File Details
Loothing v2.0.29
- R
- Apr 24, 2026
- 998.78 KB
- 27
- 12.0.5+1
- Retail
File Name
loothing_2.0.29.zip
Supported Versions
- 12.0.5
- 12.0.1
Polish pass on the picker and the loot-detection edges. Nothing blocking here — these are the "you'll notice them if they happen to you" bugs that the last two releases' fresh-eyes reviews surfaced.
Fixed
- The Loot Picker now tells you it's still waiting for roll results. Previously it opened at boss-kill with an empty list and a static "No tradable loot detected yet." message — users would see the empty state, assume the addon was broken, and dismiss before group-loot rolls resolved at T+30s. The picker now tracks the active bag-scan window and shows a live countdown ("Waiting for roll results… 27s remaining") that ticks down and replaces itself with real item rows as they arrive. Once the scan window closes with no items found, it falls back to the original static message.
- Items won by the ML during a mid-tick roster update are no longer silently dropped. The
LOOT_ITEM_ROLL_WONgate requiredLoothing.handleLootto be true at the exact instant the event fired. AGROUP_ROSTER_UPDATEfiring while rolls are resolving (someone changes spec, pet respawns, an officer dies) briefly clearsisMasterLooteras the ML-detection re-verifies, and any roll resolving in that tick would drop. The gate now relies on the canonicalIsCanonicalMLcheck first (settings-designated ML → leader-designated ML → raid leader fallback), which doesn't participate in that race. - The tooltip trade-time cache no longer permanently poisons freshly-landed items. When an item lands in your bag, the server decorates its tooltip with the "You may trade…" line shortly after. If the bag scanner probed the tooltip during that gap, it got back
math.huge(no trade line + no bound line) and cached that value, causing every subsequent probe to short-circuit to the stale result. The cache now refuses to trustmath.hugereads: the next scan tick re-probes and picks up the now-warm tooltip.
Cleanup
- Deleted the
BOSS_KILL → OnBossKillno-op registration — ENCOUNTER_END is the canonical "boss died" signal and the OnBossKill stub was documenting-intent-as-code rather than doing anything. Will re-register per-encounter only if a specific fight is found that fires BOSS_KILL without ENCOUNTER_END. - Tightened / corrected several comments that had drifted out of sync with behavior changes from 2.0.27:
preEncounterBagSnapshotkey-type comment (itemID not itemLink),OnLootReceived"fires on any player" (actually only fires on the ML's client, for any raid member's drop), and the GroupLootEvents event-landscape doc block (ENCOUNTER_LOOT_RECEIVED fires for any encounter-distributed item, not personal-loot-only — the group-loot branch just happens to resolve throughC_LootHistoryin 12.0).

