How do I debug py.test in PyCharm when coverage is enabled?
Coverage is enabled using --cov=project --cov-report=term-missing, removing this and breakpoints are hit.
Versions: pycharm 5.0.3, pytest==2.8.5, pytest-cache==1.0, pytest-cov==2.2.0, pytest-pep8==1.0.6, pytest-xdist==1.13.1, python-coveralls==2.6.0.
(thanks for jon's advice on further diagnosing the issue)
There is now a flag in py.test to disable coverage which you can activate when running tests from PyCharm.
The flag to use is --no-cov. If you want this to apply to all your test runs you can add this to the default pytest configuration as below:
Extra tip: You probably also want a -s flag in there so output isn't swallowed by py.test. See https://stackoverflow.com/a/17810324/238166 for details.
In case you receive an "unrecognized argument" error, you may need to install pytest-cov, e.g. by pip install pytest-cov.
Related
The results of python tests are cached for only a short time in the VSCode MS python extension.
Whether they succeeded or failed is only cached for a short while, and then they revert back to question marks again.
How to retain the results of the tests?
I think it has to do with this some logs that I see in the Output for "Python Test Log", which shows many (~30) of these lines:
python /home/.../.vscode/extensions/ms-python.python-2020.6.89148/pythonFiles/testing_tools/run_adapter.py discover pytest -- --rootdir /home/projectdir --cache-clear -s
In particular the --cache-clear is suspicious, however I don't know what is triggering these outputs.
So my question is: how to remove --cache-clear from the call? Or maybe something else is going on entirely?
I too found myself in need of disabling this vscode behaviour.
I'm using pytest with the pytest-html plugin, and after all tests have run, as soon as i change a test file and save, this command is automatically run by vscode:
python c:\Users\username\.vscode\extensions\ms-python.python-2021.6.944021595\pythonFiles\testing_tools\run_adapter.py discover pytest -- --rootdir c:\Users\username\app\app-test-automation -s --cache-clear --html=report.html --self-contained-html tests
Which inevitably overwrite the .html test report.
I haven't found a way to remove the --cache-clear from the call.
BUT
You can disable the auto discovery of test (guilty of rerunning the command with the --cache-clear flag) by changing the following vscode setting:
Picture of the setting from the vscode settings menu
Everything works fine with PyCharm and pytest, except if I have failing tests, then it duplicates the error output:
One of actual failures if the red one and other is white. This is really annoys, and I haven't found any way to disable such behaviour.
There is an option to disable log via py.test, however it will disable all logging.
Note: everything works as expected if I run python -m pytest test.py.
I think that is a feature not a bug. The top level is being emitted during the testing which allows you to review the failure before the testing is complete. The second copy of the results is the summary which effectively removes any of the text that was showing test progress.
You can easily view just part of the test output by clicking on the test hierarchy:
The duplicated output can be elimited by running pytest with the -q or --quiet parameter.
You can configure the parameter to be applied to all PyCharm pytest tests by setting it in Edit Configurations --> Templates --> Python Tests --> pytest --> Additional Arguments.
This will then apply those arguments to all new run configurations. If you have a bunch of existing test run configurations, deleting all of them and then re-generating them by running a test or tests using the gutter icon is the quickest way to reset the output.
I have a Python project and a tests task, set up to run pytest from the project's working directory.
Doing Run 'tests' with coverage from the Run menu successfully runs the tests, and the console results shows that coverage was measured - e.g. 53% cover for mws.py.
The automatically applied coverage (as on the right) is 0% for all files, I'm not sure why. I'm using IntelliJ 2017.2.2 EAP.
NB: there is a related five year old question here, but the top rated solution there doesn't apply. There is no error message in the results console in this case.
I had a similar issue, but the accepted solution here didn't solve it.
I had pytest automatically run coverage in its configuration file.
In PyCharm, I added a Run Configuration to run all my tests with pytest.
It seemed to work, and I saw all tests running and got their results to display in PyCharm's run window.
But soon I noticed there were two problems:
When I selected "Run with coverage" I got an error like "coverage results not found", and all files showed 0% coverage.
Breakpoints in tests were not hit when running test in Debug mode.
Both problems disappeared when I added --no-cov to the "Additional Arguments" passed to to pytest (this option is in the Run Configuration).
So It seems the fix was to tell pytest to not run coverage when running it from PyCharm. Both "Run with coverage" option and the Breakpoints in tests now work.
I think the problem lies in you use pytest-cov, so Pycharm cannot parse the result which is shown in text like 53% generated by pytest-cov;
So Changes option in pytest.ini to addopts = -s -v when you want to use Pycharm built in coverage tools.
In this command -v stands for verbosity and -s for disabling all output.
Take a look at my answer for other question about same issue: https://stackoverflow.com/a/45729723/1229510
Basically if you use symlinks - coverage display won't work.
When running tests with py.test there is a --pdb option to enter pdb on failure.
Is there a similar way to enter the debugger when running the same test from within PyCharm?
There is a py.test plugin, pytest-pycharm, that will halt the PyCharm debugger when a test emits an uncaught exception.
Follow the steps below to setup the correct run configuration:
Run > Edit Configurations...
Click the '+' button to add a new configuration.
Name the configuration and specify the configuration parameters below:
Script: The path to your py.test executable (find by typing which py.test)
Script Parameters: This parameter is your test case followed by the --pdb option (ie /Users/Johan/projects/misc/testing.py --pdb)
After setting up the configuration, you can run the test case from within PyCharm. If a test case fails, your PyCharm run window will show the pdb prompt.
I'm using Eclipse / PyDev and PyUnit on OSX for development. It was recommended to me that I use Nose to execute our suite of tests.
When I configure Nose as the test runner, however, output from the interactive console (either standalone or during debugging) disappears. I can type commands but do not see any output.
Is this normal, or am I missing some configuration?
I eventually found in the Preferences > PyDev > PyUnit menu that adding -s to the Parameters for test running stopped this. The parameter prevents the capture of stdout that nose does by default.
The alternate --nocapture parameter should work too.