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

NAME

Test::BDD::Cucumber::Manual::Tutorial - Quick Start Guide

VERSION

version 0.17

Introduction

In this article we're going to jump straight in to using Test::BDD::Cucumber to build some simple tests for Digest, a core Perl module which provides message digests.

We'll create a features/ directory, and put our first test case in it, features/basic.feature in it. The contents of it are, in their entirity:

 # Somehow I don't see this replacing the other tests this module has...
 Feature: Simple tests of Digest.pm
  As a developer planning to use Digest.pm
  I want to test the basic functionality of Digest.pm
  In order to have confidence in it

  Background:
    Given a usable Digest class

  Scenario: Check MD5
    Given a Digest MD5 object
    When I've added "foo bar baz" to the object
    And I've added "bat ban shan" to the object
    Then the hex output is "bcb56b3dd4674d5d7459c95e4c8a41d5"
    Then the base64 output is "1B2M2Y8AsgTpgAmY7PhCfg"

  Scenario: Check SHA-1
    Given a Digest SHA-1 object
    When I've added "<data>" to the object
    Then the hex output is "<output>"
    Examples:
      | data | output   |
      | foo  | 0beec7b5ea3f0fdbc95d0dd47f3c5bc275da8a33 |
      | bar  | 62cdb7020ff920e5aa642c3d4066950dd1f01f4d |
      | baz  | bbe960a25ea311d21d40669e93df2003ba9b90a2 |

  Scenario: MD5 longer data
    Given a Digest MD5 object
    When I've added the following to the object
      """
      Here is a chunk of text that works a bit like a HereDoc. We'll split
      off indenting space from the lines in it up to the indentation of the
      first \"\"\"
      """
    Then the hex output is "75ad9f578e43b863590fae52d5d19ce6"

This is a complete test, and if you run pherkin against it, you will get sane output! It just doesn't do anything ... yet.

In the features/ we'll add a step_definitions/ directory, and add our first (and again, only) step definitions features/step_definitions/basic_steps.pl file in it:

 #!perl

 use strict;
 use warnings;

 use Test::More;
 use Test::BDD::Cucumber::StepFile;
 use Method::Signatures;

 Given qr/a usable (\S+) class/, func ($c) { use_ok( $1 ); };
 Given qr/a Digest (\S+) object/, func ($c) {
    my $object = Digest->new($1);
    ok( $object, "Object created" );
    $c->stash->{'scenario'}->{'object'} = $object;
 };

 When qr/I've added "(.+)" to the object/, func ($c) {
    $c->stash->{'scenario'}->{'object'}->add( $1 );
 };

 When "I've added the following to the object", func ($c) {
    $c->stash->{'scenario'}->{'object'}->add( $c->data );
 };

 Then qr/the (.+) output is "(.+)"/, func ($c) {
    my $method = {base64 => 'b64digest', 'hex' => 'hexdigest' }->{ $1 } ||
        do { fail("Unknown output type $1"); return };
    is( $c->stash->{'scenario'}->{'object'}->$method, $2 );
 };

When you run pherkin or the Test::Builder-based test which does the same thing ("900_run_features.thttps://github.com/sheriff/test-bdd-cucumber-perl/blob/master/t/900_run_features.t" in t), we look for a features/ directory, and search for step definitions files (matched by *_steps.pl) and feature files (matched by *.feature).

The step matchers (the code that starts with Given, When and Then) are all loaded, and then we execture the feature files one by one. Let's step through the feature file, and look at how it matches up to the step definitions file.

Name and conditions of satisfaction

 Feature: Simple tests of Digest.pm
  As a developer planning to use Digest.pm
  I want to test the basic functionality of Digest.pm
  In order to have confidence in it

The first non-comment line of your feature file is a description of what you intend to do. You need to start the name itself with the string Feature:, and that should be the first line of your file, save comments (denoted by #).

Anything after that before the next new-line are your conditions of satisfaction. These aren't parsed, they're treated as human-readable text, and by convention, they're a user story.

Background

  Background:
    Given a usable Digest class

Next up, we have the Background section. The Background is a special kind of Scenario that doesn't have an explicit name, and should occur only once in your feature file. Its steps are run before the steps of every other scenario - the harnesses distributed with this distro won't display the Background section separately, they'll just subsume the steps in to the other scenarios.

