bzip2-1.0.3bzip2-1.0.3

1 files changed, 2966 insertions, 0 deletions

diff --git a/manual.xml b/manual.xml
new file mode 100644
index 0000000..1ab5bd7
--- /dev/null
+++ b/manual.xml
@@ -0,0 +1,2966 @@
+<?xml version="1.0"?> <!-- -*- sgml -*- -->
+<!DOCTYPE book PUBLIC "-//OASIS//DTD DocBook XML V4.2//EN"
+  "http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd"[
+
+<!-- various strings, dates etc. common to all docs -->
+<!ENTITY % common-ents SYSTEM "entities.xml"> %common-ents;
+]>
+
+<book lang="en" id="userman" xreflabel="bzip2 Manual">
+
+ <bookinfo>
+  <title>bzip2 and libbzip2, version 1.0.3</title>
+  <subtitle>A program and library for data compression</subtitle>
+  <copyright>
+   <year>&bz-lifespan;</year>
+   <holder>Julian Seward</holder>
+  </copyright>
+  <releaseinfo>Version &bz-version; of &bz-date;</releaseinfo>
+
+  <authorgroup>
+   <author>
+    <firstname>Julian</firstname>
+    <surname>Seward</surname>
+    <affiliation>
+     <orgname>&bz-url;</orgname>
+    </affiliation>
+   </author>
+  </authorgroup>
+
+  <legalnotice>
+
+  <para>This program, <computeroutput>bzip2</computeroutput>, the
+  associated library <computeroutput>libbzip2</computeroutput>, and
+  all documentation, are copyright &copy; &bz-lifespan; Julian Seward.
+  All rights reserved.</para>
+
+  <para>Redistribution and use in source and binary forms, with
+  or without modification, are permitted provided that the
+  following conditions are met:</para>
+
+  <itemizedlist mark='bullet'>
+
+   <listitem><para>Redistributions of source code must retain the
+   above copyright notice, this list of conditions and the
+   following disclaimer.</para></listitem>
+
+   <listitem><para>The origin of this software must not be
+   misrepresented; you must not claim that you wrote the original
+   software.  If you use this software in a product, an
+   acknowledgment in the product documentation would be
+   appreciated but is not required.</para></listitem>
+
+   <listitem><para>Altered source versions must be plainly marked
+   as such, and must not be misrepresented as being the original
+   software.</para></listitem>
+
+   <listitem><para>The name of the author may not be used to
+   endorse or promote products derived from this software without
+   specific prior written permission.</para></listitem>
+
+  </itemizedlist>
+
+  <para>THIS SOFTWARE IS PROVIDED BY THE AUTHOR "AS IS" AND ANY
+  EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO,
+  THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
+  PARTICULAR PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE
+  AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
+  EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED
+  TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+  DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+  ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+  LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
+  IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF
+  THE POSSIBILITY OF SUCH DAMAGE.</para>
+
+ <para>PATENTS: To the best of my knowledge,
+ <computeroutput>bzip2</computeroutput> and
+ <computeroutput>libbzip2</computeroutput> do not use any patented
+ algorithms.  However, I do not have the resources to carry
+ out a patent search.  Therefore I cannot give any guarantee of
+ the above statement.
+ </para>
+
+</legalnotice>
+
+</bookinfo>
+
+
+
+<chapter id="intro" xreflabel="Introduction">
+<title>Introduction</title>
+
+<para><computeroutput>bzip2</computeroutput> 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.</para>
+
+<para><computeroutput>bzip2</computeroutput> is built on top of
+<computeroutput>libbzip2</computeroutput>, a flexible library for
+handling compressed data in the
+<computeroutput>bzip2</computeroutput> format.  This manual
+describes both how to use the program and how to work with the
+library interface.  Most of the manual is devoted to this
+library, not the program, which is good news if your interest is
+only in the program.</para>
+
+<itemizedlist mark='bullet'>
+
+ <listitem><para><xref linkend="using"/> describes how to use
+ <computeroutput>bzip2</computeroutput>; this is the only part
+ you need to read if you just want to know how to operate the
+ program.</para></listitem>
+
+ <listitem><para><xref linkend="libprog"/> describes the
+ programming interfaces in detail, and</para></listitem>
+
+ <listitem><para><xref linkend="misc"/> records some
+ miscellaneous notes which I thought ought to be recorded
+ somewhere.</para></listitem>
+
+</itemizedlist>
+
+</chapter>
+
+
+<chapter id="using" xreflabel="How to use bzip2">
+<title>How to use bzip2</title>
+
+<para>This chapter contains a copy of the
+<computeroutput>bzip2</computeroutput> man page, and nothing
+else.</para>
+
+<sect1 id="name" xreflabel="NAME">
+<title>NAME</title>
+
+<itemizedlist mark='bullet'>
+
+ <listitem><para><computeroutput>bzip2</computeroutput>,
+  <computeroutput>bunzip2</computeroutput> - a block-sorting file
+  compressor, v1.0.3</para></listitem>
+
+ <listitem><para><computeroutput>bzcat</computeroutput> -
+   decompresses files to stdout</para></listitem>
+
+ <listitem><para><computeroutput>bzip2recover</computeroutput> -
+   recovers data from damaged bzip2 files</para></listitem>
+
+</itemizedlist>
+
+</sect1>
+
+
+<sect1 id="synopsis" xreflabel="SYNOPSIS">
+<title>SYNOPSIS</title>
+
+<itemizedlist mark='bullet'>
+
+ <listitem><para><computeroutput>bzip2</computeroutput> [
+  -cdfkqstvzVL123456789 ] [ filenames ...  ]</para></listitem>
+
+ <listitem><para><computeroutput>bunzip2</computeroutput> [
+  -fkvsVL ] [ filenames ...  ]</para></listitem>
+
+ <listitem><para><computeroutput>bzcat</computeroutput> [ -s ] [
+  filenames ...  ]</para></listitem>
+
+ <listitem><para><computeroutput>bzip2recover</computeroutput>
+  filename</para></listitem>
+
+</itemizedlist>
+
+</sect1>
+
+
+<sect1 id="description" xreflabel="DESCRIPTION">
+<title>DESCRIPTION</title>
+
+<para><computeroutput>bzip2</computeroutput> 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.</para>
+
+<para>The command-line options are deliberately very similar to
+those of GNU <computeroutput>gzip</computeroutput>, but they are
+not identical.</para>
+
+<para><computeroutput>bzip2</computeroutput> 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
+<computeroutput>original_name.bz2</computeroutput>.  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
+MS-DOS.</para>
+
+<para><computeroutput>bzip2</computeroutput> and
+<computeroutput>bunzip2</computeroutput> will by default not
+overwrite existing files.  If you want this to happen, specify
+the <computeroutput>-f</computeroutput> flag.</para>
+
+<para>If no file names are specified,
+<computeroutput>bzip2</computeroutput> compresses from standard
+input to standard output.  In this case,
+<computeroutput>bzip2</computeroutput> will decline to write
+compressed output to a terminal, as this would be entirely
+incomprehensible and therefore pointless.</para>
+
+<para><computeroutput>bunzip2</computeroutput> (or
+<computeroutput>bzip2 -d</computeroutput>) decompresses all
+specified files.  Files which were not created by
+<computeroutput>bzip2</computeroutput> will be detected and
+ignored, and a warning issued.
+<computeroutput>bzip2</computeroutput> attempts to guess the
+filename for the decompressed file from that of the compressed
+file as follows:</para>
+
+<itemizedlist mark='bullet'>
+
+ <listitem><para><computeroutput>filename.bz2 </computeroutput>
+  becomes
+  <computeroutput>filename</computeroutput></para></listitem>
+
+ <listitem><para><computeroutput>filename.bz </computeroutput>
+  becomes
+  <computeroutput>filename</computeroutput></para></listitem>
+
+ <listitem><para><computeroutput>filename.tbz2</computeroutput>
+  becomes
+  <computeroutput>filename.tar</computeroutput></para></listitem>
+
+ <listitem><para><computeroutput>filename.tbz </computeroutput>
+  becomes
+  <computeroutput>filename.tar</computeroutput></para></listitem>
+
+ <listitem><para><computeroutput>anyothername </computeroutput>
+  becomes
+  <computeroutput>anyothername.out</computeroutput></para></listitem>
+
+</itemizedlist>
+
+<para>If the file does not end in one of the recognised endings,
+<computeroutput>.bz2</computeroutput>,
+<computeroutput>.bz</computeroutput>,
+<computeroutput>.tbz2</computeroutput> or
+<computeroutput>.tbz</computeroutput>,
+<computeroutput>bzip2</computeroutput> complains that it cannot
+guess the name of the original file, and uses the original name
+with <computeroutput>.out</computeroutput> appended.</para>
+
+<para>As with compression, supplying no filenames causes
+decompression from standard input to standard output.</para>
+
+<para><computeroutput>bunzip2</computeroutput> 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
+(<computeroutput>-t</computeroutput>) of concatenated compressed
+files is also supported.</para>
+
+<para>You can also compress or decompress files to the standard
+output by giving the <computeroutput>-c</computeroutput> 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
+<computeroutput>bzip2</computeroutput> version 0.9.0 or later.
+Earlier versions of <computeroutput>bzip2</computeroutput> will
+stop after decompressing the first file in the stream.</para>
+
+<para><computeroutput>bzcat</computeroutput> (or
+<computeroutput>bzip2 -dc</computeroutput>) decompresses all
+specified files to the standard output.</para>
+
+<para><computeroutput>bzip2</computeroutput> will read arguments
+from the environment variables
+<computeroutput>BZIP2</computeroutput> and
+<computeroutput>BZIP</computeroutput>, in that order, and will
+process them before any arguments read from the command line.
+This gives a convenient way to supply default arguments.</para>
+
+<para>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%.</para>
+
+<para>As a self-check for your protection,
+<computeroutput>bzip2</computeroutput> 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
+<computeroutput>bzip2</computeroutput> (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
+<computeroutput>bzip2recover</computeroutput> to try to recover
+data from damaged files.</para>
+
+<para>Return values: 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 (eg, bug) which caused
+<computeroutput>bzip2</computeroutput> to panic.</para>
+
+</sect1>
+
+
+<sect1 id="options" xreflabel="OPTIONS">
+<title>OPTIONS</title>
+
+<variablelist>
+
+ <varlistentry>
+ <term><computeroutput>-c --stdout</computeroutput></term>
+ <listitem><para>Compress or decompress to standard
+  output.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><computeroutput>-d --decompress</computeroutput></term>
+ <listitem><para>Force decompression.
+  <computeroutput>bzip2</computeroutput>,
+  <computeroutput>bunzip2</computeroutput> and
+  <computeroutput>bzcat</computeroutput> 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 bzip2 to decompress.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><computeroutput>-z --compress</computeroutput></term>
+ <listitem><para>The complement to
+  <computeroutput>-d</computeroutput>: forces compression,
+  regardless of the invokation name.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><computeroutput>-t --test</computeroutput></term>
+ <listitem><para>Check integrity of the specified file(s), but
+  don't decompress them.  This really performs a trial
+  decompression and throws away the result.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><computeroutput>-f --force</computeroutput></term>
+ <listitem><para>Force overwrite of output files.  Normally,
+  <computeroutput>bzip2</computeroutput> will not overwrite
+  existing output files.  Also forces
+  <computeroutput>bzip2</computeroutput> to break hard links to
+  files, which it otherwise wouldn't do.</para>
+  <para><computeroutput>bzip2</computeroutput> normally declines
+  to decompress files which don't have the correct magic header
+  bytes. If forced (<computeroutput>-f</computeroutput>),
+  however, it will pass such files through unmodified. This is
+  how GNU <computeroutput>gzip</computeroutput> behaves.</para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><computeroutput>-k --keep</computeroutput></term>
+ <listitem><para>Keep (don't delete) input files during
+  compression or decompression.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><computeroutput>-s --small</computeroutput></term>
+ <listitem><para>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.</para>
+  <para>During compression, <computeroutput>-s</computeroutput>
+  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 <computeroutput>-s</computeroutput> for everything.  See
+  <xref linkend="memory-management"/> below.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><computeroutput>-q --quiet</computeroutput></term>
+ <listitem><para>Suppress non-essential warning messages.
+  Messages pertaining to I/O errors and other critical events
+  will not be suppressed.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><computeroutput>-v --verbose</computeroutput></term>
+ <listitem><para>Verbose mode -- show the compression ratio for
+  each file processed.  Further
+  <computeroutput>-v</computeroutput>'s increase the verbosity
+  level, spewing out lots of information which is primarily of
+  interest for diagnostic purposes.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><computeroutput>-L --license -V --version</computeroutput></term>
+ <listitem><para>Display the software version, license terms and
+  conditions.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><computeroutput>-1</computeroutput> (or
+ <computeroutput>--fast</computeroutput>) to
+ <computeroutput>-9</computeroutput> (or
+ <computeroutput>-best</computeroutput>)</term>
+ <listitem><para>Set the block size to 100 k, 200 k ...  900 k
+  when compressing.  Has no effect when decompressing.  See <xref
+  linkend="memory-management" /> below.  The
+  <computeroutput>--fast</computeroutput> and
+  <computeroutput>--best</computeroutput> aliases are primarily
+  for GNU <computeroutput>gzip</computeroutput> compatibility.
