NAME
Test::Mockingbird - Advanced mocking library for Perl with support for dependency injection and spies
VERSION
Version 0.09
SYNOPSIS
use Test::Mockingbird;
# Mocking
Test::Mockingbird::mock('My::Module', 'method', sub { return 'mocked!' });
# Spying
my $spy = Test::Mockingbird::spy('My::Module', 'method');
My::Module::method('arg1', 'arg2');
my @calls = $spy->(); # Get captured calls
# Dependency Injection
Test::Mockingbird::inject('My::Module', 'Dependency', $mock_object);
# Unmocking
Test::Mockingbird::unmock('My::Module', 'method');
# Restore everything
Test::Mockingbird::restore_all();
DESCRIPTION
Test::Mockingbird provides powerful mocking, spying, and dependency injection capabilities to streamline testing in Perl.
DIAGNOSTICS
Test::Mockingbird provides optional, non-intrusive diagnostic routines that allow inspection of the current mocking state during test execution. These routines are purely observational. They do not modify any mocking behaviour, symbol table entries, or internal state.
Diagnostics are useful when debugging complex test suites, verifying mock layering behaviour, or understanding interactions between multiple mocking primitives such as mock, spy, inject, and the sugar functions.
diagnose_mocks
Return a structured hashref describing all currently active mock layers. Each entry includes the fully qualified method name, the number of active layers, whether the original method existed, and metadata for each layer (type and installation location). See the diagnose_mocks method for full API details.
diagnose_mocks_pretty
Return a human-readable, multi-line string describing all active mock layers. This routine is intended for debugging and inspection during test development. The output format is stable for human consumption but is not guaranteed for machine parsing. See the diagnose_mocks_pretty method for full API details.
Diagnostic Metadata
Diagnostic information is recorded automatically whenever a mock layer is successfully installed. Each layer records:
* type The category of mock layer (for example: mock, spy,
inject, mock_return, mock_exception, mock_sequence,
mock_once, mock_scoped)
* installed_at The file and line number where the layer was created
This metadata is maintained in parallel with the internal mock stack and is automatically cleared when a method is fully restored via unmock or restore_all.
Diagnostics never alter the behaviour of the mocking engine and may be safely invoked at any point during a test run.
DEBUGGING EXAMPLES
This section provides practical examples of using the diagnostic routines to understand and debug complex mocking behaviour. All examples are safe to run inside test files and do not modify mocking semantics.
Example 1: Inspecting a simple mock
{
package Demo::One;
sub value { 1 }
}
mock_return 'Demo::One::value' => 42;
my $diag = diagnose_mocks();
print diagnose_mocks_pretty();
The output will resemble:
Demo::One::value:
depth: 1
original_existed: 1
- type: mock_return installed_at: t/example.t line 12
This confirms that the method has exactly one active mock layer and shows where it was installed.
Example 2: Stacked mocks
{
package Demo::Two;
sub compute { 10 }
}
mock_return 'Demo::Two::compute' => 20;
mock_exception 'Demo::Two::compute' => 'fail';
print diagnose_mocks_pretty();
Possible output:
Demo::Two::compute:
depth: 2
original_existed: 1
- type: mock_return installed_at: t/example.t line 8
- type: mock_exception installed_at: t/example.t line 9
This shows the order in which layers were applied. The most recent layer appears last.
Example 3: Spies and injected dependencies
{
package Demo::Three;
sub action { 1 }
sub dep { 2 }
}
spy 'Demo::Three::action';
inject 'Demo::Three::dep' => sub { 99 };
print diagnose_mocks_pretty();
Example output:
Demo::Three::action:
depth: 1
original_existed: 1
- type: spy installed_at: t/example.t line 7
Demo::Three::dep:
depth: 1
original_existed: 1
- type: inject installed_at: t/example.t line 8
This confirms that both the spy and the injected dependency are active.
Example 4: After restore_all
mock_return 'Demo::Four::x' => 5;
restore_all();
print diagnose_mocks_pretty();
Output:
(no output)
After restore_all, all diagnostic metadata is cleared along with the mock layers.
Example 5: Using diagnostics inside a failing test
When a test fails unexpectedly, adding the following line can help identify the active mocks:
diag diagnose_mocks_pretty();
This prints the current mocking state into the test output without affecting the test run.
