Fixed ProxyObject.__del__ to use the new Disk IO API from #791

[madsbk]

Anyone that knows how to make a pytest that tests exceptions in class destructors?

[pentschev]

rerun tests

[pentschev]

Anyone that knows how to make a pytest that tests exceptions in class destructors?

@madsbk what exactly do you want to test? You want to ensure no exceptions are raised in the destructor, or do you want to verify that the exception is indeed raised? If it's the latter, during __del__ no exceptions should be raised and thus there's a design issue.

[codecov-commenter]

Codecov Report

❗ No coverage uploaded for pull request base (branch-22.02@6aa330f). Click here to learn what that means.
The diff coverage is n/a.

@@               Coverage Diff               @@
##             branch-22.02     #802   +/-   ##
===============================================
  Coverage                ?   89.83%           
===============================================
  Files                   ?       16           
  Lines                   ?     1997           
  Branches                ?        0           
===============================================
  Hits                    ?     1794           
  Misses                  ?      203           
  Partials                ?        0           

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 6aa330f...23b6b29. Read the comment docs.

[madsbk]

This: You want to ensure no exceptions are raised in the destructor.
I have tried with the capsys and capfd fixtures but it doesn't work.

It is not blocking this PR but it would be nice to have a regression test

[pentschev]

@madsbk we can add error::pytest.PytestUnraisableExceptionWarning to https://github.com/rapidsai/dask-cuda/blob/branch-21.12/pytest.ini , that will error on uncatched exceptions in __del__.

[pentschev]

LGTM @madsbk . It's up to you whether you want to add error::pytest.PytestUnraisableExceptionWarning, or maybe even error::pytest.PytestWarning. I think it wouldn't hurt for us to be more strict on pytest warnings.

[madsbk]

LGTM @madsbk . It's up to you whether you want to add error::pytest.PytestUnraisableExceptionWarning, or maybe even error::pytest.PytestWarning. I think it wouldn't hurt for us to be more strict on pytest warnings.

Thanks @pentschev, let's merge this PR and then try a more strict warning policy in another PR.

[madsbk]

@gpucibot merge