NAME
   CPAN - query, download and build perl modules from CPAN sites

SYNOPSIS
   Interactive mode:

     perl -MCPAN -e shell;

   Batch mode:

     use CPAN;

     autobundle, clean, install, make, recompile, test

DESCRIPTION
   The CPAN module is designed to automate the make and install of
   perl modules and extensions. It includes some searching
   capabilities and knows how to use Net::FTP or LWP (or lynx or an
   external ftp client) to fetch the raw data from the net.

   Modules are fetched from one or more of the mirrored CPAN
   (Comprehensive Perl Archive Network) sites and unpacked in a
   dedicated directory.

   The CPAN module also supports the concept of named and versioned
   'bundles' of modules. Bundles simplify the handling of sets of
   related modules. See BUNDLES below.

   The package contains a session manager and a cache manager.
   There is no status retained between sessions. The session
   manager keeps track of what has been fetched, built and
   installed in the current session. The cache manager keeps track
   of the disk space occupied by the make processes and deletes
   excess space according to a simple FIFO mechanism.

   For extended searching capabilities there's a plugin for CPAN
   available, the CPAN::WAIT manpage. `CPAN::WAIT' is a full-text
   search engine that indexes all documents available in CPAN
   authors directories. If `CPAN::WAIT' is installed on your
   system, the interactive shell of <CPAN.pm> will enable the `wq',
   `wr', `wd', `wl', and `wh' commands which send queries to the
   WAIT server that has been configured for your installation.

   All other methods provided are accessible in a programmer style
   and in an interactive shell style.

 Interactive Mode

   The interactive mode is entered by running

       perl -MCPAN -e shell

   which puts you into a readline interface. You will have the most
   fun if you install Term::ReadKey and Term::ReadLine to enjoy
   both history and command completion.

   Once you are on the command line, type 'h' and the rest should
   be self-explanatory.

   The most common uses of the interactive modes are

   Searching for authors, bundles, distribution files and modules
     There are corresponding one-letter commands `a', `b', `d', and
     `m' for each of the four categories and another, `i' for any
     of the mentioned four. Each of the four entities is
     implemented as a class with slightly differing methods for
     displaying an object.

     Arguments you pass to these commands are either strings
     exactly matching the identification string of an object or
     regular expressions that are then matched case-insensitively
     against various attributes of the objects. The parser
     recognizes a regular expression only if you enclose it between
     two slashes.

     The principle is that the number of found objects influences
     how an item is displayed. If the search finds one item, the
     result is displayed as object->as_string, but if we find more
     than one, we display each as object->as_glimpse. E.g.

         cpan> a ANDK
         Author id = ANDK
             EMAIL        [email protected]
             FULLNAME     Andreas K�nig

         cpan> a /andk/
         Author id = ANDK
             EMAIL        [email protected]
             FULLNAME     Andreas K�nig

         cpan> a /and.*rt/
         Author          ANDYD (Andy Dougherty)
         Author          MERLYN (Randal L. Schwartz)

   make, test, install, clean  modules or distributions
     These commands take any number of arguments and investigates
     what is necessary to perform the action. If the argument is a
     distribution file name (recognized by embedded slashes), it is
     processed. If it is a module, CPAN determines the distribution
     file in which this module is included and processes that,
     following any dependencies named in the module's Makefile.PL
     (this behavior is controlled by *prerequisites_policy*.)

     Any `make' or `test' are run unconditionally. An

       install <distribution_file>

     also is run unconditionally. But for

       install <module>

     CPAN checks if an install is actually needed for it and prints
     *module up to date* in the case that the distribution file
     containing the module doesn't need to be updated.

     CPAN also keeps track of what it has done within the current
     session and doesn't try to build a package a second time
     regardless if it succeeded or not. The `force' command takes
     as a first argument the method to invoke (currently: `make',
     `test', or `install') and executes the command from scratch.

     Example:

         cpan> install OpenGL
         OpenGL is up to date.
         cpan> force install OpenGL
         Running make
         OpenGL-0.4/
         OpenGL-0.4/COPYRIGHT
         [...]

     A `clean' command results in a

       make clean

     being executed within the distribution file's working
     directory.

   readme, look module or distribution
     These two commands take only one argument, be it a module or a
     distribution file. `readme' unconditionally runs, displaying
     the README of the associated distribution file. `Look' gets
     and untars (if not yet done) the distribution file, changes to
     the appropriate directory and opens a subshell process in that
     directory.