+  In particular, <computeroutput>--fast</computeroutput> doesn't
+  make things significantly faster.  And
+  <computeroutput>--best</computeroutput> merely selects the
+  default behaviour.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><computeroutput>--</computeroutput></term>
+ <listitem><para>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:
+  <computeroutput>bzip2 --
+  -myfilename</computeroutput>.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><computeroutput>--repetitive-fast</computeroutput></term>
+ <term><computeroutput>--repetitive-best</computeroutput></term>
+ <listitem><para>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.</para></listitem>
+ </varlistentry>
+
+</variablelist>
+
+</sect1>
+
+
+<sect1 id="memory-management" xreflabel="MEMORY MANAGEMENT">
+<title>MEMORY MANAGEMENT</title>
+
+<para><computeroutput>bzip2</computeroutput> 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 <computeroutput>-1</computeroutput>
+through <computeroutput>-9</computeroutput> 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
+<computeroutput>bunzip2</computeroutput> then allocates itself
+just enough memory to decompress the file.  Since block sizes are
+stored in compressed files, it follows that the flags
+<computeroutput>-1</computeroutput> to
+<computeroutput>-9</computeroutput> are irrelevant to and so
+ignored during decompression.</para>
+
+<para>Compression and decompression requirements, in bytes, can be
+estimated as:</para>
+<programlisting>
+Compression:   400k + ( 8 x block size )
+
+Decompression: 100k + ( 4 x block size ), or
+               100k + ( 2.5 x block size )
+</programlisting>
+
+<para>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 <computeroutput>bzip2</computeroutput> 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.</para>
+
+<para>For files compressed with the default 900k block size,
+<computeroutput>bunzip2</computeroutput> will require about 3700
+kbytes to decompress.  To support decompression of any file on a
+4 megabyte machine, <computeroutput>bunzip2</computeroutput> 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 <computeroutput>-s</computeroutput>.</para>
+
+<para>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.</para>
+
+<para>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 <computeroutput>-9</computeroutput> 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.</para>
+
+<para>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.</para>
+
+<programlisting>
+        Compress   Decompress   Decompress   Corpus
+Flag     usage      usage       -s usage     Size
+
+ -1      1200k       500k         350k      914704
+ -2      2000k       900k         600k      877703
+ -3      2800k      1300k         850k      860338
+ -4      3600k      1700k        1100k      846899
+ -5      4400k      2100k        1350k      845160
+ -6      5200k      2500k        1600k      838626
+ -7      6100k      2900k        1850k      834096
+ -8      6800k      3300k        2100k      828642
+ -9      7600k      3700k        2350k      828642
+</programlisting>
+
+</sect1>
+
+
+<sect1 id="recovering" xreflabel="RECOVERING DATA FROM DAMAGED FILES">
+<title>RECOVERING DATA FROM DAMAGED FILES</title>
+
+<para><computeroutput>bzip2</computeroutput> compresses files in
+blocks, usually 900kbytes long.  Each block is handled
+independently.  If a media or transmission error causes a
+multi-block <computeroutput>.bz2</computeroutput> file to become
+damaged, it may be possible to recover data from the undamaged
+blocks in the file.</para>
+
+<para>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.</para>
+
+<para><computeroutput>bzip2recover</computeroutput> is a simple
+program whose purpose is to search for blocks in
+<computeroutput>.bz2</computeroutput> files, and write each block
+out into its own <computeroutput>.bz2</computeroutput> file.  You
+can then use <computeroutput>bzip2 -t</computeroutput> to test
+the integrity of the resulting files, and decompress those which
+are undamaged.</para>
+
+<para><computeroutput>bzip2recover</computeroutput> takes a
+single argument, the name of the damaged file, and writes a
+number of files <computeroutput>rec0001file.bz2</computeroutput>,
+<computeroutput>rec0002file.bz2</computeroutput>, etc, containing
+the extracted blocks.  The output filenames are designed so that
+the use of wildcards in subsequent processing -- for example,
+<computeroutput>bzip2 -dc rec*file.bz2 &#62;
+recovered_data</computeroutput> -- lists the files in the correct
+order.</para>
+
+<para><computeroutput>bzip2recover</computeroutput> should be of
+most use dealing with large <computeroutput>.bz2</computeroutput>
+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.</para>
+
+</sect1>
+
+
+<sect1 id="performance" xreflabel="PERFORMANCE NOTES">
+<title>PERFORMANCE NOTES</title>
+
+<para>The sorting phase of compression gathers together similar
+strings in the file.  Because of this, files containing very long
+runs of repeated symbols, like "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
+<computeroutput>-vvvv</computeroutput> option to monitor progress
+in great detail, if you want.</para>
+
+<para>Decompression speed is unaffected by these
+phenomena.</para>
+
+<para><computeroutput>bzip2</computeroutput> 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
+<computeroutput>bzip2</computeroutput> will perform best on
+machines with very large caches.</para>
+
+</sect1>
+
+
+
+<sect1 id="caveats" xreflabel="CAVEATS">
+<title>CAVEATS</title>
+
+<para>I/O error messages are not as helpful as they could be.
+<computeroutput>bzip2</computeroutput> tries hard to detect I/O
+errors and exit cleanly, but the details of what the problem is
+sometimes seem rather misleading.</para>
+
+<para>This manual page pertains to version &bz-version; of
+<computeroutput>bzip2</computeroutput>.  Compressed data created
+by this version is entirely forwards and backwards compatible
+with the previous public releases, versions 0.1pl2, 0.9.0 and
+0.9.5, 1.0.0, 1.0.1 and 1.0.2, 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.</para>
+
+<para><computeroutput>bzip2recover</computeroutput> versions
+prior to 1.0.2 used 32-bit integers to represent bit positions in
+compressed files, so it 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
+<computeroutput>bzip2recover</computeroutput> 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
+<computeroutput>MaybeUInt64</computeroutput> set to be an
+unsigned 64-bit integer.</para>
+
+</sect1>
+
+
+
+<sect1 id="author" xreflabel="AUTHOR">
+<title>AUTHOR</title>
+
+<para>Julian Seward,
+<computeroutput>&bz-email;</computeroutput></para>
+
+<para>The ideas embodied in
+<computeroutput>bzip2</computeroutput> are due to (at least) the
+following people: Michael Burrows and David Wheeler (for the
+block sorting transformation), David Wheeler (again, for the
+Huffman coder), Peter Fenwick (for the structured coding model in
+the original <computeroutput>bzip</computeroutput>, and many
+refinements), and Alistair Moffat, Radford Neal and Ian Witten
+(for the arithmetic coder in the original
+<computeroutput>bzip</computeroutput>).  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.
+Many people sent
+patches, helped with portability problems, lent machines, gave
+advice and were generally helpful.</para>
+
+</sect1>
+
+</chapter>
+
+
+
+<chapter id="libprog" xreflabel="Programming with libbzip2">
+<title>
+Programming with <computeroutput>libbzip2</computeroutput>
+</title>
+
+<para>This chapter describes the programming interface to
+<computeroutput>libbzip2</computeroutput>.</para>
+
+<para>For general background information, particularly about
+memory use and performance aspects, you'd be well advised to read
+<xref linkend="using"/> as well.</para>
+
+
+<sect1 id="top-level" xreflabel="Top-level structure">
+<title>Top-level structure</title>
+
+<para><computeroutput>libbzip2</computeroutput> is a flexible
+library for compressing and decompressing data in the
+<computeroutput>bzip2</computeroutput> data format.  Although
+packaged as a single entity, it helps to regard the library as
+three separate parts: the low level interface, and the high level
+interface, and some utility functions.</para>
+
+<para>The structure of
+<computeroutput>libbzip2</computeroutput>'s interfaces is similar
+to that of Jean-loup Gailly's and Mark Adler's excellent
+<computeroutput>zlib</computeroutput> library.</para>
+
+<para>All externally visible symbols have names beginning
+<computeroutput>BZ2_</computeroutput>.  This is new in version
+1.0.  The intention is to minimise pollution of the namespaces of
+library clients.</para>
+
+<para>To use any part of the library, you need to
+<computeroutput>#include &lt;bzlib.h&gt;</computeroutput>
+into your sources.</para>
+
+
+
+<sect2 id="ll-summary" xreflabel="Low-level summary">
+<title>Low-level summary</title>
+
+<para>This interface provides services for compressing and
+decompressing data in memory.  There's no provision for dealing
+with files, streams or any other I/O mechanisms, just straight
+memory-to-memory work.  In fact, this part of the library can be
+compiled without inclusion of
+<computeroutput>stdio.h</computeroutput>, which may be helpful
+for embedded applications.</para>
+
+<para>The low-level part of the library has no global variables
+and is therefore thread-safe.</para>
+
+<para>Six routines make up the low level interface:
+<computeroutput>BZ2_bzCompressInit</computeroutput>,
+<computeroutput>BZ2_bzCompress</computeroutput>, and
+<computeroutput>BZ2_bzCompressEnd</computeroutput> for
+compression, and a corresponding trio
+<computeroutput>BZ2_bzDecompressInit</computeroutput>,
+<computeroutput>BZ2_bzDecompress</computeroutput> and
+<computeroutput>BZ2_bzDecompressEnd</computeroutput> for
+decompression.  The <computeroutput>*Init</computeroutput>
+functions allocate memory for compression/decompression and do
+other initialisations, whilst the
+<computeroutput>*End</computeroutput> functions close down
+operations and release memory.</para>
+
+<para>The real work is done by
+<computeroutput>BZ2_bzCompress</computeroutput> and
+<computeroutput>BZ2_bzDecompress</computeroutput>.  These
+compress and decompress data from a user-supplied input buffer to
+a user-supplied output buffer.  These buffers can be any size;
+arbitrary quantities of data are handled by making repeated calls
+to these functions.  This is a flexible mechanism allowing a
+consumer-pull style of activity, or producer-push, or a mixture
+of both.</para>
+
+</sect2>
+
+
+<sect2 id="hl-summary" xreflabel="High-level summary">
+<title>High-level summary</title>
+
+<para>This interface provides some handy wrappers around the
+low-level interface to facilitate reading and writing
+<computeroutput>bzip2</computeroutput> format files
+(<computeroutput>.bz2</computeroutput> files).  The routines
+provide hooks to facilitate reading files in which the
+<computeroutput>bzip2</computeroutput> data stream is embedded
+within some larger-scale file structure, or where there are
+multiple <computeroutput>bzip2</computeroutput> data streams
+concatenated end-to-end.</para>
+
+<para>For reading files,
+<computeroutput>BZ2_bzReadOpen</computeroutput>,
+<computeroutput>BZ2_bzRead</computeroutput>,
+<computeroutput>BZ2_bzReadClose</computeroutput> and 
+<computeroutput>BZ2_bzReadGetUnused</computeroutput> are
+supplied.  For writing files,
+<computeroutput>BZ2_bzWriteOpen</computeroutput>,
+<computeroutput>BZ2_bzWrite</computeroutput> and
+<computeroutput>BZ2_bzWriteFinish</computeroutput> are
+available.</para>
+
+<para>As with the low-level library, no global variables are used
+so the library is per se thread-safe.  However, if I/O errors
+occur whilst reading or writing the underlying compressed files,
+you may have to consult <computeroutput>errno</computeroutput> to
+determine the cause of the error.  In that case, you'd need a C
+library which correctly supports
+<computeroutput>errno</computeroutput> in a multithreaded
+environment.</para>
+
+<para>To make the library a little simpler and more portable,
+<computeroutput>BZ2_bzReadOpen</computeroutput> and
+<computeroutput>BZ2_bzWriteOpen</computeroutput> require you to
+pass them file handles (<computeroutput>FILE*</computeroutput>s)
+which have previously been opened for reading or writing
+respectively.  That avoids portability problems associated with
+file operations and file attributes, whilst not being much of an
+imposition on the programmer.</para>
+
+</sect2>
+
+
+<sect2 id="util-fns-summary" xreflabel="Utility functions summary">
+<title>Utility functions summary</title>
+
+<para>For very simple needs,
+<computeroutput>BZ2_bzBuffToBuffCompress</computeroutput> and
+<computeroutput>BZ2_bzBuffToBuffDecompress</computeroutput> are
+provided.  These compress data in memory from one buffer to
+another buffer in a single function call.  You should assess
+whether these functions fulfill your memory-to-memory
+compression/decompression requirements before investing effort in
+understanding the more general but more complex low-level
+interface.</para>
+
+<para>Yoshioka Tsuneo
+(<computeroutput>QWF00133@niftyserve.or.jp</computeroutput> /
+<computeroutput>tsuneo-y@is.aist-nara.ac.jp</computeroutput>) has
+contributed some functions to give better
+<computeroutput>zlib</computeroutput> compatibility.  These
+functions are <computeroutput>BZ2_bzopen</computeroutput>,
+<computeroutput>BZ2_bzread</computeroutput>,
+<computeroutput>BZ2_bzwrite</computeroutput>,
+<computeroutput>BZ2_bzflush</computeroutput>,
+<computeroutput>BZ2_bzclose</computeroutput>,
+<computeroutput>BZ2_bzerror</computeroutput> and
+<computeroutput>BZ2_bzlibVersion</computeroutput>.  You may find
+these functions more convenient for simple file reading and
+writing, than those in the high-level interface.  These functions
+are not (yet) officially part of the library, and are minimally
+documented here.  If they break, you get to keep all the pieces.
+I hope to document them properly when time permits.</para>
+
+<para>Yoshioka also contributed modifications to allow the
+library to be built as a Windows DLL.</para>
+
+</sect2>
+
+</sect1>
+
+
+<sect1 id="err-handling" xreflabel="Error handling">
+<title>Error handling</title>
+
+<para>The library is designed to recover cleanly in all
+situations, including the worst-case situation of decompressing
+random data.  I'm not 100% sure that it can always do this, so
+you might want to add a signal handler to catch segmentation
+violations during decompression if you are feeling especially
+paranoid.  I would be interested in hearing more about the
+robustness of the library to corrupted compressed data.</para>
+
+<para>Version 1.0.3 more robust in this respect than any
+previous version.  Investigations with Valgrind (a tool for detecting
+problems with memory management) indicate
+that, at least for the few files I tested, all single-bit errors
+in the decompressed data are caught properly, with no
+segmentation faults, no uses of uninitialised data, no out of
+range reads or writes, and no infinite looping in the decompressor.
