eb423fb749
Traced the live remote path: the mt-6 chase runs through the verbatim MoveToManager, but every inbound mt-0 attack UM fires the unpack_movement head interrupt (bound to MoveTo.CancelMoveTo) AND installs the UM's ForwardCommand (Ready/idle for an attack-only UM), so the chase MoveTo is cancelled and the run legs are replaced by idle while the body keeps dead-reckoning -> glide + repeated idle/attack. The 15-bit ServerActionStamp gate is faithful (uniform-anim is not a stamp bug). This chase+attack coexistence is R5/MovementManager scope, so #170 sub-bugs 2/3 are downstream of the incomplete MovementManager port (R5-V4), not #159. Recorded as a hypothesis pending a live ACDREAM_DUMP_MOTION capture + retail side-by-side (no guessing a fix in this revert-prone path). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> |
||
|---|---|---|
| .. | ||
| architecture | ||
| audit | ||
| plans | ||
| research | ||
| superpowers | ||
| bugs.md | ||
| ISSUES.md | ||