Upgrade pyiron_base to 0.6.14

[jan-janssen]

No description provided.

[jan-janssen]

Waiting for conda-forge/pyiron_base-feedstock#180

[jan-janssen]

@pmrv There seems to be an issue with the new read-only objects pyiron/pyiron_base#1162

======================================================================
ERROR: test_multiple_jobs (testing.test_serialMaster.TestSerialMaster.test_multiple_jobs)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/runner/work/pyiron_atomistics/pyiron_atomistics/tests/testing/test_serialMaster.py", line 66, in test_multiple_jobs
    job_ser_reload.append(ham)
  File "/usr/share/miniconda3/envs/my-env/lib/python3.11/site-packages/pyiron_base/jobs/master/generic.py", line 171, in append
    self.server.cores = job.server.cores
    ^^^^^^^^^^^^^^^^^
  File "/usr/share/miniconda3/envs/my-env/lib/python3.11/site-packages/pyiron_base/interfaces/lockable.py", line 54, in dispatch_or_error
    raise Locked(
pyiron_base.interfaces.lockable.Locked: Object is currently locked!  Use unlocked() if you know what you are doing.

======================================================================
ERROR: test_convergence_goal (testing.test_serialMaster_non_modal.TestSerialMaster.test_convergence_goal)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/runner/work/pyiron_atomistics/pyiron_atomistics/tests/testing/test_serialMaster_non_modal.py", line 90, in test_convergence_goal
    self.project.wait_for_job(job_ser, interval_in_s=5, max_iterations=50)
  File "/usr/share/miniconda3/envs/my-env/lib/python3.11/site-packages/pyiron_base/project/generic.py", line 1457, in wait_for_job
    wait_for_job(
  File "/usr/share/miniconda3/envs/my-env/lib/python3.11/site-packages/pyiron_base/jobs/job/extension/server/queuestatus.py", line 237, in wait_for_job
    raise ValueError(
ValueError: Maximum iterations reached, but the job was not finished.

----------------------------------------------------------------------

[pmrv]

It seems the error originates here where an already run master job has new jobs appended and is re-run. I would actually argue that this is exactly the sort of problem read-only objects are supposed to catch. I could fix this in base, but I really think the test is wrong here in atomistics and should be changed. Thoughts?

[pmrv]

So from what I have reconstructed so far the test actually needs to be written in the original way, because calling run on SerialMaster only runs the last appended job, possibly in a restart loop if a convergence goal is set. This appears to be by design. The only actual implementations ConvEncutSerial/ConvergenceVolume also just restarts the same job. I have no idea what this test is even supposed to check.

I've opened a quick fix in base, however as far as I can tell these two classes are the only two serial masters in existence (create_pipeline uses FlexibleMaster), so I'd actually prefer to simply remove both classes here and SerialMasterBase from base entirely.

[niklassiemer]

I would in this case not release a new version. Even if we manage to fix the issue and release before christmas, I will not apply it on the cluster before christmas.

[pmrv]

I would prefer to have a version available that has #1249. I'd say we release the current HEAD (without this PR) and install that on the cluster, that means for now pyiron_base will still be 0.6.13 on the cluster, but I don't see a problem with that.