+So it's certainly pretty robust, although
+I wouldn't claim it to be totally bombproof.</para>
+
+<para>The file <computeroutput>bzlib.h</computeroutput> contains
+all definitions needed to use the library.  In particular, you
+should definitely not include
+<computeroutput>bzlib_private.h</computeroutput>.</para>
+
+<para>In <computeroutput>bzlib.h</computeroutput>, the various
+return values are defined.  The following list is not intended as
+an exhaustive description of the circumstances in which a given
+value may be returned -- those descriptions are given later.
+Rather, it is intended to convey the rough meaning of each return
+value.  The first five actions are normal and not intended to
+denote an error situation.</para>
+
+<variablelist>
+
+ <varlistentry>
+  <term><computeroutput>BZ_OK</computeroutput></term>
+  <listitem><para>The requested action was completed
+   successfully.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+  <term><computeroutput>BZ_RUN_OK, BZ_FLUSH_OK,
+    BZ_FINISH_OK</computeroutput></term>
+  <listitem><para>In 
+   <computeroutput>BZ2_bzCompress</computeroutput>, the requested
+   flush/finish/nothing-special action was completed
+   successfully.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+  <term><computeroutput>BZ_STREAM_END</computeroutput></term>
+  <listitem><para>Compression of data was completed, or the
+   logical stream end was detected during
+   decompression.</para></listitem>
+ </varlistentry>
+
+</variablelist>
+
+<para>The following return values indicate an error of some
+kind.</para>
+
+<variablelist>
+
+ <varlistentry>
+  <term><computeroutput>BZ_CONFIG_ERROR</computeroutput></term>
+  <listitem><para>Indicates that the library has been improperly
+   compiled on your platform -- a major configuration error.
+   Specifically, it means that
+   <computeroutput>sizeof(char)</computeroutput>,
+   <computeroutput>sizeof(short)</computeroutput> and
+   <computeroutput>sizeof(int)</computeroutput> are not 1, 2 and
+   4 respectively, as they should be.  Note that the library
+   should still work properly on 64-bit platforms which follow
+   the LP64 programming model -- that is, where
+   <computeroutput>sizeof(long)</computeroutput> and
+   <computeroutput>sizeof(void*)</computeroutput> are 8.  Under
+   LP64, <computeroutput>sizeof(int)</computeroutput> is still 4,
+   so <computeroutput>libbzip2</computeroutput>, which doesn't
+   use the <computeroutput>long</computeroutput> type, is
+   OK.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+  <term><computeroutput>BZ_SEQUENCE_ERROR</computeroutput></term>
+  <listitem><para>When using the library, it is important to call
+   the functions in the correct sequence and with data structures
+   (buffers etc) in the correct states.
+   <computeroutput>libbzip2</computeroutput> checks as much as it
+   can to ensure this is happening, and returns
+   <computeroutput>BZ_SEQUENCE_ERROR</computeroutput> if not.
+   Code which complies precisely with the function semantics, as
+   detailed below, should never receive this value; such an event
+   denotes buggy code which you should
+   investigate.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+  <term><computeroutput>BZ_PARAM_ERROR</computeroutput></term>
+  <listitem><para>Returned when a parameter to a function call is
+   out of range or otherwise manifestly incorrect.  As with
+   <computeroutput>BZ_SEQUENCE_ERROR</computeroutput>, this
+   denotes a bug in the client code.  The distinction between
+   <computeroutput>BZ_PARAM_ERROR</computeroutput> and
+   <computeroutput>BZ_SEQUENCE_ERROR</computeroutput> is a bit
+   hazy, but still worth making.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+  <term><computeroutput>BZ_MEM_ERROR</computeroutput></term>
+  <listitem><para>Returned when a request to allocate memory
+   failed.  Note that the quantity of memory needed to decompress
+   a stream cannot be determined until the stream's header has
+   been read.  So
+   <computeroutput>BZ2_bzDecompress</computeroutput> and
+   <computeroutput>BZ2_bzRead</computeroutput> may return
+   <computeroutput>BZ_MEM_ERROR</computeroutput> even though some
+   of the compressed data has been read.  The same is not true
+   for compression; once
+   <computeroutput>BZ2_bzCompressInit</computeroutput> or
+   <computeroutput>BZ2_bzWriteOpen</computeroutput> have
+   successfully completed,
+   <computeroutput>BZ_MEM_ERROR</computeroutput> cannot
+   occur.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+  <term><computeroutput>BZ_DATA_ERROR</computeroutput></term>
+  <listitem><para>Returned when a data integrity error is
+   detected during decompression.  Most importantly, this means
+   when stored and computed CRCs for the data do not match.  This
+   value is also returned upon detection of any other anomaly in
+   the compressed data.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+  <term><computeroutput>BZ_DATA_ERROR_MAGIC</computeroutput></term>
+  <listitem><para>As a special case of
+   <computeroutput>BZ_DATA_ERROR</computeroutput>, it is
+   sometimes useful to know when the compressed stream does not
+   start with the correct magic bytes (<computeroutput>'B' 'Z'
+   'h'</computeroutput>).</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+  <term><computeroutput>BZ_IO_ERROR</computeroutput></term>
+  <listitem><para>Returned by
+   <computeroutput>BZ2_bzRead</computeroutput> and
+   <computeroutput>BZ2_bzWrite</computeroutput> when there is an
+   error reading or writing in the compressed file, and by
+   <computeroutput>BZ2_bzReadOpen</computeroutput> and
+   <computeroutput>BZ2_bzWriteOpen</computeroutput> for attempts
+   to use a file for which the error indicator (viz,
+   <computeroutput>ferror(f)</computeroutput>) is set.  On
+   receipt of <computeroutput>BZ_IO_ERROR</computeroutput>, the
+   caller should consult <computeroutput>errno</computeroutput>
+   and/or <computeroutput>perror</computeroutput> to acquire
+   operating-system specific information about the
+   problem.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+  <term><computeroutput>BZ_UNEXPECTED_EOF</computeroutput></term>
+  <listitem><para>Returned by
+   <computeroutput>BZ2_bzRead</computeroutput> when the
+   compressed file finishes before the logical end of stream is
+   detected.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+  <term><computeroutput>BZ_OUTBUFF_FULL</computeroutput></term>
+  <listitem><para>Returned by
+   <computeroutput>BZ2_bzBuffToBuffCompress</computeroutput> and
+   <computeroutput>BZ2_bzBuffToBuffDecompress</computeroutput> to
+   indicate that the output data will not fit into the output
+   buffer provided.</para></listitem>
+ </varlistentry>
+
+</variablelist>
+
+</sect1>
+
+
+
+<sect1 id="low-level" xreflabel=">Low-level interface">
+<title>Low-level interface</title>
+
+
+<sect2 id="bzcompress-init" xreflabel="BZ2_bzCompressInit">
+<title><computeroutput>BZ2_bzCompressInit</computeroutput></title>
+
+<programlisting>
+typedef struct {
+  char *next_in;
+  unsigned int avail_in;
+  unsigned int total_in_lo32;
+  unsigned int total_in_hi32;
+
+  char *next_out;
+  unsigned int avail_out;
+  unsigned int total_out_lo32;
+  unsigned int total_out_hi32;
+
+  void *state;
+
+  void *(*bzalloc)(void *,int,int);
+  void (*bzfree)(void *,void *);
+  void *opaque;
+} bz_stream;
+
+int BZ2_bzCompressInit ( bz_stream *strm, 
+                         int blockSize100k, 
+                         int verbosity,
+                         int workFactor );
+</programlisting>
+
+<para>Prepares for compression.  The
+<computeroutput>bz_stream</computeroutput> structure holds all
+data pertaining to the compression activity.  A
+<computeroutput>bz_stream</computeroutput> structure should be
+allocated and initialised prior to the call.  The fields of
+<computeroutput>bz_stream</computeroutput> comprise the entirety
+of the user-visible data.  <computeroutput>state</computeroutput>
+is a pointer to the private data structures required for
+compression.</para>
+
+<para>Custom memory allocators are supported, via fields
+<computeroutput>bzalloc</computeroutput>,
+<computeroutput>bzfree</computeroutput>, and
+<computeroutput>opaque</computeroutput>.  The value
+<computeroutput>opaque</computeroutput> is passed to as the first
+argument to all calls to <computeroutput>bzalloc</computeroutput>
+and <computeroutput>bzfree</computeroutput>, but is otherwise
+ignored by the library.  The call <computeroutput>bzalloc (
+opaque, n, m )</computeroutput> is expected to return a pointer
+<computeroutput>p</computeroutput> to <computeroutput>n *
+m</computeroutput> bytes of memory, and <computeroutput>bzfree (
+opaque, p )</computeroutput> should free that memory.</para>
+
+<para>If you don't want to use a custom memory allocator, set
+<computeroutput>bzalloc</computeroutput>,
+<computeroutput>bzfree</computeroutput> and
+<computeroutput>opaque</computeroutput> to
+<computeroutput>NULL</computeroutput>, and the library will then
+use the standard <computeroutput>malloc</computeroutput> /
+<computeroutput>free</computeroutput> routines.</para>
+
+<para>Before calling
+<computeroutput>BZ2_bzCompressInit</computeroutput>, fields
+<computeroutput>bzalloc</computeroutput>,
+<computeroutput>bzfree</computeroutput> and
+<computeroutput>opaque</computeroutput> should be filled
+appropriately, as just described.  Upon return, the internal
+state will have been allocated and initialised, and
+<computeroutput>total_in_lo32</computeroutput>,
+<computeroutput>total_in_hi32</computeroutput>,
+<computeroutput>total_out_lo32</computeroutput> and
+<computeroutput>total_out_hi32</computeroutput> will have been
+set to zero.  These four fields are used by the library to inform
+the caller of the total amount of data passed into and out of the
+library, respectively.  You should not try to change them.  As of
+version 1.0, 64-bit counts are maintained, even on 32-bit
+platforms, using the <computeroutput>_hi32</computeroutput>
+fields to store the upper 32 bits of the count.  So, for example,
+the total amount of data in is <computeroutput>(total_in_hi32
+&#60;&#60; 32) + total_in_lo32</computeroutput>.</para>
+
+<para>Parameter <computeroutput>blockSize100k</computeroutput>
+specifies the block size to be used for compression.  It should
+be a value between 1 and 9 inclusive, and the actual block size
+used is 100000 x this figure.  9 gives the best compression but
+takes most memory.</para>
+
+<para>Parameter <computeroutput>verbosity</computeroutput> should
+be set to a number between 0 and 4 inclusive.  0 is silent, and
+greater numbers give increasingly verbose monitoring/debugging
+output.  If the library has been compiled with
+<computeroutput>-DBZ_NO_STDIO</computeroutput>, no such output
+will appear for any verbosity setting.</para>
+
+<para>Parameter <computeroutput>workFactor</computeroutput>
+controls how the compression phase behaves when presented with
+worst case, highly repetitive, input data.  If compression runs
+into difficulties caused by repetitive data, the library switches
+from the standard sorting algorithm to a fallback algorithm.  The
+fallback is slower than the standard algorithm by perhaps a
+factor of three, but always behaves reasonably, no matter how bad
+the input.</para>
+
+<para>Lower values of <computeroutput>workFactor</computeroutput>
+reduce the amount of effort the standard algorithm will expend
+before resorting to the fallback.  You should set this parameter
+carefully; too low, and many inputs will be handled by the
+fallback algorithm and so compress rather slowly, too high, and
+your average-to-worst case compression times can become very
+large.  The default value of 30 gives reasonable behaviour over a
+wide range of circumstances.</para>
+
+<para>Allowable values range from 0 to 250 inclusive.  0 is a
+special case, equivalent to using the default value of 30.</para>
+
+<para>Note that the compressed output generated is the same
+regardless of whether or not the fallback algorithm is
+used.</para>
+
+<para>Be aware also that this parameter may disappear entirely in
+future versions of the library.  In principle it should be
+possible to devise a good way to automatically choose which
+algorithm to use.  Such a mechanism would render the parameter
+obsolete.</para>
+
+<para>Possible return values:</para>
+
+<programlisting>
+BZ_CONFIG_ERROR
+  if the library has been mis-compiled
+BZ_PARAM_ERROR
+  if strm is NULL 
+  or blockSize < 1 or blockSize > 9
+  or verbosity < 0 or verbosity > 4
+  or workFactor < 0 or workFactor > 250
+BZ_MEM_ERROR 
+  if not enough memory is available
+BZ_OK 
+  otherwise
+</programlisting>
+
+<para>Allowable next actions:</para>
+
+<programlisting>
+BZ2_bzCompress
+  if BZ_OK is returned
+  no specific action needed in case of error
+</programlisting>
+
+</sect2>
+
+
+<sect2 id="bzCompress" xreflabel="BZ2_bzCompress">
+<title><computeroutput>BZ2_bzCompress</computeroutput></title>
+
+<programlisting>
+int BZ2_bzCompress ( bz_stream *strm, int action );
+</programlisting>
+
+<para>Provides more input and/or output buffer space for the
+library.  The caller maintains input and output buffers, and
+calls <computeroutput>BZ2_bzCompress</computeroutput> to transfer
+data between them.</para>
+
+<para>Before each call to
+<computeroutput>BZ2_bzCompress</computeroutput>,
+<computeroutput>next_in</computeroutput> should point at the data
+to be compressed, and <computeroutput>avail_in</computeroutput>
+should indicate how many bytes the library may read.
+<computeroutput>BZ2_bzCompress</computeroutput> updates
+<computeroutput>next_in</computeroutput>,
+<computeroutput>avail_in</computeroutput> and
+<computeroutput>total_in</computeroutput> to reflect the number
+of bytes it has read.</para>
+
+<para>Similarly, <computeroutput>next_out</computeroutput> should
+point to a buffer in which the compressed data is to be placed,
+with <computeroutput>avail_out</computeroutput> indicating how
+much output space is available.
