to every byte, using minimum + * number of bytes required to represent the given value. In other words, + * it puts 7-63 bits into 1-9 bytes. This implementation limits the number + * of bits used to 63, thus num must be at maximum of INT64_MAX / 2. You + * may use LZMA_VLI_VALUE_MAX for clarity. + * + * \param vli Integer to be encoded + * \param vli_pos How many bytes have already been written out. This + * must be less than 9 before calling this function. + * \param vli_size Minimum size that the variable-length representation + * must take. This is useful if you want to use + * variable-length integers as padding. Usually you want + * to set this to zero. The maximum allowed value is 9. + * \param out Beginning of the output buffer + * \param out_pos The next byte will be written to out[*out_pos]. + * \param out_size Size of the out buffer; the first byte into + * which no data is written to is out[out_size]. + * + * \return - LZMA_OK: So far all OK, but the integer is not + * completely written out yet. + * - LZMA_STREAM_END: Integer successfully encoded. + * - LZMA_BUF_ERROR: No output space (*out_pos == out_size) + * - LZMA_PROG_ERROR: Arguments are not sane. + */ +extern lzma_ret lzma_vli_encode( + lzma_vli vli, size_t *lzma_restrict vli_pos, size_t vli_size, + uint8_t *lzma_restrict out, size_t *lzma_restrict out_pos, + size_t out_size); + + +/** + * \brief Decodes variable-length integer + * + * \param vli Pointer to decoded integer. The decoder will + * initialize it to zero when *vli_pos == 0, so + * application isn't required to initialize *vli. + * \param vli_pos How many bytes have already been decoded. When + * starting to decode a new integer, *vli_pos must + * be initialized to zero. + * \param in Beginning of the input buffer + * \param in_pos The next byte will be read from in[*in_pos]. + * \param in_size Size of the input buffer; the first byte that + * won't be read is in[in_size]. + * + * \return - LZMA_OK: So far all OK, but the integer is not + * completely decoded yet. + * - LZMA_STREAM_END: Integer successfully decoded. + * - LZMA_BUF_ERROR: No input data (*in_pos == in_size) + * - LZMA_DATA_ERROR: Integer is longer than nine bytes. + * - LZMA_PROG_ERROR: Arguments are not sane. + */ +extern lzma_ret lzma_vli_decode(lzma_vli *lzma_restrict vli, + size_t *lzma_restrict vli_pos, const uint8_t *lzma_restrict in, + size_t *lzma_restrict in_pos, size_t in_size); + + +/** + * \brief Decodes variable-length integer reading buffer backwards + * + * The variable-length integer encoding is designed so that it can be read + * either from the beginning to the end, or from the end to the beginning. + * This feature is needed to make the Stream parseable backwards; + * specifically, to read the Backward Size field in Stream Footer. + * + * \param vli Pointer to variable to hold the decoded integer. + * \param in Beginning of the input buffer + * \param in_size Number of bytes available in the in[] buffer. + * On successful decoding, this is updated to match + * the number of bytes used. (in[*in_size - 1] is the + * first byte to process. After successful decoding, + * in[*in_size] will point to the first byte of the + * variable-length integer.) + * + * \return - LZMA_OK: Decoding successful + * - LZMA_DATA_ERROR: No valid variable-length integer was found. + * - LZMA_BUF_ERROR: Not enough input. Note that in practice, + * this tends to be a sign of broken input, because the + * applications usually do give as much input to this function + * as the applications have available. + */ +extern lzma_ret lzma_vli_reverse_decode( + lzma_vli *vli, const uint8_t *in, size_t *in_size); + + +/** + * \brief Gets the minimum number of bytes required to encode vli + * + * \return Number of bytes on success (1-9). If vli isn't valid, + * zero is returned. + */ +extern size_t lzma_vli_size(lzma_vli vli); |