Don't compute axle weights from a degenerate truck geometry

[HansBug]

Fixes #539

Summary

• compute_axle_weights divides by harness_rear_axle_distance and by front_axle_middle_axle_distance. A bin file that gives axle weight limits but no truck geometry leaves both at 0, and boxstacks::InstanceBuilder::read_bin_types flags every bin read from a CSV file as a semi-trailer truck, so the divisions produced an infinite rear axle weight and a NaN middle axle weight. The infinite weight exceeds any finite REAR_AXLE_MAXIMUM_WEIGHT, so Solution::feasible_axle_weights rejected every candidate and the solver returned an empty solution with exit code 0 and no message.
• Skip the axle weights the geometry cannot produce instead of computing them from a zero divisor.

Changes

include/packingsolver/algorithms/truck.hpp — two guards in compute_axle_weights. Without harness_rear_axle_distance neither weight can be computed, since both are derived from harness_weight, so the function returns {0, 0}. Without front_axle_middle_axle_distance only the middle axle weight is lost, so the rear axle stays constrained. The original expression and its variable-name comments are kept as they are.

test/algorithms/truck_test.cpp (new) — three unit tests: the complete geometry, and each of the two degenerate cases. The first passes both before and after this change, which is what shows the change is inert for a well-formed truck.

I deliberately left two nearby things alone. compute_axle_weights also divides by weight, which is 0 for an empty bin; that already yields NaN and no violation today. And read_bin_types still flags every bin from a CSV file as a semi-trailer truck, which is why the degenerate state is reachable from an ordinary bin file at all; changing that is a policy decision about existing bin files, so it is yours to make rather than mine.

Test plan

• The two degenerate tests fail on master and pass with this change; Truck.ComputeAxleWeights passes both ways.
• Full build, default configuration (-DCMAKE_BUILD_TYPE=Release, CLP and HiGHS both on, liblapack-dev and libbz2-dev installed, as in .github/workflows/build.yml).
• ctest --parallel 8 in build/test: 599 tests pass, which is the 596 on master plus the three added here.
• The reproducer from [boxstacks] A bin file with axle weight limits but no truck geometry silently returns an empty solution #539 now packs 3/3 items into one bin instead of returning an empty solution.
• Checked that no constraint that works today is silently dropped: a bin with IS_SEMI_TRAILER_TRUCK=1, a full geometry and FRONT_AXLE_MIDDLE_AXLE_DISTANCE=0 still has its rear axle limit enforced, REAR_AXLE_MAXIMUM_WEIGHT=4600 giving one bin and 4400 splitting the load into two, both before and after the change.
• Checked that a bin type providing both distances is unaffected, using the semi-trailer instance from [boxstacks] Out-of-range read in the axle-weight repair loop of sequential_onedimensional_rectangle (std::bad_array_new_length) #537: it behaves identically on this branch and on master.

Environment: Ubuntu 24.04.3 LTS, x86-64, glibc 2.39, GCC 13.3.0, CMake 4.4.3.