   Signals
     CPAN.pm installs signal handlers for SIGINT and SIGTERM. While
     you are in the cpan-shell it is intended that you can press
     `^C' anytime and return to the cpan-shell prompt. A SIGTERM
     will cause the cpan-shell to clean up and leave the shell
     loop. You can emulate the effect of a SIGTERM by sending two
     consecutive SIGINTs, which usually means by pressing `^C'
     twice.

     CPAN.pm ignores a SIGPIPE. If the user sets
     inactivity_timeout, a SIGALRM is used during the run of the
     `perl Makefile.PL' subprocess.

 CPAN::Shell

   The commands that are available in the shell interface are
   methods in the package CPAN::Shell. If you enter the shell
   command, all your input is split by the
   Text::ParseWords::shellwords() routine which acts like most
   shells do. The first word is being interpreted as the method to
   be called and the rest of the words are treated as arguments to
   this method. Continuation lines are supported if a line ends
   with a literal backslash.

 autobundle

   `autobundle' writes a bundle file into the `$CPAN::Config-
   >{cpan_home}/Bundle' directory. The file contains a list of all
   modules that are both available from CPAN and currently
   installed within @INC. The name of the bundle file is based on
   the current date and a counter.

 recompile

   recompile() is a very special command in that it takes no
   argument and runs the make/test/install cycle with brute force
   over all installed dynamically loadable extensions (aka XS
   modules) with 'force' in effect. The primary purpose of this
   command is to finish a network installation. Imagine, you have a
   common source tree for two different architectures. You decide
   to do a completely independent fresh installation. You start on
   one architecture with the help of a Bundle file produced
   earlier. CPAN installs the whole Bundle for you, but when you
   try to repeat the job on the second architecture, CPAN responds
   with a `"Foo up to date"' message for all modules. So you invoke
   CPAN's recompile on the second architecture and you're done.

   Another popular use for `recompile' is to act as a rescue in
   case your perl breaks binary compatibility. If one of the
   modules that CPAN uses is in turn depending on binary
   compatibility (so you cannot run CPAN commands), then you should
   try the CPAN::Nox module for recovery.

 The four `CPAN::*' Classes: Author, Bundle, Module, Distribution

   Although it may be considered internal, the class hierarchy does
   matter for both users and programmer. CPAN.pm deals with above
   mentioned four classes, and all those classes share a set of
   methods. A classical single polymorphism is in effect. A
   metaclass object registers all objects of all kinds and indexes
   them with a string. The strings referencing objects have a
   separated namespace (well, not completely separated):

            Namespace                         Class

      words containing a "/" (slash)      Distribution
       words starting with Bundle::          Bundle
             everything else            Module or Author

   Modules know their associated Distribution objects. They always
   refer to the most recent official release. Developers may mark
   their releases as unstable development versions (by inserting an
   underbar into the visible version number), so the really hottest
   and newest distribution file is not always the default. If a
   module Foo circulates on CPAN in both version 1.23 and 1.23_90,
   CPAN.pm offers a convenient way to install version 1.23 by
   saying

       install Foo

   This would install the complete distribution file (say BAR/Foo-
   1.23.tar.gz) with all accompanying material. But if you would
   like to install version 1.23_90, you need to know where the
   distribution file resides on CPAN relative to the authors/id/
   directory. If the author is BAR, this might be BAR/Foo-
   1.23_90.tar.gz; so you would have to say

       install BAR/Foo-1.23_90.tar.gz

   The first example will be driven by an object of the class
   CPAN::Module, the second by an object of class
   CPAN::Distribution.

 Programmer's interface

   If you do not enter the shell, the available shell commands are
   both available as methods (`CPAN::Shell->install(...)') and as
   functions in the calling package (`install(...)').

   There's currently only one class that has a stable interface -
   CPAN::Shell. All commands that are available in the CPAN shell
   are methods of the class CPAN::Shell. Each of the commands that
   produce listings of modules (`r', `autobundle', `u') also return
   a list of the IDs of all modules within the list.

   expand($type,@things)
     The IDs of all objects available within a program are strings
     that can be expanded to the corresponding real objects with
     the `CPAN::Shell->expand("Module",@things)' method. Expand
     returns a list of CPAN::Module objects according to the
     `@things' arguments given. In scalar context it only returns
     the first element of the list.

   Programming Examples
     This enables the programmer to do operations that combine
     functionalities that are available in the shell.