METHODS
mock($package, $method, $replacement)
Mocks a method in the specified package. Supports two forms:
mock('My::Module', 'method', sub { ... });
or the shorthand:
mock 'My::Module::method' => sub { ... };
unmock($package, $method)
Restores the original method for a mocked method. Supports two forms:
unmock('My::Module', 'method');
or the shorthand:
unmock 'My::Module::method';
mock_scoped
Creates a scoped mock that is automatically restored when the returned guard goes out of scope.
This behaves like mock, but instead of requiring an explicit call to unmock or restore_all, all mocked methods are reverted automatically when the guard object is destroyed.
Single-method forms
Shorthand:
my $g = mock_scoped 'My::Module::method' => sub { 'mocked' };
Longhand:
my $g = mock_scoped('My::Module', 'method', sub { ... });
Multi-method forms
Mock several methods on one package with a single guard:
my $g = mock_scoped('My::Module',
fetch => sub { 'mocked_fetch' },
save => sub { 'mocked_save' },
delete => sub { 'mocked_delete' },
);
Mock methods across different packages in one call (shorthand pairs):
my $g = mock_scoped(
'My::Module::fetch' => sub { 'mocked_fetch' },
'Other::Module::save' => sub { 'mocked_save' },
);
In both multi-method forms, every mocked method is restored when $g goes out of scope or is explicitly undefed.
Scoped lifecycle
{
my $g = mock_scoped 'My::Module::method' => sub { 'mocked' };
My::Module::method(); # returns 'mocked'
}
My::Module::method(); # original behaviour restored
Interaction with spy
A spy is not automatically restored when a mock_scoped guard goes out of scope. mock_scoped only manages the specific mock layer it installs. If you install a spy inside a scoped block, you must restore it explicitly:
{
my $g = mock_scoped 'My::Module::method' => sub { 1 };
my $spy = spy 'My::Module::method';
My::Module->method('arg');
}
# $g is destroyed here -- the mock_scoped layer is restored
# but the spy layer is still active
restore_all(); # needed to fully restore method
The safe pattern when combining mock_scoped and spy is to call restore_all at the end of the block, or to avoid combining them and use mock with an explicit restore_all instead:
spy 'My::Module::method';
My::Module->method('arg');
my @calls = $spy->();
restore_all();
Notes
If you need both a modified implementation and call recording in the same test, install the spy first and then the mock. The spy will still capture calls even when the implementation is replaced by the mock layer above it, because the spy wraps the layer below it at installation time, not the current top of the stack. To avoid confusion, prefer explicit restore_all over mock_scoped when combining with spies.
spy($package, $method)
Wraps a method so that all calls and arguments are recorded. Supports two forms:
spy('My::Module', 'method');
or the shorthand:
spy 'My::Module::method';
Returns a coderef which, when invoked, returns the list of captured calls. The original method is preserved and still executed.
Call record format
Each captured call is an arrayref with the following structure:
[ $method_name, $invocant, @arguments ]
where:
$method_name- the fully qualified method name as a string (e.g.'My::Module::method')$invocant- the first argument to the call, typically$selffor method calls or the first positional argument for function calls@arguments- the remaining arguments passed to the method, in the order they were supplied. For named-parameter calls these will be alternating key/value pairs suitable for assignment to a hash:my %args = @{$call}[2..$#{$call}]
Example
spy 'My::Module::process';
My::Module->process(name => 'foo', value => 42);
my @calls = $spy->();
my $call = $calls[0];
# $call->[0] eq 'My::Module::process'
# $call->[1] is the My::Module object
# @{$call}[2..$#{$call}] gives (name => 'foo', value => 42)
my %args = @{$call}[2..$#{$call}];
is($args{name}, 'foo', 'name arg captured');
is($args{value}, 42, 'value arg captured');
inject($package, $dependency, $mock_object)
Injects a mock dependency. Supports two forms:
inject('My::Module', 'Dependency', $mock_object);
or the shorthand:
inject 'My::Module::Dependency' => $mock_object;
The injected dependency can be restored with restore_all or unmock.
restore_all
Restores all mocked methods and injected dependencies.
