Quick Navigator

Search Site

Unix VPS
A - Starter
B - Basic
C - Preferred
D - Commercial
MPS - Dedicated
Previous VPSs
* Sign Up! *

Contact Us
Online Help
Domain Status
Man Pages

Virtual Servers

Topology Map

Server Agreement
Year 2038

USA Flag



Man Pages

Manual Reference Pages  -  TEST::SPEC (3)

.ds Aq ’


Test::Spec - Write tests in a declarative specification style



  use Test::Spec; # automatically turns on strict and warnings

  describe "A date" => sub {

    my $date;

    describe "in a leap year" => sub {

      before each => sub {
        $date = DateTime->new(year => 2000, month => 2, day => 28);

      it "should know that it is in a leap year" => sub {

      it "should recognize Feb. 29" => sub {
        is($date->add(days => 1)->day, 29);


    describe "not in a leap year" => sub {
      before each => sub {
        $date = DateTime->new(year => 2001, month => 2, day => 28);

      it "should know that it is NOT in a leap year" => sub {

      it "should NOT recognize Feb. 29" => sub {
        is($date->add(days => 1)->day, 1);


  runtests unless caller;

  # Generates the following output:
  # ok 1 - A date in a leap year should know that it is in a leap year
  # ok 2 - A date in a leap year should recognize Feb. 29
  # ok 3 - A date not in a leap year should know that it is NOT in a leap year
  # ok 4 - A date not in a leap year should NOT recognize Feb. 29
  # 1..4


This is a declarative specification-style testing system for behavior-driven development (BDD) in Perl. The tests (a.k.a. examples) are named with strings instead of subroutine names, so your fingers will suffer less fatigue from underscore-itis, with the side benefit that the test reports are more legible.

This module is inspired by and borrows heavily from RSpec (, a BDD tool for the Ruby programming language.


When given <B>no listB> (i.e. use Test::Spec;), this class will export:
o Spec definition functions

These are the functions you will use to define behaviors and run your specs: describe, it, they, before, after, runtests, share, shared_examples_for, it_should_behave_like, and spec_helper.

o The stub/mock functions in Test::Spec::Mocks.
o Everything that Test::More normally exports

This includes ok, is and friends. You’ll use these to assert correct behavior.

o Everything that Test::Deep normally exports

More assertions including cmp_deeply.

o Everything that Test::Trap normally exports

The trap() function, which let you test behaviors that call exit() and other hard things like that. A block eval on steroids.

If you specify an import list, only functions directly from Test::Spec (those documented below) are available.


runtests(@patterns) Runs all the examples whose descriptions match one of the (non case-sensitive) regular expressions in @patterns. If @patterns is not provided, runs all examples. The environment variable SPEC will be used as a default pattern if present.

If called as a function (i.e. not a method call with ->), runtests will autodetect the package from which it is called and run that package’s examples. A useful idiom is:

  runtests unless caller;

which will run the examples when the file is loaded as a script (for example, by running it from the command line), but not when it is loaded as a module (with require or use).

describe CODE Defines a specification context under which examples and more descriptions can be defined. All examples must come inside a describe block.
describe blocks can be nested to DRY up your specs. For large specifications, describe blocks can save you a lot of duplication:

  describe "A User object" => sub {
    my $user;
    before sub {
      $user = User->new;
    describe "from a web form" => sub {
      before sub {
        $user->init_from_tree({ username => "bbill", ... });
      it "should read its attributes from the form";
      describe "when saving" => sub {
        it "should require a unique username";
        it "should require a password";

The setup work done in each before block cascades from one level to the next, so you don’t have to make a call to some initialization function manually in each test. It’s done automatically based on context.

Using describe blocks improves legibility without requiring more typing. The name of the context will be included by default in the success/failure report generated by Test::Builder-based testing methods (e.g. Test::More’s ok() function). For an example like this:

  describe "An unladen swallow" => sub {
    it "has an airspeed of 11 meters per second" => sub {
      is($swallow->airspeed, "11m/s");

The output generated is:

  ok 1 - An unladen swallow has an airspeed of 11 meters per second

Contrast this to the following test case to generate the same output:

  sub unladen_swallow_airspeed : Test {
    is($swallow->airspeed, "11m/s",
       "An unladen swallow has an airspeed of 11 meters per second");

describe blocks execute in the order in which they are defined. Multiple describe blocks with the same name are allowed. They do not replace each other, rather subsequent describes extend the existing one of the same name.

