src/lockfree/ws_deque_test.ml may take long (or loop?)

[jmid]

Third and final goodie from #88:

• another MacOSX run on beta1 timed out during what I think was a src/lockfree/ws_deque_test.ml
  https://github.com/jmid/multicoretests/actions/runs/3444725863/jobs/5747632278#step:8:768

Originally posted by @jmid in #88 (comment)

By just trying multiple runs locally I was able to trigger some that took a long time, to the point that I thought it was looping - only to discover that it was just dog-slow.
Here's an example that took 10mins on my machine:

$ dune exec src/lockfree/ws_deque_test.exe -- -v -s 385505135
random seed: 385505135
generated error fail pass / total     time test name
[✓] 1000    0    0 1000 / 1000     0.3s sequential ws_deque test
[✓] 1000    0    0 1000 / 1000   605.1s parallel ws_deque test
[✓]    1    0    1    0 / 1000     4.5s parallel ws_deque test, negative

--- Info -----------------------------------------------------------------------

Negative test parallel ws_deque test, negative failed as expected (33 shrink steps):

                        |                
                     (Push 4)            
                     (Push 5)            
                        |                
             .---------------------.
             |                     |                
            Pop                   Pop               


+++ Messages ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Messages for test parallel ws_deque test, negative:

  Results incompatible with linearized model

                              |                            
                        (Push 4) : ()                      
                        (Push 5) : ()                      
                              |                            
           .------------------------------------.
           |                                    |                            
     Pop : Ok (5)                         Pop : Ok (0)                       

================================================================================
success (ran 3 tests)

I suspect this is caused by the generator checking whether all commands have their precondition satisfied on all interleavings IIRC.
However I'm unsure whether the CI hits a similar - but even slower example - or whether it is hitting a different issue since it is taking so long that it times out...

[jmid]

Closing as the lockfree tests are not part of the repo since #265