++ed by:
KWAKWA AWNCORP BENNING CKRAS PAVELSR

6 PAUSE users
8 non-PAUSE users.

Erik Huelsmann
and 21 contributors

NAME

Test::BDD::Cucumber::Manual::Steps - How to write Step Definitions

VERSION

version 0.62

INTRODUCTION

The 'code' part of a Cucumber test-suite are the Step Definition files which match steps, and execute code based on them. This document aims to give you a quick overview of those.

STARTING OFF

Most of your step files will want to start something like:

 #!perl

 use strict;
 use warnings;

 use Test::More; # 'use Test2::V0;' is also supported
 use Test::BDD::Cucumber::StepFile;

The fake shebang line gives some hints to syntax highlighters, and use strict; and use warnings; are hopefully fairly standard at this point.

Most of my Step Definition files make use of Test::More, but you can use any Test2 or Test::Builder based testing module. E.g. Test2::V0 or Test::Exception.

Test::BDD::Cucumber::StepFile gives us the functions Given(), When(), Then() and Step().

STEP DEFINITIONS

 Given qr/I have (\d+)/, sub {
    S->{'count'} += $1;
 };

 When "The count is an integer", sub {
    S->{'count'} =
        int( S->{'count'} );
 };

 Then qr/The count should be (\d+)/, sub {
    is( S->{'count'}, C->matches->[0], "Count matches" );
 };

Each of the exported verb functions accept a regular expression (or a string that's used as one), and a coderef. The coderef is passed a single argument, the Test::BDD::Cucumber::StepContext object. Before the subref is executed, localized definitions of S and C are set, such that the lines below are equivalent:

  # Access the first match
  sub { my $context = shift; print $context->matches->[0] }
  sub { C->matches->[0] }

  # Set a value in the scenario-level stash
  sub { my $context = shift; my $stash = $context->stash->{'scenario'}; $stash->{'count'} = 1 }
  sub { S->{'count'} = 1 }

We will evaluate the regex immediately before we execute the coderef, so you can use $1, $2, $etc. Similarly you can access named matches using $+{match_name}.

Accessing step, scenario and feature properties

Step functions have access to the various properties of the step, scenario and feature in which they're being used. This includes tags, line numbers, etc. E.g.:

  # Examples of step properties
  C->step->line->number
  C->step->verb
  C->step->original_verb

  # Examples of scenario properties
  C->scenario->name
  C->scenario->tags

  # Examples of feature properties
  C->feature->name
  C->feature->tags
  C->feature->language

For a full review of available properties, see Test::BDD::Cucumber::Model::Step, Test::BDD::Cucumber::Model::Scenario and Test::BDD::Cucumber::Model::Feature respectively.

Re-using step definitions

Sometimes you want to call one step from another step. You can do this via the StepContext, using the dispatch() method. For example:

  Given qr/I have entered (\d+)/, sub {
        C->dispatch( 'Given', "I have pressed $1");
        C->dispatch( 'Given', "I have pressed enter", { some => 'data' } );
  };

For more on this topic, check the Redispatching section in the documentation for Test::BDD::Cucumber::StepContext.

LOCALIZATION

Both feature files and step files can be written using non-english Gherkin keywords. A german feature file could look like the example below.

  # language: de
  Funktionalität: Grundlegende Taschenrechnerfunktionen
    Um sicherzustellen, dass ich die Calculator-Klasse korrekt programmiert habe,
    möchte ich als Entwickler einige grundlegende Funktionen prüfen,
    damit ich beruhigt meine Calculator-Klasse verwenden kann.

    Szenario: Anzeige des ersten Tastendrucks
      Gegeben sei ein neues Objekt der Klasse Calculator
      Wenn ich 1 gedrückt habe
      Dann ist auf der Anzeige 1 zu sehen

To see which keywords (and sub names) to use, ask pherkin about a specific language:

 > pherkin --i18n de
 | feature          | "Funktionalität"                             |
 | background       | "Grundlage"                                  |
 ...
 | given (code)     | "Angenommen", "Gegebensei", "Gegebenseien"   |
 | when (code)      | "Wenn"                                       |
 | then (code)      | "Dann"                                       |

The last three lines of this list show you which sub names to use in your step file as indicated by the '(code)' suffix. A corresponding step file specifying a step function for Wenn ich 1 gedrückt habe, could be:

  #!perl

  use strict;
  use warnings;
  use utf8;    # Interpret accented German chars in regexes and identifiers properly

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


  Wenn qr/^ich (.+) gedrückt habe/, sub {
      S->{'Calculator'}->press($_) for split( /(,| und) /, C->matches->[0] );
  };

For more extensive examples see examples/i18n_de/ and examples/i18n_es.

ADDITIONAL STEPS

Next to the steps that will be matched directly against feature file input, a number of additional step functions are supported:

  • Before and After

    These steps create hooks into the evaluation process of feature files. E.g.

       Before sub {  # Run before every scenario
          # ... scenario set up code
       };
    
       After sub {  # Run after every scenario
         # ... scenario tear down code
       };

    For more extensive hook functionality, see Test::BDD::Cucumber::Extension.

  • Transform

    The Transform step serves to map matched values or table rows from feature file (string) input to step input values. The step takes two arguments, same as the Given, When and Then steps: a regular expression and a code reference. E.g.

       Transform qr/^(\d+)$/, sub {
          # transform matches of digit-only strings
    
          my $rv = $1;
          # ... do something with $rv
          return $rv;
       };
    
       Transform qr/^table:col1,col2$/, sub {
          # transform tables with 2 columns, named col1 and col2 respectively
    
          my ($step_context, $data) = @_;
          # ... transform data in $data
          return $data;
       };

BEST PRACTICES

When writing step files, it's a good idea to take a few things into account.

  • Declare a package at the top of your step file

    By declaring a specific package (your own), you make sure not to step on internals of other modules. At the time of writing, the default package is Test::BDD::Cucumber::StepFile, which may lead to errors being reported in that package, even though they occur in your step file (which is confusing).

    The default package may change in the future and it will likely not be seeded with the content of the T::B::C::StepFile package.

  • Declare a different package per step file

    By using different packages per step file (or group of step files), name spaces are isolated which reduces the risk of importing functions with the same name from different packages.

    An example where this will be the case is when some of your step files are written using Test::More and some others are written using Test2::Bundle::More -- both export a function ok, but with conflicting function prototypes.

  • Don't define functions in your step file

    Especially step files provided by extensions. Step files may be loaded more than once, depending on the exact scenario in which App::pherkin is run. When the step files are being loaded multiple times, there won't be any impact on step definition, but any function definitions will cause 'function redefined' warnings.

NEXT STEPS

How step files are loaded is discussed in Test::BDD::Cucumber::Manual::Architecture, but isn't of much interest. Of far more interest should be seeing what you have available in Test::BDD::Cucumber::StepContext...

AUTHOR

Peter Sergeant pete@clueball.com

LICENSE

Copyright 2011-2019, Peter Sergeant; Licensed under the same terms as Perl