cp_model_solver_test.cc: NoOverlap2dCpModelTest.RequiresLns is flaky

[Mizux]

What version of OR-Tools and what language are you using?
Version: main and v99bugfix
Language: C++

Which solver are you using (e.g. CP-SAT, Routing Solver, GLOP, BOP, Gurobi)
CP-SAT

What operating system (Linux, Windows, ...) and version?
Linux & MacOS seems impacted

What did you do?
usual CMake based build

cmake -S. -Bbuild -DBUILD_DEPS=ON
cmake --build build --config Release -j8

# Seems to fail every 3/4 runs
while true; do (cd build && ctest -C Release -R "cxx_sat_cp_model_solver_test"); done

What did you expect to see
all test passs

What did you see instead?

#12 122.3 [----------] 1 test from NoOverlap2dCpModelTest
#12 122.3 [ RUN      ] NoOverlap2dCpModelTest.RequiresLns
#12 122.3 /home/project/ortools/sat/cp_model_solver_test.cc:2307: Failure
#12 122.3 Expected equality of these values:
#12 122.3   response.objective_value()
#12 122.3     Which is: 121
#12 122.3   120
#12 122.3 
#12 122.3 [  FAILED  ] NoOverlap2dCpModelTest.RequiresLns (31 ms)
#12 122.3 [----------] 1 test from NoOverlap2dCpModelTest (31 ms total)

trace: https://github.com/google/or-tools/actions/runs/15299444066/job/43036129045#step:7:10602

MacOS trace with 123...
https://github.com/google/or-tools/actions/runs/15389037373/job/43294059677#step:6:10548

Archlinux docker on v99bugfix branch
https://github.com/google/or-tools/actions/runs/15278476271/job/42971024148#step:7:10634

Make sure you include information that can help us debug (full error message, model Proto).
note1: I've also seen 124, 123 and 122,
note2: this check is after the Check for OPTIMAL --> missing a barrier/semaphore on the objective "best" value ?

ed on google3: cp_model_solver_test.cc;l=2293