When KOBO_BENCH=1 is set in the environment, the game becomes a
hands-off benchmark:
* Skips the intro/menu sequence and pushes st_game directly with
scene=0, skill=CLASSIC, invulnerable ship, always_fire = constant
auto-fire, countdown = 0 so "GET READY" auto-pops in 700ms.
* Seeds pubrand and gamerand from KOBO_BENCH_SEED (default 12345)
so the map, stress spawn pattern, and enemy AI are identical
run-to-run.
* Defaults KOBO_STRESS=1500 unless overridden, so the move pass has
a steady workload, and KOBO_PROF=1 so the profiler is on.
* Counts frames in run_game; after KOBO_BENCH_FRAMES (default 1800)
it dumps prof to stderr and exits cleanly.
Usage from repo root (the cwd matters — relative GFX paths only
resolve from here):
KOBO_BENCH=1 ./build/kobodl 2>bench-logs/phaseN.log
A single command produces a comparable profiler dump at every phase
of the DOD refactor — no menus to navigate, no keyboard input,
identical workload.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Adds a tiny zone-based profiler (prof.{h,cpp}) enabled via KOBO_PROF=1.
Instruments run_game with zones for myship.move, enemies.move (split into
realize and move_pass), hit_structure, and hit_bolt. Adds KOBO_STRESS=N
to keep the enemy pool topped up with rocks for stable per-frame timing.
Why: the entity/render pipeline refactor in BENCH.md needs a comparable
baseline before any layout changes. Without a stress workload the natural
gameplay enemy count is too noisy to compare across commits.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Import existing source tree; original VCS history is no longer available.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