This is matched by:

 Given qr/a usable (\S+) class/, func ($c) { use_ok( $1 ); };

Given() is a function exported by Test::BDD::Cucumber::StepFile that accepts two arguments: a regular expression (or a string when you don't need to do any smart matching) and a coderef. We use Method::Signatures to save some typing here, so note that written longer hand, the above might be written:

 Given(
     qr/a usable (\S+) class/,
     sub {
        my $c = shift;
        use_ok( $1 );
     };
 );

If you're paying attention, you might notice that use_ok comes from Test::More. Each step is run, from a Test::Builder perspective, as its own distinct test file. This happens seamlessly, so you can use any Test::Builder-based testing tools in your step definitions without really worrying about it. There's some more detail in Test::BDD::Cucumber::Manual::Steps.

The First Scenario...

  Scenario: Check MD5
    Given a Digest MD5 object
    When I've added "foo bar baz" to the object
    And I've added "bat ban shan" to the object
    Then the hex output is "bcb56b3dd4674d5d7459c95e4c8a41d5"
    Then the base64 output is "1B2M2Y8AsgTpgAmY7PhCfg"

The first scenario is delimited from the previous steps by a blank line, and it's called Check MD5. Scenarios are marked out using the Scenario: keyword, and just like the Background section before, it's a series of steps. These steps rely on the step before, which means we can examine the Test::Builder::StepContext object $c a little more closely.

 Given qr/a Digest (\S+) object/, func ($c) {
    my $object = Digest->new($1);
    ok( $object, "Object created" );
    $c->stash->{'scenario'}->{'object'} = $object;
 };

Creates a step definition. We create a new Digest object, and then use Test::More's ok() function to check that worked. We then put it in the stash for other steps to use. There are three stashes documented in Test::Builder::StepContext, feature, scenario and step. As you might expect, feature is available to all step definitions that are being executed as part of a feature, and scenario is available to all steps that are being executed as part of a feature.

The context is the single argument that gets passed to each step, and it contains evertything that step should need to execute. We'll be looking at some of the methods you can call against it as we look at other steps, and you can find complete documentation for it here: Test::Builder::StepContext.

This scenario also introduce several ways of starting a step, Given, When, and Then, as well as And. These are used to organize steps by type, with Given tending to describe setup steps, When describing the key actions that you're testing, and Then describing the outputs you're looking for. You can find more on this here: https://github.com/cucumber/cucumber/wiki/Given-When-Then.

A step definition you've declared with Given will not match a step starting with Then. You can use the keyword Step to declare general matching steps in your step definition files, although it's considered bad practice.

Finally, the keywords And and But are simply replaced with the verb on the line before them.

Scenario Outlines

  Scenario: Check SHA-1
    Given a Digest SHA-1 object
    When I've added "<data>" to the object
    Then the hex output is "<output>"
    Examples:
      | data | output   |
      | foo  | 0beec7b5ea3f0fdbc95d0dd47f3c5bc275da8a33 |
      | bar  | 62cdb7020ff920e5aa642c3d4066950dd1f01f4d |
      | baz  | bbe960a25ea311d21d40669e93df2003ba9b90a2 |

The next scenario adds one of the three ways you can provide structured data to your steps, using placeholders and a table. This scenario is run three times, one for each table row, and with the <placeholders > being replaced by the appropriate row's column. These are called Scenario Outlines.

Multiline Step Arguments

  Scenario: MD5 longer data
    Given a Digest MD5 object
    When I've added the following to the object
      """
      Here is a chunk of text that works a bit like a HereDoc. We'll split
      off indenting space from the lines in it up to the indentation of the
      first \"\"\"
      """
    Then the hex output is "75ad9f578e43b863590fae52d5d19ce6"

While before we were looking at structured data on a Scenario level, we can also provide it on a Step level, in two ways. Firstly, we can provide multi-line strings, as above, using a feature that is syntactically similar to pystrings, and conceptually similar to HEREDOCs. The contents of the string will be available to the step definition via the data() method of the context:

 When "I've added the following to the object", func ($c) {
    $c->stash->{'scenario'}->{'object'}->add( $c->data );
 };

While we don't have an example of it here, you can also provide tables to your steps, which will also be available via data():

 Scenario: Sort Employees
   Given a set of employees
     | name  | wage   | hair color |
     | Peter | 10,000 | brown      |
     | John  | 20,000 | blond      |
     | Joan  | 30,000 | green      |

You can find out more about these features in the Cucumber documentation here: https://github.com/cucumber/cucumber/wiki/Multiline-Step-Arguments.

Next Steps...

That's the tutorial done! You can find out more about writing steps in Test::BDD::Cucumber::Manual::Steps, the documentation for our simple command-line tool in App::pherkin, and how to integrate with Test::Builder in Test::BDD::Cucumber::Manual::Integration.

AUTHOR

Peter Sergeant pete@clueball.com

LICENSE

Copyright 2011, Peter Sergeant; Licensed under the same terms as Perl