Lin Bytes negative test fails to trigger on MSVC bytecode

[jmid]

#516 changed the Lin Bytes test, turning it into a negative test.

The merge to main of the PR then failed to find a counterexample within 1000 attempts on MSVC bytecode 5.3:
https://github.com/ocaml-multicore/multicoretests/actions/runs/12791598373/job/35660087124

random seed: 278988859
generated error fail pass / total     time test name

[ ]    0    0    0    0 / 1000     0.0s Lin Bytes test with Domain
[ ]    0    0    0    0 / 1000     0.0s Lin Bytes test with Domain (generating)
[ ]   75    0    0   75 / 1000    60.1s Lin Bytes test with Domain
[ ]  343    0    0  343 / 1000   121.0s Lin Bytes test with Domain
[ ]  926    0    0  926 / 1000   181.0s Lin Bytes test with Domain
[✗] 1000    0    0 1000 / 1000   188.4s Lin Bytes test with Domain

[ ]    0    0    0    0 /  250     0.0s Lin Bytes test with Thread
[ ]   69    0    0   69 /  250   243.1s Lin Bytes test with Thread
[✓]  250    0    0  250 /  250   300.4s Lin Bytes test with Thread

--- Failure --------------------------------------------------------------------

Test Lin Bytes test with Domain failed:

Negative test Lin Bytes test with Domain succeeded but was expected to fail
================================================================================
failure (1 tests failed, 0 tests errored, ran 2 tests)
File "src/bytes/dune", line 12, characters 7-16:
12 |  (name lin_tests)
            ^^^^^^^^^
(cd _build/default/src/bytes && ./lin_tests.exe --verbose)
Command exited with code 1.

We should just increase the count for a start, as it is not taking particularly long to execute - and doesn't cost anything for platforms finding a counterexample earlier 🤷

[jmid]

Reopening this as this happened on the push to main in connection with the 0.7 release for MSVC bytecode trunk:
https://github.com/ocaml-multicore/multicoretests/actions/runs/12891458000/job/35943440732

random seed: 318271471
generated error fail pass / total     time test name

[ ]    0    0    0    0 / 5000     0.0s Lin Bytes test with Domain
[ ]    0    0    0    0 / 5000     0.0s Lin Bytes test with Domain (generating)
[ ]  609    0    0  609 / 5000    88.8s Lin Bytes test with Domain
[ ]  902    0    0  902 / 5000   157.6s Lin Bytes test with Domain
[ ] 1208    0    0 1208 / 5000   258.0s Lin Bytes test with Domain
[ ] 1735    0    0 1735 / 5000   903.8s Lin Bytes test with Domain
[ ] 2176    0    0 2176 / 5000  1013.6s Lin Bytes test with Domain
[ ] 3118    0    0 3118 / 5000  1077.0s Lin Bytes test with Domain
[ ] 3493    0    0 3493 / 5000  1138.1s Lin Bytes test with Domain
[ ] 4085    0    0 4085 / 5000  1199.2s Lin Bytes test with Domain
[ ] 4247    0    0 4247 / 5000  1373.2s Lin Bytes test with Domain
[✗] 5000    0    0 5000 / 5000  1422.3s Lin Bytes test with Domain

I wonder if there's something fundamental with (MSVC) bytecode or whether it can be solved, e.g., by adjusting weights a bit like src/dynarray/lin_tests.ml does... 🤔

[jmid]

Saw this again on the scheduled 5.3 MSVC bytecode run:
https://github.com/ocaml-multicore/multicoretests/actions/runs/13489954915/job/37686516127

random seed: 478587156
generated error fail pass / total     time test name

[ ]    0    0    0    0 / 5000     0.0s Lin Bytes test with Domain
[ ]    0    0    0    0 / 5000     0.0s Lin Bytes test with Domain (generating)
[ ]  852    0    0  852 / 5000    60.0s Lin Bytes test with Domain
[ ] 1167    0    0 1167 / 5000   120.0s Lin Bytes test with Domain
[ ] 1547    0    0 1547 / 5000   180.0s Lin Bytes test with Domain
[ ] 2300    0    0 2300 / 5000   245.6s Lin Bytes test with Domain
[ ] 2949    0    0 2949 / 5000   329.9s Lin Bytes test with Domain
[ ] 3532    0    0 3532 / 5000   390.9s Lin Bytes test with Domain
[ ] 3967    0    0 3967 / 5000   460.7s Lin Bytes test with Domain
[ ] 4618    0    0 4618 / 5000   520.9s Lin Bytes test with Domain
[✗] 5000    0    0 5000 / 5000   557.5s Lin Bytes test with Domain

The common trait here seems to be MSVC and bytecode... 🤔