Files
bnstoolkit/gradle.properties
T
Matt 0b90e94079
Build / build (push) Has been cancelled
0.5.0: per-tile village supply/demand throttle for spur-paying trades
Anti-farm system that limits how much income players can extract from
villager trading in one location. Designed to make 50-librarian halls
diminish into futility without disabling villager trading altogether.

Model:
- World is partitioned into 64×64 block tiles (no POI lookups,
  villager-position rounded by integer division)
- Each tile has a CompoundTag of {item_id -> remaining_demand} at
  server.persistentData.bnsVillageDemand[dim:tileX:tileZ]
- Demand caps per item are set in VillageDemand.java; paper=20,
  wheat=25, rotten flesh=15, iron=6, diamond=1, etc.
- Demand refills linearly: full cap restored over 1 hour of real time
  (no scheduled tick — refresh computed lazily on each query from
  lastUpdated timestamp)
- A 1-tile multi-profession hall earns ~150 spurs/hour at saturation.
  Scaling income requires building halls in distant tiles — that's
  overworld logistics, not micro-exploit.

Implementation:
- VillageDemand.java: per-item cap config
- VillageDemandTracker.java: lazy-refresh + atomic decrement against
  the KubeJS server-scope NBT (existing reflection accessor)
- VillageTrackedMerchantOffer.java: MerchantOffer subclass that
  overrides increaseUses() to decrement demand on trade completion
- VillagerTradeRebalance.java: at offer-build time, queries current
  tile demand for the cost item. If saturated (<1 left), the offer
  comes back with uses=maxUses (vanilla "out of stock" UI). Otherwise
  wrapped in VillageTrackedMerchantOffer so post-trade decrement fires.
- Wandering trader: uses whatever tile he's wandered into. No special
  case — it's just whichever tile he stands in at trade time.

Wandering traders, modded merchants (entity is not Villager): same
tracking, same caps. Anything that produces a spur as a trade result
is throttled by the tile.

Hooks: zero mixins. MerchantOffer.increaseUses() is vanilla-public
and fires server-side after the trade settles, so a subclass override
is the right hook point.

Perf: lazy refresh = no scheduled ticks. Per-trade overhead is one
NBT lookup + a few timestamp ops, microseconds. Idle servers do zero
work.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-06-07 16:26:48 +00:00

24 lines
681 B
Properties

# Sets default memory used for gradle commands. Can be overridden by user or command line properties.
org.gradle.jvmargs=-Xmx2G
org.gradle.daemon=true
org.gradle.parallel=true
org.gradle.caching=true
org.gradle.configuration-cache=true
# Parchment mappings (parameter names + javadocs)
parchment_minecraft_version=1.21.1
parchment_mappings_version=2024.11.17
# Environment Properties — must match the live BNS server
minecraft_version=1.21.1
minecraft_version_range=[1.21.1]
neo_version=21.1.228
loader_version_range=[1,)
## Mod Properties
mod_id=bnstoolkit
mod_name=Brass and Sigil Toolkit
mod_license=All Rights Reserved
mod_version=0.5.0
mod_group_id=uk.sijbers.bnstoolkit