         # install everything that is outdated on my disk:
         perl -MCPAN -e 'CPAN::Shell->install(CPAN::Shell->r)'

         # install my favorite programs if necessary:
         for $mod (qw(Net::FTP MD5 Data::Dumper)){
             my $obj = CPAN::Shell->expand('Module',$mod);
             $obj->install;
         }

         # list all modules on my disk that have no VERSION number
         for $mod (CPAN::Shell->expand("Module","/./")){
             next unless $mod->inst_file;
             # MakeMaker convention for undefined $VERSION:
             next unless $mod->inst_version eq "undef";
             print "No VERSION in ", $mod->id, "\n";
         }

     Or if you want to write a cronjob to watch The CPAN, you could
     list all modules that need updating:

         perl -e 'use CPAN; CPAN::Shell->r;'

     If you don't want to get any output if all modules are up to
     date, you can parse the output of above command for the
     regular expression //modules are up to date// and decide to
     mail the output only if it doesn't match. Ick?

     If you prefer to do it more in a programmer style in one
     single process, maybe something like this suites you better:

       # list all modules on my disk that have newer versions on CPAN
       for $mod (CPAN::Shell->expand("Module","/./")){
         next unless $mod->inst_file;
         next if $mod->uptodate;
         printf "Module %s is installed as %s, could be updated to %s from CPAN\n",
             $mod->id, $mod->inst_version, $mod->cpan_version;
       }

     If that gives you too much output every day, you maybe only
     want to watch for three modules. You can write

       for $mod (CPAN::Shell->expand("Module","/Apache|LWP|CGI/")){

     as the first line instead. Or you can combine some of the
     above tricks:

       # watch only for a new mod_perl module
       $mod = CPAN::Shell->expand("Module","mod_perl");
       exit if $mod->uptodate;
       # new mod_perl arrived, let me know all update recommendations
       CPAN::Shell->r;

 Methods in the four Classes

 Cache Manager

   Currently the cache manager only keeps track of the build
   directory ($CPAN::Config->{build_dir}). It is a simple FIFO
   mechanism that deletes complete directories below `build_dir' as
   soon as the size of all directories there gets bigger than
   $CPAN::Config->{build_cache} (in MB). The contents of this cache
   may be used for later re-installations that you intend to do
   manually, but will never be trusted by CPAN itself. This is due
   to the fact that the user might use these directories for
   building modules on different architectures.

   There is another directory ($CPAN::Config->{keep_source_where})
   where the original distribution files are kept. This directory
   is not covered by the cache manager and must be controlled by
   the user. If you choose to have the same directory as build_dir
   and as keep_source_where directory, then your sources will be
   deleted with the same fifo mechanism.

 Bundles

   A bundle is just a perl module in the namespace Bundle:: that
   does not define any functions or methods. It usually only
   contains documentation.

   It starts like a perl module with a package declaration and a
   $VERSION variable. After that the pod section looks like any
   other pod with the only difference being that *one special pod
   section* exists starting with (verbatim):

           =head1 CONTENTS

   In this pod section each line obeys the format

           Module_Name [Version_String] [- optional text]

   The only required part is the first field, the name of a module
   (e.g. Foo::Bar, ie. *not* the name of the distribution file).
   The rest of the line is optional. The comment part is delimited
   by a dash just as in the man page header.

   The distribution of a bundle should follow the same convention
   as other distributions.

   Bundles are treated specially in the CPAN package. If you say
   'install Bundle::Tkkit' (assuming such a bundle exists), CPAN
   will install all the modules in the CONTENTS section of the pod.
   You can install your own Bundles locally by placing a conformant
   Bundle file somewhere into your @INC path. The autobundle()
   command which is available in the shell interface does that for
   you by including all currently installed modules in a snapshot
   bundle file.

 Prerequisites

   If you have a local mirror of CPAN and can access all files with
   "file:" URLs, then you only need a perl better than perl5.003 to
   run this module. Otherwise Net::FTP is strongly recommended. LWP
   may be required for non-UNIX systems or if your nearest CPAN
   site is associated with an URL that is not `ftp:'.

   If you have neither Net::FTP nor LWP, there is a fallback
   mechanism implemented for an external ftp command or for an
   external lynx command.

 Finding packages and VERSION

   This module presumes that all packages on CPAN

   * declare their $VERSION variable in an easy to parse manner. This
     prerequisite can hardly be relaxed because it consumes far too
     much memory to load all packages into the running program just
     to determine the $VERSION variable. Currently all programs
     that are dealing with version use something like this

         perl -MExtUtils::MakeMaker -le \
             'print MM->parse_version(shift)' filename

     If you are author of a package and wonder if your $VERSION can
     be parsed, please try the above method.

