refactor(input): own gameplay action routing

Move the sole semantic action-priority graph, combat and diagnostic commands, and retained-root item-drop lifetime behind focused typed owners. Preserve retail toggle behavior, explicit auto-wield cancellation, shutdown quiescence, and symmetric callback cleanup.

Co-authored-by: Codex <codex@openai.com>
This commit is contained in:
Erik 2026-07-22 12:43:05 +02:00
parent 8b8afeefa3
commit 4eae9b4f5a
25 changed files with 2608 additions and 418 deletions

View file

@ -27,12 +27,13 @@ What does NOT go here:
## Current queue — 2026-07-22
- **Active structural work:** `GameWindow` decomposition. Slices 17 are
complete; Slice 8 checkpoints AE now own native callbacks, live-session
complete; Slice 8 checkpoints AF now own native callbacks, live-session
composition, world-environment state, reversible device input/raw pointer,
and physical framebuffer publication outside the shell.
physical framebuffer publication, gameplay action priority, typed commands,
and the retained-root item-drop edge outside the shell.
`GameWindow.OnUpdate` and `GameWindow.OnRender` are typed orchestration
handoffs and the class is 4,266 lines / 194 fields / 65 methods. Remaining
Slice 8 action-routing/settings/composition/shutdown
handoffs and the class is 4,057 lines / 198 fields / 54 methods. Remaining
Slice 8 settings/resource/composition/shutdown
cleanup is active in
[`docs/architecture/code-structure.md`](architecture/code-structure.md).
This is the behavior-preserving prerequisite before new M4 feature bodies.