{ "content": [ { "type": "text", "text": "# bzip2 (info)\n\n## NAME\n\nbzip2, bunzip2 - a block-sorting file compressor, v1.0.8 bzcat - decompresses files to stdout bzip2recover - recovers data from damaged bzip2 files\n\n## SYNOPSIS\n\nbzip2 [ -cdfkqstvzVL123456789 ] [ filenames ... ]\nbzip2 [ -h|--help ]\nbunzip2 [ -fkvsVL ] [ filenames ... ]\nbunzip2 [ -h|--help ]\nbzcat [ -s ] [ filenames ... ]\nbzcat [ -h|--help ]\nbzip2recover filename\n\n## DESCRIPTION\n\nbzip2 compresses files using the Burrows-Wheeler block sorting text\ncompression algorithm, and Huffman coding. Compression is generally\nconsiderably better than that achieved by more conventional\nLZ77/LZ78-based compressors, and approaches the performance of the PPM\nfamily of statistical compressors.\n\n## Sections\n\n- **NAME**\n- **SYNOPSIS**\n- **DESCRIPTION**\n- **OPTIONS**\n- **MEMORY MANAGEMENT**\n- **RECOVERING DATA FROM DAMAGED FILES**\n- **PERFORMANCE NOTES**\n- **CAVEATS**\n- **AUTHOR**\n\nUse structuredContent.sections for detailed options, examples, and full documentation.\n" } ], "structuredContent": { "command": "bzip2", "section": "", "mode": "info", "summary": "bzip2, bunzip2 - a block-sorting file compressor, v1.0.8 bzcat - decompresses files to stdout bzip2recover - recovers data from damaged bzip2 files", "synopsis": "bzip2 [ -cdfkqstvzVL123456789 ] [ filenames ... ]\nbzip2 [ -h|--help ]\nbunzip2 [ -fkvsVL ] [ filenames ... ]\nbunzip2 [ -h|--help ]\nbzcat [ -s ] [ filenames ... ]\nbzcat [ -h|--help ]\nbzip2recover filename", "tldr_summary": null, "tldr_examples": [], "tldr_source": null, "flags": [], "examples": [], "see_also": [], "section_outline": [ { "name": "NAME", "lines": 4, "subsections": [] }, { "name": "SYNOPSIS", "lines": 8, "subsections": [] }, { "name": "DESCRIPTION", "lines": 90, "subsections": [] }, { "name": "OPTIONS", "lines": 79, "subsections": [] }, { "name": "MEMORY MANAGEMENT", "lines": 64, "subsections": [] }, { "name": "RECOVERING DATA FROM DAMAGED FILES", "lines": 28, "subsections": [] }, { "name": "PERFORMANCE NOTES", "lines": 20, "subsections": [] }, { "name": "CAVEATS", "lines": 21, "subsections": [] }, { "name": "AUTHOR", "lines": 20, "subsections": [] } ], "sections": { "NAME": { "content": "bzip2, bunzip2 - a block-sorting file compressor, v1.0.8\nbzcat - decompresses files to stdout\nbzip2recover - recovers data from damaged bzip2 files\n", "subsections": [] }, "SYNOPSIS": { "content": "bzip2 [ -cdfkqstvzVL123456789 ] [ filenames ... ]\nbzip2 [ -h|--help ]\nbunzip2 [ -fkvsVL ] [ filenames ... ]\nbunzip2 [ -h|--help ]\nbzcat [ -s ] [ filenames ... ]\nbzcat [ -h|--help ]\nbzip2recover filename\n", "subsections": [] }, "DESCRIPTION": { "content": "bzip2 compresses files using the Burrows-Wheeler block sorting text\ncompression algorithm, and Huffman coding. Compression is generally\nconsiderably better than that achieved by more conventional\nLZ77/LZ78-based compressors, and approaches the performance of the PPM\nfamily of statistical compressors.\n\nThe command-line options are deliberately very similar to those of GNU\ngzip, but they are not identical.\n\nbzip2 expects a list of file names to accompany the command-line flags.\nEach file is replaced by a compressed version of itself, with the name\n\"originalname.bz2\". Each compressed file has the same modification\ndate, permissions, and, when possible, ownership as the corresponding\noriginal, so that these properties can be correctly restored at decom-\npression time. File name handling is naive in the sense that there is\nno mechanism for preserving original file names, permissions, owner-\nships or dates in filesystems which lack these concepts, or have seri-\nous file name length restrictions, such as MS-DOS.\n\nbzip2 and bunzip2 will by default not overwrite existing files. If you\nwant this to happen, specify the -f flag.\n\nIf no file names are specified, bzip2 compresses from standard input to\nstandard output. In this case, bzip2 will decline to write compressed\noutput to a terminal, as this would be entirely incomprehensible and\ntherefore pointless.\n\nbunzip2 (or bzip2 -d) decompresses all specified files. Files which\nwere not created by bzip2 will be detected and ignored, and a warning\nissued. bzip2 attempts to guess the filename for the decompressed file\nfrom that of the compressed file as follows:\n\nfilename.bz2 becomes filename\nfilename.bz becomes filename\nfilename.tbz2 becomes filename.tar\nfilename.tbz becomes filename.tar\nanyothername becomes anyothername.out\n\nIf the file does not end in one of the recognised endings, .bz2, .bz,\n.tbz2 or .tbz, bzip2 complains that it cannot guess the name of the\noriginal file, and uses the original name with .out appended.\n\nAs with compression, supplying no filenames causes decompression from\nstandard input to standard output.\n\nbunzip2 will correctly decompress a file which is the concatenation of\ntwo or more compressed files. The result is the concatenation of the\ncorresponding uncompressed files. Integrity testing (-t) of concate-\nnated compressed files is also supported.\n\nYou can also compress or decompress files to the standard output by\ngiving the -c flag. Multiple files may be compressed and decompressed\nlike this. The resulting outputs are fed sequentially to stdout. Com-\npression of multiple files in this manner generates a stream containing\nmultiple compressed file representations. Such a stream can be decom-\npressed correctly only by bzip2 version 0.9.0 or later. Earlier ver-\nsions of bzip2 will stop after decompressing the first file in the\nstream.\n\nbzcat (or bzip2 -dc) decompresses all specified files to the standard\noutput.\n\nbzip2 will read arguments from the environment variables BZIP2 and\nBZIP, in that order, and will process them before any arguments read\nfrom the command line. This gives a convenient way to supply default\narguments.\n\nCompression is always performed, even if the compressed file is\nslightly larger than the original. Files of less than about one hun-\ndred bytes tend to get larger, since the compression mechanism has a\nconstant overhead in the region of 50 bytes. Random data (including\nthe output of most file compressors) is coded at about 8.05 bits per\nbyte, giving an expansion of around 0.5%.\n\nAs a self-check for your protection, bzip2 uses 32-bit CRCs to make\nsure that the decompressed version of a file is identical to the origi-\nnal. This guards against corruption of the compressed data, and\nagainst undetected bugs in bzip2 (hopefully very unlikely). The\nchances of data corruption going undetected is microscopic, about one\nchance in four billion for each file processed. Be aware, though, that\nthe check occurs upon decompression, so it can only tell you that some-\nthing is wrong. It can't help you recover the original uncompressed\ndata. You can use bzip2recover to try to recover data from damaged\nfiles.\n\nReturn values: 0 for a normal exit, 1 for environmental problems (file\nnot found, invalid flags, I/O errors, &c), 2 to indicate a corrupt com-\npressed file, 3 for an internal consistency error (eg, bug) which\ncaused bzip2 to panic.\n", "subsections": [] }, "OPTIONS": { "content": "-c --stdout\nCompress or decompress to standard output.\n\n-d --decompress\nForce decompression. bzip2, bunzip2 and bzcat are really the\nsame program, and the decision about what actions to take is\ndone on the basis of which name is used. This flag overrides\nthat mechanism, and forces bzip2 to decompress.\n\n-z --compress\nThe complement to -d: forces compression, regardless of the in-\nvocation name.\n\n-t --test\nCheck integrity of the specified file(s), but don't decompress\nthem. This really performs a trial decompression and throws\naway the result.\n\n-f --force\nForce overwrite of output files. Normally, bzip2 will not over-\nwrite existing output files. Also forces bzip2 to break hard\nlinks to files, which it otherwise wouldn't do.\n\nbzip2 normally declines to decompress files which don't have the\ncorrect magic header bytes. If forced (-f), however, it will\npass such files through unmodified. This is how GNU gzip be-\nhaves.\n\n-k --keep\nKeep (don't delete) input files during compression or decompres-\nsion.\n\n-s --small\nReduce memory usage, for compression, decompression and testing.\nFiles are decompressed and tested using a modified algorithm\nwhich only requires 2.5 bytes per block byte. This means any\nfile can be decompressed in 2300 k of memory, albeit at about\nhalf the normal speed.\n\nDuring compression, -s selects a block size of 200 k, which lim-\nits memory use to around the same figure, at the expense of your\ncompression ratio. In short, if your machine is low on memory\n(8 megabytes or less), use -s for everything. See MEMORY MAN-\nAGEMENT below.\n\n-q --quiet\nSuppress non-essential warning messages. Messages pertaining to\nI/O errors and other critical events will not be suppressed.\n\n-v --verbose\nVerbose mode -- show the compression ratio for each file pro-\ncessed. Further -v's increase the verbosity level, spewing out\nlots of information which is primarily of interest for diagnos-\ntic purposes.\n\n-h --help\nPrint a help message and exit.\n\n-L --license -V --version\nDisplay the software version, license terms and conditions.\n\n-1 (or --fast) to -9 (or --best)\nSet the block size to 100 k, 200 k ... 900 k when compressing.\nHas no effect when decompressing. See MEMORY MANAGEMENT below.\nThe --fast and --best aliases are primarily for GNU gzip compat-\nibility. In particular, --fast doesn't make things signifi-\ncantly faster. And --best merely selects the default behaviour.\n\n-- Treats all subsequent arguments as file names, even if they\nstart with a dash. This is so you can handle files with names\nbeginning with a dash, for example: bzip2 -- -myfilename.\n\n--repetitive-fast --repetitive-best\nThese flags are redundant in versions 0.9.5 and above. They\nprovided some coarse control over the behaviour of the sorting\nalgorithm in earlier versions, which was sometimes useful.\n0.9.5 and above have an improved algorithm which renders these\nflags irrelevant.\n", "subsections": [] }, "MEMORY MANAGEMENT": { "content": "bzip2 compresses large files in blocks. The block size affects both\nthe compression ratio achieved, and the amount of memory needed for\ncompression and decompression. The flags -1 through -9 specify the\nblock size to be 100,000 bytes through 900,000 bytes (the default) re-\nspectively. At decompression time, the block size used for compression\nis read from the header of the compressed file, and bunzip2 then allo-\ncates itself just enough memory to decompress the file. Since block\nsizes are stored in compressed files, it follows that the flags -1 to\n-9 are irrelevant to and so ignored during decompression.\n\nCompression and decompression requirements, in bytes, can be estimated\nas:\n\nCompression: 400 k + ( 8 x block size )\n\nDecompression: 100 k + ( 4 x block size ), or\n100 k + ( 2.5 x block size )\n\nLarger block sizes give rapidly diminishing marginal returns. Most of\nthe compression comes from the first two or three hundred k of block\nsize, a fact worth bearing in mind when using bzip2 on small machines.\nIt is also important to appreciate that the decompression memory re-\nquirement is set at compression time by the choice of block size.\n\nFor files compressed with the default 900 k block size, bunzip2 will\nrequire about 3700 kbytes to decompress. To support decompression of\nany file on a 4 megabyte machine, bunzip2 has an option to decompress\nusing approximately half this amount of memory, about 2300 kbytes. De-\ncompression speed is also halved, so you should use this option only\nwhere necessary. The relevant flag is -s.\n\nIn general, try and use the largest block size memory constraints al-\nlow, since that maximises the compression achieved. Compression and\ndecompression speed are virtually unaffected by block size.\n\nAnother significant point applies to files which fit in a single block\n-- that means most files you'd encounter using a large block size. The\namount of real memory touched is proportional to the size of the file,\nsince the file is smaller than a block. For example, compressing a\nfile 20,000 bytes long with the flag -9 will cause the compressor to\nallocate around 7600 k of memory, but only touch 400 k + 20000 * 8 =\n560 kbytes of it. Similarly, the decompressor will allocate 3700 k but\nonly touch 100 k + 20000 * 4 = 180 kbytes.\n\nHere is a table which summarises the maximum memory usage for different\nblock sizes. Also recorded is the total compressed size for 14 files\nof the Calgary Text Compression Corpus totalling 3,141,622 bytes. This\ncolumn gives some feel for how compression varies with block size.\nThese figures tend to understate the advantage of larger block sizes\nfor larger files, since the Corpus is dominated by smaller files.\n\nCompress Decompress Decompress Corpus\nFlag usage usage -s usage Size\n\n-1 1200k 500k 350k 914704\n-2 2000k 900k 600k 877703\n-3 2800k 1300k 850k 860338\n-4 3600k 1700k 1100k 846899\n-5 4400k 2100k 1350k 845160\n-6 5200k 2500k 1600k 838626\n-7 6100k 2900k 1850k 834096\n-8 6800k 3300k 2100k 828642\n-9 7600k 3700k 2350k 828642\n", "subsections": [] }, "RECOVERING DATA FROM DAMAGED FILES": { "content": "bzip2 compresses files in blocks, usually 900 kbytes long. Each block\nis handled independently. If a media or transmission error causes a\nmulti-block .bz2 file to become damaged, it may be possible to recover\ndata from the undamaged blocks in the file.\n\nThe compressed representation of each block is delimited by a 48-bit\npattern, which makes it possible to find the block boundaries with rea-\nsonable certainty. Each block also carries its own 32-bit CRC, so dam-\naged blocks can be distinguished from undamaged ones.