Frequent segfaults in CI

[fingolfin]

We've been seeing a worrying number of segfaults in CI recently. Let's start collecting some to get a better id of which ones occur and how often. Perhaps then we get a better idea how to tackle them. (I should say that I've also had the Oscar test suite segfault on my MacBook. Unfortunately never with a debugger attached and no good stack traces, and it doesn't always crash, so it's tedious to learn more this way :-(.

From this log:

      From worker 3:	GC: pause 286.04ms. collected 532.221947MB. incr 
      From worker 5:	Test Summary:          | Pass  Total  Time
      From worker 5:	Serialization.Upgrades |    3      3  3.1s
      From worker 7:	
      From worker 7:	GC: pause 98.44ms. collected 434.767586MB. incr 
      From worker 7:	
      From worker 7:	[2631225] signal (11.1): Segmentation fault
      From worker 7:	in expression starting at /home/oscarci-tester/ssd-data/ssd-runner-08/_work/Oscar.jl/Oscar.jl/experimental/GITFans/test/runtests.jl:1
      From worker 7:	__gmpn_copyi_zen at /home/oscarci-tester/ssd-data/ssd-runner-08/_work/_tool/julia/1.10.9/x64/bin/../lib/julia/libgmp.so.10 (unknown line)
      From worker 7:	Allocations: 435428402 (Pool: 435106894; Big: 321508); GC: 111
      From worker 5: