Archive::Libarchive

High-level bindings to libarchive

NAME

Archive::Libarchive - High-level bindings to libarchive

SYNOPSIS

use v6;

use Archive::Libarchive;
use Archive::Libarchive::Constants;

sub MAIN(:$file! where { .IO.f // die "file '$file' not found" })
{
  my Archive::Libarchive $a .= new:
      operation => LibarchiveExtract,
      file => $file,
      flags => ARCHIVE_EXTRACT_TIME +| ARCHIVE_EXTRACT_PERM +| ARCHIVE_EXTRACT_ACL +| ARCHIVE_EXTRACT_FFLAGS;
  try {
    $a.extract: sub (Archive::Libarchive::Entry $e --> Bool) { $e.pathname eq 'test2' };
    CATCH {
      say "Can't extract files: $_";
    }
  }
  $a.close;
}

For more examples see the example directory.

DESCRIPTION

Archive::Libarchive provides an OO interface to libarchive using Archive::Libarchive::Raw.

As the Libarchive site (http://www.libarchive.org/) states, its implementation is able to:

Read a variety of formats, including tar, pax, cpio, zip, xar, lha, ar, cab, mtree, rar, and ISO images.
Write tar, pax, cpio, zip, xar, ar, ISO, mtree, and shar archives.
Handle automatically archives compressed with gzip, bzip2, lzip, xz, lzma, or compress.

new(LibarchiveOp :file?, Int :format?, :@filters?)

Creates an Archive::Libarchive object. It takes one mandatory argument: operation, what kind of operation will be performed.

The list of possible operations is provided by the LibarchiveOp enum:

LibarchiveRead: open the archive to list its content.
LibarchiveWrite: create a new archive. The file must not be already present.
LibarchiveOverwrite: create a new archive. The file will be overwritten if present.
LibarchiveExtract: extract the archive content.

When extracting one can specify some options to be applied to the newly created files. The default options are:

ARCHIVE_EXTRACT_TIME +| ARCHIVE_EXTRACT_PERM +| ARCHIVE_EXTRACT_ACL +| ARCHIVE_EXTRACT_FFLAGS

Those constants are defined in Archive::Libarchive::Constants, part of the Archive::Libarchive::Raw distribution. More details about those operation modes can be found on the libarchive site: http://www.libarchive.org/

If the optional argument $file is provided, then it will be opened; if not provided during the initialization, the program must call the open method later.

If the optional $format argument is provided, then the object will select that specific format while dealing with the archive.

List of possible read formats:

7zip
ar
cab
cpio
empty
gnutar
iso9660
lha
mtree
rar
raw
tar
warc
xar
zip

List of possible write formats:

7zip
ar
cpio
gnutar
iso9660
mtree
pax
raw
shar
ustar
v7tar
warc
xar
zip

If the optional @filters parameter is provided, then the object will add those filter to the archive. Multiple filters can be specified, so a program can manage a file.tar.gz.uu for example. The order of the filters is significant, in order to correctly deal with such files as file.tar.uu.gz and file.tar.gz.uu.

List of possible read filters:

bzip2
compress
gzip
grzip
lrzip
lz4
lzip
lzma
lzop
none
rpm
uu
xz

List of possible write filters:

b64encode
bzip2
compress
grzip
gzip
lrzip
lz4
lzip
lzma
lzop
none
uuencode
xz

Note

Recent versions of libarchive implement an automatic way to determine the best mix of format and filters. If one's using a pretty recent libarchive, both $format and @filters may be omitted: the new method will determine automatically the right combination of parameters. Older versions though don't have that capability and the programmer has to define explicitly both parameters.

open(Str size?, :$format?, :@filters?)

open(Buf $data!)

Opens an archive; the first form is used on files, while the second one is used to open an archive that resides in memory. The first argument is always mandatory, while the other ones might been omitted. $size is the size of the internal buffer and defaults to 10240 bytes.

