Replace msm execution contexts with cudaq::dem_from_kernel in app examples

[bmhowe23]

The surface_code-2/3 C++ app examples and the surface_code_1.py Python example built their detector error models by running the kernels under the "msm_size"/"msm" execution contexts and hand-assembling the detector and observable matrices from the raw measurement syndrome matrix. The new cudaq::dem_from_kernel API supersedes that: the demo kernels now annotate themselves with cudaq::detector/detectors and cudaq::logical_observable when a declare_detectors flag is set (detectors pair each round against the previous one, with the lock-in round as the reference; prep0's deterministic first-round Z stabilizers are declared as singles where the old code used them), and the hosts parse the returned Stim DEM text with dem_from_stim_text. The runtime D_sparse now comes straight from the analysis' measurements-to-detectors (m2d) map instead of generate_timelike_sparse_detector_matrix, remapped into each decoder's own enqueue stream for surface_code-3's split Z/X decoders (which run one dem_from_kernel pass per decoder via declare_detectors_z/_x flags).

Notable constraints baked into the kernels:

• dem_from_kernel rejects kernels tagged qubitMeasurementFeedback, so the final data measurements are packed branch-free instead of via to_integer(to_bools(...)), whose discriminate-into-call flow sets the tag.
• The Python kernels avoid measure-handle lists crossing kernel-call boundaries and avoid iterating returned handle lists: both leave loops and heap traffic that do not survive the full loop unrolling the adaptive QIR profile (quantinuum target) requires. The DEM slice of demo_circuit_qpu is instead a single custom_memory_circuit_stabs call covering lock-in plus one decoder window, which is gate-for-gate identical to the live path.

The now-unused compute_msm/construct_mz_table Python bindings are removed.

Testing: all 54 app_examples ctest cases pass (local, sliding-window, cqr in-process and two-process, quantinuum-emulate at d=3/5, prep0/prepp), and the 12 surface_code-1-test.py pytest cases pass in ~10s (with /usr/local/cudaq and the built python dir on PYTHONPATH).