Allow failure for specific dependencies

[avcopan]

Context: Am am working on adapting a code that I contribute to to use HyperQueue. I build a dependency graph and then add jobs through the Python API using job.program() as you can see here.

In the workflow I am running, I have one particular step where I want to wait for a particular set of tasks to complete before moving on, but I don't care whether they fail or not (the subsequent task determines whether or not it was successful):

What I am looking for is some sort of a flag that I can pass to these tasks -- something like job.program(..., ignore_failure=True), but I haven't found anything in the API documentation, or in the discussions/issues.

[spirali]

The solution is to wrap your program by something that swallow the failure, e.g. a bash script that runs your inner command and ignores the return code.