NAME

DESCRIPTION

SYNOPSIS

    use Test2::Mock;
    use MyClass;

    my $mock = Test2::Mock->new(
        track => $BOOL, # enable call tracking if desired
        class => 'MyClass',
        override => [
            name => sub { 'fred' },
            ...
        ],
        add => [
            is_mocked => sub { 1 }
            ...
        ],
        ...
    );

    # Unmock the 'name' sub
    $mock->restore('name');

    ...

    $mock = undef; # Will remove all the mocking

CONSTRUCTION

METHODS

$mock = Test2::Mock->new(class => $CLASS, ...)

This will create a new instance of Test2::Mock that manages mocking for the specified $CLASS.

Any "Test2::Mock" method can be used as a constructor argument, each should be followed by an arrayref of arguments to be used within the method. For instance the "add()" method:

    my $mock = Test2::Mock->new(
        class => 'AClass',
        add => [foo => sub { 'foo' }],
    );
    

is identical to this:

    my $mock = Test2::Mock->new(
        class => 'AClass',
    );
    $mock->add(foo => sub { 'foo' });
    

$mock->track($bool)

Turn tracking on or off. Any sub added/overridden/set when tracking is on will log every call in a hash retrievable via "$mock->tracking". Changing the tracking toggle will not affect subs already altered, but will affect any additional alterations.

$hashref = $mock->sub_tracking

The tracking data looks like this:

    {
        sub_name => [
            {sub_name => $sub_name, sub_ref => $mock_subref, args => [... copy of @_ from the call ... ]},
            ...,
            ...,
        ],
    }
    

Unlike call_tracking, this lists all calls by sub, so you can choose to only look at the sub specific calls.

Please note: The hashref items with the subname and args are shared with call_tracking, modifying one modifies the other, so copy first!

$arrayref = $mock->call_tracking

The tracking data looks like this:

    [
        {sub_name => $sub_name, sub_ref => $mock_subref, args => [... copy of @_ from the call ... ]},
        ...,
        ...,
    ]
    

Unlike sub_tracking this lists all calls to any mocked sub, in the order they were called. To filter by sub use sub_tracking.

Please note: The hashref items with the subname and args are shared with sub_tracking, modifying one modifies the other, so copy first!

$mock->clear_sub_tracking()

$mock->clear_sub_tracking(\@subnames)

Clear tracking data. With no arguments ALL tracking data is cleared. When arguments are provided then only those specific keys will be cleared.

$mock->clear_call_tracking()

Clear all items from call_tracking.

$mock->add('symbol' => ..., 'symbol2' => ...)

$mock->override('symbol1' => ..., 'symbol2' => ...)

$mock->set('symbol1' => ..., 'symbol2' => ...)

"add()" and "override()" are the primary ways to add/modify methods for a class. Both accept the exact same type of arguments. The difference is that "override" will fail unless the symbol you are overriding already exists, "add" on the other hand will fail if the symbol does already exist.

"set()" was more recently added for cases where you may not know if the sub already exists. These cases are rare, and set should be avoided (think of it like 'no strict'). However there are valid use cases, so it was added.

Note: Think of override as a push operation. If you call override on the same symbol multiple times it will track that. You can use "restore()" as a pop operation to go back to the previous mock. "reset" can be used to remove all the mocking for a symbol.

Arguments must be a symbol name, with optional sigil, followed by a new specification of the symbol. If no sigil is specified then '&' (sub) is assumed. A simple example of overriding a sub:

    $mock->override(foo => sub { 'overridden foo' });
    my $val = $class->foo; # Runs our override
    # $val is now set to 'overridden foo'
    

You can also simply provide a value and it will be wrapped in a sub for you:

    $mock->override( foo => 'foo' );
    

The example above will generate a sub that always returns the string 'foo'.

There are three *special* values that can be used to generate accessors:

    $mock->add(
        name => 'rw',   # Generates a read/write accessor
        age  => 'ro',   # Generates a read only accessor
        size => 'wo',   # Generates a write only accessor
    );
    

If you want to have a sub that actually returns one of the three special strings, or that returns a coderef, you can use a hashref as the spec:

    my $ref = sub { 'my sub' };
    $mock->add(
        rw_string => { val => 'rw' },
        ro_string => { val => 'ro' },
        wo_string => { val => 'wo' },
        coderef   => { val => $ref }, # the coderef method returns $ref each time
    );
    

You can also override/add other symbol types, such as hash:

    package Foo;
    ...

