man CPAN::Mini (3): create a minimal mirror of CPAN

VERSION

version 1.111016

SYNOPSIS

(If you're not going to do something weird, you probably want to look at the minicpan command, instead.)

use CPAN::Mini;
CPAN::Mini->update_mirror(
remote => "http://cpan.mirrors.comintern.su",
local => "/usr/share/mirrors/cpan",
log_level => 'debug',
);

DESCRIPTION

CPAN::Mini provides a simple mechanism to build and update a minimal mirror of the CPAN on your local disk. It contains only those files needed to install the newest version of every distribution. Those files are:

01mailrc.txt.gz
02packages.details.txt.gz
03modlist.data.gz
the last non-developer release of every dist for every author

METHODS

update_mirror

  CPAN::Mini->update_mirror(
    remote => "http://cpan.mirrors.comintern.su",
    local  => "/usr/share/mirrors/cpan",
    force  => 0,
    log_level => 'debug',
  );

This is the only method that need be called from outside this module. It will update the local mirror with the files from the remote mirror.

If called as a class method, "update_mirror" creates an ephemeral CPAN::Mini object on which other methods are called. That object is used to store mirror location and state.

This method returns the number of files updated.

The following options are recognized:

"local"
This is the local file path where the mirror will be written or updated.
"remote"
This is the URL of the CPAN mirror from which to work. A reasonable default will be picked by default. A list of CPAN mirrors can be found at <http://www.cpan.org/SITES.html>
"dirmode"
Generally an octal number, this option sets the permissions of created directories. It defaults to 0711.
"exact_mirror"
If true, the "files_allowed" method will allow all extra files to be mirrored.
"ignore_source_control"
If true, CPAN::Mini will not try to remove source control files during cleanup. See "clean_unmirrored" for details.
"force"
If true, this option will cause CPAN::Mini to read the entire module list and update anything out of date, even if the module list itself wasn't out of date on this run.
"skip_perl"
If true, CPAN::Mini will skip the major language distributions: perl, parrot, and ponie. It will also skip embperl, sybperl, bioperl, and kurila.
"log_level"
This defines the minimum level of message to log: debug, info, warn, or fatal
"errors"
If true, CPAN::Mini will warn with status messages on errors. (default: true)
"path_filters"
This options provides a set of rules for filtering paths. If a distribution matches one of the rules in "path_filters", it will not be mirrored. A regex rule is matched if the path matches the regex; a code rule is matched if the code returns 1 when the path is passed to it. For example, the following setting would skip all distributions from RJBS and SUNGO:
```
 path_filters => [
   qr/RJBS/,
   sub { $_[0] =~ /SUNGO/ }
 ]
```
"module_filters"
This option provides a set of rules for filtering modules. It behaves like path_filters, but acts only on module names. (Since most modules are in distributions with more than one module, this setting will probably be less useful than "path_filters".) For example, this setting will skip any distribution containing only modules with the word ``Acme'' in them:
```
 module_filters => [ qr/Acme/i ]
```
"also_mirror"
This option should be an arrayref of extra files in the remote CPAN to mirror locally.
"skip_cleanup"
If this option is true, CPAN::Mini will not try delete unmirrored files when it has finished mirroring
"offline"
If offline, CPAN::Mini will not attempt to contact remote resources.
"no_conn_cache"
If true, no connection cache will be established. This is mostly useful as a workaround for connection cache failures.

new

  my $minicpan = CPAN::Mini->new;

This method constructs a new CPAN::Mini object. Its parameters are described above, under "update_mirror".

mirror_indices

  $minicpan->mirror_indices;

This method updates the index files from the CPAN.

mirror_file

  $minicpan->mirror_file($path, $skip_if_present)

This method will mirror the given file from the remote to the local mirror, overwriting any existing file unless $skip_if_present is true.

file_allowed

  next unless $minicpan->file_allowed($filename);

This method returns true if the given file is allowed to exist in the local mirror, even if it isn't one of the required mirror files.

By default, only dot-files are allowed. If the "exact_mirror" option is true, all files are allowed.

clean_unmirrored

  $minicpan->clean_unmirrored;

This method looks through the local mirror's files. If it finds a file that neither belongs in the mirror nor is allowed (see the "file_allowed" method), "clean_file" is called on the file.

If you set "ignore_source_control" to a true value, then this doesn't clean up files that belong to source control systems. Currently this ignores:

        .cvs .cvsignore
        .svn .svnignore
        .git .gitignore

Send patches for other source control files that you would like to have added.

clean_file

  $minicpan->clean_file($filename);

This method, called by "clean_unmirrored", deletes the named file. It returns true if the file is successfully unlinked. Otherwise, it returns false.

log_warn

log

log_debug

  $minicpan->log($message);

This will log (print) the given message unless the log level is too low.

"log", which logs at the info level, may also be called as "trace" for backward compatibility reasons.

read_config

  my %config = CPAN::Mini->read_config(\%options);

This routine returns a set of arguments that can be passed to CPAN::Mini's "new" or "update_mirror" methods. It will look for a file called .minicpanrc in the user's home directory as determined by File::HomeDir.

config_file

  my $config_file = CPAN::Mini->config_file( { options } );

This routine returns the config file name. It first looks at for the "config_file" setting, then the "CPAN_MINI_CONFIG" environment variable, then the default ~/.minicpanrc, and finally the CPAN/Mini/minicpan.conf. It uses the first defined value it finds. If the filename it selects does not exist, it returns false.

OPTIONS is an optional hash reference of the "CPAN::Mini" config hash.

remote_from

  my $remote = CPAN::Mini->remote_from( $remote_from, $orig_remote, $quiet );

This routine take an string argument and turn it into a method call to handle to retrieve the a cpan mirror url from a source. Currently supported methods:

    cpan     - fetch the first mirror from your CPAN.pm config
    cpanplus - fetch the first mirror from your CPANPLUS.pm config

remote_from_cpan

  my $remote = CPAN::Mini->remote_from_cpan;

This routine loads your CPAN.pm config and returns the first mirror in mirror list. You can set this as your default by setting remote_from:cpan in your .minicpanrc file.

remote_from_cpanplus

  my $remote = CPAN::Mini->remote_from_cpanplus;

This routine loads your CPANPLUS.pm config and returns the first mirror in mirror list. You can set this as your default by setting remote_from:cpanplus in your .minicpanrc file.

THANKS

Thanks to David Dyck for letting me know about my stupid documentation errors.

Thanks to Roy Fulbright for finding an obnoxious bug on Win32.

Thanks to Shawn Sorichetti for fixing a stupid octal-number-as-string bug.

Thanks to sungo for implementing the filters, so I can finally stop mirroring bioperl, and Robert Rothenberg for suggesting adding coderef rules.

Thanks to Adam Kennedy for noticing and complaining about a lot of stupid little design decisions.

Thanks to Michael Schwern and Jason Kohles, for pointing out missing documentation.

Thanks to David Golden for some important bugfixes and refactoring.

AUTHORS

Ricardo SIGNES <[email protected]>
Randal Schwartz <[email protected]>

COPYRIGHT AND LICENSE

This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself.