STOP! If you're just getting started writing tests, have a look at Test::Simple first. This is a drop in replacement for Test::Simple which you can switch to once you get the hang of basic testing.
The purpose of this module is to provide a wide range of testing
utilities. Various ways to say ``ok'' with better diagnostics,
facilities to skip tests, test future features and compare complicated
data structures. While you can do almost anything with a simple
ok() function, it doesn't provide good diagnostic output.
Before anything else, you need a testing plan. This basically declares
how many tests your script is going to run to protect against premature
failure.
The preferred way to do this is to declare a plan when you use Test::More.
There are rare cases when you will not know beforehand how many tests
your script is going to run. In this case, you can declare that you
have no plan. (Try to avoid using this as it weakens your test.)
Your script will declare a skip with the reason why you skipped and
exit immediately with a zero (success). See the Test::Harness manpage for
details.
If you want to control what functions Test::More will export, you
have to use the 'import' option. For example, to import everything
but 'fail', you'd do:
By convention, each test is assigned a number in order. This is
largely done automatically for you. However, it's often very useful to
assign a name to each test. Which would you rather see:
ok 4
not ok 5
ok 6
or
ok 4 - basic multi-variable
not ok 5 - simple exponential
ok 6 - force == mass * acceleration
The later gives you some idea of what failed. It also makes it easier
to find the test in your script, simply search for ``simple
exponential''.
All test functions take a name argument. It's optional, but highly
suggested that you use it.
The basic purpose of this module is to print out either ``ok #'' or ``not
ok #'' depending on if a given test succeeded or failed. Everything
else is just gravy.
All of the following print ``ok'' or ``not ok'' depending on if the test
succeeded or failed. They all also return true or false,
respectively.
This simply evaluates any expression ($this eq $that is just a
simple example) and uses that to determine if the test succeeded or
failed. A true expression passes, a false one fails. Very simple.
$test_name is a very short description of the test that will be printed
out. It makes it very easy to find a test in your script when it fails
and gives others an idea of your intentions. $test_name is optional,
but we very strongly encourage its use.
Should an ok() fail, it will produce some diagnostics:
not ok 18 - sufficient mucus
# Failed test 'sufficient mucus'
# in foo.t at line 42.
Similar to ok(), is() and isnt() compare their two arguments
with eq and ne respectively and use the result of that to
determine if the test succeeded or failed. So these:
# Is the ultimate answer 42?
is( ultimate_answer(), 42, "Meaning of Life" );
ok( ultimate_answer() eq 42, "Meaning of Life" );
ok( $foo ne '', "Got some foo" );
(Mnemonic: ``This is that.'' ``This isn't that.'')
So why use these? They produce better diagnostics on failure. ok()
cannot know what you are testing for (beyond the name), but is() and
isnt() know what the test was and why it failed. For example this
test:
my $foo = 'waffle'; my $bar = 'yarblokos';
is( $foo, $bar, 'Is foo the same as bar?' );
Will produce something like this:
not ok 17 - Is foo the same as bar?
# Failed test 'Is foo the same as bar?'
# in foo.t at line 139.
# got: 'waffle'
# expected: 'yarblokos'
So you can figure out what went wrong without rerunning the test.
You are encouraged to use is() and isnt() over ok() where possible,
however do not be tempted to use them to find out if something is
true or false!
# XXX BAD!
is( exists $brooklyn{tree}, 1, 'A tree grows in Brooklyn' );
This does not check ifexists $brooklyn{tree} is true, it checks if it returns 1. Very different. Similar caveats exist for false and 0.
In these cases, use ok().
ok( exists $brooklyn{tree}, 'A tree grows in Brooklyn' );
For those grammatical pedants out there, there's an isn't()
function which is an alias of isnt().
Similar to ok(), like() matches $this against the regex qr/that/.
So this:
like($this, qr/that/, 'this is like that');
is similar to:
ok( $this =~ /that/, 'this is like that');
(Mnemonic ``This is like that''.)
The second argument is a regular expression. It may be given as a
regex reference (i.e. qr//) or (for better compatibility with older
perls) as a string that looks like a regex (alternative delimiters are
currently not supported):
like( $this, '/that/', 'this is like that' );
Regex options may be placed on the end ('/that/i').
Its advantages over ok() are similar to that of is() and isnt(). Better diagnostics on failure.
Checks to see if the given $object->isa($class). Also checks to make
sure the object was defined in the first place. Handy for this sort of thing:
my $obj = Some::Module->new;
isa_ok( $obj, 'Some::Module' );
where you'd otherwise have to write
my $obj = Some::Module->new;
ok( defined $obj && $obj->isa('Some::Module') );
to safeguard against your test script blowing up.
It works on references, too:
isa_ok( $array_ref, 'ARRAY' );
The diagnostics of this test normally just refer to 'the object'. If
you'd like them to be more specific, you can supply an $object_name
(for example 'Test customer').
Sometimes you just want to say that the tests have passed. Usually
the case is you've got some complicated condition that is difficult to
wedge into an ok(). In this case, you can simply use pass() (to
declare the test ok) or fail (for not ok). They are synonyms for
ok(1) and ok(0).
You usually want to test if the module you're testing loads ok, rather
than just vomiting if its load fails. For such purposes we have
use_ok and require_ok.
BEGIN { use_ok($module); }
BEGIN { use_ok($module, @imports); }
These simply use the given $module and test to make sure the load
happened ok. It's recommended that you run use_ok() inside a BEGIN
block so its functions are exported at compile-time and prototypes are
properly honored.
If @imports are given, they are passed through to the use. So this:
BEGIN { use_ok('Some::Module', qw(foo bar)) }
is like doing this:
use Some::Module qw(foo bar);
Version numbers can be checked like so:
# Just like "use Some::Module 1.02"
BEGIN { use_ok('Some::Module', 1.02) }
Don't try to do this:
BEGIN {
use_ok('Some::Module');
...some code that depends on the use...
...happening at compile time...
}
because the notion of ``compile-time'' is relative. Instead, you want:
BEGIN { use_ok('Some::Module') }
BEGIN { ...some code that depends on the use... }
Not everything is a simple eq check or regex. There are times you
need to see if two data structures are equivalent. For these
instances Test::More provides a handful of useful functions.
NOTE I'm not quite sure what will happen with filehandles.
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. If the two structures are different, it will display the
place where they start differing.
is_deeply() compares the dereferenced values of references, the
references themselves (except for their type) are ignored. This means
aspects such as blessing and ties are not considered ``different''.
is_deeply() current has very limited handling of function reference
and globs. It merely checks if they have the same referent. This may
improve in the future.
Test::Differences and Test::Deep provide more in-depth functionality
along these lines.
If you pick the right test function, you'll usually get a good idea of
what went wrong when it failed. But sometimes it doesn't work out
that way. So here we have ways for you to write your own diagnostic
messages which are safer than just print STDERR.
ok( grep(/foo/, @users), "There's a foo user" ) or
diag("Since there's no foo, check that /etc/bar is set up right");
which would produce:
not ok 42 - There's a foo user
# Failed test 'There's a foo user'
# in foo.t at line 52.
# Since there's no foo, check that /etc/bar is set up right.
You might remember ok() or diag() with the mnemonic open() or
die().
NOTE The exact formatting of the diagnostic output is still
changing, but it is guaranteed that whatever you throw at it it won't
interfere with the test.
Sometimes running a test under certain conditions will cause the
test script to die. A certain function or method isn't implemented
(such as fork() on MacOS), some resource isn't available (like a
net connection) or a module isn't available. In these cases it's
necessary to skip tests, or declare that they are supposed to fail
but will work in the future (a todo test).
For more details on the mechanics of skip and todo tests see
the Test::Harness manpage.
The way Test::More handles this is with a named block. Basically, a
block of tests which can be skipped over or made todo. It's best if I
just show you...
This declares a block of tests that might be skipped, $how_many tests
there are, $why and under what $condition to skip them. An example is
the easiest way to illustrate:
my $lint = new HTML::Lint;
isa_ok( $lint, "HTML::Lint" );
$lint->parse( $html );
is( $lint->errors, 0, "No errors found in HTML" );
}
If the user does not have HTML::Lint installed, the whole block of
code won't be run at all. Test::More will output special ok's
which Test::Harness interprets as skipped, but passing, tests.
It's important that $how_many accurately reflects the number of tests
in the SKIP block so the # of tests run will match up with your plan.
If your plan is no_plan $how_many is optional and will default to 1.
It's perfectly safe to nest SKIP blocks. Each SKIP block must have
the label SKIP, or Test::More can't work its magic.
You don't skip tests which are failing because there's a bug in your
program, or for which you don't yet have code written. For that you
use TODO. Read on.
With a todo block, the tests inside are expected to fail. Test::More will run the tests normally, but print out special flags indicating
they are ``todo''. Test::Harness will interpret failures as being ok.
Should anything succeed, it will report it as an unexpected success.
You then know the thing you had todo is done and can remove the
TODO flag.
The nice part about todo tests, as opposed to simply commenting out a
block of tests, is it's like having a programmatic todo list. You know
how much work is left to be done, you're aware of what bugs there are,
and you'll know immediately when they're fixed.
Once a todo test starts succeeding, simply move it outside the block.
When the block is empty, delete it.
