* apt-pkg/contrib/fileutl.cc:

[ntk/apt.git] / doc / apt-ftparchive.1.xml

<?xml version="1.0" encoding="utf-8" standalone="no"?>
<!DOCTYPE refentry PUBLIC "-//OASIS//DTD DocBook XML V4.2//EN"
  "http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd" [

<!ENTITY % aptent SYSTEM "apt.ent">
%aptent;

<!ENTITY % aptverbatiment SYSTEM "apt-verbatim.ent">
%aptverbatiment;

]>

<refentry>

 <refentryinfo>
   &apt-author.jgunthorpe;
   &apt-author.team;
   &apt-email;
   &apt-product;
   <!-- The last update date -->
   <date>2009-08-17T00:00:00Z</date>
 </refentryinfo>
 
 <refmeta>
   <refentrytitle>apt-ftparchive</refentrytitle>
   <manvolnum>1</manvolnum>
   <refmiscinfo class="manual">APT</refmiscinfo>
 </refmeta>
 
 <!-- Man page title -->
 <refnamediv>
    <refname>apt-ftparchive</refname>
    <refpurpose>Utility to generate index files</refpurpose>
 </refnamediv>

 &synopsis-command-apt-ftparchive;

 <refsect1><title>Description</title>
   <para><command>apt-ftparchive</command> is the command line tool that generates the index 
   files that APT uses to access a distribution source. The index files should 
   be generated on the origin site based on the content of that site.</para>

   <para><command>apt-ftparchive</command> is a superset of the &dpkg-scanpackages; program,
   incorporating its entire functionality via the <literal>packages</literal> command.
   It also contains a contents file generator, <literal>contents</literal>, and an 
   elaborate means to 'script' the generation process for a complete 
   archive.</para>

   <para>Internally <command>apt-ftparchive</command> can make use of binary databases to 
   cache the contents of a .deb file and it does not rely on any external 
   programs aside from &gzip;. When doing a full generate it automatically 
   performs file-change checks and builds the desired compressed output files.</para>
   
   <para>Unless the <option>-h</option>, or <option>--help</option> option is given one of the
   commands below must be present.</para>
   
   <variablelist>
     <varlistentry><term><option>packages</option></term>
     <listitem><para>
     The packages command generates a package file from a directory tree. It
     takes the given directory and recursively searches it for .deb files, 
     emitting a package record to stdout for each. This command is 
     approximately equivalent to &dpkg-scanpackages;.</para>

     <para>The option <option>--db</option> can be used to specify a binary caching DB.</para></listitem>
     </varlistentry>
     
     <varlistentry><term><option>sources</option></term>
     <listitem><para>
     The <literal>sources</literal> command generates a source index file from a directory tree. 
     It takes the given directory and recursively searches it for .dsc files,
     emitting a source record to stdout for each. This command is approximately
     equivalent to &dpkg-scansources;.</para>
     <para>
     If an override file is specified then a source override file will be 
     looked for with an extension of .src. The --source-override option can be 
     used to change the source override file that will be used.</para></listitem>
     </varlistentry>

     <varlistentry><term><option>contents</option></term>
     <listitem><para>
     The <literal>contents</literal> command generates a contents file from a directory tree. It
     takes the given directory and recursively searches it for .deb files, 
     and reads the file list from each file. It then sorts and writes to stdout
     the list of files matched to packages. Directories are not written to 
     the output. If multiple packages own the same file then each package is
     separated by a comma in the output.</para>
     <para>
     The option <option>--db</option> can be used to specify a binary caching DB.</para></listitem>
     </varlistentry>

     <varlistentry><term><option>release</option></term>
     <listitem><para>
     The <literal>release</literal> command generates a Release file from a
     directory tree. It recursively searches the given directory for uncompressed
     <filename>Packages</filename> and <filename>Sources</filename> files and the ones
     compressed with <command>gzip</command>, <command>bzip2</command> or <command>lzma</command>
     as well as <filename>Release</filename> and <filename>md5sum.txt</filename> files by default
     (<literal>APT::FTPArchive::Release::Default-Patterns</literal>). Additional filename patterns
     can be added by listing them in <literal>APT::FTPArchive::Release::Patterns</literal>.
     It then writes to stdout a Release file containing a MD5, SHA1 and SHA256 digest
     for each file.</para>
     <para>
     Values for the additional metadata fields in the Release file are
     taken from the corresponding variables under
     <literal>APT::FTPArchive::Release</literal>,
     e.g. <literal>APT::FTPArchive::Release::Origin</literal>.  The supported fields
     are: <literal>Origin</literal>, <literal>Label</literal>, <literal>Suite</literal>,
     <literal>Version</literal>, <literal>Codename</literal>, <literal>Date</literal>,
     <literal>Valid-Until</literal>, <literal>Architectures</literal>,
     <literal>Components</literal>, <literal>Description</literal>.</para></listitem>