+<computeroutput>BZ2_bzCompress</computeroutput> updates
+<computeroutput>next_out</computeroutput>,
+<computeroutput>avail_out</computeroutput> and
+<computeroutput>total_out</computeroutput> to reflect the number
+of bytes output.</para>
+
+<para>You may provide and remove as little or as much data as you
+like on each call of
+<computeroutput>BZ2_bzCompress</computeroutput>.  In the limit,
+it is acceptable to supply and remove data one byte at a time,
+although this would be terribly inefficient.  You should always
+ensure that at least one byte of output space is available at
+each call.</para>
+
+<para>A second purpose of
+<computeroutput>BZ2_bzCompress</computeroutput> is to request a
+change of mode of the compressed stream.</para>
+
+<para>Conceptually, a compressed stream can be in one of four
+states: IDLE, RUNNING, FLUSHING and FINISHING.  Before
+initialisation
+(<computeroutput>BZ2_bzCompressInit</computeroutput>) and after
+termination (<computeroutput>BZ2_bzCompressEnd</computeroutput>),
+a stream is regarded as IDLE.</para>
+
+<para>Upon initialisation
+(<computeroutput>BZ2_bzCompressInit</computeroutput>), the stream
+is placed in the RUNNING state.  Subsequent calls to
+<computeroutput>BZ2_bzCompress</computeroutput> should pass
+<computeroutput>BZ_RUN</computeroutput> as the requested action;
+other actions are illegal and will result in
+<computeroutput>BZ_SEQUENCE_ERROR</computeroutput>.</para>
+
+<para>At some point, the calling program will have provided all
+the input data it wants to.  It will then want to finish up -- in
+effect, asking the library to process any data it might have
+buffered internally.  In this state,
+<computeroutput>BZ2_bzCompress</computeroutput> will no longer
+attempt to read data from
+<computeroutput>next_in</computeroutput>, but it will want to
+write data to <computeroutput>next_out</computeroutput>.  Because
+the output buffer supplied by the user can be arbitrarily small,
+the finishing-up operation cannot necessarily be done with a
+single call of
+<computeroutput>BZ2_bzCompress</computeroutput>.</para>
+
+<para>Instead, the calling program passes
+<computeroutput>BZ_FINISH</computeroutput> as an action to
+<computeroutput>BZ2_bzCompress</computeroutput>.  This changes
+the stream's state to FINISHING.  Any remaining input (ie,
+<computeroutput>next_in[0 .. avail_in-1]</computeroutput>) is
+compressed and transferred to the output buffer.  To do this,
+<computeroutput>BZ2_bzCompress</computeroutput> must be called
+repeatedly until all the output has been consumed.  At that
+point, <computeroutput>BZ2_bzCompress</computeroutput> returns
+<computeroutput>BZ_STREAM_END</computeroutput>, and the stream's
+state is set back to IDLE.
+<computeroutput>BZ2_bzCompressEnd</computeroutput> should then be
+called.</para>
+
+<para>Just to make sure the calling program does not cheat, the
+library makes a note of <computeroutput>avail_in</computeroutput>
+at the time of the first call to
+<computeroutput>BZ2_bzCompress</computeroutput> which has
+<computeroutput>BZ_FINISH</computeroutput> as an action (ie, at
+the time the program has announced its intention to not supply
+any more input).  By comparing this value with that of
+<computeroutput>avail_in</computeroutput> over subsequent calls
+to <computeroutput>BZ2_bzCompress</computeroutput>, the library
+can detect any attempts to slip in more data to compress.  Any
+calls for which this is detected will return
+<computeroutput>BZ_SEQUENCE_ERROR</computeroutput>.  This
+indicates a programming mistake which should be corrected.</para>
+
+<para>Instead of asking to finish, the calling program may ask
+<computeroutput>BZ2_bzCompress</computeroutput> to take all the
+remaining input, compress it and terminate the current
+(Burrows-Wheeler) compression block.  This could be useful for
+error control purposes.  The mechanism is analogous to that for
+finishing: call <computeroutput>BZ2_bzCompress</computeroutput>
+with an action of <computeroutput>BZ_FLUSH</computeroutput>,
+remove output data, and persist with the
+<computeroutput>BZ_FLUSH</computeroutput> action until the value
+<computeroutput>BZ_RUN</computeroutput> is returned.  As with
+finishing, <computeroutput>BZ2_bzCompress</computeroutput>
+detects any attempt to provide more input data once the flush has
+begun.</para>
+
+<para>Once the flush is complete, the stream returns to the
+normal RUNNING state.</para>
+
+<para>This all sounds pretty complex, but isn't really.  Here's a
+table which shows which actions are allowable in each state, what
+action will be taken, what the next state is, and what the
+non-error return values are.  Note that you can't explicitly ask
+what state the stream is in, but nor do you need to -- it can be
+inferred from the values returned by
+<computeroutput>BZ2_bzCompress</computeroutput>.</para>
+
+<programlisting>
+IDLE/any
+  Illegal.  IDLE state only exists after BZ2_bzCompressEnd or
+  before BZ2_bzCompressInit.
+  Return value = BZ_SEQUENCE_ERROR
+
+RUNNING/BZ_RUN
+  Compress from next_in to next_out as much as possible.
+  Next state = RUNNING
+  Return value = BZ_RUN_OK
+
+RUNNING/BZ_FLUSH
+  Remember current value of next_in. Compress from next_in
+  to next_out as much as possible, but do not accept any more input.
+  Next state = FLUSHING
+  Return value = BZ_FLUSH_OK
+
+RUNNING/BZ_FINISH
+  Remember current value of next_in. Compress from next_in
+  to next_out as much as possible, but do not accept any more input.
+  Next state = FINISHING
+  Return value = BZ_FINISH_OK
+
+FLUSHING/BZ_FLUSH
+  Compress from next_in to next_out as much as possible, 
+  but do not accept any more input.
+  If all the existing input has been used up and all compressed
+  output has been removed
+    Next state = RUNNING; Return value = BZ_RUN_OK
+  else
+    Next state = FLUSHING; Return value = BZ_FLUSH_OK
+
+FLUSHING/other     
+  Illegal.
+  Return value = BZ_SEQUENCE_ERROR
+
+FINISHING/BZ_FINISH
+  Compress from next_in to next_out as much as possible,
+  but to not accept any more input.  
+  If all the existing input has been used up and all compressed
+  output has been removed
+    Next state = IDLE; Return value = BZ_STREAM_END
+  else
+    Next state = FINISHING; Return value = BZ_FINISHING
+
+FINISHING/other
+  Illegal.
+  Return value = BZ_SEQUENCE_ERROR
+</programlisting>
+
+
+<para>That still looks complicated?  Well, fair enough.  The
+usual sequence of calls for compressing a load of data is:</para>
+
+<orderedlist>
+
+ <listitem><para>Get started with
+  <computeroutput>BZ2_bzCompressInit</computeroutput>.</para></listitem>
+
+ <listitem><para>Shovel data in and shlurp out its compressed form
+  using zero or more calls of
+  <computeroutput>BZ2_bzCompress</computeroutput> with action =
+  <computeroutput>BZ_RUN</computeroutput>.</para></listitem>
+
+ <listitem><para>Finish up. Repeatedly call
+  <computeroutput>BZ2_bzCompress</computeroutput> with action =
+  <computeroutput>BZ_FINISH</computeroutput>, copying out the
+  compressed output, until
+  <computeroutput>BZ_STREAM_END</computeroutput> is
+  returned.</para></listitem> <listitem><para>Close up and go home.  Call
+  <computeroutput>BZ2_bzCompressEnd</computeroutput>.</para></listitem>
+
+</orderedlist>
+
+<para>If the data you want to compress fits into your input
+buffer all at once, you can skip the calls of
+<computeroutput>BZ2_bzCompress ( ..., BZ_RUN )</computeroutput>
+and just do the <computeroutput>BZ2_bzCompress ( ..., BZ_FINISH
+)</computeroutput> calls.</para>
+
+<para>All required memory is allocated by
+<computeroutput>BZ2_bzCompressInit</computeroutput>.  The
+compression library can accept any data at all (obviously).  So
+you shouldn't get any error return values from the
+<computeroutput>BZ2_bzCompress</computeroutput> calls.  If you
+do, they will be
+<computeroutput>BZ_SEQUENCE_ERROR</computeroutput>, and indicate
+a bug in your programming.</para>
+
+<para>Trivial other possible return values:</para>
+
+<programlisting>
+BZ_PARAM_ERROR
+  if strm is NULL, or strm->s is NULL
+</programlisting>
+
+</sect2>
+
+
+<sect2 id="bzCompress-end" xreflabel="BZ2_bzCompressEnd">
+<title><computeroutput>BZ2_bzCompressEnd</computeroutput></title>
+
+<programlisting>
+int BZ2_bzCompressEnd ( bz_stream *strm );
+</programlisting>
+
+<para>Releases all memory associated with a compression
+stream.</para>
+
+<para>Possible return values:</para>
+
+<programlisting>
+BZ_PARAM_ERROR  if strm is NULL or strm->s is NULL
+BZ_OK           otherwise
+</programlisting>
+
+</sect2>
+
+
+<sect2 id="bzDecompress-init" xreflabel="BZ2_bzDecompressInit">
+<title><computeroutput>BZ2_bzDecompressInit</computeroutput></title>
+
+<programlisting>
+int BZ2_bzDecompressInit ( bz_stream *strm, int verbosity, int small );
+</programlisting>
+
+<para>Prepares for decompression.  As with
+<computeroutput>BZ2_bzCompressInit</computeroutput>, a
+<computeroutput>bz_stream</computeroutput> record should be
+allocated and initialised before the call.  Fields
+<computeroutput>bzalloc</computeroutput>,
+<computeroutput>bzfree</computeroutput> and
+<computeroutput>opaque</computeroutput> should be set if a custom
+memory allocator is required, or made
+<computeroutput>NULL</computeroutput> for the normal
+<computeroutput>malloc</computeroutput> /
+<computeroutput>free</computeroutput> routines.  Upon return, the
+internal state will have been initialised, and
+<computeroutput>total_in</computeroutput> and
+<computeroutput>total_out</computeroutput> will be zero.</para>
+
+<para>For the meaning of parameter
+<computeroutput>verbosity</computeroutput>, see
+<computeroutput>BZ2_bzCompressInit</computeroutput>.</para>
+
+<para>If <computeroutput>small</computeroutput> is nonzero, the
+library will use an alternative decompression algorithm which
+uses less memory but at the cost of decompressing more slowly
+(roughly speaking, half the speed, but the maximum memory
+requirement drops to around 2300k).  See <xref linkend="using"/>
+for more information on memory management.</para>
+
+<para>Note that the amount of memory needed to decompress a
+stream cannot be determined until the stream's header has been
+read, so even if
+<computeroutput>BZ2_bzDecompressInit</computeroutput> succeeds, a
+subsequent <computeroutput>BZ2_bzDecompress</computeroutput>
+could fail with
+<computeroutput>BZ_MEM_ERROR</computeroutput>.</para>
+
+<para>Possible return values:</para>
+
+<programlisting>
+BZ_CONFIG_ERROR
+  if the library has been mis-compiled
+BZ_PARAM_ERROR
+  if ( small != 0 && small != 1 )
+  or (verbosity <; 0 || verbosity > 4)
+BZ_MEM_ERROR
+  if insufficient memory is available
+</programlisting>
+
+<para>Allowable next actions:</para>
+
+<programlisting>
+BZ2_bzDecompress
+  if BZ_OK was returned
+  no specific action required in case of error
+</programlisting>
+
+</sect2>
+
+
+<sect2 id="bzDecompress" xreflabel="BZ2_bzDecompress">
+<title><computeroutput>BZ2_bzDecompress</computeroutput></title>
+
+<programlisting>
+int BZ2_bzDecompress ( bz_stream *strm );
+</programlisting>
+
+<para>Provides more input and/out output buffer space for the
+library.  The caller maintains input and output buffers, and uses
+<computeroutput>BZ2_bzDecompress</computeroutput> to transfer
+data between them.</para>
+
+<para>Before each call to
+<computeroutput>BZ2_bzDecompress</computeroutput>,
+<computeroutput>next_in</computeroutput> should point at the
+compressed data, and <computeroutput>avail_in</computeroutput>
+should indicate how many bytes the library may read.
+<computeroutput>BZ2_bzDecompress</computeroutput> updates
+<computeroutput>next_in</computeroutput>,
+<computeroutput>avail_in</computeroutput> and
+<computeroutput>total_in</computeroutput> to reflect the number
+of bytes it has read.</para>
+
+<para>Similarly, <computeroutput>next_out</computeroutput> should
+point to a buffer in which the uncompressed output is to be
+placed, with <computeroutput>avail_out</computeroutput>
+indicating how much output space is available.
+<computeroutput>BZ2_bzCompress</computeroutput> updates
+<computeroutput>next_out</computeroutput>,
+<computeroutput>avail_out</computeroutput> and
+<computeroutput>total_out</computeroutput> to reflect the number
+of bytes output.</para>
+
+<para>You may provide and remove as little or as much data as you
+like on each call of
+<computeroutput>BZ2_bzDecompress</computeroutput>.  In the limit,
+it is acceptable to supply and remove data one byte at a time,
+although this would be terribly inefficient.  You should always
+ensure that at least one byte of output space is available at
+each call.</para>
+
+<para>Use of <computeroutput>BZ2_bzDecompress</computeroutput> is
+simpler than
+<computeroutput>BZ2_bzCompress</computeroutput>.</para>
+
+<para>You should provide input and remove output as described
+above, and repeatedly call
+<computeroutput>BZ2_bzDecompress</computeroutput> until
+<computeroutput>BZ_STREAM_END</computeroutput> is returned.
+Appearance of <computeroutput>BZ_STREAM_END</computeroutput>
+denotes that <computeroutput>BZ2_bzDecompress</computeroutput>
+has detected the logical end of the compressed stream.