With todo tests, it's best to have the tests actually run. That way
you'll know when they start passing. Sometimes this isn't possible.
Often a failing test will cause the whole program to die or hang, even
inside an eval BLOCK with and using alarm. In these extreme
cases you have no choice but to skip over the broken tests entirely.
The syntax and behavior is similar to a SKIP: BLOCK except the
tests will be marked as failing but todo. Test::Harness will
interpret them as passing.
If it's something the user might not be able to do, use SKIP.
This includes optional modules that aren't installed, running under
an OS that doesn't have some feature (like fork() or symlinks), or maybe
you need an Internet connection and one isn't available.
If it's something the programmer hasn't done yet, use TODO. This
is for any code you haven't written yet, or bugs you have yet to fix,
but want to put tests in your testing script (always a good idea).
Incidates to the harness that things are going so badly all testing
should terminate. This includes the running any additional test scripts.
This is typically used when testing cannot continue such as a critical
module failing to compile or a necessary external utility not being
available such as a database connection failing.
The use of the following functions is discouraged as they are not
actually testing functions and produce no diagnostics to help figure
out what went wrong. They were written before is_deeply() existed
because I couldn't figure out how to display a useful diff of two
arbitrary data structures.
These functions are usually used inside an ok().
ok( eq_array(\@this, \@that) );
is_deeply() can do that better and with diagnostics.
Similar to eq_array(), except the order of the elements is not
important. This is a deep check, but the irrelevancy of order only
applies to the top level.
ok( eq_set(\@this, \@that) );
Is better written:
is_deeply( [sort @this], [sort @that] );
NOTE By historical accident, this is not a true set comparison.
While the order of elements does not matter, duplicate elements do.
NOTEeq_set() does not know how to deal with references at the top
level. The following is an example of a comparison which might not work:
eq_set([\1, \2], [\2, \1]);
Test::Deep contains much better set comparison functions.
Sometimes the Test::More interface isn't quite enough. Fortunately, Test::More is built on top of Test::Builder which provides a single,
unified backend for any test library to use. This means two test
libraries which both use Test::Buildercan be used together in the
same program.
If you simply want to do a little tweaking of how the tests behave,
you can access the underlying Test::Builder object like so:
If all your tests passed, Test::Builder will exit with zero (which is
normal). If anything failed it will exit with how many failed. If
you run less (or more) tests than you planned, the missing (or extras)
will be considered failures. If no tests were ever run Test::Builder will throw a warning and exit with 255. If the test died, even after
having successfully completed all its tests, it will still be
considered a failure and will exit with 255.
So the exit codes are...
0 all tests successful
255 test died or all passed but wrong # of tests run
any other number how many failed (including missing or extras)
If you fail more than 254 tests, it will be reported as 254.
NOTE This behavior may go away in future versions.
String overloaded objects are compared as strings (or in cmp_ok()'s
case, strings or numbers as appropriate to the comparison op). This
prevents Test::More from piercing an object's interface allowing
better blackbox testing. So if a function starts returning overloaded
objects instead of bare strings your tests won't notice the
difference. This is good.
However, it does mean that functions like is_deeply() cannot be used to
test the internals of string overloaded objects. In this case I would
suggest Test::Deep which contains more flexible testing functions for
complex data structures.
no_plan and todo depend on new Test::Harness features and fixes. If
you're going to distribute tests that use no_plan or todo your
end-users will have to upgrade Test::Harness to the latest one on
CPAN. If you avoid no_plan and TODO tests, the stock Test::Harness will work fine.
Installing Test::More should also upgrade Test::Harness.
This is a case of convergent evolution with Joshua Pritikin's Test module. I was largely unaware of its existence when I'd first
written my own ok() routines. This module exists because I can't
figure out how to easily wedge test names into Test's interface (along
with a few other problems).
The goal here is to have a testing utility that's simple to learn,
quick to use and difficult to trip yourself up with while still
providing more flexibility than the existing Test.pm. As such, the
names of the most common routines are kept tiny, special cases and
magic side-effects are kept to a minimum. WYSIWYG.
the Test::Simple manpage if all this confuses you and you just want to write
some tests. You can upgrade to Test::More later (it's forward
compatible).
perldoc2tree.cgi: /usr/lib/perl5/5.8.8/Test/More.pm: cannot resolve L in paragraph 414.
Test is the old testing module. Its main benefit is that it has
been distributed with Perl since 5.004_05.
the Test::Harness manpage for details on how your test results are interpreted
by Perl.
Michael G Schwern <schwern@pobox.com> with much inspiration
from Joshua Pritikin's Test module and lots of help from Barrie
Slaymaker, Tony Bowden, blackstar.co.uk, chromatic, Fergal Daly and
the perl-qa gang.
