The Perl Toolchain Summit needs more sponsors. If your company depends on Perl, please support this very important event.

NAME

Archive::Builder - File generation and archiving framework

SYNOPSIS

  # Make a builder with one section, and some files
  my $Builder = Archive::Builder->new;
  my $Section = $Builder->new_section( 'html' );
  $Section->add_file( 'one.html', 'string', qq~<html><body>
        Hello World!
        </body></html>~ );
  $Section->add_file( 'two.html', 'file', './source/file.html' );
  $Section->add_file( 'three.html', 'Custom::function', @args );

  # Generate and save to disk
  $Builder->save( './somewhere' );

  # Create an zip file from it and save it.
  my $Archive = $Builder->archive( 'zip' ).
  $Archive->save( 'foo.zip' );

  # Create a tar.gz file of just one section
  my $Tar = $Section->archive( 'tar.gz' );

DESCRIPTION

Perl is often used for applications that generate large numbers of files, and Archive::Builder is designed to assist in these sorts of tasks.

It provides a framework for defining a set of files, and how they will be generated, and a series of methods for turning them into an Archive of varying types, or saving directly to disk.

Structure

Each Archive::Builder object consists of one or more Archive::Builder::Sections, which contain one or more Archive::Builder::Files. Each of these files know their location within the section, and are given a generation function, with a set of arguments specific to each generator. Some simple generators are provided built-in, or you can provide an function name as the generator.

Generating Archives

Once a Archive::Builder is fully defined, you can save it to disk, or get an archive, containing the generated files, in one of several formats. You can also save and get the archive any of the individual sections within the builder.

During the generation process for an entire Archive::Builder a subdirectory is created for each section matching the name of the section. So, for a builder with a Section name 'one', containing a single file 'two.txt', and a section 'three', containing files 'four.html' and 'five.dat', the following file structure would result

  one/two.txt
  three/four.html
  three/five.dat

Caching

Caching is dont of the generated files on a per-file basis. Two calls to the content() method of an Archive::Builder::File object will only result in the file being generated once, and the same contents returned twice.

Generation on Demand Caveats

During an output action, such as a save() or archive() method call, the contents of each file are generated only as needed. This means that if the generation of a file fails, an action may have already been taken ( especially in the case of save(), where you may end up with only part of the files written to disk.

To avoid this, in most cases you should test() the Archive or Section first. This will generate all of the files, and cache them. A save() or archive() done after this will be done with the cached generated content.

This should be done whenever you have a large of complex generation tree, that you consider has a non-zero chance of one of the files failing to generate correctly.

METHODS

The methods for the three main classes involved in generation trees, Archive::Builder, Archive::Builder::Section and Archive::Builder::File are documented below. For the archive handlers the builders generate, see Archive::Builder::Archive. For information on the built-in generators, and how to write your own generators, see Archive::Builder::Generators.

Common Error Handlers

Errors from any object of any type below Archive::Builder are set in the global $Archive::Builder::errstr variable. They can also be retrieved from the errstr() method called on any object.

Archive::Builder

new

The new() constructor takes no arguments, and returns a new Archive::Builder object.

add_section $Archive::Builder::Section

The add_section method takes as an argument an Archive::Builder::Section object, and adds it to the builder.

Returns true if the section is added successfully. Returns undef on error, for example if another Section with the same name has already been added.

new_section $name

Creates a new Archive::Builder::Section object with the name provided, and immediately adds it to the builder. Returns the Section created.

Returns the new Section object on success. Returns undef on error.

new_sections $name [, $name, ... ]

new_sections \%names_to_paths

Primarily used for initial set up of Builder objects, the new_sections method adds a number of sections at the same time.

It accepts as argument either the names of the section to be created, with the paths of them to be the same as their names, or alternatively, a reference to a HASH with the keys as section names, and the values as section paths.

section $name

Finds and returns a Archive::Builder::Section object with the provided name within the builder and returns it. Returns undef if passed name does not exist.

sections

Returns a hash containing all the sections, indexed by name. Returns 0 if no sections have been created in the builder.

section_list

Returns all the sections in the builder as a list, sorted by section name. Returns a null list () if no sections are defined in the builder.

remove_section $name

Removes a section of a given name from the builder, if it exists. Returns undef if no such section exists.

file_count

Returns the total number of files in all sections in the builder

files

Returns a HASH reference containing all of the Achive::Builder::File object in the Archive::Builder, keyed by full path name.

save $directory

Generates the file tree for the entire builder and attempts to save it below a given directory. The passed directory does not have to exist, it will be created on demand.

Returns true if all files were generated and saved successfully. Returns undef if an error occurs, or the directory is bad.

delete

Because of the structure used to support the parent methods, you should probably explicitly delete Builds when you are done with them to avoid memory leaks due to circular dependencies.

The delete method always returns true.

reset

If the contents of any of the files in the Archive::Builder has been generated ( and thus cached ), the reset method will remove any cached content from the files, forcing them to be generated again.

The reset method always returns true.

archive $type

Creates a handle to an archive of a specified type related to the builder. Types can only be used if the modules that support them are installed. The following types are supported, and their prerequisites are listed.

  zip    - Archive::Zip
  tar    - Archive::Tar
  tar.gz - Archive::Tar
  tgz    - Archive::Tar

The tar.gz and tgz are aliases that produce the same thing with a different file extension.

The archive method only returns a Archive::Builder::Archive handle to the object, not the object itself. Also, the files are not generated at the time that the archive is created, so generation errors cannot be guarenteed to have occurred by this time.

To save or otherwise act on the archive, see the Archive::Builder::Archive section below.

