Test - Test support module for perl6
use v6-alpha; require Test;
plan 10; force_todo(1, 3 .. 5, 9);
use_ok('Some::Module');
use_ok('Some::Other::Module', todo => 1);
ok(2 + 2 == 4, '2 and 2 make 4'); is(2 + 2, 4, '2 and 2 make 4'); isa_ok([1, 2, 3], 'List');
ok(2 + 2 == 5, '2 and 2 make 5', :todo(1));
is(2 + 2, 5, desc => '2 and 2 make 5', todo => 1);
isa_ok({'one' => 1}, 'Hash', :todo(1));
use_ok('My::Module');
pass('This test passed');
flunk('This test failed');
skip('skip this test for now');
flunk('this fails, but might work soon', :todo(1));
diag('some misc comments and documentation');
# TODO the next test with respect to the "deadline" specified. todo :pugs<6.2.13>, :foo<1.23>; is foo, bar, '...';
This module was built to facilitate the Pugs test suite. It has the distinction of being the very first module written for Pugs.
It provides a simple set of common test utility functions, and is an implementation of the TAP protocol.
This module, like Pugs, is a work in progress. As new features are added to Pugs, new test functions will be defined to facilitate the testing of those features. For more information see the FUTURE PLANS section of this document.
plan (Int $number_of_tests) returns Void
All tests need a plan. A plan is simply the number of tests which are expected to run. This should be specified at the very top of your tests.
force_todo (*@todo_tests) returns Void
If you have some tests which you would like to force into being TODO tests then you can pass them through this function. This is primarily a release tool, but can be useful in other contexts as well.
use_ok (Str $module, Bool :$todo, Str :$depends) returns Bool
ok (Bool $cond, Str $desc?, Bool :$todo, Str :$depends) returns Bool
is (Str $got, Str $expected, Str $desc?, Bool :$todo, Str :$depends) returns Bool isnt (Str $got, Str $expected, Str $desc?, Bool :$todo, Str :$depends) returns Bool
like (Str $got, Rule $expected, Str $desc?, Bool :$todo, Str :$depends) returns Bool is export unlike (Str $got, Rule $expected, Str $desc?, Bool :$todo, Str :$depends) returns Bool is export
These functions should work with most reg-exps, but given that they are still a somewhat experimental feature in Pugs, it is suggested you don't try anything too funky.
cmp_ok (Str $got, Code &compare_func, Str $expected, Str $desc?, Bool :$todo, Str :$depends) returns Bool
This function will compare $got and $expected using &compare_func. This will
eventually allow Test::More-style cmp_ok() though the following syntax:
cmp_ok('test', &infix:<gt>, 'me', '... testing gt on two strings');
However the &infix:<gt> is currently not implemented, so you will have to wait
a little while. Until then, you can just write your own functions like this:
cmp_ok('test', sub ($a, $b) { ?($a gt $b) }, 'me', '... testing gt on two strings');
isa_ok ($ref, Str $expected_type, Str $desc?, Bool :$todo, Str :$depends) returns Bool
This function currently on checks with WHAT() since we do not yet have
object support. Once object support is created, we will add it here, and
maintain backwards compatibility as well.
eval_ok (Str $code, Str $desc?, Bool :$todo, Str :$depends) returns Bool eval_is (Str $code, Str $expected, Str $desc?, Bool :$todo, Str :$depends) returns Bool eval_dies_ok (Str $code, Str $desc?, Bool :$todo, Str :$depends) returns Bool
These are the functions to use if you have a piece of code that would otherwise
failed to be parsed. If the code parses, but may die at run time, consider
using dies_ok or lives_ok, which have lower overhead.