+<computeroutput>BZ2_bzDecompress</computeroutput> will not
+produce <computeroutput>BZ_STREAM_END</computeroutput> until all
+output data has been placed into the output buffer, so once
+<computeroutput>BZ_STREAM_END</computeroutput> appears, you are
+guaranteed to have available all the decompressed output, and
+<computeroutput>BZ2_bzDecompressEnd</computeroutput> can safely
+be called.</para>
+
+<para>If case of an error return value, you should call
+<computeroutput>BZ2_bzDecompressEnd</computeroutput> to clean up
+and release memory.</para>
+
+<para>Possible return values:</para>
+
+<programlisting>
+BZ_PARAM_ERROR
+  if strm is NULL or strm->s is NULL
+  or strm->avail_out < 1
+BZ_DATA_ERROR
+  if a data integrity error is detected in the compressed stream
+BZ_DATA_ERROR_MAGIC
+  if the compressed stream doesn't begin with the right magic bytes
+BZ_MEM_ERROR
+  if there wasn't enough memory available
+BZ_STREAM_END
+  if the logical end of the data stream was detected and all
+  output in has been consumed, eg s-->avail_out > 0
+BZ_OK
+  otherwise
+</programlisting>
+
+<para>Allowable next actions:</para>
+
+<programlisting>
+BZ2_bzDecompress
+  if BZ_OK was returned
+BZ2_bzDecompressEnd
+  otherwise
+</programlisting>
+
+</sect2>
+
+
+<sect2 id="bzDecompress-end" xreflabel="BZ2_bzDecompressEnd">
+<title><computeroutput>BZ2_bzDecompressEnd</computeroutput></title>
+
+<programlisting>
+int BZ2_bzDecompressEnd ( bz_stream *strm );
+</programlisting>
+
+<para>Releases all memory associated with a decompression
+stream.</para>
+
+<para>Possible return values:</para>
+
+<programlisting>
+BZ_PARAM_ERROR
+  if strm is NULL or strm->s is NULL
+BZ_OK
+  otherwise
+</programlisting>
+
+<para>Allowable next actions:</para>
+
+<programlisting>
+  None.
+</programlisting>
+
+</sect2>
+
+</sect1>
+
+
+<sect1 id="hl-interface" xreflabel="High-level interface">
+<title>High-level interface</title>
+
+<para>This interface provides functions for reading and writing
+<computeroutput>bzip2</computeroutput> format files.  First, some
+general points.</para>
+
+<itemizedlist mark='bullet'>
+
+ <listitem><para>All of the functions take an
+  <computeroutput>int*</computeroutput> first argument,
+  <computeroutput>bzerror</computeroutput>.  After each call,
+  <computeroutput>bzerror</computeroutput> should be consulted
+  first to determine the outcome of the call.  If
+  <computeroutput>bzerror</computeroutput> is
+  <computeroutput>BZ_OK</computeroutput>, the call completed
+  successfully, and only then should the return value of the
+  function (if any) be consulted.  If
+  <computeroutput>bzerror</computeroutput> is
+  <computeroutput>BZ_IO_ERROR</computeroutput>, there was an
+  error reading/writing the underlying compressed file, and you
+  should then consult <computeroutput>errno</computeroutput> /
+  <computeroutput>perror</computeroutput> to determine the cause
+  of the difficulty.  <computeroutput>bzerror</computeroutput>
+  may also be set to various other values; precise details are
+  given on a per-function basis below.</para></listitem>
+
+ <listitem><para>If <computeroutput>bzerror</computeroutput> indicates
+  an error (ie, anything except
+  <computeroutput>BZ_OK</computeroutput> and
+  <computeroutput>BZ_STREAM_END</computeroutput>), you should
+  immediately call
+  <computeroutput>BZ2_bzReadClose</computeroutput> (or
+  <computeroutput>BZ2_bzWriteClose</computeroutput>, depending on
+  whether you are attempting to read or to write) to free up all
+  resources associated with the stream.  Once an error has been
+  indicated, behaviour of all calls except
+  <computeroutput>BZ2_bzReadClose</computeroutput>
+  (<computeroutput>BZ2_bzWriteClose</computeroutput>) is
+  undefined.  The implication is that (1)
+  <computeroutput>bzerror</computeroutput> should be checked
+  after each call, and (2) if
+  <computeroutput>bzerror</computeroutput> indicates an error,
+  <computeroutput>BZ2_bzReadClose</computeroutput>
+  (<computeroutput>BZ2_bzWriteClose</computeroutput>) should then
+  be called to clean up.</para></listitem>
+
+ <listitem><para>The <computeroutput>FILE*</computeroutput> arguments
+  passed to <computeroutput>BZ2_bzReadOpen</computeroutput> /
+  <computeroutput>BZ2_bzWriteOpen</computeroutput> should be set
+  to binary mode.  Most Unix systems will do this by default, but
+  other platforms, including Windows and Mac, will not.  If you
+  omit this, you may encounter problems when moving code to new
+  platforms.</para></listitem>
+
+ <listitem><para>Memory allocation requests are handled by
+  <computeroutput>malloc</computeroutput> /
+  <computeroutput>free</computeroutput>.  At present there is no
+  facility for user-defined memory allocators in the file I/O
+  functions (could easily be added, though).</para></listitem>
+
+</itemizedlist>
+
+
+
+<sect2 id="bzreadopen" xreflabel="BZ2_bzReadOpen">
+<title><computeroutput>BZ2_bzReadOpen</computeroutput></title>
+
+<programlisting>
+typedef void BZFILE;
+
+BZFILE *BZ2_bzReadOpen( int *bzerror, FILE *f, 
+                        int verbosity, int small,
+                        void *unused, int nUnused );
+</programlisting>
+
+<para>Prepare to read compressed data from file handle
+<computeroutput>f</computeroutput>.
+<computeroutput>f</computeroutput> should refer to a file which
+has been opened for reading, and for which the error indicator
+(<computeroutput>ferror(f)</computeroutput>)is not set.  If
+<computeroutput>small</computeroutput> is 1, the library will try
+to decompress using less memory, at the expense of speed.</para>
+
+<para>For reasons explained below,
+<computeroutput>BZ2_bzRead</computeroutput> will decompress the
+<computeroutput>nUnused</computeroutput> bytes starting at
+<computeroutput>unused</computeroutput>, before starting to read
+from the file <computeroutput>f</computeroutput>.  At most
+<computeroutput>BZ_MAX_UNUSED</computeroutput> bytes may be
+supplied like this.  If this facility is not required, you should
+pass <computeroutput>NULL</computeroutput> and
+<computeroutput>0</computeroutput> for
+<computeroutput>unused</computeroutput> and
+n<computeroutput>Unused</computeroutput> respectively.</para>
+
+<para>For the meaning of parameters
+<computeroutput>small</computeroutput> and
+<computeroutput>verbosity</computeroutput>, see
+<computeroutput>BZ2_bzDecompressInit</computeroutput>.</para>
+
+<para>The amount of memory needed to decompress a file cannot be
+determined until the file's header has been read.  So it is
+possible that <computeroutput>BZ2_bzReadOpen</computeroutput>
+returns <computeroutput>BZ_OK</computeroutput> but a subsequent
+call of <computeroutput>BZ2_bzRead</computeroutput> will return
+<computeroutput>BZ_MEM_ERROR</computeroutput>.</para>
+
+<para>Possible assignments to
+<computeroutput>bzerror</computeroutput>:</para>
+
+<programlisting>
+BZ_CONFIG_ERROR
+  if the library has been mis-compiled
+BZ_PARAM_ERROR
+  if f is NULL
+  or small is neither 0 nor 1
+  or ( unused == NULL && nUnused != 0 )
+  or ( unused != NULL && !(0 <= nUnused <= BZ_MAX_UNUSED) )
+BZ_IO_ERROR
+  if ferror(f) is nonzero
+BZ_MEM_ERROR
+  if insufficient memory is available
+BZ_OK
+  otherwise.
+</programlisting>
+
+<para>Possible return values:</para>
+
+<programlisting>
+Pointer to an abstract BZFILE
+  if bzerror is BZ_OK
+NULL
+  otherwise
+</programlisting>
+
+<para>Allowable next actions:</para>
+
+<programlisting>
+BZ2_bzRead
+  if bzerror is BZ_OK
+BZ2_bzClose
+  otherwise
+</programlisting>
+
+</sect2>
+
+
+<sect2 id="bzread" xreflabel="BZ2_bzRead">
+<title><computeroutput>BZ2_bzRead</computeroutput></title>
+
+<programlisting>
+int BZ2_bzRead ( int *bzerror, BZFILE *b, void *buf, int len );
+</programlisting>
+
+<para>Reads up to <computeroutput>len</computeroutput>
+(uncompressed) bytes from the compressed file
+<computeroutput>b</computeroutput> into the buffer
+<computeroutput>buf</computeroutput>.  If the read was
+successful, <computeroutput>bzerror</computeroutput> is set to
+<computeroutput>BZ_OK</computeroutput> and the number of bytes
+read is returned.  If the logical end-of-stream was detected,
+<computeroutput>bzerror</computeroutput> will be set to
+<computeroutput>BZ_STREAM_END</computeroutput>, and the number of
+bytes read is returned.  All other
+<computeroutput>bzerror</computeroutput> values denote an
+error.</para>
+
+<para><computeroutput>BZ2_bzRead</computeroutput> will supply
+<computeroutput>len</computeroutput> bytes, unless the logical
+stream end is detected or an error occurs.  Because of this, it
+is possible to detect the stream end by observing when the number
+of bytes returned is less than the number requested.
+Nevertheless, this is regarded as inadvisable; you should instead
+check <computeroutput>bzerror</computeroutput> after every call
+and watch out for
+<computeroutput>BZ_STREAM_END</computeroutput>.</para>
+
+<para>Internally, <computeroutput>BZ2_bzRead</computeroutput>
+copies data from the compressed file in chunks of size
+<computeroutput>BZ_MAX_UNUSED</computeroutput> bytes before
+decompressing it.  If the file contains more bytes than strictly
+needed to reach the logical end-of-stream,
+<computeroutput>BZ2_bzRead</computeroutput> will almost certainly
+read some of the trailing data before signalling
+<computeroutput>BZ_SEQUENCE_END</computeroutput>.  To collect the
+read but unused data once
+<computeroutput>BZ_SEQUENCE_END</computeroutput> has appeared,
+call <computeroutput>BZ2_bzReadGetUnused</computeroutput>
+immediately before
+<computeroutput>BZ2_bzReadClose</computeroutput>.</para>
+
+<para>Possible assignments to
+<computeroutput>bzerror</computeroutput>:</para>
+
+<programlisting>
+BZ_PARAM_ERROR
+  if b is NULL or buf is NULL or len < 0
+BZ_SEQUENCE_ERROR
+  if b was opened with BZ2_bzWriteOpen
+BZ_IO_ERROR
+  if there is an error reading from the compressed file
+BZ_UNEXPECTED_EOF
+  if the compressed file ended before 
+  the logical end-of-stream was detected
+BZ_DATA_ERROR
+  if a data integrity error was detected in the compressed stream
+BZ_DATA_ERROR_MAGIC
+  if the stream does not begin with the requisite header bytes 
+  (ie, is not a bzip2 data file).  This is really 
+  a special case of BZ_DATA_ERROR.
+BZ_MEM_ERROR
+  if insufficient memory was available
+BZ_STREAM_END
+  if the logical end of stream was detected.
+BZ_OK
+  otherwise.
+</programlisting>
+
+<para>Possible return values:</para>
+
+<programlisting>
+number of bytes read
+  if bzerror is BZ_OK or BZ_STREAM_END
+undefined
+  otherwise
+</programlisting>
+
+<para>Allowable next actions:</para>
+
+<programlisting>
+collect data from buf, then BZ2_bzRead or BZ2_bzReadClose
+  if bzerror is BZ_OK
+collect data from buf, then BZ2_bzReadClose or BZ2_bzReadGetUnused
+  if bzerror is BZ_SEQUENCE_END
+BZ2_bzReadClose
+  otherwise
+</programlisting>
+
+</sect2>
+
+
+<sect2 id="bzreadgetunused" xreflabel="BZ2_bzReadGetUnused">
+<title><computeroutput>BZ2_bzReadGetUnused</computeroutput></title>
+
+<programlisting>
+void BZ2_bzReadGetUnused( int* bzerror, BZFILE *b, 
+                          void** unused, int* nUnused );
+</programlisting>
+
+<para>Returns data which was read from the compressed file but
+was not needed to get to the logical end-of-stream.
+<computeroutput>*unused</computeroutput> is set to the address of
+the data, and <computeroutput>*nUnused</computeroutput> to the
+number of bytes.  <computeroutput>*nUnused</computeroutput> will
+be set to a value between <computeroutput>0</computeroutput> and
+<computeroutput>BZ_MAX_UNUSED</computeroutput> inclusive.</para>
+
+<para>This function may only be called once
+<computeroutput>BZ2_bzRead</computeroutput> has signalled
+<computeroutput>BZ_STREAM_END</computeroutput> but before
+<computeroutput>BZ2_bzReadClose</computeroutput>.</para>
+
+<para>Possible assignments to
+<computeroutput>bzerror</computeroutput>:</para>
+
+<programlisting>
+BZ_PARAM_ERROR
+  if b is NULL
+  or unused is NULL or nUnused is NULL
+BZ_SEQUENCE_ERROR
+  if BZ_STREAM_END has not been signalled
+  or if b was opened with BZ2_bzWriteOpen
+BZ_OK
+  otherwise
+</programlisting>
+
+<para>Allowable next actions:</para>
+
+<programlisting>
+BZ2_bzReadClose
+</programlisting>
+
+</sect2>
+
+
+<sect2 id="bzreadclose" xreflabel="BZ2_bzReadClose">
+<title><computeroutput>BZ2_bzReadClose</computeroutput></title>
+
+<programlisting>
+void BZ2_bzReadClose ( int *bzerror, BZFILE *b );
+</programlisting>
+
+<para>Releases all memory pertaining to the compressed file
+<computeroutput>b</computeroutput>.
+<computeroutput>BZ2_bzReadClose</computeroutput> does not call
+<computeroutput>fclose</computeroutput> on the underlying file
+handle, so you should do that yourself if appropriate.
+<computeroutput>BZ2_bzReadClose</computeroutput> should be called
+to clean up after all error situations.</para>
+
+<para>Possible assignments to
+<computeroutput>bzerror</computeroutput>:</para>
+
+<programlisting>
+BZ_SEQUENCE_ERROR
+  if b was opened with BZ2_bzOpenWrite
+BZ_OK
+  otherwise
+</programlisting>
+
+<para>Allowable next actions:</para>
+
+<programlisting>
+none
+</programlisting>
+
+</sect2>
+
+
+<sect2 id="bzwriteopen" xreflabel="BZ2_bzWriteOpen">
+<title><computeroutput>BZ2_bzWriteOpen</computeroutput></title>
+
+<programlisting>
+BZFILE *BZ2_bzWriteOpen( int *bzerror, FILE *f, 
+                         int blockSize100k, int verbosity,
+                         int workFactor );
+</programlisting>
+
+<para>Prepare to write compressed data to file handle
+<computeroutput>f</computeroutput>.
+<computeroutput>f</computeroutput> should refer to a file which
+has been opened for writing, and for which the error indicator
+(<computeroutput>ferror(f)</computeroutput>)is not set.</para>
+
+<para>For the meaning of parameters
+<computeroutput>blockSize100k</computeroutput>,
+<computeroutput>verbosity</computeroutput> and
+<computeroutput>workFactor</computeroutput>, see
+<computeroutput>BZ2_bzCompressInit</computeroutput>.</para>
+
+<para>All required memory is allocated at this stage, so if the
+call completes successfully,
+<computeroutput>BZ_MEM_ERROR</computeroutput> cannot be signalled
+by a subsequent call to
+<computeroutput>BZ2_bzWrite</computeroutput>.</para>
+
+<para>Possible assignments to
+<computeroutput>bzerror</computeroutput>:</para>
+
+<programlisting>
+BZ_CONFIG_ERROR
+  if the library has been mis-compiled
+BZ_PARAM_ERROR
+  if f is NULL
+  or blockSize100k < 1 or blockSize100k > 9
+BZ_IO_ERROR
+  if ferror(f) is nonzero
+BZ_MEM_ERROR
+  if insufficient memory is available
+BZ_OK
+  otherwise
+</programlisting>
+
+<para>Possible return values:</para>
+
+<programlisting>
+Pointer to an abstract BZFILE
+  if bzerror is BZ_OK
+NULL
+  otherwise
+</programlisting>
+
+<para>Allowable next actions:</para>
+
+<programlisting>
+BZ2_bzWrite
+  if bzerror is BZ_OK
+  (you could go directly to BZ2_bzWriteClose, but this would be pretty pointless)
+BZ2_bzWriteClose
+  otherwise
+</programlisting>
+
+</sect2>
+
+
+<sect2 id="bzwrite" xreflabel="BZ2_bzWrite">
+<title><computeroutput>BZ2_bzWrite</computeroutput></title>
+
+<programlisting>
+void BZ2_bzWrite ( int *bzerror, BZFILE *b, void *buf, int len );
+</programlisting>
+
+<para>Absorbs <computeroutput>len</computeroutput> bytes from the
+buffer <computeroutput>buf</computeroutput>, eventually to be
+compressed and written to the file.</para>
+
+<para>Possible assignments to
+<computeroutput>bzerror</computeroutput>:</para>
+
+<programlisting>
+BZ_PARAM_ERROR
+  if b is NULL or buf is NULL or len < 0
+BZ_SEQUENCE_ERROR
+  if b was opened with BZ2_bzReadOpen
+BZ_IO_ERROR
+  if there is an error writing the compressed file.
+BZ_OK
+  otherwise
+</programlisting>
+
+</sect2>
+
+
+<sect2 id="bzwriteclose" xreflabel="BZ2_bzWriteClose">
+<title><computeroutput>BZ2_bzWriteClose</computeroutput></title>
+
+<programlisting>
+void BZ2_bzWriteClose( int *bzerror, BZFILE* f,
+                       int abandon,
+                       unsigned int* nbytes_in,
+                       unsigned int* nbytes_out );
+
+void BZ2_bzWriteClose64( int *bzerror, BZFILE* f,
+                         int abandon,
+                         unsigned int* nbytes_in_lo32,
+                         unsigned int* nbytes_in_hi32,
+                         unsigned int* nbytes_out_lo32,
+                         unsigned int* nbytes_out_hi32 );
+</programlisting>
+
+<para>Compresses and flushes to the compressed file all data so
+far supplied by <computeroutput>BZ2_bzWrite</computeroutput>.
+The logical end-of-stream markers are also written, so subsequent
+calls to <computeroutput>BZ2_bzWrite</computeroutput> are
+illegal.  All memory associated with the compressed file
+<computeroutput>b</computeroutput> is released.
+<computeroutput>fflush</computeroutput> is called on the
+compressed file, but it is not
+<computeroutput>fclose</computeroutput>'d.</para>
+
+<para>If <computeroutput>BZ2_bzWriteClose</computeroutput> is
+called to clean up after an error, the only action is to release
+the memory.  The library records the error codes issued by
+previous calls, so this situation will be detected automatically.
+There is no attempt to complete the compression operation, nor to
+<computeroutput>fflush</computeroutput> the compressed file.  You
+can force this behaviour to happen even in the case of no error,
+by passing a nonzero value to
+<computeroutput>abandon</computeroutput>.</para>
+
+<para>If <computeroutput>nbytes_in</computeroutput> is non-null,
+<computeroutput>*nbytes_in</computeroutput> will be set to be the
+total volume of uncompressed data handled.  Similarly,
+<computeroutput>nbytes_out</computeroutput> will be set to the
+total volume of compressed data written.  For compatibility with
+older versions of the library,
+<computeroutput>BZ2_bzWriteClose</computeroutput> only yields the
+lower 32 bits of these counts.  Use
+<computeroutput>BZ2_bzWriteClose64</computeroutput> if you want
+the full 64 bit counts.  These two functions are otherwise
+absolutely identical.</para>
+
+<para>Possible assignments to
+<computeroutput>bzerror</computeroutput>:</para>
+
+<programlisting>
+BZ_SEQUENCE_ERROR
+  if b was opened with BZ2_bzReadOpen
+BZ_IO_ERROR
+  if there is an error writing the compressed file
+BZ_OK
+  otherwise
+</programlisting>
+
+</sect2>
+
+
+<sect2 id="embed" xreflabel="Handling embedded compressed data streams">
+<title>Handling embedded compressed data streams</title>
+
+<para>The high-level library facilitates use of
+<computeroutput>bzip2</computeroutput> data streams which form
+some part of a surrounding, larger data stream.</para>
+
+<itemizedlist mark='bullet'>
+
+ <listitem><para>For writing, the library takes an open file handle,
+  writes compressed data to it,
+  <computeroutput>fflush</computeroutput>es it but does not
+  <computeroutput>fclose</computeroutput> it.  The calling
+  application can write its own data before and after the
+  compressed data stream, using that same file handle.</para></listitem>
+
+ <listitem><para>Reading is more complex, and the facilities are not as
+  general as they could be since generality is hard to reconcile
+  with efficiency.  <computeroutput>BZ2_bzRead</computeroutput>
+  reads from the compressed file in blocks of size
+  <computeroutput>BZ_MAX_UNUSED</computeroutput> bytes, and in
+  doing so probably will overshoot the logical end of compressed
+  stream.  To recover this data once decompression has ended,
+  call <computeroutput>BZ2_bzReadGetUnused</computeroutput> after
+  the last call of <computeroutput>BZ2_bzRead</computeroutput>
+  (the one returning
+  <computeroutput>BZ_STREAM_END</computeroutput>) but before
+  calling
+  <computeroutput>BZ2_bzReadClose</computeroutput>.</para></listitem>
+
+</itemizedlist>
+
+<para>This mechanism makes it easy to decompress multiple
+<computeroutput>bzip2</computeroutput> streams placed end-to-end.
+As the end of one stream, when
+<computeroutput>BZ2_bzRead</computeroutput> returns
+<computeroutput>BZ_STREAM_END</computeroutput>, call
+<computeroutput>BZ2_bzReadGetUnused</computeroutput> to collect
+the unused data (copy it into your own buffer somewhere).  That
+data forms the start of the next compressed stream.  To start
+uncompressing that next stream, call
+<computeroutput>BZ2_bzReadOpen</computeroutput> again, feeding in
+the unused data via the <computeroutput>unused</computeroutput> /
+<computeroutput>nUnused</computeroutput> parameters.  Keep doing
+this until <computeroutput>BZ_STREAM_END</computeroutput> return
+coincides with the physical end of file
+(<computeroutput>feof(f)</computeroutput>).  In this situation
+<computeroutput>BZ2_bzReadGetUnused</computeroutput> will of
+course return no data.</para>
+
+<para>This should give some feel for how the high-level interface
+can be used.  If you require extra flexibility, you'll have to
+bite the bullet and get to grips with the low-level
+interface.</para>
+
+</sect2>
+
+
+<sect2 id="std-rdwr" xreflabel="Standard file-reading/writing code">
+<title>Standard file-reading/writing code</title>
+
+<para>Here's how you'd write data to a compressed file:</para>
+
+<programlisting>
+FILE*   f;
+BZFILE* b;
+int     nBuf;
+char    buf[ /* whatever size you like */ ];
+int     bzerror;
+int     nWritten;
+
+f = fopen ( "myfile.bz2", "w" );
+if ( !f ) {
+ /* handle error */
+}
+b = BZ2_bzWriteOpen( &bzerror, f, 9 );
+if (bzerror != BZ_OK) {
+ BZ2_bzWriteClose ( b );
+ /* handle error */
+}
+
+while ( /* condition */ ) {
+ /* get data to write into buf, and set nBuf appropriately */
+ nWritten = BZ2_bzWrite ( &bzerror, b, buf, nBuf );
+ if (bzerror == BZ_IO_ERROR) { 
+   BZ2_bzWriteClose ( &bzerror, b );
+   /* handle error */
+ }
+}
+
+BZ2_bzWriteClose( &bzerror, b );
+if (bzerror == BZ_IO_ERROR) {
+ /* handle error */
+}
+</programlisting>
+
+<para>And to read from a compressed file:</para>
+
+<programlisting>
+FILE*   f;
+BZFILE* b;
+int     nBuf;
+char    buf[ /* whatever size you like */ ];
+int     bzerror;
+int     nWritten;
+
+f = fopen ( "myfile.bz2", "r" );
+if ( !f ) {
+  /* handle error */
+}
+b = BZ2_bzReadOpen ( &bzerror, f, 0, NULL, 0 );
+if ( bzerror != BZ_OK ) {
+  BZ2_bzReadClose ( &bzerror, b );
+  /* handle error */
+}
+
+bzerror = BZ_OK;
+while ( bzerror == BZ_OK && /* arbitrary other conditions */) {
+  nBuf = BZ2_bzRead ( &bzerror, b, buf, /* size of buf */ );
+  if ( bzerror == BZ_OK ) {
+    /* do something with buf[0 .. nBuf-1] */
+  }
+}
+if ( bzerror != BZ_STREAM_END ) {
+   BZ2_bzReadClose ( &bzerror, b );
+   /* handle error */
+} else {
+   BZ2_bzReadClose ( &bzerror );
+}
+</programlisting>
+
+</sect2>
+
+</sect1>
+
+
+<sect1 id="util-fns" xreflabel="Utility functions">
+<title>Utility functions</title>
+
+
+<sect2 id="bzbufftobuffcompress" xreflabel="BZ2_bzBuffToBuffCompress">
+<title><computeroutput>BZ2_bzBuffToBuffCompress</computeroutput></title>
+
+<programlisting>
+int BZ2_bzBuffToBuffCompress( char*         dest,
+                              unsigned int* destLen,
+                              char*         source,
+                              unsigned int  sourceLen,
+                              int           blockSize100k,
+                              int           verbosity,
+                              int           workFactor );
+</programlisting>
+
+<para>Attempts to compress the data in <computeroutput>source[0
+.. sourceLen-1]</computeroutput> into the destination buffer,
+<computeroutput>dest[0 .. *destLen-1]</computeroutput>.  If the
+destination buffer is big enough,
+<computeroutput>*destLen</computeroutput> is set to the size of
+the compressed data, and <computeroutput>BZ_OK</computeroutput>
+is returned.  If the compressed data won't fit,
+<computeroutput>*destLen</computeroutput> is unchanged, and
+<computeroutput>BZ_OUTBUFF_FULL</computeroutput> is
+returned.</para>
+
+<para>Compression in this manner is a one-shot event, done with a
+single call to this function.  The resulting compressed data is a
+complete <computeroutput>bzip2</computeroutput> format data
+stream.  There is no mechanism for making additional calls to
+provide extra input data.  If you want that kind of mechanism,
+use the low-level interface.</para>
+
+<para>For the meaning of parameters
+<computeroutput>blockSize100k</computeroutput>,
+<computeroutput>verbosity</computeroutput> and
+<computeroutput>workFactor</computeroutput>, see
+<computeroutput>BZ2_bzCompressInit</computeroutput>.</para>
+
+<para>To guarantee that the compressed data will fit in its
+buffer, allocate an output buffer of size 1% larger than the
+uncompressed data, plus six hundred extra bytes.</para>
+
+<para><computeroutput>BZ2_bzBuffToBuffDecompress</computeroutput>
+will not write data at or beyond
+<computeroutput>dest[*destLen]</computeroutput>, even in case of
+buffer overflow.</para>
+
+<para>Possible return values:</para>
+
+<programlisting>
+BZ_CONFIG_ERROR
+  if the library has been mis-compiled
+BZ_PARAM_ERROR
+  if dest is NULL or destLen is NULL
+  or blockSize100k < 1 or blockSize100k > 9
+  or verbosity < 0 or verbosity > 4
+  or workFactor < 0 or workFactor > 250
+BZ_MEM_ERROR
+  if insufficient memory is available 
+BZ_OUTBUFF_FULL
+  if the size of the compressed data exceeds *destLen
+BZ_OK
+  otherwise
+</programlisting>
+
+</sect2>
+
+
+<sect2 id="bzbufftobuffdecompress" xreflabel="BZ2_bzBuffToBuffDecompress">
+<title><computeroutput>BZ2_bzBuffToBuffDecompress</computeroutput></title>
+
+<programlisting>
+int BZ2_bzBuffToBuffDecompress( char*         dest,
+                                unsigned int* destLen,
+                                char*         source,
+                                unsigned int  sourceLen,
+                                int           small,
+                                int           verbosity );
+</programlisting>
+
+<para>Attempts to decompress the data in <computeroutput>source[0
+.. sourceLen-1]</computeroutput> into the destination buffer,
+<computeroutput>dest[0 .. *destLen-1]</computeroutput>.  If the
+destination buffer is big enough,
+<computeroutput>*destLen</computeroutput> is set to the size of
+the uncompressed data, and <computeroutput>BZ_OK</computeroutput>
+is returned.  If the compressed data won't fit,
+<computeroutput>*destLen</computeroutput> is unchanged, and
+<computeroutput>BZ_OUTBUFF_FULL</computeroutput> is
+returned.</para>
+
+<para><computeroutput>source</computeroutput> is assumed to hold
+a complete <computeroutput>bzip2</computeroutput> format data
+stream.
