Allow example tests to be run on subprocesses

[leofang]

Right now we just exec all tests in the main test process

cuda-python/cuda_core/tests/example_tests/utils.py

Line 35 in 213baf4

exec(script, env if env else {})

which is fine if there is no subtle bugs in the examples. However, it is not the end user experience (ex: from user perspective CUDA contexts are implicitly created/tear down between each sample runs). We should allow configuring it in the future.

[cpcloud]

Is this still in use?

It seems like there's not much going on in that script beyond the standard python script.py run. Happy to take a crack at deleting this file and replacing whatever is invoking it with something a bit less custom.

[rwgk]

JIC/FYI: In the meantime I added this for testing cuda_pathfinder dynamic library loading in independent processes:

cuda-python/cuda_pathfinder/tests/spawned_process_runner.py

Lines 69 to 77 in ecb8113

	"""Run `target` in a spawned child process, capturing stdout/stderr.
	The provided `target` must be defined at the top level of a module, and must
	be importable in the spawned child process. Lambdas, closures, or interactively
	defined functions (e.g., in Jupyter notebooks) will not work.
	If `rethrow=True` and the child process exits with a nonzero code,
	raises ChildProcessError with the captured stderr.
	"""

I guess it could also be useful to isolate the cuda_core examples from each other?

[leofang]

I think this is done by Mike in #1800 and #1842. Closing!