They eval a string, and then pass the result to is or ok
on success, or report that the eval was not successful on failure. In the case of
eval_dies_ok, unsuccessful or failure was expected.
throws_ok (Code &code, Any $expected, Str $desc?, Bool :$todo, Str :$depends) returns Bool
This function takes a block of code and runs it. It then smart-matches (~~) any $!
value with the $expected value.
dies_ok (Code &code, Str $desc?, Bool :$todo, Str :$depends) returns Bool lives_ok (Code &code, Str $desc?, Bool :$todo, Str :$depends) returns Bool
These functions both take blocks of code, and test whether they live or die using try
The code must at least be parsable. If the code might not parse, see eval_ok above.
is_deeply(Any $got, Any $expected, Str $desc?, :$todo, :$depends) returns Bool
Similar to is(), except that if $this and $that are references, it does a deep comparison walking each data structure to see if they are equivalent.
todo (*%deadline) returns Bool is export
If and only if the deadline has been hit (or passed), the next one test will be marked as TODO.
For example:
todo :pugs<6.28.0>; is($got, $expected, $desc);
The call to the todo function will mark the next one test as TODO if
and only if the current compiler's $?COMPILER holds a string whose lowercase
version equals to 'pugs' and $?VERSION holds a value less than '6.28.0'.
The todo functuion will perform partial ordering comparison between version
numbers.
More implementation-specific deadlines can be appended to a single todo call:
todo :pugs<6.28.0>, :p6p5<0.011>, :parrot<0.45>;
skip (Str $reason?) returns Bool skip (Int $count, Str $reason?) returns Bool
If for some reason a test is to be skipped, you can use this function to do so.
pass (Str $desc?) returns Bool
Sometimes what you need to test does not fit into one of the standard
testing functions. In that case, you can use the rather blunt pass()
functions and its compliment the flunk() function.
flunk (Str $desc?, Bool :$todo) returns Bool
This is the opposite of pass()
diag (Str $diag)
This will print each string with a '#' character appended to it, this is ignored by the TAP protocol.
Sometimes a test is broken because something is not implemented yet. So
in order to still allow that to be tested, and those tests to knowingly
fail, we provide the todo function to TODO the next one test according
to a given deadline. (See below.)
Take the Pugs implementation. When TODOing failing tests before the Pugs release (say, 6.2.12), the following form of todo should be used:
todo :pugs<6.2.13>; # version of the next point release
By doing this, temporarily TODO'd tests can get unTODO'd automatically once the the release is done (and the version number gets updated).
The version number fed to todo is optional. If omitted,
the corresponding tests won't get expired unless we unTODO them manually.
It is also possible to use the force_todo() function to do large scale
TODO-ing of tests.
The :depends("string") parameter to most of the functions is a way
to provide a comment that refers to another file or test which must be
made to pass before this test can pass (or before an implementation
could be started). This is most useful when writing modules and you
find there is some language feature missing, or core bug that needs to
be sorted out before you can continue.
This module is still a work in progress. As Pugs grows, so will it's testing needs. This module will be the code support for those needs. The following is a list of future features planned for this module.
- better error handling for cmp_ok
The error handling capabilities need to be expanded more to handle the
error reporting needs of the cmp_ok() function.
Setting the environment variable TEST_ALWAYS_CALLER to force Test.pm to always
append the caller information to the test's $desc.
The Perl 5 Test modules
- Test
- Test::More
Information about the TAP protocol can be found in the Test::Harness distribution.
Audrey Tang <autrijus@autrijus.org>
Benjamin Smith
Norman Nunley
Steve Peters
Stevan Little <stevan@iinteractive.com>
Brian Ingerson <ingy@cpan.org>
Jesse Vincent <jesse@bestpractical.com>
Yuval Kogman <nothingmuch@woobling.org>
Darren Duncan <perl@DarrenDuncan.net>
Nathan Gray <kolibrie@graystudios.org>
Max Maischein <corion@cpan.org>
Ingo Blechschmidt <iblech@web.de>
Gaal Yahas <gaal@forum2.org>
Mark Stosberg
Simon Sun <dolmens@gmail.com>
Copyright (c) 2005, 2006. Audrey Tang. All rights reserved.
This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself.