.\"	$NetBSD: bzip2.1,v 1.5 2019/07/21 21:07:12 wiz Exp $
.\"
.Dd July 13, 2019
.Dt BZIP2 1
.Os
.Sh NAME
.Nm bzip2 ,
.Nm bunzip2 ,
.Nm bzcat ,
.Nm bzip2recover
.Nd block-sorting file compressor
.Sh SYNOPSIS
.Nm bzip2
.Op Fl 123456789cdfkLqstVvz
.Op Ar filename Ar
.Pp
.Nm bunzip2
.Op Fl fkLVvs
.Op Ar filename Ar
.Pp
.Nm bzcat
.Op Fl s
.Op Ar filename Ar
.Pp
.Nm bzip2recover
.Ar filename
.Sh DESCRIPTION
.Nm bzip2
compresses files using the Burrows-Wheeler block sorting
text compression algorithm, and Huffman coding.
Compression is generally considerably better than that achieved by
more conventional LZ77/LZ78-based compressors, and approaches the
performance of the PPM family of statistical compressors.
.Pp
.Nm bzcat
decompresses files to stdout, and
.Nm bzip2recover
recovers data from damaged bzip2 files.
.Pp
The command-line options are deliberately very similar to
those of
.Xr gzip 1 ,
but they are not identical.
.Pp
.Nm bzip2
expects a list of file names to accompany the command-line flags.
Each file is replaced by a compressed version of
itself, with the name
.Dq Pa original_name.bz2 .
Each compressed file has the same modification date, permissions, and,
when possible, ownership as the corresponding original, so that these
properties can be correctly restored at decompression time.
File name handling is naive in the sense that there is no mechanism
for preserving original file names, permissions, ownerships or dates
in filesystems which lack these concepts, or have serious file name
length restrictions, such as
.Tn MS-DOS .
.Nm bzip2
and
.Nm bunzip2
will by default not overwrite existing files.
If you want this to happen, specify the
.Fl f
flag.
.Pp
If no file names are specified,
.Nm bzip2
compresses from standard input to standard output.
In this case,
.Nm bzip2
will decline to write compressed output to a terminal, as this would
be entirely incomprehensible and therefore pointless.
.Pp
.Nm bunzip2
(or
.Nm bzip2 Fl d )
decompresses all specified files.
Files which were not created by
.Nm bzip2
will be detected and ignored, and a warning issued.
.Nm bzip2
attempts to guess the filename for the decompressed file
from that of the compressed file as follows:
.Bl -column "filename.tbz2" "becomes" -offset indent
.It Pa filename.bz2  Ta becomes Ta Pa filename
.It Pa filename.bz   Ta becomes Ta Pa filename
.It Pa filename.tbz2 Ta becomes Ta Pa filename.tar
.It Pa filename.tbz  Ta becomes Ta Pa filename.tar
.It Pa anyothername  Ta becomes Ta Pa anyothername.out
.El
.Pp
If the file does not end in one of the recognised endings,
.Pa .bz2 ,
.Pa .bz ,
.Pa .tbz2 ,
or
.Pa .tbz ,
.Nm bzip2
complains that it cannot guess the name of the original file, and uses
the original name with
.Pa .out
appended.
.Pp
As with compression, supplying no filenames causes decompression from
standard input to standard output.
.Pp
.Nm bunzip2
will correctly decompress a file which is the concatenation of two or
more compressed files.
The result is the concatenation of the corresponding uncompressed
files.
Integrity testing
.Pq Fl t
of concatenated compressed files is also supported.
.Pp
You can also compress or decompress files to the standard output by
giving the
.Fl c
flag.
Multiple files may be compressed and decompressed like this.
The resulting outputs are fed sequentially to stdout.
Compression of multiple files in this manner generates a stream
containing multiple compressed file representations.
Such a stream can be decompressed correctly only by
.Nm bzip2
version 0.9.0 or later.
Earlier versions of
.Nm bzip2
will stop after decompressing
the first file in the stream.
.Pp
.Nm bzcat
(or
.Nm bzip2 Fl dc )
decompresses all specified files to the standard output.
.Pp
Compression is always performed, even if the compressed file is
slightly larger than the original.
Files of less than about one hundred bytes tend to get larger, since
the compression mechanism has a constant overhead in the region of 50
bytes.
Random data (including the output of most file compressors) is coded
at about 8.05 bits per byte, giving an expansion of around 0.5%.
.Pp
As a self-check for your protection,
.Nm bzip2
uses 32-bit CRCs to make sure that the decompressed version of a file
is identical to the original.
This guards against corruption of the compressed data, and against
undetected bugs in
.Nm bzip2
(hopefully very unlikely).
The chances of data corruption going undetected is microscopic, about
one chance in four billion for each file processed.
Be aware, though, that the check occurs upon decompression, so it can
only tell you that something is wrong.
It can't help you recover the original uncompressed data.
You can use
.Nm bzip2recover
to try to recover data from
damaged files.
.Sh OPTIONS
.Bl -tag -width "XXrepetitiveXfastXX"
.It Fl Fl
Treats all subsequent arguments as file names, even if they start with
a dash.
This is so you can handle files with names beginning with a dash, for
example:
.Dl bzip2 -- -myfilename .
.It Fl 1 , Fl Fl fast
to
.It Fl 9 , Fl Fl best
Set the block size to 100 k, 200 k ... 900 k when compressing.
Has no effect when decompressing.
See
.Sx MEMORY MANAGEMENT
below.
The
.Fl Fl fast
and
.Fl Fl best
aliases are primarily for GNU
.Xr gzip 1
compatibility.
In particular,
.Fl Fl fast
doesn't make things significantly faster, and
.Fl Fl best
merely selects the default behaviour.
.It Fl c , Fl Fl stdout
Compress or decompress to standard output.
.It Fl d , Fl Fl decompress
Force decompression.
.Nm bzip2 ,
.Nm bunzip2 ,
and
.Nm bzcat
are really the same program, and the decision about what actions to
take is done on the basis of which name is used.
This flag overrides that mechanism, and forces
.Nm bzip2
to decompress.
.It Fl f , Fl Fl force
Force overwrite of output files.
Normally,
.Nm bzip2
will not overwrite existing output files.
Also forces
.Nm bzip2
to break hard links
to files, which it otherwise wouldn't do.
.Pp
.Nm bzip2
normally declines to decompress files which don't have the correct
magic header bytes.
If forced
.Pq Fl f ,
however, it will pass such files through unmodified.
This is how GNU
.Xr gzip 1
behaves.
.It Fl k , Fl Fl keep
Keep (don't delete) input files during compression
or decompression.
.It Fl L , Fl Fl license
Display the license terms and conditions.
.It Fl q , Fl Fl quiet
Suppress non-essential warning messages.
Messages pertaining to I/O errors and other critical events will not
be suppressed.
.It Fl Fl repetitive-fast
.It Fl Fl repetitive-best
These flags are redundant in versions 0.9.5 and above.
They provided some coarse control over the behaviour of the sorting
algorithm in earlier versions, which was sometimes useful.
0.9.5 and above have an improved algorithm which renders these flags
irrelevant.
.It Fl s , Fl Fl small
Reduce memory usage, for compression, decompression and testing.
Files are decompressed and tested using a modified algorithm which
only requires 2.5 bytes per block byte.
This means any file can be decompressed in 2300k of memory, albeit at
about half the normal speed.
During compression,
.Fl s
selects a block size of 200k, which limits memory use to around the
same figure, at the expense of your compression ratio.
In short, if your machine is low on memory (8 megabytes or less), use
.Fl s
for everything.
See
.Sx MEMORY MANAGEMENT
below.
.It Fl t , Fl Fl test
Check integrity of the specified file(s), but don't decompress them.
This really performs a trial decompression and throws away the result.
.It Fl V , Fl Fl version
Display the software version.
.It Fl v , Fl Fl verbose
Verbose mode: show the compression ratio for each file processed.
Further
.Fl v Ap s
increase the verbosity level, spewing out lots of information which is
primarily of interest for diagnostic purposes.
.It Fl z , Fl Fl compress
The complement to
Fl d :
forces compression, regardless of the invocation name.
.El
.Ss MEMORY MANAGEMENT
.Nm bzip2
compresses large files in blocks.
The block size affects both the compression ratio achieved, and the
amount of memory needed for compression and decompression.
The flags
.Fl 1
through
.Fl 9
specify the block size to be 100,000 bytes through 900,000 bytes (the
default) respectively.