     </varlistentry>

     <varlistentry><term><option>generate</option></term>
     <listitem><para>
     The <literal>generate</literal> command is designed to be runnable from a cron script and
     builds indexes according to the given config file. The config language
     provides a flexible means of specifying which index files are built from 
     which directories, as well as providing a simple means of maintaining the
     required settings.</para></listitem>
     </varlistentry>

     <varlistentry><term><option>clean</option></term>
     <listitem><para>
     The <literal>clean</literal> command tidies the databases used by the given 
     configuration file by removing any records that are no longer necessary.</para></listitem>
     </varlistentry>     
   </variablelist>  
 </refsect1>

 <refsect1><title>The Generate Configuration</title>
   <para>
   The <literal>generate</literal> command uses a configuration file to describe the 
   archives that are going to be generated. It follows the typical ISC 
   configuration format as seen in ISC tools like bind 8 and dhcpd. 
   &apt-conf; contains a description of the syntax. Note that the generate 
   configuration is parsed in sectional manner, but &apt-conf; is parsed in a
   tree manner. This only effects how the scope tag is handled.</para>

   <para>
   The generate configuration has 4 separate sections, each described below.</para>

   <refsect2><title>Dir Section</title>
     <para>
     The <literal>Dir</literal> section defines the standard directories needed to 
     locate the files required during the generation process. These 
     directories are prepended certain relative paths defined in later 
     sections to produce a complete an absolute path.</para>
     <variablelist>     
      <varlistentry><term><option>ArchiveDir</option></term>
      <listitem><para>
      Specifies the root of the FTP archive, in a standard
      Debian configuration this is the directory that contains the 
      <filename>ls-LR</filename> and dist nodes.</para></listitem>
      </varlistentry>

      <varlistentry><term><option>OverrideDir</option></term>
      <listitem><para>
      Specifies the location of the override files.</para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>CacheDir</option></term>
      <listitem><para>
      Specifies the location of the cache files</para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>FileListDir</option></term>
      <listitem><para>
      Specifies the location of the file list files, 
      if the <literal>FileList</literal> setting is used below.</para></listitem>
      </varlistentry>
     </variablelist>
   </refsect2>
   
   <refsect2><title>Default Section</title>
     <para>
     The <literal>Default</literal> section specifies default values, and settings 
     that control the operation of the generator. Other sections may override 
     these defaults with a per-section setting.</para>
     <variablelist>     
      <varlistentry><term><option>Packages::Compress</option></term>
      <listitem><para>
      Sets the default compression schemes to use 
      for the Package index files. It is a string that contains a space 
      separated list of at least one of: '.' (no compression), 'gzip' and 
      'bzip2'. The default for all compression schemes is '. gzip'.</para></listitem>
      </varlistentry>

      <varlistentry><term><option>Packages::Extensions</option></term>
      <listitem><para>
      Sets the default list of file extensions that are package files.
      This defaults to '.deb'.</para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>Sources::Compress</option></term>
      <listitem><para>
      This is similar to <literal>Packages::Compress</literal> 
      except that it controls the compression for the Sources files.</para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>Sources::Extensions</option></term>
      <listitem><para>
      Sets the default list of file extensions that are source files.
      This defaults to '.dsc'.</para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>Contents::Compress</option></term>
      <listitem><para>
      This is similar to <literal>Packages::Compress</literal> 
      except that it controls the compression for the Contents files.</para></listitem>
      </varlistentry>

      <varlistentry><term><option>Translation::Compress</option></term>
      <listitem><para>
      This is similar to <literal>Packages::Compress</literal> 
      except that it controls the compression for the Translation-en master file.</para></listitem>
      </varlistentry>

      <varlistentry><term><option>DeLinkLimit</option></term>
      <listitem><para>
      Specifies the number of kilobytes to delink (and 
      replace with hard links) per run. This is used in conjunction with the 
      per-section <literal>External-Links</literal> setting.</para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>FileMode</option></term>
      <listitem><para>
      Specifies the mode of all created index files. It 
      defaults to 0644. All index files are set to this mode with no regard 
      to the umask.</para></listitem>
      </varlistentry>

      <varlistentry><term><option>LongDescription</option></term>
      <listitem><para>
      Sets if long descriptions should be included in the Packages file or split
      out into a master Translation-en file.</para></listitem>
      </varlistentry>
     </variablelist>
   </refsect2>
   