   * come as compressed or gzipped tarfiles or as zip files and contain
     a Makefile.PL (well, we try to handle a bit more, but without
     much enthusiasm).

 Debugging

   The debugging of this module is pretty difficult, because we
   have interferences of the software producing the indices on
   CPAN, of the mirroring process on CPAN, of packaging, of
   configuration, of synchronicity, and of bugs within CPAN.pm.

   In interactive mode you can try "o debug" which will list
   options for debugging the various parts of the package. The
   output may not be very useful for you as it's just a by-product
   of my own testing, but if you have an idea which part of the
   package may have a bug, it's sometimes worth to give it a try
   and send me more specific output. You should know that "o debug"
   has built-in completion support.

 Floppy, Zip, Offline Mode

   CPAN.pm works nicely without network too. If you maintain
   machines that are not networked at all, you should consider
   working with file: URLs. Of course, you have to collect your
   modules somewhere first. So you might use CPAN.pm to put
   together all you need on a networked machine. Then copy the
   $CPAN::Config->{keep_source_where} (but not $CPAN::Config-
   >{build_dir}) directory on a floppy. This floppy is kind of a
   personal CPAN. CPAN.pm on the non-networked machines works
   nicely with this floppy. See also below the paragraph about CD-
   ROM support.

CONFIGURATION
   When the CPAN module is installed, a site wide configuration
   file is created as CPAN/Config.pm. The default values defined
   there can be overridden in another configuration file:
   CPAN/MyConfig.pm. You can store this file in
   $HOME/.cpan/CPAN/MyConfig.pm if you want, because $HOME/.cpan is
   added to the search path of the CPAN module before the use() or
   require() statements.

   Currently the following keys in the hash reference $CPAN::Config
   are defined:

     build_cache        size of cache for directories to build modules
     build_dir          locally accessible directory to build modules
     index_expire       after this many days refetch index files
     cpan_home          local directory reserved for this package
     gzip               location of external program gzip
     inactivity_timeout breaks interactive Makefile.PLs after this
                        many seconds inactivity. Set to 0 to never break.
     inhibit_startup_message
                        if true, does not print the startup message
     keep_source_where  directory in which to keep the source (if we do)
     make               location of external make program
     make_arg           arguments that should always be passed to 'make'
     make_install_arg   same as make_arg for 'make install'
     makepl_arg         arguments passed to 'perl Makefile.PL'
     pager              location of external program more (or any pager)
     prerequisites_policy
                        what to do if you are missing module prerequisites
                        ('follow' automatically, 'ask' me, or 'ignore')
     scan_cache         controls scanning of cache ('atstart' or 'never')
     tar                location of external program tar
     unzip              location of external program unzip
     urllist            arrayref to nearby CPAN sites (or equivalent locations)
     wait_list          arrayref to a wait server to try (See CPAN::WAIT)
     ftp_proxy,      }  the three usual variables for configuring
       http_proxy,   }  proxy requests. Both as CPAN::Config variables
       no_proxy      }  and as environment variables configurable.

   You can set and query each of these options interactively in the
   cpan shell with the command set defined within the `o conf'
   command:

   o conf <scalar option>
     prints the current value of the *scalar option*

   o conf <scalar option> <value>
     Sets the value of the *scalar option* to *value*

   o conf <list option>
     prints the current value of the *list option* in MakeMaker's
     neatvalue format.

   o conf <list option> [shift|pop]
     shifts or pops the array in the *list option* variable

   o conf <list option> [unshift|push|splice] <list>
     works like the corresponding perl commands.

 Note on urllist parameter's format

   urllist parameters are URLs according to RFC 1738. We do a
   little guessing if your URL is not compliant, but if you have
   problems with file URLs, please try the correct format. Either:

       file://localhost/whatever/ftp/pub/CPAN/

   or

       file:///home/ftp/pub/CPAN/

 urllist parameter has CD-ROM support

   The `urllist' parameter of the configuration table contains a
   list of URLs that are to be used for downloading. If the list
   contains any `file' URLs, CPAN always tries to get files from
   there first. This feature is disabled for index files. So the
   recommendation for the owner of a CD-ROM with CPAN contents is:
   include your local, possibly outdated CD-ROM as a `file' URL at
   the end of urllist, e.g.

     o conf urllist push file://localhost/CDROM/CPAN

   CPAN.pm will then fetch the index files from one of the CPAN
   sites that come at the beginning of urllist. It will later check
   for each module if there is a local copy of the most recent
   version.