context An alias for describe().
xdescribe Specification contexts may be disabled by calling xdescribe instead of describe(). All examples inside an xdescribe are reported as # TODO (disabled), which prevents Test::Harness/prove from counting them as failures.
xcontext An alias for xdescribe().
it TODO_SPECIFICATION Defines an example to be tested. Despite its awkward name, it allows a natural (in my opinion) way to describe expected behavior:

  describe "A captive of Buffalo Bill" => sub {
    it "puts the lotion on its skin" => sub {
    it "puts the lotion in the basket"; # TODO

If a code reference is not passed, the specification is assumed to be unimplemented and will be reported as TODO (unimplemented) in the test results (see todo_skip in Test::Builder. TODO tests report as skipped, not failed.

they CODE
they TODO_SPECIFICATION An alias for it. This is useful for describing behavior for groups of items, so the verb agrees with the noun:

  describe "Captives of Buffalo Bill" => sub {
    they "put the lotion on their skin" => sub {
    they "put the lotion in the basket"; # TODO

xit/xthey Examples may be disabled by calling xit()/xthey() instead of it()/they(). These examples are reported as # TODO (disabled), which prevents Test::Harness/prove from counting them as failures.
before each => CODE
before all => CODE
before CODE Defines code to be run before tests in the current describe block are run. If each is specified, CODE will be re-executed for every test in the context. If all is specified, CODE will only be executed before the first test.

The default is each, due to this logic presented in RSpec’s documentation:

"It is very tempting to use before(:all) and after(:all) for situations in which it is not appropriate. before(:all) shares some (not all) state across multiple examples. This means that the examples become bound together, which is an absolute no-no in testing. You should really only ever use before(:all) to set up things that are global collaborators but not the things that you are describing in the examples.

The most common cases of abuse are database access and/or fixture setup. Every example that accesses the database should start with a clean slate, otherwise the examples become brittle and start to lose their value with false negatives and, worse, false positives."


There is no restriction on having multiple before blocks. They will run in sequence within their respective each or all groups. before "all" blocks run before before "each" blocks.

after each => CODE
after all => CODE
after CODE Like before, but backwards. Runs CODE after each or all tests, respectively. The default is each.

after "all" blocks run after after "each" blocks.

shared_examples_for DESCRIPTION => CODE Defines a group of examples that can later be included in describe blocks or other shared_examples_for blocks. See Shared example groups.

Example group names are <B>globalB>, but example groups can be defined at any level (i.e. they can be defined in the global context, or inside a describe block).

  my $browser;
  shared_examples_for "all browsers" => sub {
    it "should open a URL" => sub { ok($browser->open("")) };
  describe "Firefox" => sub {
    before all => sub { $browser = Firefox->new };
    it_should_behave_like "all browsers";
    it "should have firefox features";
  describe "Safari" => sub {
    before all => sub { $browser = Safari->new };
    it_should_behave_like "all browsers";
    it "should have safari features";

it_should_behave_like DESCRIPTION Asserts that the thing currently being tested passes all the tests in the example group identified by DESCRIPTION (having previously been defined with a shared_examples_for block). In essence, this is like copying all the tests from the named shared_examples_for block into the current context. See Shared example groups and shared_examples_for.
share %HASH Registers %HASH for sharing data between tests and example groups. This lets you share variables with code in different lexical scopes without resorting to using package (i.e. global) variables or jumping through other hoops to circumvent scope problems.

Every hash that is shared refers to the <B>same dataB>. Sharing a hash will make its existing contents inaccessible, because afterwards it contains the same data that all other shared hashes contain. The result is that you get a hash with global semantics but with lexical scope (assuming %HASH is a lexical variable).

There are a few benefits of using share over using a regular global hash. First, you don’t have to decide what package the hash will belong to, which is annoying when you have specs in several packages referencing the same shared examples. You also don’t have to clutter your examples with colons for fully-qualified names. For example, at my company our specs go in the ICA::TestCase hierarchy, and $ICA::TestCase::Some::Package::variable is exhausting to both the eyes and the hands. Lastly, using share allows Test::Spec to provide this functionality without deciding on the variable name for you (and thereby potentially clobbering one of your variables).

  share %vars;      # %vars now refers to the global share
  share my %vars;   # declare and share %vars in one step

spec_helper FILESPEC Loads the Perl source in FILESPEC into the current spec’s package. If FILESPEC is relative (no leading slash), it is treated as relative to the spec file (i.e. <B>notB> the currently running script). This lets you keep helper scripts near the specs they are used by without exercising your File::Spec skills in your specs.

