Walkable lit field validation, 2026-10-24
Environment: Ubuntu 24.04 x86_64 cloud container, cc 13.3.0, Python 3.13,
SDL 2.30.0, dummy video with the software renderer. Not run on the local WSL
host or native Windows. Tested hashes are in validation.json.

1. make -C native clean; make -C native all test test-runtime
   Passed. 16 native C suites. Python core 43; runtime: texture viewer 7, BCH
   texture viewer 6, model viewer 10, BCH model viewer 9, BCH textured 5, BCH
   combiner 4, BCH lighting 2, BCH walk 2 (new), field viewer 9. 97 Python
   tests in total.
2. make -C reconstructed test: passed.
3. python3 -m unittest discover -s tests: 38 tests, 2 optional skips.
4. ASan/UBSan with leak detection: the sanitized viewer under the walk,
   lighting, combiner, textured, BCH-model and model viewer suites and a
   real-input --walk run passed without diagnostics; capture identical.
5. All 539 reconstructed and 85 earlier native reproduction entries match;
   the stage-0, combiner and lit day/night captures are byte-identical.
6. python3 -I -B native/reproduction/bch-walk/verify_walk.py --romfs <RomFS>
   --viewer native/build/pokemon_model_viewer
   Passed: 52 standard cells, all reachable from (3,10); at each, the viewer's
   cell, center and moves match and its collision height (0) equals a flat
   model surface; 25 carpet cells also have a surface at 0.5. Capture
   (--steps DDDRRR) sha256 55f1b9b264592a85a153c985617ee87729f10ea600955401a897a3aa095fdb5e.

Not established: the game's movement and collision rules, entry, player
model, field camera, culling and any running-game comparison.
