[ANN] Test-Data-Split - split data-driven tests into several test scripts.

[Shlomi Fish]

Hi all,

I am proud to announce Test-Data-Split:

* https://metacpan.org/release/Test-Data-Split

* https://github.com/shlomif/perl-Test-Data-Split

This is a small Perl module that allows one to automatically generate test
files based on data, that can later be run individually or parallelised using
prove -j$N .

Here's an example. Given this generator file:

< CODE >

#!/usr/bin/perl

use strict;
use warnings;
use autodie;

use FindBin;
use lib "$FindBin::Bin/../t/t/lib";

use Test::Data::Split;
use FC_Solve::Test::Valgrind;

Test::Data::Split->new(
    {
        target_dir => 't',
        filename_cb => sub {
            my ($self, $args) = @_;

            return "valgrind--$args->{id}.t";
        },
        contents_cb => sub {
            my ($self, $args) = @_;
            my $id_quoted = quotemeta($args->{id});
            return <<"EOF";
#!/usr/bin/perl

use strict;
use warnings;

use Test::More tests => 1;

use FC_Solve::Test::Valgrind;

# TEST
FC_Solve::Test::Valgrind->new->run_valgrind_id_test({ id => qq/$id_quoted/, });

EOF
        },
        data_obj => FC_Solve::Test::Valgrind->new,
    }
)->run;
  
< / CODE >

And a FC_Solve::Test::Valgrind module like this one:

< CODE >

package FC_Solve::Test::Valgrind;

use strict;
use warnings;

use parent 'Test::Data::Split::Backend::Hash';

use Test::More ();
use Carp ();
use File::Spec ();
use File::Temp qw( tempdir );

my %valgrind_tests =
(
    'dbm_fc_solver_1' =>
    {
        prog => "dbm_fc_solver",
        argv =>
        [
            '--offload-dir-path', {type => 'tempdir', },
            { type => 'catfile',
                args => [{ type => 'ENV', arg => 'FCS_SRC_PATH'},
                    't', 't', 'data',
                    'sample-boards', '2freecells-24-mid-with-colons.board'
                ],
            }
        ],
        blurb => qq{dbm_fc_solver from 24-mid-with-colons.},
    },

[snipped]
);

sub get_hash
{
    return \%valgrind_tests;
}

[snipped]

sub run_valgrind_id_test
{
    local $Test::Builder::Level = $Test::Builder::Level + 1;

    my ($self, $args) = @_;

    my $id = $args->{id};

    return _test_using_valgrind($id, $self->lookup_data($id))
}

< / CODE >

Then I get individual test scripts like:

<CODE>
#!/usr/bin/perl

use strict;
use warnings;

use Test::More tests => 1;

use FC_Solve::Test::Valgrind;

# TEST
FC_Solve::Test::Valgrind->new->run_valgrind_id_test({ id =>
qq/board_gen__aisleriot__t_only/, });


</CODE>

<CODE>
#!/usr/bin/perl

use strict;
use warnings;

use Test::More tests => 1;

use FC_Solve::Test::Valgrind;

# TEST
FC_Solve::Test::Valgrind->new->run_valgrind_id_test({ id =>
qq/dbm_fc_solver_1/, });


</CODE>

Etc.

=============

Some plans for the future are:

1. Code several different backends: DBI/SQL, BDB/Google LevelDB , etc.

2. Allow a way to aggregate several IDs into an individual script. 

3. Perhaps create an ::Easy interface with more magic.

Comments are welcome. Enjoy and Shabbath Shalom.

Regards,

        Shlomi Fish

-- 
-----------------------------------------------------------------
Shlomi Fish       http://www.shlomifish.org/
Understand what Open Source is - http://shlom.in/oss-fs

Chuck Norris can only convince you that you're deceased. Summer Glau can
also convince you that you're alive, which is much harder.
    — http://www.shlomifish.org/humour/bits/facts/Summer-Glau/

Please reply to list if it's a mailing list post - http://shlom.in/reply .