  # in foo/spec.t
  spec_helper "";          # loads foo/
  spec_helper "helpers/";  # loads foo/helpers/
  spec_helper "/path/to/"; # loads /path/to/

    Shared example groups

This feature comes straight out of RSpec, as does this documentation:

You can create shared example groups and include those groups into other groups.

Suppose you have some behavior that applies to all editions of your product, both large and small.

First, factor out the shared behavior:

  shared_examples_for "all editions" => sub {
    it "should behave like all editions" => sub {

then when you need to define the behavior for the Large and Small editions, reference the shared behavior using the it_should_behave_like() function.

  describe "SmallEdition" => sub {
    it_should_behave_like "all editions";

  describe "LargeEdition" => sub {
    it_should_behave_like "all editions";
    it "should also behave like a large edition" => sub {

it_should_behave_like will search for an example group by its description string, in this case, all editions.

Shared example groups may be included in other shared groups:

  shared_examples_for "All Employees" => sub {
    it "should be payable" => sub {

  shared_examples_for "All Managers" => sub {
    it_should_behave_like "All Employees";
    it "should be bonusable" => sub {

  describe Officer => sub {
    it_should_behave_like "All Managers";
    it "should be optionable";

  # generates:
  ok 1 - Officer should be optionable
  ok 2 - Officer should be bonusable
  ok 3 - Officer should be payable

Refactoring into files

If you want to factor specs into separate files, variable scopes can be tricky. This is especially true if you follow the recommended pattern and give each spec its own package name. Test::Spec offers a couple of functions that ease this process considerably: share and spec_helper.

Consider the browsers example from shared_examples_for. A real browser specification would be large, so putting the specs for all browsers in the same file would be a bad idea. So let’s say we create for the shared examples, and give Safari and Firefox safari.t and firefox.t, respectively.

The problem then becomes: how does the code in access the $browser variable? In the example code, $browser is a lexical variable that is in scope for all the examples. But once those examples are split into multiple files, you would have to use either package global variables or worse, come up with some other hack. This is where share and spec_helper come in.

  # safari.t
  package Testcase::Safari;
  use Test::Spec;

  describe "Safari" => sub {
    share my %vars;
    before all => sub { $vars{browser} = Safari->new };
    it_should_behave_like "all browsers";
    it "should have safari features";

  # firefox.t
  package Testcase::Firefox;
  use Test::Spec;

  describe "Firefox" => sub {
    share my %vars;
    before all => sub { $vars{browser} = Firefox->new };
    it_should_behave_like "all browsers";
    it "should have firefox features";

  # in
  shared_examples_for "all browsers" => sub {
    # doesnt have to be the same name!
    share my %t;
    it "should open a URL" => sub {
      ok $t{browser}->open("");

    Order of execution

This example, shamelessly adapted from the RSpec website, gives an overview of the order in which examples run, with particular attention to before and after.

  describe Thing => sub {
    before all => sub {
      # This is run once and only once, before all of the examples
      # and before any before("each") blocks.

    before each => sub {
      # This is run before each example.

    before sub {
      # "each" is the default, so this is the same as before("each")

    it "should do stuff" => sub {

    it "should do more stuff" => sub {

    after each => sub {
      # this is run after each example

    after sub {
      # "each" is the default, so this is the same as after("each")

    after all => sub {
      # this is run once and only once after all of the examples
      # and after any after("each") blocks



RSpec (, Test::More, Test::Deep, Test::Trap, Test::Builder.

The mocking and stubbing tools are in Test::Spec::Mocks.


Philip Garrett <>


The source code for Test::Spec lives on github:

If you want to contribute a patch, fork my repository, make your change, and send me a pull request.


If you have found a defect or have a feature request please report an issue at For help using the module, standard Perl support channels like Stack Overflow <> and comp.lang.perl.misc <> are probably your best bet.


Copyright (c) 2010-2011 by Informatics Corporation of America.

This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself.

Search for    or go to Top of page |  Section 3 |  Main Index

perl v5.20.3 TEST::SPEC (3) 2015-07-27

Powered by GSP Visit the GSP FreeBSD Man Page Interface.
Output converted with manServer 1.07.