Archive::Builder::Section

new name

Creates a new Archive::Builder::Section object of a given name. Although meant to be used in an Archive::Builder object, they can still be used effectively standalone, as they have both save and archive methods.

Returns undef is an invalid section name is given. A section name must contain only word ( \w ) characters and be 1 to 31 characters long.

As a side note, the reason that Sections exist at all is so that Builders can be defined containing multiple sections, where the sections will be saved to different locations, but should still be passed around as a single entity.

name

Returns the name of the Section.

path [ path ]

When used within the context of a Builder object, and set to the same value as the section's name by default, this method returns the path below the Builder root that will be used, or if passed a relative path, will set the path to a new value. You are not likely to need this, as in general, the same value will suffice for both the name and path.

Builder

If the Section has been added to a Builder, the Builder method will return it.

Returns a Archive::Builder object if added, or undef if not.

add_file $Archive::Builder::File

Adds an existing Archive::Builder::File object to the section.

Returns true on success. Returns undef on error, or if the path of the file clashes with an existing file in the Section.

This could happen if you try to add a file with the same name, of if your path contains a directory that is already in the Section as a file. For example, the two files could not exist in the same Section.

  first/second
  first/second/third

Creation of the directory first/second would be blocked by the existing file first/second ( or vica versa ). This issue is caught for you now, rather than wait until we are halfway through writing the files to disk to find out.

new_file $path, $generator [, @arguments ]

Creates a new file, using the arguments provided, and immediately adds it to the current section. See the new method for Archive::Builder::File below for more details on the arguments.

Returns true if the file is created and added successfuly. Returns undef if an error occurs during either the creation or addition of the file.

file $path

Finds the Archive::Builder::File object with the given path and returns it. Returns undef if no such file exists.

files

Returns a reference to a hash containing all of the files objects, keyed by their paths. Returns 0 if no files exist within the section.

file_list

Returns a list of all the file objects, sorted by path. Returns a null array () if no files exist within the section.

remove_file $path

Removes the file object with the given path from the section. Returns undef if no such path exists within the section.

file_count

Returns the number of files contained in the section

save $directory

The save() method works the same as the Archive::Builder save method, generating the files and saving them below the directory provided. Again, the directory is created on demand.

Returns undef if an error during generation or saving occurs.

delete

The delete method deletes a Section, removing it from its parent Builder if applicable, and removing all child Files from the Section.

The delete method always returns true.

reset

If the contents of any of the files in the Section has been generated ( and thus cached ), the reset method will remove any cached content from the files, forcing them to be generated again.

The reset method always returns true.

archive $type

As for the Archive::Builder acrhive method, creates an archive handle of the given type. Returns a Archive::Builder::Archive object on success. Returns undef on error.

Archive::Builder::File

new $path, $generator [, @arguments ]

Creates a new Archive::Builder::File object and returns it. This method is not normally used directly, with the Archive::Builder::Section method new_file() being more typically used.

The path argument should be a valid looking relative path. That is, it cannot start with /. For safety, the use of escaping slashes and relative '..' paths are restricted for safety.

The generator should be a string containing the name of the function to be used to generate the file contents. A check will be done to ensure that the module containing the function is installed, although the existance of the function itself will not be tested. For example, for the generator function Foo::Bar::makeme, a test to make sure Foo::Bar is installed will be done.

To specify a function in the the main package ( say in a script ), the format main::function MUST be used. A generator value that does not contain a package seperator will be assumed to be one of the default generators. The list of default generators, and instructions on how to write your own generators, are in the Archive::Builder::Generators documentation.

Anything passed after the generator are assumed to be arguments to the generator function, and will be stored and passed as needed. Note that the arguments are not copied or cloned, so any objects passed as arguments and later modified, will be generated using the modified values. This is considered a feature. If you need to freeze a copy of the object for the generation, you are recommended to Clone it before passing.

path

Returns the path for the file. This cannot be changed after creation.

generator

Returns the generator for the file. This cannot be changed after creation

arguments

Returns a reference to an array containing the arguments to be passed to the generator, or 0 if there are no arguments. The list of arguments cannot be changed after creation ( although of course objects passed can be changed outside the scope of this API ).

binary

This method will analyse the file contents ( generating if needed ) to determine if the file is a binary file. While not 100% accurate, it should be good enough for most situations.

executable

Calling this method will add a hint to the file that it should be considered as an executable file, should the need arise. This is most likely used in situations where permissions need to be set after generation.

Section

If added to a Section, the Section method returns the Section to which we have been added.

Returns a Archive::Builder::Section object if the File is added to one, or undef if not.

contents

Generates and returns the contents of the file as a scalar reference. Returns undef if a generation error occurs.

save $filename

Bypassing the normal generation process and path name, the save method allows you to generate a single file object and save it to a specific filename. Any directories need to write the file will be created on demand. Returns undef if a generation permissions error occurs.

delete

If added to a Section, the delete method allows us to remove and delete the file from the parent Section. Always returns true.

reset

If the file has been generated ( and thus cached ), the reset method will remove any cached content from the files, forcing it to be generated again.

The reset method always returns true.

TODO

Better control over caching, more archive types, pre-generation testing.

SUPPORT

Bugs should always be submitted via the CPAN bug tracker.

http://rt.cpan.org/NoAuth/ReportBug.html?Queue=Archive-Builder

For other issues, contact the maintainer.

AUTHOR

Adam Kennedy <adamk@cpan.org>

SEE ALSO

Archive::Builder::Archive, Archive::Builder::Generators, Archive::Tar, Archive::Zip.

COPYRIGHT

Copyright 2002 - 2011 Adam Kennedy.

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

The full text of the license can be found in the LICENSE file included with this module.