Ben Sherman

They are different kinds of limits, and I don't think it's possible to collapse them into one setting. This PR for CPU request/limit is intended to recover an edge case...

Let's see how these examples look like with my proposal ### Example 1 System-wide CPU limit, no K8s limit ```groovy process.resourceLimits = [cpus: 8] process.cpus = 2 ``` ### Example...

Both approaches are complex but I think mine is still less confusing: - There is exactly one way to specify a system-wide limit, the `resourceLimits` directive - Being a process...

Fixed by https://github.com/nextflow-io/nextflow/commit/42504d3c83145e16179e22df83e9d35ecc995eca

Not quite. This PR pertains only to nullable paths, whereas that issue pertains to optional inputs in general, and I don't think we have fully worked out the meaning of...

@pditommaso I cleaned up this PR and updated the e2e tests. Also fixed a bug with the nullable input path. I think this PR is ready to go. Here's the...

In case the unit tests don't run, I confirmed that they pass locally: ```console $ make smoke NXF_SMOKE=1 ./gradlew test BUILD SUCCESSFUL in 3m 14s 91 actionable tasks: 39 executed,...

While I understand the difference between optional and nullable, I'm wondering if we shouldn't just extend `optional` to be allowed within tuple elements and handle it accordingly, rather than adding...

I would avoid that issue by having the nullable behavior only apply to nested params in a tuple, that way there is no distinction between `val(x), optional: true` and `val(x,...

I do need to fix the merge conflict, but otherwise it's just waiting for @pditommaso 's approval.