   <refsect2><title>TreeDefault Section</title>
     <para>
     Sets defaults specific to <literal>Tree</literal> sections. All of these
     variables are substitution variables and have the strings $(DIST), 
     $(SECTION) and $(ARCH) replaced with their respective values.</para>
     
     <variablelist>     
      <varlistentry><term><option>MaxContentsChange</option></term>
      <listitem><para>
      Sets  the number of kilobytes of contents 
      files that are generated each day. The contents files are round-robined
      so that over several days they will all be rebuilt.</para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>ContentsAge</option></term>
      <listitem><para>
      Controls the number of days a contents file is allowed
      to be checked without changing. If this limit is passed the mtime of the 
      contents file is updated. This case can occur if the package file is 
      changed in such a way that does not result in a new contents file 
      [override edit for instance]. A hold off is allowed in hopes that new 
      .debs will be installed, requiring a new file anyhow. The default is 10, 
      the units are in days.</para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>Directory</option></term>
      <listitem><para>
      Sets the top of the .deb directory tree. Defaults to
      <filename>$(DIST)/$(SECTION)/binary-$(ARCH)/</filename></para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>SrcDirectory</option></term>
      <listitem><para>
      Sets the top of the source package directory tree. Defaults to
      <filename>$(DIST)/$(SECTION)/source/</filename></para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>Packages</option></term>
      <listitem><para>
      Sets the output Packages file. Defaults to 
      <filename>$(DIST)/$(SECTION)/binary-$(ARCH)/Packages</filename></para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>Sources</option></term>
      <listitem><para>
      Sets the output Sources file. Defaults to 
      <filename>$(DIST)/$(SECTION)/source/Sources</filename></para></listitem>
      </varlistentry>

      <varlistentry><term><option>Translation</option></term>
      <listitem><para>
      Set the output Translation-en master file with the long descriptions if they
      should be not included in the Packages file. Defaults to
      <filename>$(DIST)/$(SECTION)/i18n/Translation-en</filename></para></listitem>
      </varlistentry>

      <varlistentry><term><option>InternalPrefix</option></term>
      <listitem><para>
      Sets the path prefix that causes a symlink to be
      considered an internal link instead of an external link. Defaults to
      <filename>$(DIST)/$(SECTION)/</filename></para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>Contents</option></term>
      <listitem><para>
      Sets the output Contents file. Defaults to
      <filename>$(DIST)/Contents-$(ARCH)</filename>. If this setting causes multiple 
      Packages files to map onto a single Contents file (such as the default) 
      then <command>apt-ftparchive</command> will integrate those package files 
      together automatically.</para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>Contents::Header</option></term>
      <listitem><para>
      Sets header file to prepend to the contents output.</para></listitem>
      </varlistentry>

      <varlistentry><term><option>BinCacheDB</option></term>
      <listitem><para>
      Sets the binary cache database to use for this 
      section. Multiple sections can share the same database.</para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>FileList</option></term>
      <listitem><para>
      Specifies that instead of walking the directory tree, 
      <command>apt-ftparchive</command> should read the list of files from the given 
      file. Relative files names are prefixed with the archive directory.</para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>SourceFileList</option></term>
      <listitem><para>
      Specifies that instead of walking the directory tree, 
      <command>apt-ftparchive</command> should read the list of files from the given 
      file. Relative files names are prefixed with the archive directory. 
      This is used when processing source indexes.</para></listitem>
      </varlistentry>
     </variablelist>     
   </refsect2>
   
   <refsect2><title>Tree Section</title>
     <para>
     The <literal>Tree</literal> section defines a standard Debian file tree which 
     consists of a base directory, then multiple sections in that base 
     directory and finally multiple Architectures in each section. The exact 
     pathing used is defined by the <literal>Directory</literal> substitution variable.</para> 
     <para>
     The <literal>Tree</literal> section takes a scope tag which sets the 
     <literal>$(DIST)</literal> variable and defines the root of the tree 
     (the path is prefixed by <literal>ArchiveDir</literal>).
     Typically this is a setting such as <filename>dists/&stable-codename;</filename>.</para>
     <para>
     All of the settings defined in the <literal>TreeDefault</literal> section can be
     use in a <literal>Tree</literal> section as well as three new variables.</para>
     <para>
     When processing a <literal>Tree</literal> section <command>apt-ftparchive</command> 
     performs an operation similar to:
     <programlisting>
for i in Sections do 
   for j in Architectures do
      Generate for DIST=scope SECTION=i ARCH=j
     </programlisting></para>

