# ABSTRACT: pull archives from upstream repositories package App::Pinto::Command::pull; use strict; use warnings; #------------------------------------------------------------------------------ use base 'App::Pinto::Command'; #------------------------------------------------------------------------------ our $VERSION = '0.14'; # VERSION #----------------------------------------------------------------------------- sub opt_spec { my ( $self, $app ) = @_; return ( [ 'cascade' => 'Always pick latest upstream package' ], [ 'diff-style=s' => 'Set style of diff reports' ], [ 'dry-run' => 'Do not commit any changes' ], [ 'message|m=s' => 'Message to describe the change' ], [ 'no-fail' => 'Do not fail when there is an error' ], [ 'recurse!' => 'Recursively pull prereqs (negatable)' ], [ 'pin' => 'Pin the packages to the stack' ], [ 'skip-missing-prerequisite|k=s@' => 'Skip missing prereq (repeatable)' ], [ 'skip-all-missing-prerequisites|K' => 'Skip all missing prereqs' ], [ 'stack|s=s' => 'Put packages into this stack' ], [ 'use-default-message|M' => 'Use the generated message' ], [ 'with-development-prerequisites|wd' => 'Also pull prereqs for development' ], ); } #------------------------------------------------------------------------------ sub args_attribute { return 'targets' } #------------------------------------------------------------------------------ sub args_from_stdin { return 1 } #------------------------------------------------------------------------------ 1; __END__ =pod =encoding UTF-8 =for :stopwords Jeffrey Ryan Thalhammer norecurse =head1 NAME App::Pinto::Command::pull - pull archives from upstream repositories =head1 VERSION version 0.14 =head1 SYNOPSIS pinto --root=REPOSITORY_ROOT pull [OPTIONS] TARGET ... =head1 DESCRIPTION This command locates packages in your upstream repositories and then pulls the distributions providing those packages into your repository and registers them on a stack. Then it recursively locates and pulls all the distributions that are necessary to satisfy their prerequisites. You can also request to directly pull particular distributions. When locating packages, Pinto first looks at the packages that already exist in the local repository, then Pinto looks at the packages that are available on the upstream repositories. =head1 COMMAND ARGUMENTS Arguments are the targets that you want to pull. Targets can be specified as packages (with or without a version specification) or as distributions. Targets can be expressed in a number of ways, so please see L below for more information. You can also pipe arguments to this command over STDIN. In that case, blank lines and lines that look like comments (i.e. starting with "#" or ';') will be ignored. =head1 COMMAND OPTIONS =over 4 =item --cascade !! THIS OPTION IS EXPERIMENTAL !! When searching for a package (or one of its prerequisites), always take the latest satisfactory version of the package found amongst B the upstream repositories, rather than just taking the B satisfactory version that is found. Remember that Pinto only searches the upstream repositories when the local repository does not already contain a satisfactory version of the package. =item --diff-style=STYLE Controls the style of the diff reports. STYLE must be either C or C. Concise reports show only one record for each distribution added or deleted. Detailed reports show one record for every package added or deleted. The default style is C. However, the default style can changed by setting the C environment variable to your preferred STYLE. This variable affects the default style for diff reports generated by all other commands too. =item --dry-run Go through all the motions, but do not actually commit any changes to the repository. At the conclusion, a diff showing the changes that would have been made will be displayed. Use this option to see how upgrades would potentially impact the stack. =item --no-fail !! THIS OPTION IS EXPERIMENTAL !! Normally, failure to pull a target (or its prerequisites) causes the command to immediately abort and rollback the changes to the repository. But if C<--no-fail> is set, then only the changes caused by the failed target (and its prerequisites) will be rolled back and the command will continue processing the remaining targets. This option is useful if you want to throw a list of targets into a repository and see which ones are problematic. Once you've fixed the broken ones, you can throw the whole list at the repository again. =item --message=TEXT =item -m TEXT Use TEXT as the revision history log message. If you do not use the C<--message> option or the C<--use-default-message> option, then you will be prompted to enter the message via your text editor. Use the C or C or C environment variables to control which editor is used. A log message is not required whenever the C<--dry-run> option is set, or if the action did not yield any changes to the repository. =item --pin Pins the packages to the stack, so they cannot be changed until you unpin them. Only the packages in the requested targets will be pinned -- packages in prerequisites will not be pinned. However, you may pin them separately with the L command if you so desire. =item --recurse =item --no-recurse Recursively pull any distributions required to satisfy prerequisites for the targets. The default value for this option can be configured in the F configuration file for the repository (it is usually set to 1). To disable recursion, use C<--no-recurse>. =item --skip-missing-prerequisite=PACKAGE =item -k PACKAGE !! THIS OPTION IS EXPERIMENTAL !! Skip any prerequisite with name PACKAGE if a satisfactory version cannot be found. However, a warning will be given whenever this occurrs. This option only has effect when recursively fetching prerequisites for the targets (See also the C<--recurse> option). This option can be repeated. =item --skip-all-missing-prerequisites =item -K !! THIS OPTION IS EXPERIMENTAL !! Skips all missing prerequisites if a satisfactory version cannot be found. However, a warning will be given whenever this occurrs. This option will silently override the C<--skip-missing-prerequisite> option and only has effect when recursively fetching prerequisites for the targets (See also the C<--recurse> option). =item --stack=NAME =item -s NAME Puts all the packages onto the stack with the given NAME. Defaults to the name of whichever stack is currently marked as the default stack. Use the L command to see the stacks in the repository. =item --use-default-message =item -M Use the default value for the revision history log message. Pinto will generate a semi-informative log message just based on the command and its arguments. If you set an explicit message with C<--message>, the C<--use- default-message> option will be silently ignored. =item --with-development-prerequisites =item --wd Also pull development prerequisites so you'll have everything you need to work on those distributions, in the event that you need to patch them in the future. Be aware that most distributions do not actually declare their development prerequisites. =back =head1 TARGETS Targets are a compact notation that identifies the things you want to pull into your repository. Targets come in two flavors: package targets and distribution targets. =head2 Package Targets A package target consists of a package name and (optionally) a version specification. Here are some examples: Foo::Bar # Any version of Foo::Bar Foo::Bar~1.2 # Foo::Bar version 1.2 or higher Foo::Bar==1.2 # Only version 1.2 of Foo::Bar Foo::Bar<1,2!=1.3,<=1.9 # Complex version range Package names are case-sensitive, and the version specification must follow the format used by L. All whitespace within the target will be discarded. If your version specification contains any special shell characters, take care to quote or escape them in your command. In all cases, pinto queries the local repository and then each upstream repository in order, and pulls the first distribution it can find that provides a package which satisfies the version specification. =head2 Distribution Targets A distribution target consists of an author ID, zero or more subdirectories, and the distribution name and version number. This corresponds to the actual path where the distribution archive lives in the repository or CPAN mirror. Here are some examples. SHAKESPEARE/King-Lear-1.2.tar.gz # A specific distribution SHAKESPEARE/tragedies/Hamlet-4.2.tar.gz # Same, but with a subdirectory The author ID will always be forced to uppercase, but the reset of the path is case-sensitive. =head2 Caveats L has no strict rules on how packages are versioned. It is quite common to see a package with the same verison number (or no version at all) in many releases of a distribution. So when you specify a package target with a precise version or version range, what you actually get is the latest distribution (chronologically) that has a package which satisfies the target. Most of the time this works out fine because you usally pull the "main module" of the distribution and authors always increment that version in each release. Since most CPAN mirrors only report the latest version of a package they have, they often cannot satisfy package targets that have a precise version specification. However, the mirror at L is special and can locate a precise version of any package. Package targets always resolve to production releases, unless you specify a precise developer release version (e.g. C). But since most CPAN mirrors do not index developer releases, this only works when using the mirror at L. However, you can usually pull a developer release from any mirror by using a distribution target. Remember that developer releases are those with an underscore in the version number. For repositories created with Pinto version 0.098 or later, the first upstream source is C (unless you configure it otherwise). For repositories created with older versions, you can manually add C to the C parameter in the configuration file located at F<.pinto/config/pinto.ini> within the repository. =head1 AUTHOR Jeffrey Ryan Thalhammer =head1 COPYRIGHT AND LICENSE This software is copyright (c) 2015 by Jeffrey Ryan Thalhammer. This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself. =cut