9482b8a2bf
faking a crash Audit item 8, plus two bugs the last commit introduced and one it exposed. REQUESTPOINTERLOCK RETURNS A PROMISE. Both new call sites wrapped it in try/catch, which catches nothing — the rejection went to window.onunhandledrejection, which the previous commit had just taught to paint a full-screen FEED LOST. So a browser declining a lock for any of its ordinary reasons (document not focused; Chrome's ~1s lockout after the player leaves a lock with Escape) presented to the player as the game crashing. Caught by clicking: "the root document of this element is not valid for pointer lock" in 15px type over a game running perfectly well underneath. All three sites — title dismissal, resume, and input.js's canvas click — now swallow the rejection. A refused lock costs one click, which is what input.js has always been for. PAUSE ON BLUR HAD NO CARD AND TWO OWNERS. boot.js emitted ui:pause directly, so the loop halted while cards.js's own `paused` stayed false: the game stopped with nothing on screen to say why, and the next P press raised a card that then took a SECOND press to leave. Moved to cards.js, which owns the pause state — setPaused raises the card and emits. It already refuses while the title or debrief is up, so alt-tabbing away from a card correctly does nothing. Verified: blur raises PAUSED, one P resumes, exactly one true and one false on the bus. THE PAUSE CARD WAS NOT MODAL. #ui is pointer-events:none and it inherits, so a click anywhere but the RESUME button fell through to the canvas — whose mousedown handler re-requests pointer lock whenever it is not held. Clicking beside the button captured the pointer, hid the cursor, and left the player paused with nothing to press RESUME with. The dimmer takes pointer-events now (our own handlers are capture-phase on window, so they still see every click; only the canvas stops seeing them). The title keeps no dimmer and stays click-through by design. Pausing also now releases the lock and resuming takes it back. Pausing with P used to keep the pointer captured — a RESUME button that could not be pointed at — while Escape looked fine only because the browser drops the lock for you. Same key, two behaviours, no stated reason. And losing the lock now drops the queued EDGES, not just held keys: the Space you pressed to reach the RESUME button was still in the queue when the card went, so the run restarted by spending a boost you did not ask for. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> |
||
|---|---|---|
| .claude | ||
| docs | ||
| pipeline | ||
| tools | ||
| web | ||
| .gitignore | ||
| README.md | ||
GUTS — a fantastic voyage through the alimentary canal
Browser flight-combat game (three.js, zero build step). You pilot ENDO-1, a microscopic probe, mouth → rectum through five anatomically-grounded biomes: ride peristalsis down the esophagus, cross the acid sea of the stomach, thread the villi forest of the small intestine, survive the gas swamps of the colon. Fantastic Voyage × Star Fox × Descent.
- Design: docs/GDD.md · Look: docs/ART_BIBLE.md
- Architecture + contracts: docs/TECH.md
- Asset generation: docs/PIPELINE.md
- How we work (lanes/rounds): docs/PROCESS.md — read this first if you are a lane
- Current round: docs/LANES/ROUND1_INSTRUCTIONS.md
Run it
cd web && python3 -m http.server 8140
# http://localhost:8140 — the game
# http://localhost:8140/?stub=1&dbg=1 — stub world + debug HUD (works from day 0)
No bundler, no npm, no build. ES modules + import map; three.js r175 vendored in web/vendor/
(never CDN — the game must run offline). Before you finish a session: bash tools/qa.sh.
The one-paragraph tech pitch
The whole canal is a spline. World position is (s, θ, ρ) — distance along the canal,
angle and radius in the cross-section. Tube walls are chunked extrusions along
parallel-transport frames; peristalsis and breathing are vertex-shader displacement
(cheap, squishy, everywhere); collision is an analytic radius test (no physics engine).
Caverns (mouth, stomach) switch the same controller to free 6DOF. Look is synthetic
scanner — dark endoscope void, SEM-style monochrome walls with FLUX-generated detail
textures, emissive-coded enemies. Everything is seeded + deterministic, and every asset is
optional at runtime — the game boots and plays asset-free with procedural fallbacks.
Hard laws (all lanes)
- Determinism. No
Math.randomoutsidejs/core/rng.js. Seeded RNG only. qa.sh greps. - Assets are optional. Missing texture/model/audio ⇒ procedural fallback, never a crash.
- Eyeball everything. Visual work isn't done until a screenshot is in
docs/shots/laneX/. - Perf budget: ≤300 draw calls, 60 fps on M-series / mid laptop. Measure via
window.DBG. - Free-first. Local MODELBEAST generation before any paid API; fal.ai is a gated fallback (John's go). See PIPELINE.md.
- Git: stage only your exact paths, commit + push each session, never
git add ..
Deploy
Live: https://partly.party/gutsy/ — bash tools/deploy.sh (QA gate → rsync web/ minus
dev/ → docker cp into forum-nginx → verify). No build step; web/ ships as-is.
Two things to know before you touch it, both documented at length in the script header:
- It's ephemeral. Content lives in the container's own layer (like blobbo/glytch), so a
docker compose up --force-recreateor nginx image bump wipes it — re-run the script. The durable fix is a bind mount inforum/docker-compose.yml(what cratewars/roguelike use), which costs a forum-nginx recreate and brief partly.party downtime. - A 404 there returns HTTP 200. nginx's catch-all
try_files ... /index.htmlserves the arcade index for any missing file. Verify on content, never on status codes.
Origin: ssh://git@100.71.119.27:222/monster/guts.git.