Job: target_node claim so pawns don't cluster on the same target

Visible bug: with 1 wall ghost queued, all 3 pawns picked the same site;
2 stood idle while 1 built. Same shape would affect chop/mine/haul/etc.

Design: Job carries an untyped target_node ref (the tree/rock/build-site/
crop/item/patient/workbench/etc the job is acting on). Job.is_target_taken_
by_other(target, excluding_pawn) does an O(pawns) scan of live job state to
ask 'is anyone else already working this?'. Each WorkProvider's find_best_
for() now skips claimed targets in its scan and sets j.target_node before
returning. No per-entity claim state, no .claim()/.release() bookkeeping,
no save-format change — target_node is intentionally not serialized
because pawns re-decide and re-bind naturally after load.

Providers updated: construction / chop / mine / plant (harvest path) /
hauling (item AND corpse) / cleaning (target is Vector2i tile not Node;
field is untyped, doc'd) / doctor / crafting (workbench).

Not touched: rest (everyone shares the rest tile, that's fine), eat /
sleep (food and beds have their own availability gates; flagged as a
followup if multi-pawn food contention surfaces).

Verified MCP runtime: fresh boot, 3 pawns picked 3 distinct wall sites
(44,28)/(45,28)/(44,27) with distinct target_node refs.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This commit is contained in:
megaproxy 2026-05-12 13:45:44 +01:00
parent 07fb66608a
commit e7a2407af2
9 changed files with 70 additions and 1 deletions

View file

@ -14,6 +14,40 @@ var label: String = ""
var toils: Array[Toil] = []
var current_toil_index: int = 0
## The world-space entity this job is acting on (tree, rock, build-site,
## crop, etc.). Set by the WorkProvider that built the job. Untyped to
## avoid class_name cycles. Read by sibling providers via
## Job.is_target_taken_by_other() to prevent multiple pawns claiming
## the same target.
##
## NOT serialized: target_node holds a live Node ref (or Vector2i for tile
## targets). After load, pawns re-decide and re-bind claims naturally based
## on restored JobRunner state.
var target_node = null
# ── claim helpers ────────────────────────────────────────────────────────────
## Returns true if `target` is the active target_node of any pawn's current
## job, excluding `excluding_pawn` (the pawn that's about to claim it).
## O(pawns) scan — single-threaded sim, no race within a tick because
## pawns iterate sequentially (earlier pawn commits its job before the
## next pawn's decision runs).
##
## Accepts any value for target (Node, Vector2i, etc.) — uses == comparison.
static func is_target_taken_by_other(target, excluding_pawn) -> bool:
if target == null:
return false
for p in World.pawns:
if p == excluding_pawn:
continue
if p.job_runner == null or not p.job_runner.has_job():
continue
if p.job_runner.job.target_node == target:
Audit.log("decision", "claim skip: %s already targeting %s" % [p.pawn_name, str(target)])
return true
return false
# ── queries ──────────────────────────────────────────────────────────────────