Called with no arguments, restores everything that has been mocked in the current test run:
restore_all();
Called with a package name, restores only the mocks whose fully qualified names begin with that package:
restore_all 'My::Module';
This is useful when a test installs mocks across multiple packages and needs to tear down only one package's mocks without disturbing the others:
mock 'My::Module::fetch' => sub { 'mocked_fetch' };
mock 'Other::Module::save' => sub { 'mocked_save' };
# Tear down only My::Module mocks
restore_all 'My::Module';
# Other::Module::save is still mocked here
restore_all(); # now everything is restored
Notes
Restoring a package that was never mocked is a no-op and does not warn or croak.
mock_return
Mock a method so that it always returns a fixed value.
Takes a single target (either 'Pkg::method' or ('Pkg','method')) and a value to return. Returns nothing. Side effects: installs a mock layer using "mock".
API specification
Input
Params::Validate::Strict schema:
- target: required, scalar, string; method target in shorthand or longhand form - value: required, any type; value to be returned by the mock
Output
Returns::Set schema:
- return: undef
mock_exception
Mock a method so that it always throws an exception.
Takes a single target (either 'Pkg::method' or ('Pkg','method')) and an exception message. Returns nothing. Side effects: installs a mock layer using "mock".
API specification
Input
Params::Validate::Strict schema:
- target: required, scalar, string; method target in shorthand or longhand form - message: required, scalar, string; exception text to croak with
Output
Returns::Set schema:
- return: undef
mock_sequence
Mock a method so that it returns a sequence of values over successive calls.
Takes a single target (either 'Pkg::method' or ('Pkg','method')) and one or more values. Returns nothing. Side effects: installs a mock layer using "mock". When the sequence is exhausted, the last value is repeated.
API specification
Input
Params::Validate::Strict schema:
- target: required, scalar, string; method target in shorthand or longhand form - values: required, array; one or more values to be returned in order
Output
Returns::Set schema:
- return: undef
mock_once
Install a mock that is executed exactly once. After the first call, the previous implementation is automatically restored. This is useful for testing retry logic, fallback behaviour, and state transitions.
API specification
Input (Params::Validate::Strict schema)
- target: required, scalar, string; method target in shorthand or longhand form - code: required, coderef; mock implementation to run once
Output (Returns::Set schema)
- return: undef
restore
Restore all mock layers for a single method target. This is similar to restore_all, but applies only to one method. If the method was never mocked, this routine has no effect.
API specification
Input (Params::Validate::Strict schema)
- target: required, scalar, string; method target in shorthand or longhand form
Output (Returns::Set schema)
- return: undef
diagnose_mocks
Return a structured hashref describing all currently active mock layers. This routine is purely observational and does not modify any state.
API specification
Input
Params::Validate::Strict schema:
- none
Output
Returns::Set schema:
- return: hashref; keys are fully qualified method names, values are hashrefs containing: - depth: integer; number of active mock layers - layers: arrayref of hashrefs; each layer has: - type: string - installed_at: string - original_existed: boolean
diagnose_mocks_pretty
Return a human-readable string describing all currently active mock layers. This routine is purely observational and does not modify any state.
API specification
Input
Params::Validate::Strict schema:
- none
Output
Returns::Set schema:
- return: scalar string; formatted multi-line description of all active mock layers, including: - fully qualified method name - depth (number of active layers) - whether the original method existed - each layer's type and installation location
Behaviour
Entry
- No arguments are accepted.
Exit
- Returns a formatted string describing the current mocking state.
Side effects
- None. This routine does not modify %mocked, %mock_meta, or any symbol table entries.
Notes
- This routine is intended for debugging and diagnostics. It is safe to call at any point during a test run. - The output format is stable and suitable for human inspection, but not guaranteed to remain fixed for machine parsing.
DESTROY
If Test::Mockingbird goes out of scope, restore everything.
SUPPORT
This module is provided as-is without any warranty.
Please report any bugs or feature requests to bug-test-mockingbird at rt.cpan.org, or through the web interface at http://rt.cpan.org/NoAuth/ReportBug.html?Queue=Test-Mockingbird. I will be notified, and then you'll automatically be notified of progress on your bug as I make changes.
You can find documentation for this module with the perldoc command.
perldoc Test::Mockingbird
AUTHOR
Nigel Horne, <njh at nigelhorne.com>
BUGS
SEE ALSO
REPOSITORY
https://github.com/nigelhorne/Test-Mockingbird
SUPPORT
This module is provided as-is without any warranty.
LICENCE AND COPYRIGHT
Copyright 2025-2026 Nigel Horne.
Usage is subject to GPL2 licence terms. If you use it, please let me know.