Update computation code retrieval logic

[jrbourbeau]

This PR implements the fix proposed in #5224 (comment) for when __name__ is not available. In this case, we will not attempt to retrieve the user code for a computation and will instead return "<Code not available>"

• Closes KeyError when using distributed scheduler with __array_function__ #5224
• Tests added / passed
• Passes black distributed / flake8 distributed / isort distributed

[jrbourbeau]

It looks like gpuCI is failing due to a segfault?

13:33:36 docker: Error response from daemon: OCI runtime create failed: container_linux.go:380: starting container process caused: process_linux.go:545: container init caused: Running hook #0:: error running hook: signal: segmentation fault (core dumped), stdout: , stderr:: unknown.

13:33:36 FATAL: Failed to start docker image

13:33:36 java.io.IOException: Failed to start docker image

13:33:36 	at com.gpuopenanalytics.jenkins.remotedocker.DockerState.launchContainer(DockerState.java:298)

13:33:36 	at com.gpuopenanalytics.jenkins.remotedocker.DockerState.launchContainers(DockerState.java:199)

13:33:36 	at com.gpuopenanalytics.jenkins.remotedocker.RemoteDockerBuildWrapper.setUp(RemoteDockerBuildWrapper.java:175)

13:33:36 	at hudson.model.Build$BuildExecution.doRun(Build.java:157)

13:33:36 	at hudson.model.AbstractBuild$AbstractBuildExecution.run(AbstractBuild.java:513)

13:33:36 	at hudson.model.Run.execute(Run.java:1907)

13:33:36 	at hudson.matrix.MatrixRun.run(MatrixRun.java:153)

13:33:36 	at hudson.model.ResourceController.execute(ResourceController.java:97)

13:33:36 	at hudson.model.Executor.run(Executor.java:429)

13:33:36 Recording test results

13:33:36 [Checks API] No suitable checks publisher found.

13:33:36 Finished: FAILURE

cc @quasiben @charlesbluca

EDIT: Also appearing over in #5233, so I'll assume it's unrelated to the specific changes in this PR

[charlesbluca]

Yup, this definitely looks like a gpuCI related issue - I've pinged our ops team

[charlesbluca]

rerun tests

[pentschev]

rerun tests

[pentschev]

gpuCI is back and tests pass there. It seems that other failures are not related to this PR.

[pentschev]

Thanks for the fix @jrbourbeau !

[jrbourbeau]

Thanks for the quick gpuCI fix!