Closed stehessel closed 2 years ago
So I'd like to get a bit more insight into the reason for this feature. In my mind the only way to do this (without making the entire interface for adapters more complex) would be to run each selected test as an individual process. Therefore you may as well just run each one separately.
Can you (or others) describe where you see the advantage in terms of workflow?
The use case I envision would be:
This is not about performance. How the tests are executed and whether they are run concurrently would be an implementation detail, you could also execute the tests serially in a single thread. It's about running the tests you want to run, even when they are scattered in different files.
One corner case would be parametrized tests - e.g. using pytest. If I remember correctly PyCharm treats parameters as separate tests and allows to select individual selection for them. I think that's ideal, but it might be easier to select only based on test and file names and then run all parameters.
OK that all makes sense, just wanted to make sure I understood the need properly :+1:
If I remember correctly PyCharm treats parameters as separate tests and allows to select individual selection for them
That'd be a feature request for neotest-python to detect them as separate tests but that sounds very cool
OK I've added the ability to mark tests in the summary and run them. See :h neotest.summary.run_marked()
and :h neotest.Config.summary.mappings
for details
Awesome, thank you!
It would be cool if there was a way to mark tests for execution - for example via a toggle in the summary window - and then run them with a command like
run.run_selected()
. This would increase the flexibility of test execution.