Slurm resource named `clusters` in docs, and `cluster` in `getClusters`

[jemus42]

I can't quite decide which way it is.

In clusterFunctionsSlurm.R, the getClusters function is defined which looks for a resource named cluster (singular). https://github.com/mllg/batchtools/blob/1196047ed5115d54bde2923848c1f3ec11fda6d2/R/clusterFunctionsSlurm.R#L38-L42

The documentation of submitJobs refers to clusters (plural): https://github.com/mllg/batchtools/blob/1196047ed5115d54bde2923848c1f3ec11fda6d2/R/submitJobs.R#L25

The problem:

if this resource / getClusters is NULL, there's no cluster specified and functions like findRunning() etc. will return an empty table despite jobs are running. For months I was wondering why my queued Slurm jobs where listed as expired and I assumed I was doing something wrong or my template was outdated. Took me a moment to figure out that I was accidentally correctly misspecifying clusters in my resources = list(...) call 🥴

EDIT: Oh, and the template I'm using relies on resources$clusters.

Suggested fix

I'm not sure. I was about to prepare a PR when I realized that I don't want to

• change the internal getClusters function to look for clusters instead of cluster because that will silently break peoples existing batchtools configs and they might not notice.
• change the docs for submitJobs to (wrongly?) suggest that the resource should be named cluster, implying that multiple clusters can not be specified (which I never tried)

At the very least I thought about introducing an assertion on provided resources to shield against this sort of thing, but I'm not sure how to go about that yet.