Skip to content


Subversion checkout URL

You can clone with
Download ZIP
189 lines (159 sloc) 7.367 kB
Release 0.9.7
- small update to add support for TeamCity / PyCharm test runner.
Release 0.9.6
- fix staticmethod mocking on instances
- fix comparison of kwargs ordering issues
- fix ReturnValue.__str__
Release 0.9.5
- bugfix: stop enforcing argument signatures on flexmock objects
Release 0.9.4
- add support for stubbing return values on getter properties
- add custom matcher object support to with_args
- add support for striter function signature checks
- add support for non-callable attributes
- add support chained attributes (thanks Bryce Covert!)
- add iter support to Mock objects
- add PyPy support
- add Jython support
- fix should_call to work with class mocks
- fix and_return to return None by default
- fix MRO issues on builtin methods on 2.7+/3.2+
- imporove defaults: partial mocks created using the func=return_value
style now default to replace_with instead of should_receive for callables
Release 0.9.3
- add python 3.3 test target
- add proper handling of ordered() expectation across different methods
- add property support on fake objects
- fix compatibility with pytest 2.2 (thanks jpvanhal!)
- fix insidious bug with mocking subclasses of str class
- fix tuple handling when formatting arguments
- fix reseting subclass methods
Release 0.9.2
- fix mocking builtins by reseting expectation when raising exceptions
- fix mocking private methods on classes with leading underscores
- limit the damage of "from flexmock import *" by limiting to just flexmock()
- ensure _pre_flexmock_success is cleaned up after each test
Release 0.9.1
- adding support for more test runners:
* unittest2
* django
* twisted/trial
* zope.testrunner
* subunit
* testtools
Release 0.9.0
- adding state machine support using when()
- make expectation fail as soon as number of expected calls is exceeded
- flexmock_teardown no longer returns a function
- allow should_call on class and static methods
- disallow should_call on class mocks
- fixing unicode args handling
- fixing issues with @property methods misbehaving in the debugger
- fixing pytest integration and instance teardown
- fixing private method handling
Release 0.8.1
- fixing pytest and doctest integration to always call flexmock_teardown
- fixing flexmock_teardown to return a function as before so it can be used as a decorator
Release 0.8.0
- big changes in runner integration support (no more stack examination or sketchy teardown replacement)
- doctest integration
- fixing ordering verification when the method has a default stub
- fixing calling with_args() without arguments to match exactly no arguments (thanks jerico-dev!)
- 20% performance improvement
- make sure to return object itself when partial mocking instances unless the object already has some of the methods
- ensure consecutive calls return same mock object
- adding regex support for arg matching and spy return values
- enabling replace_with for class mocks
- disabling expectation checking if an exception has already been raised
- massive refactoring of the way flexmock does monkey patching
- Fixing replace_with to work properly like and_execute
- (and_execute will be deprecated in next release!)
Release 0.7.4
- Fixed exception type check when no message specified
- Make properties work optionally with parentheses
- Make sure should_receive does not replace flexmock methods
- Removed new_instances= param in favor of new_instances() method
- Refactoring to move all state to FlexmockContainer class
Release 0.7.3
- Added new_instances method (new_instances param will be deprecated in next release!)
- Added replace_with to enable returning results of custom functions
- Added "with" support for FlexMock objects
- Moved tests to their own directory
- Lots of documentation cleanup and updates
Release 0.7.2
- Added support for chained methods
- Moved flexmock_teardown to module level to expose it for other test runners
- Added py.test support (thanks to derdon)
- Lots of test refactoring and improvements for multiple test runner support
- Fix loop in teardown
- Fix should_call for same method with different args
Release 0.7.1
- Fix bug with "never" not working when the expectation is not met
- Fix bug in duplicate calls to original method in pass_thru mode (thanks sagara-!)
- Fix bug in handling unicode characters in ReturnValue
Release 0.7.0
- Better error handling for trying to mock builtins
- Added simple test harness for running on multiple versions / test runners
- Fixed unicode arg formatting (thanks to sagara-!)
- Made it impossible to mock non-existent methods
- Ensure flexmock teardown takes varargs (for better runner integration)
Release 0.6.9
- Initial nose integration (still no support for generated tests)
- Fixing private class methods
- Some test refactoring to support different test runners
Release 0.6.8
- Add should_call() alias for should_receive().and_execute
- Ensure new_instances can't be used with expectation modifiers
- Make and_execute match return value by class in addition to value
- Support for mocking out static methods
- Bit of test fixage (thanks to derdon)
Release 0.6.7
- Fixing clobbering of original method by multiple flexmock calls
- Making and_raise work properly with exception classes and args
- Proper exception matching with and_execute
- Fix mocking same class twice
Release 0.6.6
- Removing extra args from should_receive
- Making and_execute check return/raise value of original method
- Refactoring FlexMock constructor into factory method
- Fixing new_instances to accept multiple args instead of just none
- Raising an exception when and_execute is set on class mock
Release 0.6.5
- Adding support for multiple flexmock() calls on same object
- Adding error detection on and_execute for missing or unbound methods
- Make sure empty args don't include None
Release 0.6.4
- Fixing up teardown cleanup code after an exception is raised in tests
- Fixing and_yield to return proper generator
- Adding and_yield returning a predefined generator
- Replacing and_passthru with and_execute
- Make it easier to mock private methods
Release 0.6.3
- Adding keyword argument expectation matching
Release 0.6.2
- Changing and_return(multiple=True) to one_by_one
- Making it possible to supply multiple args to and_return instead of a tuple
- Changing default mock behavior to create attributes instead of methods
- FIX teardown for python3
Release 0.6.1
- Make it even easier to integrate with new test runners
- Adding support for mixing returns and raises in return values
Release 0.6
- Adding support for multiple arg type matches
- Pulling out the entry point code from constructor into its own method.
Release 0.5
- FIX: ensuring that mocks are cleaned up properly between tests
- BROKEN: part1 on ensuring mocking multiple objects works correctly
- Make sure pass_thru doesn't try to call a non-existent method
- Fixing up copyright notice
- Adding some missing pydocs
Release 0.4
- Fixing tests and ensuring mock methods really get created properly
- Making sure shortcuts create methods rather than attributes
- Fixing doc strings
- Removing the new-style/old-style convert code, it's stupid
Release 0.3
- Making Expectation.mock into a property so that it shows up in pydoc
- Adding proxying/spying and at_least/at_most expectation modifiers
Jump to Line
Something went wrong with that request. Please try again.