Make req_memory_per_node hook print a warning if memory resource not configured in ReFrame config

[casparvl]

Fixes #181

Haven't had time to test yet...

[casparvl]

Looks as intended when running on Azure MC cluster:

$ reframe -n ESPRESSO.*P3M -t "1_node" --run
...
[ RUN      ] EESSI_ESPRESSO_P3M_IONIC_CRYSTALS %module_name=ESPResSo/4.2.1-foss-2023a %device_type=cpu %scale=1_node /4a9e965c @Magic_Castle_Azure:aarch64-neoverse-N1-16c-62gb+default
WARNING: Your ReFrame configuration file does not specify any resource called 'memory' for this partition  (aarch64-neoverse-N1-16c-62gb). Without this, an explicit memory request cannot be made from the scheduler. This test will run, but it may result in an out of memory error. Please specify how to requst memory (per node) from your resource scheduler by defining a resource 'memory' in your ReFrame configuration file for this partition. For a SLURM system, one would e.g. define: 'resources': [{'name': 'memory', 'options': ['--mem={size}']}]

[casparvl]

On a system that does have the memory specified as resources:

$ reframe -n ESPRESSO.*P3M -t "1_node" --run
...
[----------] start processing checks
[ RUN      ] EESSI_ESPRESSO_P3M_IONIC_CRYSTALS %module_name=ESPResSo/4.2.2-foss-2023b %device_type=cpu %scale=1_node /3b90ffd6 @vega:cpu+default
[ RUN      ] EESSI_ESPRESSO_P3M_IONIC_CRYSTALS %module_name=ESPResSo/4.2.2-foss-2023a %device_type=cpu %scale=1_node /f1621689 @vega:cpu+default
[ RUN      ] EESSI_ESPRESSO_P3M_IONIC_CRYSTALS %module_name=ESPResSo/4.2.1-foss-2023a %device_type=cpu %scale=1_node /4a9e965c @vega:cpu+default

I.e. it doesn't print a warning - as expected.

Imho this PR is ready to be merged.

[casparvl]

Failing CI is fixed in #183, that should be merged first.

Edit: was just merged by @bedroge . Rerunning CI check now, I'm assuming it'll pass now.

[satishskamath]

This warning is required. lgtm!