Longest Subarray - Implementation

Maintain maxDeque (decreasing) and minDeque (increasing). Both store indices. On adding index $r$: pop maxDeque back while $arr[back] < arr[r]$. Pop minDeque back while $arr[back] > arr[r]$. On shrinking from $l$: if maxDeque front = $l$, pop front. Same for minDeque. Check: $arr[maxDeque.front()] - arr[minDeque.front()] \leq limit$. If not, increment $l$ and repeat. Both queues shrink from the front when elements leave the window. Track indices, not values.

Time complexity: $O(n)$ since each element enters and leaves each deque at most once.

Space complexity: $O(n)$ for the two deques.