| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
--no-cover-print was added after 1.3.7.
|
|\
| |
| | |
Make tests work with Python ≥ 3.5
|
| | |
|
|/
|
|
|
|
| |
To make output on Python ≥ 3.5 the same as on previous Python versions.
This fixes #928.
|
|\
| |
| | |
ENH: Allow not printing coverage report
|
| | |
|
| | |
|
|/ |
|
|\
| |
| | |
Use a new test loader for each queue item.
|
|/
|
|
|
|
|
|
|
| |
The loader's ContextSuiteFactory keeps a reference to all of the test
suites, even after they've been run, which consumes more and more
memory over time.
This change creates a new loader for each suite that's dispatched to
the worker, which allows the GC to free the old test suites.
|
|\
| |
| |
| | |
Adds an option to specify the location of the coverage config file.
|
| | |
|
| | |
|
|\ \
| |/
|/| |
Adds option to allow prefixing the class name in an xunit report
|
| | |
|
|/ |
|
| |
|
| |
|
|\
| |
| |
| | |
Fake stdout should always have an encoding attribute.
|
| | |
|
|/
|
|
| |
Some programs expect this attribute, and use it to encode strings before writing to stdout. If it's not there, they fail with an AttributeError.
|
|\
| |
| |
| | |
Tee doesn't have encoding and errors attributes.
|
| | |
|
| |
| |
| |
| |
| |
| | |
On Python3, distutils.log uses the encoding and errors attribute of
sys.stdout and stderr. Tests error out when nose is run with xunit and
it replaces the stdout/stderr with a Tee.
|
|\ \
| | |
| | | |
Fix logcapture plugin to capture output from non-propagating loggers,…
|
| |/
| |
| |
| | |
not output 'no handlers could be found' message
|
|\ \
| | |
| | |
| | | |
Ensure idfile is closed if an error occurs.
|
| | | |
|
| | | |
|
|\ \ \
| |/ /
|/| | |
docs: add information about how to (continuously) run a single test suite
|
| | | |
|
|\ \ \
| | | |
| | | | |
Fix test repr with mutable args
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Previously, both `FunctionTestCase` and `MethodTestCase` created local
`arg` variables in their `__str__` methods. These variables were simply
references to `self.arg_repr`, so this commit deobfuscates this reference
by modifying the `_descriptors` method for each so that *only the test
name* is returned and not the `arg_repr` also (renaming `_descriptors` to
`_descriptor` in the process), and then referring to `self.arg_repr`
directly.
This makes the `__str__` easier to understand because it is immediately
obvious that we're only dealing with the string-representation of the args
and *not* the actual args themselves.
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Tests were added in the previous commit which showed that modifying a
mutable object given as an argument to a test case during the execution of
the test case results in the __repr__ of the test case also changing (if
the __repr__ of the object also changed). This was caused by creating a
__repr__ of the test args on the fly, and because the summary line is
printed *before* the test is executed and failure details are printed
*after* the test is executed, this can result in the two lines having
different __repr__s of the passed in arguments (see the previous commit
message for an example of the output).
This commit fixes the issue by storing a __repr__ of the argument tuple
when the *TestCase is instantiated and using that to create all __repr__s
of the *TestCase. Note that the `if` for when no args are given had to be
changed because `repr(tuple())` is '()', which is truthy.
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
At present, if a mutable argument is changed during the execution of test
case, then any printing of the test case object which occurs after the
test has run will result in the argument being printed with a __repr__ of
it's *new* state (assuming that its __repr__ is state-dependant). This is
particularly confusing when debugging failing test cases because the
summary lines and the failure-detail lines will show *different* __repr__s
of the args, e.g.
nose_weirdness.test_nose_is_weird(Foo(10), 0) ... FAIL
======================================================================
FAIL: nose_weirdness.test_nose_is_weird(Foo(0), 0)
----------------------------------------------------------------------
...
Ran 1 test in 0.001s
FAILED (failures=1)
This issue will be solved in the next commit.
|
|\ \ \ \
| |_|/ /
|/| | | |
Fix NOSE_TESTMATCH default regex.
|
| | | | |
|
| | | |
| | | |
| | | |
| | | |
| | | | |
This fixes the -m flag documentation for both the README.txt and
nosetests.1 manpage.
|
| | | |
| | | |
| | | |
| | | | |
This fixes the top half of nosetests.1 and README.txt.
|
| | | | |
|
| | |/
| |/|
| | |
| | | |
See bug #11.
|
| | | |
|
| | | |
|
|\ \ \
| | | |
| | | | |
nose.suite: Clean up forgotten set_trace()
|
| |/ / |
|
| | | |
|
|\ \ \
| |_|/
|/| | |
Fix #931: duplicate __init__ modules can be picked up
|
|/ /
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The short form here is that when setting up ignores, you may allow
__init__.py to be included unexpectedly. Since we generally use
loadTestsFromName(), the package should be imported first which will
handle the __init__.py case. Therefore, let's always ignore it in the
loadTestsFromDir() method. Just to make sure that we ignore the correct
files, we need to run the entry name through src() to make sure we
handle the situation correctly on other platforms (like Jython).
|
|\ \
| |/
|/| |
[fix travis] Coverage 4.0 is available and is not supported in python 3.2
|
| |
| |
| |
| | |
levels.
|