+<computeroutput>BZ2_bzBuffToBuffDecompress</computeroutput> tries
+to decompress the entirety of the stream into the output
+buffer.</para>
+
+<para>For the meaning of parameters
+<computeroutput>small</computeroutput> and
+<computeroutput>verbosity</computeroutput>, see
+<computeroutput>BZ2_bzDecompressInit</computeroutput>.</para>
+
+<para>Because the compression ratio of the compressed data cannot
+be known in advance, there is no easy way to guarantee that the
+output buffer will be big enough.  You may of course make
+arrangements in your code to record the size of the uncompressed
+data, but such a mechanism is beyond the scope of this
+library.</para>
+
+<para><computeroutput>BZ2_bzBuffToBuffDecompress</computeroutput>
+will not write data at or beyond
+<computeroutput>dest[*destLen]</computeroutput>, even in case of
+buffer overflow.</para>
+
+<para>Possible return values:</para>
+
+<programlisting>
+BZ_CONFIG_ERROR
+  if the library has been mis-compiled
+BZ_PARAM_ERROR
+  if dest is NULL or destLen is NULL
+  or small != 0 && small != 1
+  or verbosity < 0 or verbosity > 4
+BZ_MEM_ERROR
+  if insufficient memory is available 
+BZ_OUTBUFF_FULL
+  if the size of the compressed data exceeds *destLen
+BZ_DATA_ERROR
+  if a data integrity error was detected in the compressed data
+BZ_DATA_ERROR_MAGIC
+  if the compressed data doesn't begin with the right magic bytes
+BZ_UNEXPECTED_EOF
+  if the compressed data ends unexpectedly
+BZ_OK
+  otherwise
+</programlisting>
+
+</sect2>
+
+</sect1>
+
+
+<sect1 id="zlib-compat" xreflabel="zlib compatibility functions">
+<title><computeroutput>zlib</computeroutput> compatibility functions</title>
+
+<para>Yoshioka Tsuneo has contributed some functions to give
+better <computeroutput>zlib</computeroutput> compatibility.
+These functions are <computeroutput>BZ2_bzopen</computeroutput>,
+<computeroutput>BZ2_bzread</computeroutput>,
+<computeroutput>BZ2_bzwrite</computeroutput>,
+<computeroutput>BZ2_bzflush</computeroutput>,
+<computeroutput>BZ2_bzclose</computeroutput>,
+<computeroutput>BZ2_bzerror</computeroutput> and
+<computeroutput>BZ2_bzlibVersion</computeroutput>.  These
+functions are not (yet) officially part of the library.  If they
+break, you get to keep all the pieces.  Nevertheless, I think
+they work ok.</para>
+
+<programlisting>
+typedef void BZFILE;
+
+const char * BZ2_bzlibVersion ( void );
+</programlisting>
+
+<para>Returns a string indicating the library version.</para>
+
+<programlisting>
+BZFILE * BZ2_bzopen  ( const char *path, const char *mode );
+BZFILE * BZ2_bzdopen ( int        fd,    const char *mode );
+</programlisting>
+
+<para>Opens a <computeroutput>.bz2</computeroutput> file for
+reading or writing, using either its name or a pre-existing file
+descriptor.  Analogous to <computeroutput>fopen</computeroutput>
+and <computeroutput>fdopen</computeroutput>.</para>
+
+<programlisting>
+int BZ2_bzread  ( BZFILE* b, void* buf, int len );
+int BZ2_bzwrite ( BZFILE* b, void* buf, int len );
+</programlisting>
+
+<para>Reads/writes data from/to a previously opened
+<computeroutput>BZFILE</computeroutput>.  Analogous to
+<computeroutput>fread</computeroutput> and
+<computeroutput>fwrite</computeroutput>.</para>
+
+<programlisting>
+int  BZ2_bzflush ( BZFILE* b );
+void BZ2_bzclose ( BZFILE* b );
+</programlisting>
+
+<para>Flushes/closes a <computeroutput>BZFILE</computeroutput>.
+<computeroutput>BZ2_bzflush</computeroutput> doesn't actually do
+anything.  Analogous to <computeroutput>fflush</computeroutput>
+and <computeroutput>fclose</computeroutput>.</para>
+
+<programlisting>
+const char * BZ2_bzerror ( BZFILE *b, int *errnum )
+</programlisting>
+
+<para>Returns a string describing the more recent error status of
+<computeroutput>b</computeroutput>, and also sets
+<computeroutput>*errnum</computeroutput> to its numerical
+value.</para>
+
+</sect1>
+
+
+<sect1 id="stdio-free" 
+       xreflabel="Using the library in a stdio-free environment">
+<title>Using the library in a <computeroutput>stdio</computeroutput>-free environment</title>
+
+
+<sect2 id="stdio-bye" xreflabel="Getting rid of stdio">
+<title>Getting rid of <computeroutput>stdio</computeroutput></title>
+
+<para>In a deeply embedded application, you might want to use
+just the memory-to-memory functions.  You can do this
+conveniently by compiling the library with preprocessor symbol
+<computeroutput>BZ_NO_STDIO</computeroutput> defined.  Doing this
+gives you a library containing only the following eight
+functions:</para>
+
+<para><computeroutput>BZ2_bzCompressInit</computeroutput>,
+<computeroutput>BZ2_bzCompress</computeroutput>,
+<computeroutput>BZ2_bzCompressEnd</computeroutput>
+<computeroutput>BZ2_bzDecompressInit</computeroutput>,
+<computeroutput>BZ2_bzDecompress</computeroutput>,
+<computeroutput>BZ2_bzDecompressEnd</computeroutput>
+<computeroutput>BZ2_bzBuffToBuffCompress</computeroutput>,
+<computeroutput>BZ2_bzBuffToBuffDecompress</computeroutput></para>
+
+<para>When compiled like this, all functions will ignore
+<computeroutput>verbosity</computeroutput> settings.</para>
+
+</sect2>
+
+
+<sect2 id="critical-error" xreflabel="Critical error handling">
+<title>Critical error handling</title>
+
+<para><computeroutput>libbzip2</computeroutput> contains a number
+of internal assertion checks which should, needless to say, never
+be activated.  Nevertheless, if an assertion should fail,
+behaviour depends on whether or not the library was compiled with
+<computeroutput>BZ_NO_STDIO</computeroutput> set.</para>
+
+<para>For a normal compile, an assertion failure yields the
+message:</para>
+
+<blockquote>
+<para>bzip2/libbzip2: internal error number N.</para>
+<para>This is a bug in bzip2/libbzip2, &bz-version; of &bz-date;.
+Please report it to me at: &bz-email;.  If this happened
+when you were using some program which uses libbzip2 as a
+component, you should also report this bug to the author(s)
+of that program.  Please make an effort to report this bug;
+timely and accurate bug reports eventually lead to higher
+quality software.  Thanks.  Julian Seward, &bz-date;.
+</para></blockquote>
+
+<para>where <computeroutput>N</computeroutput> is some error code
+number.  If <computeroutput>N == 1007</computeroutput>, it also
+prints some extra text advising the reader that unreliable memory
+is often associated with internal error 1007. (This is a
+frequently-observed-phenomenon with versions 1.0.0/1.0.1).</para>
+
+<para><computeroutput>exit(3)</computeroutput> is then
+called.</para>
+
+<para>For a <computeroutput>stdio</computeroutput>-free library,
+assertion failures result in a call to a function declared
+as:</para>
+
+<programlisting>
+extern void bz_internal_error ( int errcode );
+</programlisting>
+
+<para>The relevant code is passed as a parameter.  You should
+supply such a function.</para>
+
+<para>In either case, once an assertion failure has occurred, any
+<computeroutput>bz_stream</computeroutput> records involved can
+be regarded as invalid.  You should not attempt to resume normal
+operation with them.</para>
+
+<para>You may, of course, change critical error handling to suit
+your needs.  As I said above, critical errors indicate bugs in
+the library and should not occur.  All "normal" error situations
+are indicated via error return codes from functions, and can be
+recovered from.</para>
+
+</sect2>
+
+</sect1>
+
+
+<sect1 id="win-dll" xreflabel="Making a Windows DLL">
+<title>Making a Windows DLL</title>
+
+<para>Everything related to Windows has been contributed by
+Yoshioka Tsuneo
+(<computeroutput>QWF00133@niftyserve.or.jp</computeroutput> /
+<computeroutput>tsuneo-y@is.aist-nara.ac.jp</computeroutput>), so
+you should send your queries to him (but perhaps Cc: me,
+<computeroutput>&bz-email;</computeroutput>).</para>
+
+<para>My vague understanding of what to do is: using Visual C++
+5.0, open the project file
+<computeroutput>libbz2.dsp</computeroutput>, and build.  That's
+all.</para>
+
+<para>If you can't open the project file for some reason, make a
+new one, naming these files:
+<computeroutput>blocksort.c</computeroutput>,
+<computeroutput>bzlib.c</computeroutput>,
+<computeroutput>compress.c</computeroutput>,
+<computeroutput>crctable.c</computeroutput>,
+<computeroutput>decompress.c</computeroutput>,
+<computeroutput>huffman.c</computeroutput>,
+<computeroutput>randtable.c</computeroutput> and
+<computeroutput>libbz2.def</computeroutput>.  You will also need
+to name the header files <computeroutput>bzlib.h</computeroutput>
+and <computeroutput>bzlib_private.h</computeroutput>.</para>
+
+<para>If you don't use VC++, you may need to define the
+proprocessor symbol
+<computeroutput>_WIN32</computeroutput>.</para>
+
+<para>Finally, <computeroutput>dlltest.c</computeroutput> is a
+sample program using the DLL.  It has a project file,
+<computeroutput>dlltest.dsp</computeroutput>.</para>
+
+<para>If you just want a makefile for Visual C, have a look at
+<computeroutput>makefile.msc</computeroutput>.</para>
+
+<para>Be aware that if you compile
+<computeroutput>bzip2</computeroutput> itself on Win32, you must
+set <computeroutput>BZ_UNIX</computeroutput> to 0 and
+<computeroutput>BZ_LCCWIN32</computeroutput> to 1, in the file
+<computeroutput>bzip2.c</computeroutput>, before compiling.