    $mock->add('%foo' => {a => 1});

    print $Foo::foo{a}; # prints '1'
    

You can also tell mock to deduce the symbol type for the add/override from the reference, rules are similar to glob assignments:

    $mock->add(
        -foo => sub { 'foo' },     # Adds the &foo sub to the package
        -foo => { foo => 1 },      # Adds the %foo hash to the package
        -foo => [ 'f', 'o', 'o' ], # Adds the @foo array to the package
        -foo => \"foo",            # Adds the $foo scalar to the package
    );
    

$mock->restore($SYMBOL)

Restore the symbol to what it was before the last override. If the symbol was recently added this will remove it. If the symbol has been overridden multiple times this will ONLY restore it to the previous state. Think of "override" as a push operation, and "restore" as the pop operation.

$mock->reset($SYMBOL)

Remove all mocking of the symbol and restore the original symbol. If the symbol was initially added then it will be completely removed.

$mock->orig($SYMBOL)

This will return the original symbol, before any mocking. For symbols that were added this will return undef.

$mock->current($SYMBOL)

This will return the current symbol.

$mock->reset_all

Remove all added symbols, and restore all overridden symbols to their originals.

$mock->add_constructor($NAME => $TYPE)

$mock->override_constructor($NAME => $TYPE)

This can be used to inject constructors. The first argument should be the name of the constructor. The second argument specifies the constructor type.

The "hash" type is the most common, all arguments are used to create a new hash that is blessed.

    hash => sub  {
        my ($class, %params) = @_;
        return bless \%params, $class;
    };
    

The "array" type is similar to the hash type, but accepts a list instead of key/value pairs:

    array => sub {
        my ($class, @params) = @_;
        return bless \@params, $class;
    };
    

The "ref" type takes a reference and blesses it. This will modify your original input argument.

    ref => sub {
        my ($class, $params) = @_;
        return bless $params, $class;
    };
    

The "ref_copy" type will copy your reference and bless the copy:

    ref_copy => sub {
        my ($class, $params) = @_;
        my $type = reftype($params);

        return bless {%$params}, $class
            if $type eq 'HASH';

        return bless [@$params], $class
            if $type eq 'ARRAY';

        croak "Not sure how to construct a '$class' from '$params'";
    };
    

$mock->before($NAME, sub { ... })

This will replace the original sub $NAME with a new sub that calls your custom code just before calling the original method. The return from your custom sub is ignored. Your sub and the original both get the unmodified arguments.

$mock->after($NAME, sub { ... })

This is similar to before, except your callback runs after the original code. The return from your callback is ignored.

$mock->around($NAME, sub { ... })

This gives you the chance to wrap the original sub:

    $mock->around(foo => sub {
        my $orig = shift;
        my $self = shift;
        my (@args) = @_;

        ...
        $self->$orig(@args);
        ...

        return ...;
    });
    

The original sub is passed in as the first argument, even before $self. You are responsible for making sure your wrapper sub returns the correct thing.

$mock->autoload

This will inject an "AUTOLOAD" sub into the class. This autoload will automatically generate read-write accessors for any sub called that does not already exist.

$mock->block_load

This will prevent the real class from loading until the mock is destroyed. This will fail if the class is already loaded. This will let you mock a class completely without loading the original module.

$pm_file = $mock->file

This returns the relative path to the file for the module. This corresponds to the %INC entry.

$bool = $mock->purge_on_destroy($bool)

When true, this will cause the package stash to be completely obliterated when the mock object falls out of scope or is otherwise destroyed. You do not normally want this.

$stash = $mock->stash

This returns the stash for the class being mocked. This is the equivalent of:

    my $stash = \%{"${class}\::"};
    

This saves you from needing to turn off strict.

$class = $mock->class

The class being mocked by this instance.

$p = $mock->parent

If you mock a class twice the first instance is the parent, the second is the child. This prevents the parent from being destroyed before the child, which would lead to a very unpleasant situation.

$c = $mock->child

Returns the child mock, if any.

SOURCE

MAINTAINERS

Chad Granum <exodist@cpan.org>

AUTHORS

Chad Granum <exodist@cpan.org>

COPYRIGHT

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

See <https://dev.perl.org/licenses/>