At decompression time, the block size used for compression is read
from the header of the compressed file, and
.Nm bunzip2
then allocates itself just enough memory to decompress the file.
Since block sizes are stored in compressed files, it follows that the
flags
.Fl 1
to
.Fl 9
are irrelevant to and so ignored during decompression.
.Pp
Compression and decompression requirements, in bytes, can be estimated
as:
.Bl -tag -width "Decompression:" -offset indent
.It Compression :
400k + ( 8 x block size )
.It Decompression :
100k + ( 4 x block size ), or 100k + ( 2.5 x block size )
.El
Larger block sizes give rapidly diminishing marginal returns.
Most of the compression comes from the first two or three hundred k of
block size, a fact worth bearing in mind when using
.Nm bzip2
on small machines.
It is also important to appreciate that the decompression memory
requirement is set at compression time by the choice of block size.
.Pp
For files compressed with the default 900k block size,
.Nm bunzip2
will require about 3700 kbytes to decompress.
To support decompression of any file on a 4 megabyte machine,
.Nm bunzip2
has an option to decompress using approximately half this amount of
memory, about 2300 kbytes.
Decompression speed is also halved, so you should use this option only
where necessary.
The relevant flag is
.Fl s .
.Pp
In general, try and use the largest block size memory constraints
allow, since that maximises the compression achieved.
Compression and decompression speed are virtually unaffected by block
size.
.Pp
Another significant point applies to files which fit in a single block
-- that means most files you'd encounter using a large block size.
The amount of real memory touched is proportional to the size of the
file, since the file is smaller than a block.
For example, compressing a file 20,000 bytes long with the flag
.Fl 9
will cause the compressor to allocate around 7600k of memory, but only
touch 400k + 20000 * 8 = 560 kbytes of it.
Similarly, the decompressor will allocate 3700k but only touch 100k +
20000 * 4 = 180 kbytes.
.Pp
Here is a table which summarises the maximum memory usage for different
block sizes.
Also recorded is the total compressed size for 14 files of the Calgary
Text Compression Corpus totalling 3,141,622 bytes.
This column gives some feel for how compression varies with block size.
These figures tend to understate the advantage of larger block sizes
for larger files, since the Corpus is dominated by smaller files.
.Bl -column "Flag" "Compression" "Decompression" "DecompressionXXs" "Corpus size"
.It Sy Flag Ta Sy Compression Ta Sy Decompression Ta Sy Decompression Fl s Ta Sy Corpus size
.It -1 Ta 1200k Ta  500k Ta  350k Ta 914704
.It -2 Ta 2000k Ta  900k Ta  600k Ta 877703
.It -3 Ta 2800k Ta 1300k Ta  850k Ta 860338
.It -4 Ta 3600k Ta 1700k Ta 1100k Ta 846899
.It -5 Ta 4400k Ta 2100k Ta 1350k Ta 845160
.It -6 Ta 5200k Ta 2500k Ta 1600k Ta 838626
.It -7 Ta 6100k Ta 2900k Ta 1850k Ta 834096
.It -8 Ta 6800k Ta 3300k Ta 2100k Ta 828642
.It -9 Ta 7600k Ta 3700k Ta 2350k Ta 828642
.El
.Ss RECOVERING DATA FROM DAMAGED FILES
.Nm bzip2
compresses files in blocks, usually 900kbytes long.
Each block is handled independently.
If a media or transmission error causes a multi-block
.Pa .bz2
file to become damaged, it may be possible to recover data from the
undamaged blocks in the file.
.Pp
The compressed representation of each block is delimited by a 48-bit
pattern, which makes it possible to find the block boundaries with
reasonable certainty.
Each block also carries its own 32-bit CRC, so damaged blocks can be
distinguished from undamaged ones.
.Pp
.Nm bzip2recover
is a simple program whose purpose is to search for blocks in
.Pa .bz2
files, and write each block out into its own
.Pa .bz2
file.
You can then use
.Nm bzip2
.Fl t
to test the integrity of the resulting files, and decompress those
which are undamaged.
.Pp
.Nm bzip2recover
takes a single argument, the name of the damaged file, and writes a
number of files
.Dq Pa rec00001file.bz2 ,
.Dq Pa rec00002file.bz2 ,
etc., containing the extracted blocks.