     <variablelist>     
      <varlistentry><term><option>Sections</option></term>
      <listitem><para>
      This is a space separated list of sections which appear 
      under the distribution, typically this is something like 
      <literal>main contrib non-free</literal></para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>Architectures</option></term>
      <listitem><para>
      This is a space separated list of all the 
      architectures that appear under search section. The special architecture 
      'source' is used to indicate that this tree has a source archive.</para></listitem>
      </varlistentry>

      <varlistentry><term><option>LongDescription</option></term>
      <listitem><para>
      Sets if long descriptions should be included in the Packages file or split
      out into a master Translation-en file.</para></listitem>
      </varlistentry>

      <varlistentry><term><option>BinOverride</option></term>
      <listitem><para>
      Sets the binary override file. The override file 
      contains section, priority and maintainer address information.</para></listitem>
      </varlistentry>

      <varlistentry><term><option>SrcOverride</option></term>
      <listitem><para>
      Sets the source override file. The override file 
      contains section information.</para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>ExtraOverride</option></term>
      <listitem><para>
      Sets the binary extra override file.</para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>SrcExtraOverride</option></term>
      <listitem><para>
      Sets the source extra override file.</para></listitem> 
      </varlistentry>
     </variablelist>
   </refsect2>
   
   <refsect2><title>BinDirectory Section</title>
     <para>
     The <literal>bindirectory</literal> section defines a binary directory tree 
     with no special structure. The scope tag specifies the location of 
     the binary directory and the settings are similar to the <literal>Tree</literal> 
     section with no substitution variables or
     <literal>Section</literal><literal>Architecture</literal> settings.</para>
     <variablelist>
      <varlistentry><term><option>Packages</option></term>
      <listitem><para>
      Sets the Packages file output.</para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>Sources</option></term>
      <listitem><para>
      Sets the Sources file output. At least one of
      <literal>Packages</literal> or <literal>Sources</literal> is required.</para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>Contents</option></term>
      <listitem><para>
      Sets the Contents file output. (optional)</para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>BinOverride</option></term>
      <listitem><para>
      Sets the binary override file.</para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>SrcOverride</option></term>
      <listitem><para>
      Sets the source override file.</para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>ExtraOverride</option></term>
      <listitem><para>
      Sets the binary extra override file.</para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>SrcExtraOverride</option></term>
      <listitem><para>
      Sets the source extra override file.</para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>BinCacheDB</option></term>
      <listitem><para>
      Sets the cache DB.</para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>PathPrefix</option></term>
      <listitem><para>
      Appends a path to all the output paths.</para></listitem>
      </varlistentry>
      
      <varlistentry><term><option>FileList</option></term><term><option>SourceFileList</option></term>
      <listitem><para>
      Specifies the file list file.</para></listitem>
      </varlistentry>
     </variablelist>
   </refsect2>
 </refsect1>


 <refsect1><title>The Binary Override File</title>
   <para>The binary override file is fully compatible with &dpkg-scanpackages;. It
   contains 4 fields separated by spaces. The first field is the package name,
   the second is the priority to force that package to, the third is the
   the section to force that package to and the final field is the maintainer 
   permutation field.</para>
   <para>The general form of the maintainer field is:
   <literallayout>old [// oldn]* => new</literallayout>
   or simply,
   <literallayout>new</literallayout>
   The first form allows a double-slash separated list of old email addresses
   to be specified. If any of those are found then new is substituted for the
   maintainer field. The second form unconditionally substitutes the 
   maintainer field.</para>
 </refsect1>

 
 <refsect1><title>The Source Override File</title>
   <para>
   The source override file is fully compatible with &dpkg-scansources;. It
   contains 2 fields separated by spaces. The first fields is the source 
   package name, the second is the section to assign it.</para>
 </refsect1>   

 <refsect1><title>The Extra Override File</title>
   <para>
   The extra override file allows any arbitrary tag to be added or replaced
   in the output. It has 3 columns, the first is the package, the second is
   the tag and the remainder of the line is the new value.</para>
 </refsect1>   

 <refsect1><title>options</title>
   &apt-cmdblurb;
   
   <variablelist>
     <varlistentry><term><option>--md5</option></term><term><option>--sha1</option></term><term><option>--sha256</option></term>
     <listitem><para>
     Generate the given checksum. These options default to on, when turned off the generated
     index files will not have the checksum fields where possible.
     Configuration Items: <literal>APT::FTPArchive::<replaceable>Checksum</replaceable></literal> and
     <literal>APT::FTPArchive::<replaceable>Index</replaceable>::<replaceable>Checksum</replaceable></literal> where
     <literal><replaceable>Index</replaceable></literal> can be <literal>Packages</literal>, <literal>Sources</literal> or
     <literal>Release</literal> and <literal><replaceable>Checksum</replaceable></literal> can be <literal>MD5</literal>,
     <literal>SHA1</literal> or <literal>SHA256</literal>.</para></listitem>
     </varlistentry>