   Another peculiarity of urllist is that the site that we could
   successfully fetch the last file from automatically gets a
   preference token and is tried as the first site for the next
   request. So if you add a new site at runtime it may happen that
   the previously preferred site will be tried another time. This
   means that if you want to disallow a site for the next transfer,
   it must be explicitly removed from urllist.

SECURITY
   There's no strong security layer in CPAN.pm. CPAN.pm helps you
   to install foreign, unmasked, unsigned code on your machine. We
   compare to a checksum that comes from the net just as the
   distribution file itself. If somebody has managed to tamper with
   the distribution file, they may have as well tampered with the
   CHECKSUMS file. Future development will go towards strong
   authentication.

EXPORT
   Most functions in package CPAN are exported per default. The
   reason for this is that the primary use is intended for the cpan
   shell or for oneliners.

POPULATE AN INSTALLATION WITH LOTS OF MODULES
   To populate a freshly installed perl with my favorite modules is
   pretty easiest by maintaining a private bundle definition file.
   To get a useful blueprint of a bundle definition file, the
   command autobundle can be used on the CPAN shell command line.
   This command writes a bundle definition file for all modules
   that are installed for the currently running perl interpreter.
   It's recommended to run this command only once and from then on
   maintain the file manually under a private name, say
   Bundle/my_bundle.pm. With a clever bundle file you can then
   simply say

       cpan> install Bundle::my_bundle

   then answer a few questions and then go out for a coffee.

   Maintaining a bundle definition file means to keep track of two
   things: dependencies and interactivity. CPAN.pm sometimes fails
   on calculating dependencies because not all modules define all
   MakeMaker attributes correctly, so a bundle definition file
   should specify prerequisites as early as possible. On the other
   hand, it's a bit annoying that many distributions need some
   interactive configuring. So what I try to accomplish in my
   private bundle file is to have the packages that need to be
   configured early in the file and the gentle ones later, so I can
   go out after a few minutes and leave CPAN.pm unattained.

WORKING WITH CPAN.pm BEHIND FIREWALLS
   Thanks to Graham Barr for contributing the following paragraphs
   about the interaction between perl, and various firewall
   configurations.

   Firewalls can be categorized into three basic types.

   http firewall
       This is where the firewall machine runs a web server and to
       access the outside world you must do it via the web server.
       If you set environment variables like http_proxy or
       ftp_proxy to a values beginning with http:// or in your web
       browser you have to set proxy information then you know you
       are running a http firewall.

       To access servers outside these types of firewalls with perl
       (even for ftp) you will need to use LWP.

   ftp firewall
       This where the firewall machine runs a ftp server. This kind
       of firewall will only let you access ftp serves outside the
       firewall. This is usually done by connecting to the firewall
       with ftp, then entering a username like
       "[email protected]"

       To access servers outside these type of firewalls with perl
       you will need to use Net::FTP.

   One way visibility
       I say one way visibility as these firewalls try to make
       themselve look invisible to the users inside the firewall.
       An FTP data connection is normally created by sending the
       remote server your IP address and then listening for the
       connection. But the remote server will not be able to
       connect to you because of the firewall. So for these types
       of firewall FTP connections need to be done in a passive
       mode.

       There are two that I can think off.

   SOCKS   If you are using a SOCKS firewall you will need to compile
           perl and link it with the SOCKS library, this is what is
           normally called a ``socksified'' perl. With this
           executable you will be able to connect to servers
           outside the firewall as if it is not there.

   IP Masquerade
           This is the firewall implemented in the Linux kernel, it
           allows you to hide a complete network behind one IP
           address. With this firewall no special compiling is need
           as you can access hosts directly.

BUGS
   We should give coverage for all of the CPAN and not just the
   PAUSE part, right? In this discussion CPAN and PAUSE have become
   equal -- but they are not. PAUSE is authors/ and modules/. CPAN
   is PAUSE plus the clpa/, doc/, misc/, ports/, src/, scripts/.

   Future development should be directed towards a better
   integration of the other parts.

   If a Makefile.PL requires special customization of libraries,
   prompts the user for special input, etc. then you may find CPAN
   is not able to build the distribution. In that case, you should
   attempt the traditional method of building a Perl module package
   from a shell.

AUTHOR
   Andreas K�nig <[email protected]>

SEE ALSO
   perl(1), CPAN::Nox(3)

You are viewing proxied material from ftp.icm.edu.pl. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.