tests: harden memory leak check in test_siblings_without_eval

[booxter]

When running the test suite on darwin as part of nixpkgs package build,
we often hit the following error:

       >         mx.synchronize()
       >         t()
       >         gc.collect()
       >         expected = get_mem()
       >         for _ in range(100):
       >             t()
       >         used = get_mem()
       > >       self.assertEqual(expected, used)
       > E       AssertionError: 273678336 != 273694720

You can find an example of the full failure log here:
https://hydra.nixos.org/build/320731525

The difference between actual and expected is 16k (one page size on
aarch64-darwin).

The test seems brittle as it depends on peak memory, not real memory
at the moment of measurement, so anything during the run of the process
that could push the allocator to allocate another page can break the
test.

This patch switches the test case to measure real memory using psutil on
all platforms.

Signed-off-by: Ihar Hrachyshka ihar.hrachyshka@gmail.com

Proposed changes

Please include a description of the problem or feature this PR is addressing. If there is a corresponding issue, include the issue #.

Checklist

Put an x in the boxes that apply.

• I have read the CONTRIBUTING document
• I have run pre-commit run --all-files to format my code / installed pre-commit prior to committing changes
• I have added tests that prove my fix is effective or that my feature works
• I have updated the necessary documentation (if needed)

[awni]

Looks good, thanks!

[booxter]

Looks like I broke something that was not present before I rebased to master. Will fix.