     <varlistentry><term><option>-d</option></term><term><option>--db</option></term>
     <listitem><para>
     Use a binary caching DB. This has no effect on the generate command.
     Configuration Item: <literal>APT::FTPArchive::DB</literal>.</para></listitem>
     </varlistentry>

     <varlistentry><term><option>-q</option></term><term><option>--quiet</option></term>
     <listitem><para>
     Quiet; produces output suitable for logging, omitting progress indicators.
     More q's will produce more quiet up to a maximum of 2. You can also use
     <option>-q=#</option> to set the quiet level, overriding the configuration file. 
     Configuration Item: <literal>quiet</literal>.</para></listitem>
     </varlistentry>

     <varlistentry><term><option>--delink</option></term>
     <listitem><para>
     Perform Delinking. If the <literal>External-Links</literal> setting is used then 
     this option actually enables delinking of the files. It defaults to on and 
     can be turned off with <option>--no-delink</option>.
     Configuration Item: <literal>APT::FTPArchive::DeLinkAct</literal>.</para></listitem>
     </varlistentry>

     <varlistentry><term><option>--contents</option></term>
     <listitem><para>
     Perform contents generation. When this option is set and package indexes
     are being generated with a cache DB then the file listing will also be
     extracted and stored in the DB for later use. When using the generate 
     command this option also allows the creation of any Contents files. The 
     default is on.
     Configuration Item: <literal>APT::FTPArchive::Contents</literal>.</para></listitem>
     </varlistentry>

     <varlistentry><term><option>-s</option></term><term><option>--source-override</option></term>
     <listitem><para>
     Select the source override file to use with the <literal>sources</literal> command.
     Configuration Item: <literal>APT::FTPArchive::SourceOverride</literal>.</para></listitem>
     </varlistentry>

     <varlistentry><term><option>--readonly</option></term>
     <listitem><para>
     Make the caching databases read only. 
     Configuration Item: <literal>APT::FTPArchive::ReadOnlyDB</literal>.</para></listitem>
     </varlistentry>

     <varlistentry><term><option>-a</option></term><term><option>--arch</option></term>
     <listitem><para>Accept in the <literal>packages</literal> and <literal>contents</literal>
     commands only package files matching <literal>*_arch.deb</literal> or
     <literal>*_all.deb</literal> instead of all package files in the given path.
     Configuration Item: <literal>APT::FTPArchive::Architecture</literal>.</para></listitem>
     </varlistentry>

     <varlistentry><term><option>APT::FTPArchive::AlwaysStat</option></term>
     <listitem><para>
     &apt-ftparchive; caches as much as possible of metadata in a cachedb. If packages
     are recompiled and/or republished with the same version again, this will lead to problems
     as the now outdated cached metadata like size and checksums will be used. With this option
     enabled this will no longer happen as it will be checked if the file was changed.
     Note that this option is set to "<literal>false</literal>" by default as it is not recommend
     to upload multiply versions/builds of a package with the same versionnumber, so in theory
     nobody will have these problems and therefore all these extra checks are useless.
     </para></listitem>
     </varlistentry>

     <varlistentry><term><option>APT::FTPArchive::LongDescription</option></term>
     <listitem><para>
     This configuration option defaults to "<literal>true</literal>" and should only be set to
     <literal>"false"</literal> if the Archive generated with &apt-ftparchive; also provides
     <filename>Translation</filename> files. Note that the <filename>Translation-en</filename>
     master file can only be created in the generate command.
     </para></listitem>
     </varlistentry>

     &apt-commonoptions;
     
   </variablelist>
 </refsect1>

<refsect1><title>Examples</title>

<para>To create a compressed Packages file for a directory containing
binary packages (.deb):

<programlisting>
<command>apt-ftparchive</command> packages <replaceable>directory</replaceable> | <command>gzip</command> > <filename>Packages.gz</filename>
</programlisting></para>

</refsect1>

 <refsect1><title>See Also</title>
   <para>&apt-conf;</para>
 </refsect1>

 <refsect1><title>Diagnostics</title>
   <para><command>apt-ftparchive</command> returns zero on normal operation, decimal 100 on error.</para>
 </refsect1>

 &manbugs;
 
</refentry>