perl-App-Prove-Remote/bin/rprove

#!/usr/bin/env perl
package App::Prove::Remote::rprove;

use strict;
use warnings;

use App::Prove::Remote::Connector ();
use Getopt::Long                  ();

exit run() unless caller();

sub run {
    my ( $host, $verbosity, $workdir ) = ( '127.0.0.1', 0, '', 'perl' );
    Getopt::Long::Configure( 'bundling', 'auto_help', 'pass_through' );
    Getopt::Long::GetOptions(
        'host|h=s'      => \$host,
        'verbosity|v=i' => \$verbosity,
		'workdir=s'     => \$workdir,
		'interpreter=s' => \$interpreter,
    );
    my $conn = App::Prove::Remote::Connector->new(
		$host, $verbosity, $workdir, $interpreter
	);
    return 0;
}

1;

__END__

=head1 NAME

rprove - Prove wrapper which executes your tests on the remote host

=head1 SYNOPSIS

rprove [options] [file ...]

 Options:
   -help            You are reading it!
   -host            Host to connect to. Defaults to 127.0.0.1.
   -verbosity       How verbose you want this (and the SSH connection) to be

=head1 OPTIONS

=over 8

=item B<-help>

Print a brief help message and exits.

=item B<-host>

Host to connect to. Defaults to 127.0.0.1

=item B<-verbosity>

How verbose you want this to be. Useful if you need to debug
some strange SSH behavior.

=item B<-wordkir>

Directory to execute the test from. Useful to set if the test requires it.

=item B<-interpreter>

Path to the interpreter to run your tests with. Default is /usr/bin/perl.
Useful to set if you have perl in a nonstandard location
or non-perl tests to execute that still emit valid TAP.

=back

=head1 DESCRIPTION

B<rprove> will locally run prove with a --exec argument which is a shim.
This (remote_shim.pl) will connect to the remote host for the test in
question and run it on the host.

Why do this? Because sometimes testing certain scenarios is better done
on a disposable remote environment instead of on the local environment.

If someone has a "smoker" like environment (Jenkins, some other CI) which
also runs your tests, this could also be of use from the orchestrator's
end.

Anyways, the user is responsible for ensuring the test (and code under
test) has been properly deployed to the remote system under test, so
make sure that's done first if you want this approach to work.

Output of the script is then read by the TAP parser as is expected for
a seamless testing experience *as if you had ran the test locally*.

=cut