\n\nbzip2recover is a simple program whose purpose is to search for blocks\nin .bz2 files, and write each block out into its own .bz2 file. You\ncan then use bzip2 -t to test the integrity of the resulting files, and\ndecompress those which are undamaged.\n\nbzip2recover takes a single argument, the name of the damaged file, and\nwrites a number of files \"rec00001file.bz2\", \"rec00002file.bz2\", etc.,\ncontaining the extracted blocks. The output filenames are designed\nso that the use of wildcards in subsequent processing -- for example,\n\"bzip2 -dc rec*file.bz2 > recovereddata\" -- processes the files in the\ncorrect order.\n\nbzip2recover should be of most use dealing with large .bz2 files, as\nthese will contain many blocks. It is clearly futile to use it on dam-\naged single-block files, since a damaged block cannot be recovered. If\nyou wish to minimise any potential data loss through media or transmis-\nsion errors, you might consider compressing with a smaller block size.\n", "subsections": [] }, "PERFORMANCE NOTES": { "content": "The sorting phase of compression gathers together similar strings in\nthe file. Because of this, files containing very long runs of repeated\nsymbols, like \"aabaabaabaab ...\" (repeated several hundred times) may\ncompress more slowly than normal. Versions 0.9.5 and above fare much\nbetter than previous versions in this respect. The ratio between\nworst-case and average-case compression time is in the region of 10:1.\nFor previous versions, this figure was more like 100:1. You can use\nthe -vvvv option to monitor progress in great detail, if you want.\n\nDecompression speed is unaffected by these phenomena.\n\nbzip2 usually allocates several megabytes of memory to operate in, and\nthen charges all over it in a fairly random fashion. This means that\nperformance, both for compressing and decompressing, is largely deter-\nmined by the speed at which your machine can service cache misses. Be-\ncause of this, small changes to the code to reduce the miss rate have\nbeen observed to give disproportionately large performance improve-\nments. I imagine bzip2 will perform best on machines with very large\ncaches.\n", "subsections": [] }, "CAVEATS": { "content": "I/O error messages are not as helpful as they could be. bzip2 tries\nhard to detect I/O errors and exit cleanly, but the details of what the\nproblem is sometimes seem rather misleading.\n\nThis manual page pertains to version 1.0.8 of bzip2. Compressed data\ncreated by this version is entirely forwards and backwards compatible\nwith the previous public releases, versions 0.1pl2, 0.9.0, 0.9.5,\n1.0.0, 1.0.1, 1.0.2 and above, but with the following exception: 0.9.0\nand above can correctly decompress multiple concatenated compressed\nfiles. 0.1pl2 cannot do this; it will stop after decompressing just\nthe first file in the stream.\n\nbzip2recover versions prior to 1.0.2 used 32-bit integers to represent\nbit positions in compressed files, so they could not handle compressed\nfiles more than 512 megabytes long. Versions 1.0.2 and above use\n64-bit ints on some platforms which support them (GNU supported tar-\ngets, and Windows). To establish whether or not bzip2recover was built\nwith such a limitation, run it without arguments. In any event you can\nbuild yourself an unlimited version if you can recompile it with May-\nbeUInt64 set to be an unsigned 64-bit integer.\n", "subsections": [] }, "AUTHOR": { "content": "Julian Seward, jseward@acm.org.\n\nhttps://sourceware.org/bzip2/\n\nThe ideas embodied in bzip2 are due to (at least) the following people:\nMichael Burrows and David Wheeler (for the block sorting transforma-\ntion), David Wheeler (again, for the Huffman coder), Peter Fenwick (for\nthe structured coding model in the original bzip, and many refine-\nments), and Alistair Moffat, Radford Neal and Ian Witten (for the\narithmetic coder in the original bzip). I am much indebted for their\nhelp, support and advice. See the manual in the source distribution\nfor pointers to sources of documentation. Christian von Roques encour-\naged me to look for faster sorting algorithms, so as to speed up com-\npression. Bela Lubkin encouraged me to improve the worst-case compres-\nsion performance. Donna Robinson XMLised the documentation. The bz*\nscripts are derived from those of GNU gzip. Many people sent patches,\nhelped with portability problems, lent machines, gave advice and were\ngenerally helpful.\n\nbzip2(1)", "subsections": [] } } } }