Test::Subs - Test your modules with a lightweight syntax based on anonymous block


  use SomeModule;
  use Test::Subs;
  test { 1 == 2 };


This module provide a very lightweight syntax to run Test::Harness or Tap::Harness compliant test on your code.

As opposed to other similar packages, the two main functionnalities of Test::Subs are that the tests are anonymous code block (rather than list of values), which are (subjectively) cleaner and easier to read, and that you do not need to pre-declare the number of tests that are going to be run (so all modifications in a test file are local).

Using this module is just a matter of use Test::Subs followed by the declaration of your tests with the functions described below. All tests are then run at the end of the execution of your test file.

As a protection against some error, if the compilation phase fail, the output of the test file will be one failed pseudo-test.


This is a list of the public function of this library. Functions not listed here are for internal use only by this module and should not be used in any external code unless .

All the functions described below are automatically exported into your package except if you explicitely request to opposite with use Test::Subs ();.

Finally, these function must all be called from the top-level and not inside of the code of another test function. That is because the library must know the number of test before their execution.


  test { CODE };
  test { CODE } DESCR;

This function register a code-block containing a test. During the execution of the test, the code will be run and the test will be deemed successful if the returned value is true.

The optionnal DESCR is a string (or an expression returning a string) which will be added as a comment to the result of this test. If this string contains a printf conversion (e.g. %s or %d) it will be replaced by the result of the code block. If the description is omitted, it will be replaced by the filename and line number of the test. You can use an empty string '' to deactivate completely the output of a comment to the test.


  todo { CODE };
  todo { CODE } DESCR;

This function is the same as the function test, except that the test will be registered as to-do. So a failure of this test will be ignored when your test is run inside a test plan by Test::Harness or Tap::Harness.


  match { CODE } REGEXP;
  match { CODE } REGEXP, DESCR;

This function declares a test which will succeed if the result of the code block match the given regular expression.

The regexp may be given as a scalar string or as a qr encoded regexp.


  not_ok { CODE };
  not_ok { CODE } DESCR;

This function is exactly the opposite of the test one. The test that it declares will succeed if the code block return a false value.


  fail { CODE };
  fail { CODE } DESCR;

This function declares a test that will succeed if its code block die (raise any exception).


  failwith { CODE } REGEXP;
  failwith { CODE } REGEXP, DESCR;

As for the fail function, this function declares a test which expects that its code block die. Except that the test will succeed only if the raised exception (the content of the $@ variable) match the given regular expression.

The regexp may be given as a scalar string or as a qr encoded regexp.


  comment { CODE };

This function evaluate its code and display the resulting value on the standard error handle. The buffering on STDOUT and STDERR is deactivated when Test::Subs is used and the output of this function should appear in between the result of the test when the test file is run stand-alone.

This function must be used outside of the code of the other functions described above. To output comment to STDERR inside a test, just use the print or printf function. The default output has been select-ed to STDERR so the result of the test will not be altered.

skip (new in 0.07)

  skip 'reason' unless eval 'use Foo::Bar';

This function allows to skip a test file. It must be used outside of test subs of the other functions. You will typically use it to disable a test file if the current version of Perl is missing some required functionnalities for the tests.

The argument for the function is a string explaining the reason why the tests have been skipped. This reasion will be reported in the output of a Test::Harness run.

test_pod (new in 0.04)


This function takes a list of module name and registers one test for each given module. The test will run the module file through Pod::Checker and fail if there is errors in the POD of the file. Moreover, in debug mode, all errors and warnings are printed to STDERR.


  debug { CODE } DESCR;

This function register and executes a dummy test: the CODE is executed and error messages (if any) are written on STDERR. The test will succeed under the same condition as with the test function.

Usefull when a test fail to quickly see what is going on.


Debug mode (new in 0.03)

You can pass a debug argument to the package when you are using it:

  use Test::Subs debug => 1;

If the value supplied to this option is true then all call to the test functions will behave like calls to the debug function. Also, most of the function of this library will give more output (on STDERR) if their test fails.

Path to the library files (new in 0.05)

By default, if you specify a 'My::Module' module as a target of the test_pod function, the file for this module will be searched in lib/My/ relatively to the current working directory. This should work for standard distribution. Yau can modify this behaviour with the lib option as argument to the package when you are using it:

  use Test::Subs lib => '../lib';

The supplied path will serve as the base directory to look for the module file (e.g. My/, relatively to the the test script directory (and not to the current working directory as in the default case).

Warning level for POD Checking (new in 0.05)

You can tune the number of warning generated by the test_pod function using a pod_warn argument to the package when you are using it:

  use Test::Subs pod_warn => 0;

This option expects an integer value. A value of '0' will deactivates all warnings, a value of '1' will activates most warnings and a value of '2' will activates some additionnals warnings. More details on the available warnings can be found in the POD::Checker documentation.

Note that, in any case, the warnings will only be printed in debug mode.


Here is an example of a small test file using this module.

  use strict;
  use warnings;
  use Test::Subs debug => 1, lib => '../lib';
  use My::Module;
  test { My::Module::init() } 'This is the first test';
  todo { My::Module::make_coffee() };
  not_ok { 0 };
  fail { die "fail" };
  test_pod('My::Module', 'My::Module::Internal');

Run through Test::Harness this file will pass, with only the second test failing (but marked todo so that's OK).


This package does not use the Test::Builder facility and as such is not compatible with other testing modules are using Test::Builder. This may be changed in a future release.

The standard set by Test::Harness is that all output to STDOUT is interpreted by the test parser. So a test file should write additional output only to STDERR. This is what will be done by the comment fonction. To help with this, during the execution of your test file, the STDERR file-handle will be select-ed. So any un-qualified print or printf call will end in STDERR.

This package use source filtering (with Filter::Simple). The filter applied is very simple, but there is a slight possibility that it is incompatible with other source filters. If so, do not hesitate to report this as a bug.


Please report any bugs or feature requests to, or through the web interface at


Test, Test::Tiny, Test::Lite, Test::Simple


Mathias Kende (


Version 0.08 (March 2013)


Copyright 2013 © Mathias Kende. All rights reserved.

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