* Flip equality to use mock calls' __eq__
* bpo-37555: Regression test demonstrating assert_has_calls not working with ANY and spec_set
Co-authored-by: Neal Finne <neal@nealfinne.com>
* Revert "Flip equality to use mock calls' __eq__"
This reverts commit 94ddf54c5a.
* bpo-37555: Add regression tests for mock ANY ordering issues
Add regression tests for whether __eq__ is order agnostic on _Call and _CallList, which is useful for comparisons involving ANY, especially if the ANY comparison is to a class not defaulting __eq__ to NotImplemented.
Co-authored-by: Neal Finne <neal@nealfinne.com>
* bpo-37555: Fix _CallList and _Call order sensitivity
_Call and _CallList depend on ordering to correctly process that an object being compared to ANY with __eq__ should return True. This fix updates the comparison to check both a == b and b == a and return True if either condition is met, fixing situations from the tests in the previous two commits where assertEqual would not be commutative if checking _Call or _CallList objects. This seems like a reasonable fix considering that the Python data model specifies that if an object doesn't know how to compare itself to another object it should return NotImplemented, and that on getting NotImplemented from a == b, it should try b == a, implying that good behavior for __eq__ is commutative. This also flips the order of comparison in _CallList's __contains__ method, guaranteeing ANY will be on the left and have it's __eq__ called for equality checking, fixing the interaction between assert_has_calls and ANY.
Co-author: Neal Finne <neal@neal.finne.com>
* bpo-37555: Ensure _call_matcher returns _Call object
* Adding ACK and news entry
* bpo-37555: Replacing __eq__ with == to sidestep NotImplemented
bool(NotImplemented) returns True, so it's necessary to use ==
instead of __eq__ in this comparison.
* bpo-37555: cleaning up changes unnecessary to the final product
* bpo-37555: Fixed call on bound arguments to respect args and kwargs
* Revert "bpo-37555: Add regression tests for mock ANY ordering issues"
This reverts commit 49c5310ad4.
* Revert "bpo-37555: cleaning up changes unnecessary to the final product"
This reverts commit 18e964ba01.
* Revert "bpo-37555: Replacing __eq__ with == to sidestep NotImplemented"
This reverts commit f295eaca5b.
* Revert "bpo-37555: Fix _CallList and _Call order sensitivity"
This reverts commit 874fb697b8.
* Updated NEWS.d
* bpo-37555: Add tests checking every function using _call_matcher both with and without spec
* bpo-37555: Ensure all assert methods using _call_matcher are actually passing calls
* Remove AnyCompare and use call objects everywhere.
* Revert "Remove AnyCompare and use call objects everywhere."
This reverts commit 24973c0b32.
* Check for exception in assert_any_await
* bpo-37972: unittest.mock._Call now passes on __getitem__ to the __getattr__ chaining so that call() can be subscriptable
* 📜🤖 Added by blurb_it.
* Update 2019-08-28-21-40-12.bpo-37972.kP-n4L.rst
added name of the contributor
* bpo-37972: made all dunder methods chainable for _Call
* bpo-37972: delegate only attributes of tuple instead to __getattr__
* Fix call_matcher for mock when using methods
* Add NEWS entry
* Use None check and convert doctest to unittest
* Use better name for mock in tests. Handle _SpecState when the attribute was not accessed and add tests.
* Use reset_mock instead of reinitialization. Change inner class constructor signature for check
* Reword comment regarding call object lookup logic
* Clear name and parent of mock in autospecced objects used with attach_mock
* Add NEWS entry
* Fix reversed order of comparison
* Test child and standalone function calls
* Use a helper function extracting mock to avoid code duplication and refactor tests.
Add explicit `asyncSetUp` and `asyncTearDown` methods.
The rest is the same as for #13228
`AsyncTestCase` create a loop instance for every test for the sake of test isolation.
Sometimes a loop shared between all tests can speed up tests execution time a lot but it requires control of closed resources after every test finish. Basically, it requires nested supervisors support that was discussed with @1st1 many times. Sorry, asyncio supervisors have no chance to land on Python 3.8.
