package Test::WWW::Declare; use warnings; use strict; use base 'Test::More'; use Test::WWW::Mechanize; use Test::Builder; our $VERSION = '0.02'; our @EXPORT = qw(flow run get session check mech match follow_link content should shouldnt click href button fill form SKIP _twd_dummy title equal caselessly contain matches equals contains never always lack lacks url uri); our $BUILDER = Test::Builder->new(); our $WWW_MECHANIZE; our $IN_FLOW; our %mechs; =begin private =head2 import_extra Called by L's C code when L is first C'd, it asks Test::More to export its symbols to the namespace that C'd this one. =end private =cut sub import_extra { Test::More->export_to_level(2); } =head1 NAME Test::WWW::Declare - declarative testing for your web app =head1 SYNOPSIS use Test::WWW::Declare tests => 3; use Your::Web::App::Test; Your::Web::App::Test->start_server; session 'testuser' => run { flow 'log in and out' => check { flow 'log in' => check { get 'http://localhost/'; fill form 'login' => { username => 'testuser', password => 'drowssap', }; content should contain 'log out'; }; flow 'log out' => check { get 'http://localhost/'; click href 'log out'; }; }; }; =head1 DESCRIPTION Often in web apps, tests are very dependent on the state set up by previous tests. If one test fails (e.g. "follow the link to the admin page") then it's likely there will be many more failures. This module aims to alleviate this problem, as well as provide a nicer interface to L. The central idea is that of "flow". Each flow is a sequence of commands ("fill in this form") and assertions ("content should contain 'testuser'"). If any of these commands or assertions fail then the flow is aborted. Only that one failure is reported to the test harness and user. Flows may also contain other flows. If an inner flow fails, then the outer flow fails as well. =head1 FLOWS AND SESSIONS =head2 session NAME => run { CODE } Sessions are a way of associating a set of flows with a L instance. A session is mostly equivalent with a user interacting with your web app. Within a session, every command (C, C, etc) is operating on that session's L instance. You may have multiple sessions in one test file. Two sessions with the same name are in fact the same session. This lets you write code like the following, simplified slightly: session 'first user' => run { get "$URL/give?task=1&victim=other"; session 'other user' => run { get "$URL/tasks"; content should match qr/task 1/; # this is the same session/mech as the outermost 'first user' session 'first user' => run { get "$URL/tasks"; content shouldnt match qr/task 1/; }; }; }; =head2 flow NAME => check { CODE } A flow encompasses a single test. As described above, each flow is a sequence of commands, assertions, and other flows. If any of the components of a flow fail, the rest of the flow is aborted and one or more test failures are reported to the test harness. =head1 COMMANDS =head2 get URL =head2 click button =head2 click href =head2 follow_link =head2 fill form NAME => {FIELD1 => VALUE1, FIELD2 => VALUE2} =head1 ASSERTIONS Every assertion has two parts: a subject and a verb. =head2 SUBJECTS =head3 content =head3 title =head3 url =head2 VERBS =head3 should(nt) (caselessly) match REGEX =head3 should(nt) (caselessly) contain STRING =head3 should(nt) (caselessly) lack STRING =head3 should(nt) (caselessly) equal STRING =cut # DSLey functions sub to($) { return $_[0] } sub _args { my $args = shift; return $args if ref($args) eq 'HASH'; return {expected => $args}; } sub should ($) { return _args(shift); } sub shouldnt ($) { my $args = _args(shift); $args->{negative} = 1; return $args; } sub match ($) { my $args = _args(shift); $args->{match} = 'regex'; return $args; } sub equal ($) { my $args = _args(shift); $args->{match} = 'equality'; return $args; } sub contain ($) { my $args = _args(shift); $args->{match} = 'index'; return $args; } sub lack ($) { my $args = _args(shift); $args->{match} = 'index'; $args->{negative} = 1; return $args; } sub caselessly ($) { my $args = _args(shift); $args->{case_insensitive} = 1; return $args; } sub check (&) { my $coderef = shift; return $coderef; } sub run (&) { my $coderef = shift; return $coderef; } # alternates (e.g. "foo matches bar" instead of "foo should match bar") sub contains ($) { contain $_[0] } sub equals ($) { equal $_[0] } sub matches ($) { match $_[0] } sub lacks ($) { lack $_[0] } sub always ($) { should $_[0] } sub never ($) { shouldnt $_[0] } # Mech interactions sub mech(;$) { my $name = shift; return defined $name ? $mechs{$name} : $WWW_MECHANIZE; } sub get { my $url = shift; mech()->get($url); if (!