Note: this module does't apply $*CWD to the file name under the hood, so this will create a file in the original directory.

use Archive::Libarchive;

my Archive::Libarchive $a .= new: operation => LibarchiveWrite;
chdir 'subdir';
$a.open: 'file.tar.gz', format => 'gnutar', filters => ['gzip'];
…

close

Closes the internal archive object, frees the memory and cleans up.

extract-opts(Int $flags?)

Sets the options for the files created when extracting files from an archive. The default options are:

ARCHIVE_EXTRACT_TIME +| ARCHIVE_EXTRACT_PERM +| ARCHIVE_EXTRACT_ACL +| ARCHIVE_EXTRACT_FFLAGS

next-header(Archive::Libarchive::Entry:D $e! --> Bool)

When reading an archive this method fills the Entry object and returns True till it reaches the end of the archive.

The Entry object is pubblicly defined inside the Archive::Libarchive module. It's initialized this way:

my Archive::Libarchive::Entry $e .= new;

So a complete archive lister can be implemented in few lines:

use Archive::Libarchive;

sub MAIN(Str :$file! where { .IO.f // die "file '$file' not found" })
{
  my Archive::Libarchive $a .= new: operation => LibarchiveRead, file => $file;
  my Archive::Libarchive::Entry $e .= new;
  while $a.next-header($e) {
    $e.pathname.say;
    $a.data-skip;
  }
  $a.close;
}

data-skip(--> Int)

When reading an archive this method skips file data to jump to the next header. It returns ARCHIVE_OK or ARCHIVE_EOF (defined in Archive::Libarchive::Constants)

read-file-content(Archive::Libarchive::Entry $e! --> Buf)

This method reads the content of a file represented by its Entry object and returns it.

write-header(Str pathname?, Int :filetype?, Int :atime?, Int :ctime?, Int :uid?, Int :uname?, Str :$gname? --> Bool)

When creating an archive this method writes the header entry for the file being inserted into the archive. The only mandatory argument is the file name, every other argument has a reasonable default. If the being inserted into the archive is a symbolic link, the target will be composed as a pathname relative to the base directory of the file, not as a full pathname. More details can be found on the libarchive site.

Each optional argument is available as a method of the Archive::Libarchive::Entry object and it can be set when needed.

Note write-header has a lot of optional arguments whose values are collected from the file one is adding to the archive. When using the second form of write-data one has to provide at least these arguments:

$size
$atime
$mtime
$ctime

For example:

$a.write-header($filename,
                :size($buffer.bytes),
                :atime(now.Int),
                :mtime(now.Int),
                :ctime(now.Int));

write-data(Str $path --> Bool)

write-data(Buf $data --> Bool)

When creating an archive this method writes the data for the file being inserted into the archive. $path is the pathname of the file to be archived, while $data is a data buffer.

extract(Str $destpath? --> Bool)

extract(&callback:(Archive::Libarchive::Entry destpath? --> Bool)

When extracting files from an archive this method does all the dirty work. If used in the first form it extracts all the files. The second form takes a callback function, which receives a Archive::Libarchive::Entry object.

For example, this will extract only the file whose name is test2:

$a.extract: sub (Archive::Libarchive::Entry $e --> Bool) { $e.pathname eq 'test2' };

In both cases one can specify the directory into which the files will be extracted.

lib-version

Returns a hash with the version number of libarchive and of each library used internally.

Errors

When the underlying library returns an error condition, the methods will return a Failure object, which can be trapped and the exception can be analyzed and acted upon.

The exception object has two fields: $errno and $error, and return a message stating the error number and the associated message as delivered by libarchive.

Prerequisites

This module requires the libarchive library to be installed. Please follow the instructions below based on your platform:

Debian Linux

sudo apt-get install libarchive13

The module uses Archive::Libarchive::Raw which looks for a library called libarchive.so.

Installation

To install it using zef (a module management tool):

$ zef update
$ zef install Archive::Libarchive

Author

Fernando Santagata

Contributions

Many thanks to Haythem Elganiny for implementing some multi methods in the Entry class.

License The Artistic License 2.0