update: document repair movement lock

This commit is contained in:
Tom Boullay
2026-05-11 13:12:37 +02:00
parent 49d1a7324c
commit 00b1780b10
3 changed files with 7 additions and 6 deletions
+2 -1
View File
@@ -18,6 +18,7 @@ This document lists features that are implemented in the current codebase.
- Pointer lock mouse look
- Movement with `ZQSD`
- Jumping
- Movement lock during active repair steps, while keeping trigger interactions available
- Octree-based collision against dedicated map collision nodes, currently scoped to `terrain`
## Interactions
@@ -33,7 +34,7 @@ This document lists features that are implemented in the current codebase.
- Reusable production `RepairGame` mounted for `bike`, `pylone`, and `ferme` mission states
- Debug physics playground mounts the same reusable `RepairGame` in `Bike`, `Pylone`, and `Farm` zones so each state can be tuned with isolated positioning before moving placement into the production map
- Repair mission config shared through `src/data/gameplay/repairMissions.ts`, including per-mission broken nodes, placeholder targets, scan timing, and reassembly timing
- Repair-game flow supports `waiting -> inspected -> fragmented -> scanning -> repairing -> reassembling -> done -> next mission` with `.webm` prompts, repair case spawn/opening/exit, focused repair-case view, repair-case trigger interaction, case placeholder traversal, snap-to-placeholder placement, broken-part deposit feedback, `E`, two-fists hold input, exploded and inverse reassembly transitions, completion particles, per-part scan visuals, persistent red broken-part markers, centered broken-part UI videos, multiple grabbable replacement choices, correct-part install validation feedback, and mission completion
- Repair-game flow supports `waiting -> inspected -> fragmented -> scanning -> repairing -> reassembling -> done -> next mission` with `.webm` prompts, repair case spawn/opening/exit, focused repair-case view, movement lock during active repair, repair-case trigger interaction, case placeholder traversal, snap-to-placeholder placement, broken-part deposit feedback, `E`, two-fists hold input, exploded and inverse reassembly transitions, completion particles, per-part scan visuals, persistent red broken-part markers, centered broken-part UI videos, multiple grabbable replacement choices, correct-part install validation feedback, and mission completion
## Audio
+4 -4
View File
@@ -12,7 +12,7 @@ The current user flow is:
2. Move close to the active repair object in the game scene.
3. Aim at the object and press the interaction key when prompted.
4. The mission step moves from `waiting` to `inspected`.
5. The repair case appears near the mission object and can float when the player approaches it.
5. The repair case appears near the mission object, the player movement controls are locked, and the case can float when the player approaches it.
6. Aim at the repair case and press `E`, or hold both fists closed for one second, to move from `inspected` to `fragmented`.
7. The mission object uses an exploded-model transition, then moves to `scanning`.
8. The scan visual moves across the fragmented model one part at a time and keeps a red marker plus the `cassé.webm` prompt centered on any configured broken part once it has been found.
@@ -20,7 +20,7 @@ The current user flow is:
10. Move the correct replacement part close to a placeholder. When released near a placeholder, it snaps into place with a short animation.
11. Move each scanned broken part into a compatible placeholder so the damaged parts are stored in the case.
12. Press `E` on the green install target to move to `reassembling`. Wrong parts turn the target red and cannot finish the repair.
13. The exploded object animates back into its assembled form with completion particles, then moves to `done`.
13. The exploded object animates back into its assembled form with completion particles, then moves to `done` and restores player movement controls.
14. Press `E` on the completion target. The repair case closes, returns to the ground, disappears, then `completeMission` moves to the next mission or to `outro` after `ferme`.
## Why It Matters
@@ -31,11 +31,11 @@ This feature validates the repair loop before a full mission system exists. It t
In `waiting`, the active mission renders its repair object and the `interagir.webm` prompt in the game scene. The interaction uses the shared focus/raycast interaction system, so the player still gets the normal `E` prompt.