$IN_FLOW) { $BUILDER->ok(mech->success, "navigated to $url"); } return if mech->success; Carp::croak mech->status . (mech->response ? ' - ' . mech->response->message : '') } sub href ($) { return (shift, 'href'); } sub button ($) { return (shift, 'button'); } sub click { my $link = shift; my $type = shift; if ($type eq 'button') { my $ok = mech()->click_button(value => $link); $ok = $ok->is_success if $ok; my $verb = ref($link) eq 'Regexp' ? "matching " : ""; $BUILDER->ok($ok, "Clicked button $verb$link") if !$IN_FLOW; return $ok; } else { if (ref $link ne 'Regexp') { Carp::croak "click doesn't know what to do with a link type of " . ref($link); } my $ok; my $response = mech()->follow_link(text_regex => $link); $ok = 1 if $response && $response->is_success; $BUILDER->ok($ok, "Clicked link matching $link") if !$IN_FLOW; Carp::croak($response ? $response->as_string : "No link matching $link found") if !$ok; return $ok; } } sub follow_link { my $ret = mech()->follow_link(@_); if (!$ret) { Carp::croak "follow_link couldn't find a link matching " . "(" . join(', ', @_) . ")"; } } sub content ($) { _magic_match({got => mech()->content, name => "Content", %{shift @_}}); } sub title ($) { my $title = mech()->title; _magic_match({got => $title, name => "Title '$title'", %{shift @_}}); } sub url ($) { my $url = mech()->uri; _magic_match({got => $url, name => "URL '$url'", %{shift @_}}); } *uri = \&url; # yes, there's a little too much logic in here. that's why it's magic sub _magic_match { my $orig = shift @_; my %args = %$orig; my $match; my @output; $args{negative} ||= 0; push @output, $args{name}; push @output, $args{negative} ? () : "does not"; if ($args{match} eq 'equality') { if ($args{case_insensitive}) { push @output, "caselessly"; $args{got} = lc $args{got}; $args{expected} = lc $args{expected}; } push @output, $args{negative} ? "equals" : "equal"; push @output, $orig->{expected}; $match = $args{got} eq $args{expected}; } elsif ($args{match} eq 'index') { if ($args{case_insensitive}) { push @output, "caselessly"; $args{got} = lc $args{got}; $args{expected} = lc $args{expected}; } push @output, $args{negative} ? "contains" : "contain"; push @output, $orig->{expected}; $match = index($args{got}, $args{expected}) >= 0; } elsif ($args{match} eq 'regex') { if ($args{case_insensitive}) { push @output, "caselessly"; push @output, $args{expected}; $args{expected} = "(?i:$args{expected})"; } push @output, $args{negative} ? "matches" : "match"; push @output, $orig->{expected}; $match = $args{got} =~ $args{expected}; } else { Carp::croak "No \$args{match} (yes this error needs to be fixed)"; } my $ok = ($match ? 1 : 0) ^ $args{negative}; if (!$IN_FLOW) { $BUILDER->ok($ok, join(' ', @output)); return $ok; } return 1 if $ok; Carp::croak join(' ', @output); } sub form ($$) { my $form_name = shift; my $data = shift; my $form = mech()->form_name($form_name); if (!defined($form)) { Carp::croak "There is no form named '$form_name'"; } return $data; } sub fill { my $data = shift; Carp::croak "fill expects a hashref" if ref($data) ne 'HASH'; mech()->set_fields(%{$data}); } # the meat of the module sub SKIP ($) { my $reason = shift; Carp::croak "SKIP: $reason"; } sub flow ($$) { my $name = shift; my $coderef = shift; eval { local $IN_FLOW = 1; $coderef->() }; if ($@ =~ /^SKIP: (.*)$/) { my $reason = $1; $BUILDER->skip($reason); } elsif ($@) { if ($IN_FLOW) { if ($@ =~ /^Flow '/) { die $@; } die "Flow '$name' failed: $@"; } $BUILDER->ok(0, $name); if ($@ =~ /^Flow '/) { $BUILDER->diag($@); } else { $BUILDER->diag("Flow '$name' failed: $@"); } } else { $BUILDER->ok(1, $name); } } sub session ($$) { my $title = shift; my $coderef = shift; $mechs{$title} ||= Test::WWW::Mechanize->new(quiet => 1); local $WWW_MECHANIZE = $mechs{$title}; $coderef->(); if ($@ =~ /^SKIP: (.*)$/) { my $reason = $1; $BUILDER->skip($reason); } elsif ($@ =~ /^Flow '/) { # flow already displayed the error } elsif ($@) { $BUILDER->diag($@); } } sub dump($) { my $file = shift; mech->save_content($file); } # used only for testing that we got T:W:D's goods sub _twd_dummy { "XYZZY" } =head1 SUBCLASSING One of the goals of this module is to let you subclass it to provide extra features, such as automatically logging in a user each time a session is created. =head1 CAVEATS If you fail any tests, then the actual number of tests run may be fewer than you have in your file. This is because when a flow fails, it immediately aborts the rest of its body (which may include other flows). So if you're setting the number of tests based on how many ran, make sure that all tests passed. =head1 BUGS Hopefully few. We'd like to know about any of them. Please report them to C. =head1 SEE ALSO L, L. =head1 MAINTAINER Shawn M Moore C<< >> =head1 ORIGINAL AUTHOR Jesse Vincent C<< >> =head1 COPYRIGHT Copyright 2007-2008 Best Practical Solutions, LLC This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself. =cut 1;