- The roof is rescanned only in area columns that changed since the last scan: a quiet Huge
world goes from 4.60 to 1.88 ms per scan (every 20 ticks).
- The heat pass skips react() for solids and powders below their first reaction threshold:
Huge world digging 2.54 -> 2.34 ms/tick, quiet 1.39 -> 1.18.
- Every suite's output is identical apart from timing lines; seeds make the same worlds.
- test/t_sim_optimizations.js checks the dirty roof against a full scan and the heat loop
against the reference loop byte for byte.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The game as of artifact version 24, ready for local development with
coding agents.
- Source split from three large files into 52 small ones under src/sim,
src/render and src/ui, joined in the order in src/manifest.json by
build.py. The joined code is unchanged apart from comments; every
headless suite gives the same output, and screenshots in both looks
are pixel-identical to before.
- Light tidying with no change in behaviour: one movement scan loop,
shared hazard-stripe and belt shader helpers, unused code removed, a
shared check helper for the test suites.
- World files can be exported as an ordinary download outside claude.ai.
- Tests: broken old scripts removed, experiments moved to tools/,
test/run_all.py added. Tools: shot.py (repeatable screenshots),
where.py, map.py.
- Docs: AGENTS.md, CLAUDE.md, NOTES.md (the owner's to-do inbox),
docs/PROJECT_MAP.md, DESIGN.md, TESTING.md, systems notes, the
owner's requests and an agent log.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01K9kLmYADKSEjuwUN3rFxDA
44r0n7andClaude Opus 5.5