App-Fetchware

 view release on metacpan or  search on metacpan

bin/fetchware  view on Meta::CPAN

Fetchware's database is simply the directory C</var/log/fetchware> (on Unix when
run as root), or whatever File::HomeDir recommends. When packages are installed
the final version of that package that ends with C<.fpkg> is copied to your
fetchware database path. So after you install apache your fetchware database
will look like:

    ls /var/log/fetchware
    httpd-2.4.3.fpkg

It's not a real database or anything cool like that. It is simply a directory
containting a list of fetchware packages that have been installed. However, this
directory is managed by fetchware, and should not be messed with unless you are
sure of what you are doing.

=head2 What exactly is a fetchware package?

A fetchware package is a gziped tar archive with its file extension changed to
C<.fpkg>. This archive consists of the package that was downloaded in addition
to your Fetchwarefile. For example.

    tar tvf httpd-2.4.3.fpkg

lib/App/Fetchware.pm  view on Meta::CPAN

=back

Verifies the downloaded package stored in $package_path by downloading
$download_path.{asc,sha1,md5}> and comparing the two together. Uses the
helper subroutines C<{gpg,sha1,md5,digest}_verify()>.

=over

=item LIMITATIONS
Uses gpg command line, and the interface to gpg is a little brittle.
Crypt::OpenPGP is buggy and not currently maintainted again, so fetchware cannot
make use of it, so were stuck with using the command line gpg program.

=back

=over

=item drop_privs() NOTES

This section notes whatever problems you might come accross implementing and
debugging your Fetchware extension due to fetchware's drop_privs mechanism.

lib/App/Fetchware/Util.pm  view on Meta::CPAN

###BUGALERT### Add support for dry-run functionality!!!!
sub run_prog {
    my (@args) = @_;

    # Kill weird "Insecure dependency in system while running with -T switch."
    # fatal exceptions by clearing the taint flag with a regex. I'm not actually
    # running in taint mode, but it bizarrely thinks I am.
    for my $arg (@args) {
        if ($arg =~ /(.*)/) {
            $arg = $1;
        } else {
            die <<EOD;
php.Fetchwarefile: Match anything pattern match failed! Huh! This shouldn't
happen, and is probably a bug.
EOD
        }
    }

lib/App/Fetchware/Util.pm  view on Meta::CPAN

    $url =~ s!^file://!!; # Strip useless URL scheme.
    
    # Prepend original_cwd() only if the $url is *not* absolute, which will mess
    # it up.
    $url = catdir(original_cwd(), $url) unless file_name_is_absolute($url);

    # Download the file:// URL to the current directory, which should already be
    # in $temp_dir, because of start()'s chdir().
    #
    # Don't forget to clear taint. Fetchware does *not* run in taint mode, but
    # for some reason, bug?, File::Copy checks if data is tainted, and then
    # retaints it if it is already tainted, but for some reason I get "Insecure
    # dependency" taint failure exceptions when drop priving. The fix is to
    # always untaint my data as done below.
    ###BUGALERT### Investigate this as a possible taint bug in perl or just
    #File::Copy. Perhaps the cause is using File::Copy::cp(copy) after drop
    #priving with data from root?
    $url =~ /(.*)/;
    my $untainted_url = $1;
    my $cwd = cwd();
    $cwd =~ /(.*)/;
    my $untainted_cwd = $1;
    cp($untainted_url, $untainted_cwd) or die <<EOD;
App::Fetchware: run-time error. Fetchware failed to copy the download URL
[$untainted_url] to the working directory [$untainted_cwd]. Os error [$!].
EOD

    # Return just file filename of the downloaded file.
    return file($url)->basename();
}

lib/App/Fetchware/Util.pm  view on Meta::CPAN

EOD

    my @variables;
    for my $variable (split(/$MAGIC_NUMBER/, $$output)) {
        # And some error handling just in case.
        die <<EOD if not defined $variable;
fetchware: Huh? The child failed to write the proper variable back to the
parent! The variable is [$variable]. This should be defined but it is 
not!
EOD
        # Clear possibly tainted variables. It's a weird bug that makes no
        # sense. I don't turn -t or -T on, so what gives??? If you're curious
        # try commenting out the taint clearing code below, and running the
        # t/bin-fetchware-install.t test file (Or any other ones that call
        # drop_privs().).
        my $untainted;
        # Need the m//ms options to match strings with newlines in them.
        if ($variable =~ /(.*)/ms) {
            $untainted = $1;
        } else {
            die <<EOD;
App::Fetchware::Util: Untaint failed! Huh! This just shouldn't happen! It's
probably a bug. 
EOD
        }

        # Push $untainted instead of just $variable, because I want to return
        # untatined data instead of potentially tainted data.
        push @variables, $untainted;
    }

    return @variables;
}
###BUGALERT### Add some pipe parsers that use Storable too.

} # End $MAGIC_NUMBER bare block.

t/App-Fetchware-Util.t  view on Meta::CPAN

    ok(run_prog("$Config{perlpath}", '-e print "Testing 1...2...3!!!\n"') >> 8 == 0,
        'test run_prog() success');

    # Set bin/fetchware's $quiet to false.
    $fetchware::quiet = 0;
};


subtest 'test create_tempdir()' => sub {
    # Create my own original_cwd(), because it gets tainted, because I chdir
    # more than just once.
    my $original_cwd = cwd();
    # Test create_tempdir() successes.
    my $temp_dir = create_tempdir();
    ok(-e $temp_dir, 'checked create_tempdir() success.');
    ok(-e 'fetchware.sem', 'checked fetchware semaphore creation.');

    # chdir back to original_cwd(), so that this tempdir can be deleted.
    chdir original_cwd() or fail("Failed to chdir back to original_cwd()!");

t/App-FetchwareX-HTMLPageSync.t  view on Meta::CPAN

    expect_quit();
};



##BROKEN####TEST### t/App-FetchwareX-HTMLPageSync-new works great at testing new(), but for some
##BROKEN####TEST### reason I can't figure out Expect screws up  this longer test of new() just as
##BROKEN####TEST### it screws it my Expect test for App::Fetchware's new(). So, this test is being
##BROKEN####TEST### commented out too. Consider reating a Test::IO::React, but IO::React has had
##BROKEN####TEST### no new releases for like a decade, so its either unmaintainted and broken, or
##BROKEN####TEST### perhaps unmaintainted, and still works. Don't know...
##BROKEN####TEST##subtest 'test HTMLPageSync new() success' => sub {
##BROKEN####TEST##    skip_all_unless_release_testing();
##BROKEN####TEST##
##BROKEN####TEST##    plan(skip_all => 'Optional Test::Expect testing module not installed.')
##BROKEN####TEST##        unless eval {require Test::Expect; Test::Expect->import(); 1;};
##BROKEN####TEST##
##BROKEN####TEST##    # Have Expect tell me what it's doing for easier debugging.
##BROKEN####TEST##    $Expect::Exp_Internal = 1;
##BROKEN####TEST##
##BROKEN####TEST##    # Test saying n.