use v5.14; use warnings; package Test::BDD::Cucumber::StepContext 0.87; use Moo; use Types::Standard qw( Bool Str HashRef ArrayRef InstanceOf ); use List::Util qw( first ); =head1 NAME Test::BDD::Cucumber::StepContext - Data made available to step definitions =head1 VERSION version 0.87 =head1 DESCRIPTION The coderefs in Step Definitions have a single argument passed to them, a C object. This is an attribute-only class, populated by L. When steps are run normally, C is set directly before execution to return the context; this allows you to do: sub { return C->columns } instead of: sub { my $c = shift; return $c->columns; } =head1 ATTRIBUTES =head2 columns If the step-specific data supplied is a table, the this attribute will contain the column names in the order they appeared. =cut has 'columns' => ( is => 'ro', isa => ArrayRef ); =head2 _data Step-specific data. Will either be a text string in the case of a """ string, or an arrayref of hashrefs if the step had an associated table. See the C method below. =cut has '_data' => ( is => 'ro', isa => Str|ArrayRef, init_arg => 'data', default => '' ); =head2 stash A hash of hashes, containing two keys, C, C. The stash allows you to persist data across features or scenarios. The scenario-level stash is also available to steps by calling C, making the following two lines of code equivalent: sub { my $context = shift; my $stash = $context->stash->{'scenario'}; $stash->{'count'} = 1 } sub { S->{'count'} = 1 } =cut has 'stash' => ( is => 'ro', required => 1, isa => HashRef ); =head2 feature =head2 scenario =head2 step Links to the L, L, and L objects respectively. =cut has 'feature' => ( is => 'ro', required => 1, isa => InstanceOf['Test::BDD::Cucumber::Model::Feature'] ); has 'scenario' => ( is => 'ro', required => 1, isa => InstanceOf['Test::BDD::Cucumber::Model::Scenario'] ); has 'step' => ( is => 'ro', required => 0, isa => InstanceOf['Test::BDD::Cucumber::Model::Step'] ); =head2 verb The lower-cased verb a Step Definition was called with. =cut has 'verb' => ( is => 'ro', required => 1, isa => Str ); =head2 text The text of the step, minus the verb. Placeholders will have already been multiplied out at this point. =cut has 'text' => ( is => 'ro', required => 1, isa => Str, default => '' ); =head2 harness The L harness being used by the executor. =cut has 'harness' => ( is => 'ro', required => 1, isa => InstanceOf['Test::BDD::Cucumber::Harness'] ); =head2 executor Weak reference to the L being used - this allows for step redispatch. =cut has 'executor' => ( is => 'ro', required => 1, isa => InstanceOf['Test::BDD::Cucumber::Executor'], weak_ref => 1 ); =head2 matches Any matches caught by the Step Definition's regex. These are also available as C<$1>, C<$2> etc as appropriate. =cut has '_matches' => ( is => 'rw', isa => ArrayRef, init_arg => 'matches', default => sub { [] } ); has 'transformers' => ( is => 'ro', isa => ArrayRef, predicate => 'has_transformers', ); has '_transformed_matches' => ( is => 'ro', isa => ArrayRef, lazy => 1, builder => '_build_transformed_matches', clearer => '_clear_transformed_matches', ); has '_transformed_data' => ( is => 'ro', isa => Str|ArrayRef, lazy => 1, builder => '_build_transformed_data', clearer => '_clear_transformed_data', ); =head2 is_hook The harness processing the output can decide whether to shop information for this step which is actually an internal hook, i.e. a Before or After step =cut has 'is_hook' => ( is => 'ro', isa => Bool, lazy => 1, builder => '_build_is_hook' ); =head2 parent If a step redispatches to another step, the child step will have a link back to its parent step here; otherwise undef. See L. =cut has 'parent' => ( is => 'ro', isa => InstanceOf['Test::BDD::Cucumber::StepContext'] ); =head1 METHODS =head2 background Boolean for "is this step being run as part of the background section?". Currently implemented by asking the linked Scenario object... =cut sub background { my $self = shift; return $self->scenario->background } =head2 data See the C<_data> attribute above. Calling this method will return either the """ string, or a possibly Transform-ed set of table data. =cut sub data { my $self = shift; if (@_) { $self->_data(@_); $self->_clear_transformed_data; return; } return $self->_transformed_data; } =head2 matches See the C<_matches> attribute above. Call this method will return the possibly Transform-ed matches . =cut sub matches { my $self = shift; if (@_) { $self->_matches(@_); $self->_clear_transformed_matches; return; } return $self->_transformed_matches; } =head2 transform Used internally to transform data and placeholders, but it can also be called from within your Given/When/Then code. =cut sub transform { my $self = shift; my $value = shift; defined $value or return $value; TRANSFORM: for my $transformer ( @{ $self->transformers } ) { # turn off this warning so undef can be set in the following regex no warnings 'uninitialized'; # uses the same magic as other steps # and puts any matches into $1, $2, etc. # and calls the Transform step # also, if the transformer code ref returns undef, this will be coerced # into an empty string, so need to mark it as something else # and then turn it into proper undef if ( $value =~ s/$transformer->[0]/ my $value = $transformer->[2]->( $self ); defined $value ? $value : '__UNDEF__' /e ) { # if we matched then stop processing this match return $value eq '__UNDEF__' ? undef : $value; } } # if we're here, the value will be returned unchanged return $value; } =head1 Redispatching Sometimes you want to call one step from another step. You can do this via the I, using the C method. For example: Given qr/I have entered (\d+)/, sub { C->dispatch( 'Given', "I have pressed $1"); C_>dispatch( 'Given', "I have passed-in data", C->data ); C->dispatch( 'Given', "I have pressed enter", { some => 'data' } ); }; You redispatch step will have its own, new step context with almost everything copied from the parent step context. However, specifically not copied are: C, C, the C object, and of course the C and the C. If you want to pass data to your child step, you should IDEALLY do it via the text of the step itself, or failing that, through the scenario-level stash. Otherwise it'd make more sense just to be calling some subroutine... But you B pass in a third argument - a hashref which will be used as C. The data in that third argument can be one of: =over =item * a string This scenario corresponds with having a C<""" ... """> string argument to the step. It's passed to the child step verbatim. =item * a hash reference (deprecated) This scenario corresponds with the third example above and has been supported historically. There is no good reason to use this type of argument passing, because there is no way for a feature to pass data to the step. When you need to use this scenario, please consider implementing a separate subroutine instead. =item * a reference to an array of hashes This scenario corresponsds with a data table argument to the step. The names of the columns are taken from the first hash in the array (the first row in the data table). No transformations are applied to the table passed in to prevent duplicate transformations being applied. =back The value of the third argument will be used as the C<< C->data >> value for the C of the child step. All values passed in, will be passed to the child without applying C declarations. That way, double transformation is prevented. If the step you dispatch to doesn't pass for any reason (can't be found, dies, fails, whatever), it'll throw an exception. This will get caught by the parent step, which will then fail, and show debugging output. B =head2 dispatch C->dispatch( 'Then', "the page has loaded successfully"); See the paragraphs immediately above this =cut sub dispatch { my ( $self, $verb, $text, $data ) = @_; my $step = Test::BDD::Cucumber::Model::Step->new( { text => $text, verb => $verb, line => Test::BDD::Cucumber::Model::Line->new( { number => $self->step->line->number, raw_content => "[Redispatched step: $verb $text]", document => $self->step->line->document, } ), } ); my $columns; if ($data) { if ( ref $data eq 'HASH' ) { $columns = [ sort keys %$data ]; } elsif ( ref $data eq 'ARRAY' and (scalar @{ $data } > 0) and ref $data->[0] eq 'HASH' ) { $columns = [ sort keys %{ $data->[0] } ]; } } my $new_context = $self->new( { executor => $self->executor, ( $data ? ( data => $data ) : () ), ( $data ? ( _transformed_data => $data ) : () ), ( $columns ? ( columns => $columns ) : () ), stash => { feature => $self->stash->{'feature'}, scenario => $self->stash->{'scenario'}, step => {}, }, feature => $self->feature, scenario => $self->scenario, harness => $self->harness, transformers => $self->transformers, step => $step, verb => lc($verb), text => $text, } ); my $result = $self->executor->find_and_dispatch( $new_context, 0, 1 ); # If it didn't pass, short-circuit the rest unless ( $result->result eq 'passing' ) { my $error = "Redispatched step didn't pass:\n"; $error .= "\tStatus: " . $result->result . "\n"; $error .= "\tOutput: " . $result->output . "\n"; $error .= "Failure to redispatch a step causes the parent to fail\n"; die $error; } return $result; } # the builder for the is_hook attribute sub _build_is_hook { my $self = shift; return ( $self->verb eq 'before' or $self->verb eq 'after' ) ? 1 : 0; } # the builder for the _transformed_matches attribute sub _build_transformed_matches { my $self = shift; my @transformed_matches = @{ $self->_matches }; # this stops it recursing forever... # and only Transform if there are any to process if ( $self->verb ne 'transform' and $self->has_transformers ) { @transformed_matches = map { my $match = $_; $match = $self->transform($match); } @transformed_matches; } return \@transformed_matches; } # the builder for the _transformed_data attribute sub _build_transformed_data { my $self = shift; my $transformed_data = $self->_data; # again stop recursing # only transform table data # and only Transform if there are any to process if ( $self->verb ne 'transform' and ref $transformed_data and $self->has_transformers ) { # build the string that a Transform is looking for # table:column1,column2,column3 my $table_text = 'table:' . join( ',', @{ $self->columns } ); if ( my $transformer = first { $table_text =~ $_->[0] } @{ $self->transformers } ) { # call the Transform step $transformer->[2]->( $self, $transformed_data ); } } return $transformed_data; } =head1 AUTHOR Peter Sergeant C =head1 LICENSE Copyright 2019-2023, Erik Huelsmann Copyright 2011-2019, Peter Sergeant; Licensed under the same terms as Perl =cut 1;