The output filenames are designed so that the use of wildcards in
subsequent processing -- for example,
.Dl bzip2 -dc rec*file.bz2 \*[Gt] recovered_data
-- processes the files in the correct order.
.Pp
.Nm bzip2recover
should be of most use dealing with large
.Pa .bz2
files, as these will contain many blocks.
It is clearly futile to use it on damaged single-block files, since a
damaged block cannot be recovered.
If you wish to minimise any potential data loss through media or
transmission errors, you might consider compressing with a smaller
block size.
.Ss PERFORMANCE NOTES
The sorting phase of compression gathers together similar strings in
the file.
Because of this, files containing very long runs of repeated
symbols, like
.Dq aabaabaabaab...
(repeated several hundred times) may compress more slowly than normal.
Versions 0.9.5 and above fare much better than previous versions in
this respect.
The ratio between worst-case and average-case compression time is in
the region of 10:1.
For previous versions, this figure was more like 100:1.
You can use the
.Fl vvvv
option to monitor progress in great detail, if you want.
.Pp
Decompression speed is unaffected by these phenomena.
.Pp
.Nm bzip2
usually allocates several megabytes of memory to operate in, and then
charges all over it in a fairly random fashion.
This means that performance, both for compressing and decompressing,
is largely determined by the speed at which your machine can service
cache misses.
Because of this, small changes to the code to reduce the miss rate
have been observed to give disproportionately large performance
improvements.
I imagine
.Nm bzip2
will perform best on machines with very large caches.
.Sh ENVIRONMENT
.Nm bzip2
will read arguments from the environment variables
.Ev BZIP2
and
.Ev BZIP ,
in that order, and will process them before any arguments read from
the command line.
This gives a convenient way to supply default arguments.
.Sh EXIT STATUS
0 for a normal exit, 1 for environmental problems (file not found,
invalid flags, I/O errors, etc.), 2 to indicate a corrupt compressed
file, 3 for an internal consistency error (e.g., bug) which caused
.Nm bzip2
to panic.
.Sh AUTHORS
.An -nosplit
.An Julian Seward
.Aq jseward@bzip.org
.Pp
.Pa http://www.bzip.org
.Pp
The ideas embodied in
.Nm bzip2
are due to (at least) the following people:
.An Michael Burrows
and
.An David Wheeler
(for the block sorting transformation),
.An David Wheeler
(again, for the Huffman coder),
.An Peter Fenwick
(for the structured coding model in the original
.Nm bzip ,
and many refinements), and
.An Alistair Moffat ,
.An Radford Neal ,
and
.An Ian Witten
(for the arithmetic coder in the original
.Nm bzip ) .
I am much indebted for their help, support and advice.
See the manual in the source distribution for pointers to sources of
documentation.
Christian von Roques encouraged me to look for faster sorting
algorithms, so as to speed up compression.
Bela Lubkin encouraged me to improve the worst-case compression
performance.
Donna Robinson XMLised the documentation.
The bz* scripts are derived from those of GNU gzip.
Many people sent patches, helped with portability problems, lent
machines, gave advice and were generally helpful.
.Sh CAVEATS
I/O error messages are not as helpful as they could be.
.Nm bzip2
tries hard to detect I/O errors and exit cleanly, but the details of
what the problem is sometimes seem rather misleading.
.Pp
This manual page pertains to version 1.0.8 of
.Nm bzip2 .
Compressed data created by this version is entirely forwards and
backwards compatible with the previous public releases, versions
0.1pl2, 0.9.0, 0.9.5, 1.0.0, 1.0.1, 1.0.2 and above, but with the
following exception: 0.9.0 and above can correctly decompress multiple
concatenated compressed files.
0.1pl2 cannot do this; it will stop after decompressing just the first
file in the stream.
.Pp
.Nm bzip2recover
versions prior to 1.0.2 used 32-bit integers to represent bit
positions in compressed files, so they could not handle compressed
files more than 512 megabytes long.
Versions 1.0.2 and above use 64-bit ints on some platforms which
support them (GNU supported targets, and Windows).
To establish whether or not
.Nm bzip2recover
was built with such a limitation, run it without arguments.
In any event you can build yourself an unlimited version if you can
recompile it with MaybeUInt64 set to be an unsigned 64-bit integer.