cpan - easily interact with CPAN from the command line
# with arguments and no switches, installs specified modules
cpan module_name [ module_name ... ]
# with switches, installs modules with extra behavior
cpan [-cfgimtTw] module_name [ module_name ... ]
# with just the dot, install from the distribution in the
# current directory
cpan .
# without arguments, starts CPAN.pm shell
cpan
# force install modules (usually those that fail tests)
cpan -f module_name [ module_name ... ]
# install modules but without testing them
cpan -T module_name [ module_name ... ]
# dump the configuration
cpan -J
# load a different configuration to install Module::Foo
cpan -j some/other/file Module::Foo
# without arguments, but some switches
cpan [-ahrvACDlLO]
This script provides a command interface (not a shell) to CPAN. At the moment it uses CPAN.pm to do the work, but it is not a one-shot command runner for CPAN.pm.
Creates a CPAN.pm autobundle with CPAN::Shell->autobundle.
Shows the primary maintainers for the specified modules.
Runs a `make clean` in the specified module's directories.
Show the Changes files for the specified modules
Show the module details.
Force the specified action, when it normally would have failed. Use this to install a module even if its tests fail. When you use this option, -i is not optional for installing a module when you need to force it:
% cpan -f -i Module::Foo
Turn off CPAN.pm's attempts to lock anything. You should be careful with this since you might end up with multiple scripts trying to muck in the same directory. This isn't so much of a concern if you're loading a special config with -j
, and that config sets up its own work directories.
Downloads to the current directory the latest distribution of the module.
UNIMPLEMENTED
Download to the current directory the latest distribution of the modules, unpack each distribution, and create a git repository for each distribution.
If you want this feature, check out Yanick Champoux's Git::CPAN::Patch
distribution.
Print a help message and exit. When you specify -h
, it ignores all of the other options and arguments.
Install the specified modules.
Load local::lib
(think like -I
for loading lib paths).
Load the file that has the CPAN configuration data. This should have the same format as the standard CPAN/Config.pm file, which defines $CPAN::Config
as an anonymous hash.
Dump the configuration in the same format that CPAN.pm uses. This is useful for checking the configuration as well as using the dump as a starting point for a new, custom configuration.
List all installed modules with their versions
List the modules by the specified authors.
Make the specified modules.
Show the out-of-date modules.
Ping the configured mirrors
Find the best mirrors you could be using (but doesn't configure them just yet)
Recompiles dynamically loaded modules with CPAN::Shell->recompile.
Run a `make test` on the specified modules.
Do not test modules. Simply install them.
Upgrade all installed modules. Blindly doing this can really break things, so keep a backup.
Print the script version and CPAN.pm version then exit.
Print detailed information about the cpan client.
UNIMPLEMENTED
Turn on cpan warnings. This checks various things, like directory permissions, and tells you about problems you might have.
# print a help message
cpan -h
# print the version numbers
cpan -v
# create an autobundle
cpan -a
# recompile modules
cpan -r
# upgrade all installed modules
cpan -u
# install modules ( sole -i is optional )
cpan -i Netscape::Booksmarks Business::ISBN
# force install modules ( must use -i )
cpan -fi CGI::Minimal URI
There are several components in CPAN.pm that use environment variables. The build tools, ExtUtils::MakeMaker and Module::Build use some, while others matter to the levels above them. Some of these are specified by the Perl Toolchain Gang:
Lancaster Concensus: https://github.com/Perl-Toolchain-Gang/toolchain-site/blob/master/lancaster-consensus.md
Oslo Concensus: https://github.com/Perl-Toolchain-Gang/toolchain-site/blob/master/oslo-consensus.md
cpan
splits this variable on whitespace and prepends that list to @ARGV
before it processes the command-line arguments. For instance, if you always want to use local:lib
, you can set CPAN_OPTS
to -I
.
The log level to use, with either the embedded, minimal logger or Log::Log4perl if it is installed. Possible values are the same as the Log::Log4perl
levels: TRACE
, DEBUG
, INFO
, WARN
, ERROR
, and FATAL
. The default is INFO
.
The path to the git
binary to use for the Git features. The default is /usr/local/bin/git
.
Assume no one is paying attention and skips prompts for distributions that do that correctly. cpan(1)
sets this to 1
unless it already has a value (even if that value is false).
Use the default answer for a prompted questions. cpan(1)
sets this to 1
unless it already has a value (even if that value is false).
The script exits with zero if it thinks that everything worked, or a positive number if it thinks that something failed. Note, however, that in some cases it has to divine a failure by the output of things it does not control. For now, the exit codes are vague:
1 An unknown error
2 The was an external problem
4 There was an internal problem with the script
8 A module failed to install
* one shot configuration values from the command line
* none noted
Most behaviour, including environment variables and configuration, comes directly from CPAN.pm.
This code is in Github in the CPAN.pm repository:
https://github.com/andk/cpanpm
The source used to be tracked separately in another GitHub repo, but the canonical source is now in the above repo.
Japheth Cleaver added the bits to allow a forced install (-f).
Jim Brandt suggest and provided the initial implementation for the up-to-date and Changes features.
Adam Kennedy pointed out that exit() causes problems on Windows where this script ends up with a .bat extension
brian d foy, <bdfoy@cpan.org>
Copyright (c) 2001-2014, brian d foy, All Rights Reserved.
You may redistribute this under the same terms as Perl itself.