The PR intentionally does not provide API for changing the used event loop or getting the test loop: use `asyncio.set_event_loop_policy()` and `asyncio.get_event_loop()` instead.
The PR adds four overridable methods to base `unittest.TestCase` class:
```
def _callSetUp(self):
self.setUp()
def _callTestMethod(self, method):
method()
def _callTearDown(self):
self.tearDown()
def _callCleanup(self, function, /, *args, **kwargs):
function(*args, **kwargs)
```
It allows using asyncio facilities with minimal influence on the unittest code.
The last but not least: the PR respects contextvars. The context variable installed by `asyncSetUp` is available on test, `tearDown` and a coroutine scheduled by `addCleanup`.
https://bugs.python.org/issue32972
* Fix the implicit string concatenation in `assert_has_awaits` error message.
* Use "await" instead of "call" in `assert_awaited_with` error message.
https://bugs.python.org/issue37075
Return a coroutine while patching async functions with a decorator.
Co-authored-by: Andrew Svetlov <andrew.svetlov@gmail.com>
https://bugs.python.org/issue36996
This was achieved by:
* moving many pass statements in tests onto their own lines, so they pass line coverage and can match an easy ignore pattern if branch coverage is added later.
* removing code that cannot be reached.
* removing long-disabled tests.
* removing unused code.
* adding tests for uncovered code
It turned out that removing `if __name__ == '__main__'` blocks that run unittest.main() at the bottom of test files was surprisingly contentious, so they remain and can be filtered out with an appropriate .coveragerc.
When an attribute is deleted from a Mock, a sentinel is added rather
than just deleting the attribute. This commit checks for such sentinels
when returning the child mocks in the __dir__ method as users won't
expect deleted attributes to appear when performing dir(mock).
In Python having a trace function in effect while mock is imported causes isinstance to be wrong for MagicMocks. This is due to the usage of super() in some class methods, as this sets the __class__ attribute. To avoid this, as a workaround, alias the usage of super .
Deprecated passing the following arguments as keyword arguments:
- "func" in functools.partialmethod(), weakref.finalize(),
profile.Profile.runcall(), cProfile.Profile.runcall(),
bdb.Bdb.runcall(), trace.Trace.runfunc() and
curses.wrapper().
- "function" in unittest.addModuleCleanup() and
unittest.TestCase.addCleanup().
- "fn" in the submit() method of concurrent.futures.ThreadPoolExecutor
and concurrent.futures.ProcessPoolExecutor.
- "callback" in contextlib.ExitStack.callback(),
contextlib.AsyncExitStack.callback() and
contextlib.AsyncExitStack.push_async_callback().
- "c" and "typeid" in the create() method of multiprocessing.managers.Server
and multiprocessing.managers.SharedMemoryServer.
- "obj" in weakref.finalize().
Also allowed to pass arbitrary keyword arguments (even "self" and "func")
if the above arguments are passed as positional argument.
Return None after calling unittest.mock.patch.object.stop() regardless of whether the object was started. This makes the method idempotent.
https://bugs.python.org/issue36366
* Resolve string target to patch.dict during function call
* Add NEWS entry
* Remove unneeded call
* Restore original value for support.target and refactor assertions
* Add extra assertion to verify unpatched dict
TextTestRunner of unittest.runner now uses time.perf_counter() rather
than time.time() to measure the execution time of a test: time.time()
can go backwards, whereas time.perf_counter() is monotonic.
Similar change made in libregrtest, pprint and random.
* Fix partial and partial method signatures in mock
* Add more calls
* Add NEWS entry
* Use assertEquals and fix markup in NEWS
* Refactor branching and add markup reference for functools
* Revert partial object related changes and fix pr comments
* tests: Further validate `wraps` functionality in `unittest.mock.Mock`
Add more tests to validate how `wraps` interacts with other features of
mocks.
* Don't call the wrapped object if `side_effect` is set
When a object is wrapped using `Mock(wraps=...)`, if an user sets a
`side_effect` in one of their methods, return the value of `side_effect`
and don't call the original object.
* Refactor what to be called on `mock_call`
When a `Mock` is called, it should return looking up in the following
order: `side_effect`, `return_value`, `wraps`. If any of the first two
return `mock.DEFAULT`, lookup in the next option.
It makes no sense to check for `wraps` returning default, as it is
supposed to be the original implementation and there is nothing to
fallback to.
Two kind of mistakes:
1. Missed space. After concatenating there is no space between words.
2. Missed comma. Causes unintentional concatenating in a list of strings.
The MagicMock class supports many magic methods, but not __fspath__. To ease
testing with modules such as os.path, this function is now supported by default.
Only make sure that the result is in unittest.signals._results, don't
check the full content of unittest.signals._results.
support._run_suite() uses TextTestRunner in verbose mode, but
TextTestRunner.run() calls registerResult(result) which made the test
fail with "odd object in result set".
Call also removeResult() to restore unittest.signals._results to
avoid test side effect.
* Fix multiple typos in code comments
* Add spacing in comments (test_logging.py, test_math.py)
* Fix spaces at the beginning of comments in test_logging.py
Rather than supporting dev mode directly in the warnings module, this
instead adjusts the initialisation code to add an extra 'default'
entry to sys.warnoptions when dev mode is enabled.
This ensures that dev mode behaves *exactly* as if `-Wdefault` had
been passed on the command line, including in the way it interacts
with `sys.warnoptions`, and with other command line flags like `-bb`.
Fix also bpo-20361: have -b & -bb options take precedence over any
other warnings options.
Patch written by Nick Coghlan, with minor modifications of Victor Stinner.
The new method allows the developer to control when to stop the
feature of mocks that automagically creates new mocks when accessing
an attribute that was not declared before
Signed-off-by: Mario Corchero <mariocj89@gmail.com>
Cython will, in the right circumstances, offer a MethodType instance
where im_func is a builtin function. Any instance of MethodType is
automatically assumed to be a Python-defined function (more
specifically, a function that has an inspectable signature), but
_set_signature was still conservative in its assumptions. As a result
_set_signature would return early with None instead of a mock since
the im_func had no inspectable signature. This causes problems
deeper inside mock, as _set_signature is assumed to _always_
return a mock, and nothing checked its return value.
In similar corner cases, autospec will simply not check the spec of the
function, so _set_signature is amended to now return early with the
original, not-wrapped mock object.
Patch by Aaron Gallagher.
bpo-11798, bpo-16662, bpo-16935, bpo-30813: Skip
test_discover_with_module_that_raises_SkipTest_on_import() and
test_discover_with_init_module_that_raises_SkipTest_on_import() of
test_unittest when hunting reference leaks using regrtest.
* #30190 / unittest / assertAlmostEqual(delta=...) / error msg: show the difference between the 2 numbers in case of failure
* safe_repr() diff
* also show difference when passing 'places' argument
* refactoring
* update Misc/NEWS
The concept of .pyo files no longer exists. Now .pyc files have an
optional `opt-` tag which specifies if any extra optimizations beyond
the peepholer were applied.
When used in the real world it can under some situations trigger
" assert not _relpath.startswith('..'), "Path must be within the project"
AssertionError: Path must be within the project
"
Because _get_name_from_path was not expecting to be called with the top level
directory.
We were not honouring load_tests in a package/__init__.py when that was the
start_dir parameter, though we do when it is a child package. The fix required
a little care since it introduces the possibility of infinite recursion.
Previously the ImportError was only shown if the top level containing
package failed to import, with other ImportErrors showing up as
AttributeError - hiding the real cause. As part of this,
`TestLoader.loadTestsFromNames` now captures errors to self.errors.
These tests were undebuggable as written, and there's no testing fallacy
involved in using the method we're testing to test the output of that method,
so switch to that.
This makes it possible to examine the errors from unittest discovery
without executing the test suite - important when the test suite may
be very large, or when enumerating the test ids from a test suite.