v1.1.2

REFACTOR

• task.go: Refactored taskHeap from a slice type to a struct, added tasks and minCap fields to support dynamic minimum capacity.
• task.go: Updated Len, Less, Swap, Push, and Pop methods of taskHeap to fit the new struct structure.
• pending.go: Fixed newPending initialization to calculate dynamic minimum capacity based on workers and size (max(16, min(size/8, size/workers))).

PERF

• task.go: Optimized heap shrink logic, changed trigger condition from capacity > taskHeapMinCap && length < capacity/4 to capacity > minCap*4 && length < capacity/taskHeapMinCapRatio.
• task.go: Improved shrink capacity calculation, using max(capacity/4, minCap) to ensure it does not fall below the minimum capacity.
• pending.go: Optimized promoteLocked access efficiency by retrieving the tasks field first to avoid repeated access.

UPDATE

• new.go: Fixed newPending call, added workers parameter.

FIX

• new.go: Added debug log for timeout trigger (task.timeout_triggered), recording id, preset, and timeout information.

---

REFACTOR

• task.go: 將 taskHeap 從 slice 型別重構為 struct，新增 tasks 與 minCap 欄位，支援動態最小容量
• task.go: 調整 taskHeap 的 Len、Less、Swap、Push、Pop 方法，適配新的 struct 結構
• pending.go: 修正 newPending 初始化，根據 workers 與 size 計算動態最小容量（max(16, min(size/8, size/workers))）

PERF

• task.go: 優化 heap shrink 邏輯，將觸發條件由 capacity > taskHeapMinCap && length < capacity/4 改為 capacity > minCap*4 && length < capacity/taskHeapMinCapRatio
• task.go: 改進縮減容量計算，使用 max(capacity/4, minCap) 確保不低於最小容量
• pending.go: 優化 promoteLocked 存取效率，先取得 tasks 欄位避免重複存取

UPDATE

• new.go: 修正 newPending 呼叫，新增 workers 參數

FIX

• new.go: 新增 timeout 觸發時的 debug 日誌（task.timeout_triggered），記錄 id、preset、timeout 資訊