+Otherwise the resulting binary won't work correctly.</para>
+
+<para>I haven't tried any of this stuff myself, but it all looks
+plausible.</para>
+
+</sect1>
+
+</chapter>
+
+
+
+<chapter id="misc" xreflabel="Miscellanea">
+<title>Miscellanea</title>
+
+<para>These are just some random thoughts of mine.  Your mileage
+may vary.</para>
+
+
+<sect1 id="limits" xreflabel="Limitations of the compressed file format">
+<title>Limitations of the compressed file format</title>
+
+<para><computeroutput>bzip2-1.0.X</computeroutput>,
+<computeroutput>0.9.5</computeroutput> and
+<computeroutput>0.9.0</computeroutput> use exactly the same file
+format as the original version,
+<computeroutput>bzip2-0.1</computeroutput>.  This decision was
+made in the interests of stability.  Creating yet another
+incompatible compressed file format would create further
+confusion and disruption for users.</para>
+
+<para>Nevertheless, this is not a painless decision.  Development
+work since the release of
+<computeroutput>bzip2-0.1</computeroutput> in August 1997 has
+shown complexities in the file format which slow down
+decompression and, in retrospect, are unnecessary.  These
+are:</para>
+
+<itemizedlist mark='bullet'>
+
+ <listitem><para>The run-length encoder, which is the first of the
+   compression transformations, is entirely irrelevant.  The
+   original purpose was to protect the sorting algorithm from the
+   very worst case input: a string of repeated symbols.  But
+   algorithm steps Q6a and Q6b in the original Burrows-Wheeler
+   technical report (SRC-124) show how repeats can be handled
+   without difficulty in block sorting.</para></listitem>
+
+ <listitem><para>The randomisation mechanism doesn't really need to be
+   there.  Udi Manber and Gene Myers published a suffix array
+   construction algorithm a few years back, which can be employed
+   to sort any block, no matter how repetitive, in O(N log N)
+   time.  Subsequent work by Kunihiko Sadakane has produced a
+   derivative O(N (log N)^2) algorithm which usually outperforms
+   the Manber-Myers algorithm.</para>
+
+   <para>I could have changed to Sadakane's algorithm, but I find
+   it to be slower than <computeroutput>bzip2</computeroutput>'s
+   existing algorithm for most inputs, and the randomisation
+   mechanism protects adequately against bad cases.  I didn't
+   think it was a good tradeoff to make.  Partly this is due to
+   the fact that I was not flooded with email complaints about
+   <computeroutput>bzip2-0.1</computeroutput>'s performance on
+   repetitive data, so perhaps it isn't a problem for real
+   inputs.</para>
+
+   <para>Probably the best long-term solution, and the one I have
+   incorporated into 0.9.5 and above, is to use the existing
+   sorting algorithm initially, and fall back to a O(N (log N)^2)
+   algorithm if the standard algorithm gets into
+   difficulties.</para></listitem>
+
+  <listitem><para>The compressed file format was never designed to be
+   handled by a library, and I have had to jump though some hoops
+   to produce an efficient implementation of decompression.  It's
+   a bit hairy.  Try passing
+   <computeroutput>decompress.c</computeroutput> through the C
+   preprocessor and you'll see what I mean.  Much of this
+   complexity could have been avoided if the compressed size of
+   each block of data was recorded in the data stream.</para></listitem>
+
+ <listitem><para>An Adler-32 checksum, rather than a CRC32 checksum,
+   would be faster to compute.</para></listitem>
+
+</itemizedlist>
+
+<para>It would be fair to say that the
+<computeroutput>bzip2</computeroutput> format was frozen before I
+properly and fully understood the performance consequences of
+doing so.</para>
+
+<para>Improvements which I was able to incorporate into 0.9.0,
+despite using the same file format, are:</para>
+
+<itemizedlist mark='bullet'>
+
+ <listitem><para>Single array implementation of the inverse BWT.  This
+  significantly speeds up decompression, presumably because it
+  reduces the number of cache misses.</para></listitem>
+
+ <listitem><para>Faster inverse MTF transform for large MTF values.
+  The new implementation is based on the notion of sliding blocks
+  of values.</para></listitem>
+
+ <listitem><para><computeroutput>bzip2-0.9.0</computeroutput> now reads
+  and writes files with <computeroutput>fread</computeroutput>
+  and <computeroutput>fwrite</computeroutput>; version 0.1 used
+  <computeroutput>putc</computeroutput> and
+  <computeroutput>getc</computeroutput>.  Duh!  Well, you live
+  and learn.</para></listitem>
+
+</itemizedlist>
+
+<para>Further ahead, it would be nice to be able to do random
+access into files.  This will require some careful design of
+compressed file formats.</para>
+
+</sect1>
+
+
+<sect1 id="port-issues" xreflabel="Portability issues">
+<title>Portability issues</title>
+
+<para>After some consideration, I have decided not to use GNU
+<computeroutput>autoconf</computeroutput> to configure 0.9.5 or
+1.0.</para>
+
+<para><computeroutput>autoconf</computeroutput>, admirable and
+wonderful though it is, mainly assists with portability problems
+between Unix-like platforms.  But
+<computeroutput>bzip2</computeroutput> doesn't have much in the
+way of portability problems on Unix; most of the difficulties
+appear when porting to the Mac, or to Microsoft's operating
+systems.  <computeroutput>autoconf</computeroutput> doesn't help
+in those cases, and brings in a whole load of new
+complexity.</para>
+
+<para>Most people should be able to compile the library and
+program under Unix straight out-of-the-box, so to speak,
+especially if you have a version of GNU C available.</para>
+
+<para>There are a couple of
+<computeroutput>__inline__</computeroutput> directives in the
+code.  GNU C (<computeroutput>gcc</computeroutput>) should be
+able to handle them.  If you're not using GNU C, your C compiler
+shouldn't see them at all.  If your compiler does, for some
+reason, see them and doesn't like them, just
+<computeroutput>#define</computeroutput>
+<computeroutput>__inline__</computeroutput> to be
+<computeroutput>/* */</computeroutput>.  One easy way to do this
+is to compile with the flag
+<computeroutput>-D__inline__=</computeroutput>, which should be
+understood by most Unix compilers.</para>
+
+<para>If you still have difficulties, try compiling with the
+macro <computeroutput>BZ_STRICT_ANSI</computeroutput> defined.
+This should enable you to build the library in a strictly ANSI
+compliant environment.  Building the program itself like this is
+dangerous and not supported, since you remove
+<computeroutput>bzip2</computeroutput>'s checks against
+compressing directories, symbolic links, devices, and other
+not-really-a-file entities.  This could cause filesystem
+corruption!</para>
+
+<para>One other thing: if you create a
+<computeroutput>bzip2</computeroutput> binary for public distribution,
+please consider linking it statically (<computeroutput>gcc
+-static</computeroutput>).  This avoids all sorts of library-version
+issues that others may encounter later on.</para>
+
+<para>If you build <computeroutput>bzip2</computeroutput> on
+Win32, you must set <computeroutput>BZ_UNIX</computeroutput> to 0
+and <computeroutput>BZ_LCCWIN32</computeroutput> to 1, in the
+file <computeroutput>bzip2.c</computeroutput>, before compiling.
+Otherwise the resulting binary won't work correctly.</para>
+
+</sect1>
+
+
+<sect1 id="bugs" xreflabel="Reporting bugs">
+<title>Reporting bugs</title>
+
+<para>I tried pretty hard to make sure
+<computeroutput>bzip2</computeroutput> is bug free, both by
+design and by testing.  Hopefully you'll never need to read this
+section for real.</para>
+
+<para>Nevertheless, if <computeroutput>bzip2</computeroutput> dies
+with a segmentation fault, a bus error or an internal assertion
+failure, it will ask you to email me a bug report.  Experience from
+years of feedback of bzip2 users indicates that almost all these
+problems can be traced to either compiler bugs or hardware
+problems.</para>
+
+<itemizedlist mark='bullet'>
+
+ <listitem><para>Recompile the program with no optimisation, and
+  see if it works.  And/or try a different compiler.  I heard all
+  sorts of stories about various flavours of GNU C (and other
+  compilers) generating bad code for
+  <computeroutput>bzip2</computeroutput>, and I've run across two
+  such examples myself.</para>
+
+  <para>2.7.X versions of GNU C are known to generate bad code
+  from time to time, at high optimisation levels.  If you get
+  problems, try using the flags
+  <computeroutput>-O2</computeroutput>
+  <computeroutput>-fomit-frame-pointer</computeroutput>
+  <computeroutput>-fno-strength-reduce</computeroutput>.  You
+  should specifically <emphasis>not</emphasis> use
+  <computeroutput>-funroll-loops</computeroutput>.</para>
+
+  <para>You may notice that the Makefile runs six tests as part
+  of the build process.  If the program passes all of these, it's
+  a pretty good (but not 100%) indication that the compiler has
+  done its job correctly.</para></listitem>
+
+ <listitem><para>If <computeroutput>bzip2</computeroutput>
+  crashes randomly, and the crashes are not repeatable, you may
+  have a flaky memory subsystem.
+  <computeroutput>bzip2</computeroutput> really hammers your
+  memory hierarchy, and if it's a bit marginal, you may get these
+  problems.  Ditto if your disk or I/O subsystem is slowly
+  failing.  Yup, this really does happen.</para>
+
+  <para>Try using a different machine of the same type, and see
+  if you can repeat the problem.</para></listitem>
+
+  <listitem><para>This isn't really a bug, but ... If
+  <computeroutput>bzip2</computeroutput> tells you your file is
+  corrupted on decompression, and you obtained the file via FTP,
+  there is a possibility that you forgot to tell FTP to do a
+  binary mode transfer.  That absolutely will cause the file to
+  be non-decompressible.  You'll have to transfer it
+  again.</para></listitem>
+
+</itemizedlist>
+
+<para>If you've incorporated
+<computeroutput>libbzip2</computeroutput> into your own program
+and are getting problems, please, please, please, check that the
+parameters you are passing in calls to the library, are correct,
+and in accordance with what the documentation says is allowable.
+I have tried to make the library robust against such problems,
+but I'm sure I haven't succeeded.</para>
+
+<para>Finally, if the above comments don't help, you'll have to
+send me a bug report.  Now, it's just amazing how many people
+will send me a bug report saying something like:</para>
+
+<programlisting>
+bzip2 crashed with segmentation fault on my machine
+</programlisting>
+
+<para>and absolutely nothing else.  Needless to say, a such a
+report is <emphasis>totally, utterly, completely and
+comprehensively 100% useless; a waste of your time, my time, and
+net bandwidth</emphasis>.  With no details at all, there's no way
+I can possibly begin to figure out what the problem is.</para>
+
+<para>The rules of the game are: facts, facts, facts.  Don't omit
+them because "oh, they won't be relevant".  At the bare
+minimum:</para>
+
+<programlisting>
+Machine type.  Operating system version.  
+Exact version of bzip2 (do bzip2 -V).  
+Exact version of the compiler used.  
+Flags passed to the compiler.
+</programlisting>
+
+<para>However, the most important single thing that will help me
+is the file that you were trying to compress or decompress at the
+time the problem happened.  Without that, my ability to do
+anything more than speculate about the cause, is limited.</para>
+
+</sect1>
+
+
+<sect1 id="package" xreflabel="Did you get the right package?">
+<title>Did you get the right package?</title>
+
+<para><computeroutput>bzip2</computeroutput> is a resource hog.
+It soaks up large amounts of CPU cycles and memory.  Also, it
+gives very large latencies.  In the worst case, you can feed many
+megabytes of uncompressed data into the library before getting
+any compressed output, so this probably rules out applications
+requiring interactive behaviour.</para>
+
+<para>These aren't faults of my implementation, I hope, but more
+an intrinsic property of the Burrows-Wheeler transform
+(unfortunately).  Maybe this isn't what you want.</para>
+
+<para>If you want a compressor and/or library which is faster,
+uses less memory but gets pretty good compression, and has
+minimal latency, consider Jean-loup Gailly's and Mark Adler's
+work, <computeroutput>zlib-1.2.1</computeroutput> and
+<computeroutput>gzip-1.2.4</computeroutput>.  Look for them at 
+<ulink url="http://www.zlib.org">http://www.zlib.org</ulink> and 
+<ulink url="http://www.gzip.org">http://www.gzip.org</ulink>
+respectively.</para>
+
+<para>For something faster and lighter still, you might try Markus F
+X J Oberhumer's <computeroutput>LZO</computeroutput> real-time
+compression/decompression library, at 
+<ulink url="http://www.oberhumer.com/opensource">http://www.oberhumer.com/opensource</ulink>.</para>
+
+</sect1>
+
+
+
+<sect1 id="reading" xreflabel="Further Reading">
+<title>Further Reading</title>
+
+<para><computeroutput>bzip2</computeroutput> is not research
+work, in the sense that it doesn't present any new ideas.
+Rather, it's an engineering exercise based on existing
+ideas.</para>
+
+<para>Four documents describe essentially all the ideas behind
+<computeroutput>bzip2</computeroutput>:</para>
+
+<literallayout>Michael Burrows and D. J. Wheeler:
+  "A block-sorting lossless data compression algorithm"
+   10th May 1994. 
+   Digital SRC Research Report 124.
+   ftp://ftp.digital.com/pub/DEC/SRC/research-reports/SRC-124.ps.gz
+   If you have trouble finding it, try searching at the
+   New Zealand Digital Library, http://www.nzdl.org.
+
+Daniel S. Hirschberg and Debra A. LeLewer
+  "Efficient Decoding of Prefix Codes"
+   Communications of the ACM, April 1990, Vol 33, Number 4.
+   You might be able to get an electronic copy of this
+   from the ACM Digital Library.
+
+David J. Wheeler
+   Program bred3.c and accompanying document bred3.ps.
+   This contains the idea behind the multi-table Huffman coding scheme.
+   ftp://ftp.cl.cam.ac.uk/users/djw3/
+
+Jon L. Bentley and Robert Sedgewick
+  "Fast Algorithms for Sorting and Searching Strings"
+   Available from Sedgewick's web page,
+   www.cs.princeton.edu/~rs
+</literallayout>
+
+<para>The following paper gives valuable additional insights into
+the algorithm, but is not immediately the basis of any code used
+in bzip2.</para>
+
+<literallayout>Peter Fenwick:
+   Block Sorting Text Compression
+   Proceedings of the 19th Australasian Computer Science Conference,
+     Melbourne, Australia.  Jan 31 - Feb 2, 1996.
+   ftp://ftp.cs.auckland.ac.nz/pub/peter-f/ACSC96paper.ps</literallayout>
+
+<para>Kunihiko Sadakane's sorting algorithm, mentioned above, is
+available from:</para>
+
+<literallayout>http://naomi.is.s.u-tokyo.ac.jp/~sada/papers/Sada98b.ps.gz
+</literallayout>
+
+<para>The Manber-Myers suffix array construction algorithm is
+described in a paper available from:</para>
+
+<literallayout>http://www.cs.arizona.edu/people/gene/PAPERS/suffix.ps
+</literallayout>
+
+<para>Finally, the following papers document some
+investigations I made into the performance of sorting
+and decompression algorithms:</para>
+
+<literallayout>Julian Seward
+   On the Performance of BWT Sorting Algorithms
+   Proceedings of the IEEE Data Compression Conference 2000
+     Snowbird, Utah.  28-30 March 2000.
+
+Julian Seward
+   Space-time Tradeoffs in the Inverse B-W Transform
+   Proceedings of the IEEE Data Compression Conference 2001
+     Snowbird, Utah.  27-29 March 2001.
+</literallayout>
+
+</sect1>
+
+</chapter>
+
+</book>