When the player inspects the object, `RepairGame` writes `inspected` through the generic mission store action. The repair case then appears from the mission config with a small pop animation. When the player is close enough, the existing case model floats upward and rotates gently to signal interactivity.
When the player inspects the object, `RepairGame` writes `inspected` through the generic mission store action. The repair case then appears from the mission config with a small pop animation, and player movement is locked while the repair sequence is active. When the player is close enough, the existing case model floats upward and rotates gently to signal interactivity.
In `inspected`, `RepairGame` can also move to `fragmented`. Keyboard input goes through the shared focus/raycast interaction system on the repair case, so the player must be close enough and aim at the case before pressing `E`. The hand-tracking path still uses a two-fists hold gesture and is state-based, so it does not depend on being inside a local object interaction radius.
In `fragmented`, the repair object is rendered with `ExplodableModel`, then automatically advances to `scanning`. In `scanning`, the exploded model remains visible, a blue scan visual moves from part to part, and a red halo/wire marker plus the configured broken UI video stay attached to configured broken parts after the scanner reaches them. The scan can match a specific `nodeName` when mission data provides one, otherwise it falls back to the first scanned parts as placeholder broken parts. In `repairing`, the case opens in a larger focused transform, `RepairCaseModel` traverses the case GLTF for empty nodes named `placeholder_*`, several grabbable replacement parts appear on those placeholder positions, and releasing a part near a placeholder snaps it into place with a short GSAP animation. Scanned broken parts are also rendered as grabbable objects and must be deposited into a compatible placeholder before the final install target validates. If `brokenParts[].placeholderName` is configured, that broken part snaps only to the matching placeholder; otherwise it can use any available placeholder. If the current case asset has no placeholder nodes, the flow keeps using fallback focus positions. Replacement parts show green or red placement feedback after snapping, broken parts show stored feedback after deposit, and the install target gives a short blocked feedback if the player tries to validate too early. The install target only validates when the configured correct replacement part is placed and all scanned broken parts have been deposited. In `reassembling`, the exploded model animates back into its assembled position with green completion particles before the flow moves to `done`. In `done`, the repaired object remains visible with a completion target; validating closes the repair case first, then plays the case exit animation before advancing the global mission progression.
In `fragmented`, the repair object is rendered with `ExplodableModel`, then automatically advances to `scanning`. In `scanning`, the exploded model remains visible, a blue scan visual moves from part to part, and a red halo/wire marker plus the configured broken UI video stay attached to configured broken parts after the scanner reaches them. The scan can match a specific `nodeName` when mission data provides one, otherwise it falls back to the first scanned parts as placeholder broken parts. In `repairing`, the case opens in a larger focused transform, `RepairCaseModel` traverses the case GLTF for empty nodes named `placeholder_*`, several grabbable replacement parts appear on those placeholder positions, and releasing a part near a placeholder snaps it into place with a short GSAP animation. Scanned broken parts are also rendered as grabbable objects and must be deposited into a compatible placeholder before the final install target validates. If `brokenParts[].placeholderName` is configured, that broken part snaps only to the matching placeholder; otherwise it can use any available placeholder. If the current case asset has no placeholder nodes, the flow keeps using fallback focus positions. Replacement parts show green or red placement feedback after snapping, broken parts show stored feedback after deposit, and the install target gives a short blocked feedback if the player tries to validate too early. The install target only validates when the configured correct replacement part is placed and all scanned broken parts have been deposited. Player movement stays locked through `inspected`, `fragmented`, `scanning`, `repairing`, and `reassembling`, while trigger interactions remain available. In `reassembling`, the exploded model animates back into its assembled position with green completion particles before the flow moves to `done`. In `done`, player movement is available again and the repaired object remains visible with a completion target; validating closes the repair case first, then plays the case exit animation before advancing the global mission progression.
The mission config now carries the mission-specific variations. `bike` repairs one cooling core, `pylone` scans and stores both the lamp relay and a damaged panel with slower scan/reassembly timing, and `ferme` scans and stores an irrigation pump plus humidity sensor with faster scan/reassembly timing.