*This is the easy version of the problem. The only difference between the two versions is the constraint on $x$. In this version, $x$ is always $1$.*
It's raining cats and corgis! ... and a whole bunch of other random animals.
As a zoology-meteorology-geology triple major, you immediately rush to the field to study the phenomenon. The field is an array of $N$ positions, where the $i^(t h)$ position has elevation $h_i$.
As you observe the unusual precipitation of animals, you notice that the animals fall from the sky and stack on top of each other. Each animal adds $1$ unit of elevation to the position where it lands. Animals continue to stack on top of each other until all valleys in the landscape are filled. Formally, a valley exists if there exist positions $i, j, k$ such that $1 <= i < j < k <= N$ and $h_i > h_j < h_k$ (Visual example provided below).
You wonder that if you increase the height of the landscape at position $i$ by $x$ units (*where $x$ is always $1$*), how many animals will fill the resulting landscape.
The first line of input contains 2 space-separated integers $N, Q " "(1 <= N, Q <= 10^5) -$ the size of the landscape and the number of queries to process, respectively.
The second line of input contains $N$ integers $h_i " "(0 <= h_i <= 10^5)$ separated by spaces, representing the initial elevation at the $i^(t h)$ position in the landscape array.
Each of the following $Q$ lines contains two integers $p_i " "(1 <= p_i <= N)$ and $x_i " "bold((x_i = 1))$, representing the position in the landscape array where the elevation will be increased by $x$ units.
*Note*: The changes made in each query should be preserved and taken into account when processing subsequent queries.
Please output $Q$ lines, where the $i^(t h)$ line contains the answer to the $i^(t h)$ query.
## Input
The first line of input contains 2 space-separated integers $N, Q " "(1 <= N, Q <= 10^5) -$ the size of the landscape and the number of queries to process, respectively.The second line of input contains $N$ integers $h_i " "(0 <= h_i <= 10^5)$ separated by spaces, representing the initial elevation at the $i^(t h)$ position in the landscape array.Each of the following $Q$ lines contains two integers $p_i " "(1 <= p_i <= N)$ and $x_i " "bold((x_i = 1))$, representing the position in the landscape array where the elevation will be increased by $x$ units.*Note*: The changes made in each query should be preserved and taken into account when processing subsequent queries.
## Output
Please output $Q$ lines, where the $i^(t h)$ line contains the answer to the $i^(t h)$ query.
[samples]
[{"iden":"statement","content":"*这是本题的简单版本。两个版本之间的唯一区别在于 $x$ 的约束。在本版本中，$x$ 始终为 $1$。*\n\n天上正下着猫和柯基！…… 还有一大堆其他随机动物。\n\n作为一名动物学-气象学-地质学三重专业学生，你立即冲到野外研究这一现象。该野外区域是一个包含 $N$ 个位置的数组，其中第 $i^{\\text{th}}$ 个位置的海拔为 $h_i$。\n\n当你观察这些异常的动物降水时，你注意到动物从天而降并堆叠在一起。每只动物在其落地位置增加 $1$ 单位的海拔。动物会持续堆叠，直到景观中的所有山谷都被填满。形式化地，若存在位置 $i, j, k$ 满足 $1 \\lt eq i < j < k \\lt eq N$ 且 $h_i > h_j < h_k$，则存在一个山谷（下方提供了视觉示例）。\n\n你好奇的是，若将位置 $i$ 的景观高度增加 $x$ 单位（*其中 $x$ 始终为 $1$*），将有多少只动物填满由此产生的景观。\n\n输入的第一行包含两个用空格分隔的整数 $N, Q \\ (1 \\lt eq N, Q \\lt eq 10^5)$ —— 分别表示景观的大小和待处理的查询数量。\n\n第二行输入包含 $N$ 个用空格分隔的整数 $h_i \\ (0 \\lt eq h_i \\lt eq 10^5)$，表示景观数组中第 $i^{\\text{th}}$ 个位置的初始海拔。\n\n接下来的 $Q$ 行中，每行包含两个整数 $p_i \\ (1 \\lt eq p_i \\lt eq N)$ 和 $x_i \\ \\text{bold}((x_i = 1))$，表示将景观数组中第 $p_i$ 个位置的海拔增加 $x$ 单位。\n\n*注意*：每个查询所做的更改应被保留，并在处理后续查询时予以考虑。\n\n请输出 $Q$ 行，其中第 $i^{\\text{th}}$ 行包含第 $i^{\\text{th}}$ 个查询的答案。\n"},{"iden":"input","content":"输入的第一行包含两个用空格分隔的整数 $N, Q \\ (1 \\lt eq N, Q \\lt eq 10^5)$ —— 分别表示景观的大小和待处理的查询数量。第二行输入包含 $N$ 个用空格分隔的整数 $h_i \\ (0 \\lt eq h_i \\lt eq 10^5)$，表示景观数组中第 $i^{\\text{th}}$ 个位置的初始海拔。接下来的 $Q$ 行中，每行包含两个整数 $p_i \\ (1 \\lt eq p_i \\lt eq N)$ 和 $x_i \\ \\text{bold}((x_i = 1))$，表示将景观数组中第 $p_i$ 个位置的海拔增加 $x$ 单位。*注意*：每个查询所做的更改应被保留，并在处理后续查询时予以考虑。"},{"iden":"output","content":"请输出 $Q$ 行，其中第 $i^{\\text{th}}$ 行包含第 $i^{\\text{th}}$ 个查询的答案。"},{"iden":"examples","content":"输入\n10 4\n2 1 4 4 5 2 3 4 5 3\n3 1\n7 1\n4 1\n5 1\n输出\n8\n7\n6\n6\n\n输入\n5 4\n0 5 0 5 0\n2 1\n4 1\n3 1\n1 1\n输出\n5\n6\n5\n5\n\n输入\n6 2\n0 0 0 0 0 0\n2 1\n4 1\n输出\n0\n1\n"},{"iden":"note","content":" "}]}
**Definitions:**
- Let $ h = [h_1, h_2, \dots, h_N] $ be the array of initial elevations.
- A **valley** is a triple of indices $ (i, j, k) $ such that $ 1 \leq i < j < k \leq N $ and $ h_i > h_j < h_k $.
- Each animal increases the elevation of a position by exactly 1 unit.
- Animals stack until **all valleys are filled**, i.e., no such triple $ (i, j, k) $ exists with $ h_i > h_j < h_k $.
- After each query, the elevation at position $ p_i $ is increased by $ x_i = 1 $, and the change persists.
**Given:**
- $ N, Q \in \mathbb{Z}^+ $, $ 1 \leq N, Q \leq 10^5 $
- Initial array $ h \in \mathbb{Z}_{\geq 0}^N $
- For each query $ q \in \{1, \dots, Q\} $: update $ h_{p_q} \leftarrow h_{p_q} + 1 $
**Objective:**
For each query $ q $, compute the **minimum number of animals** (i.e., total elevation increases) required to eliminate **all valleys** in the current landscape $ h $, starting from the state after the previous update (or initial state for $ q=1 $).
**Note:** The number of animals needed is the **minimum total elevation increase** across all positions such that the resulting array $ h' $ satisfies:
$$
\not\exists \, i, j, k \text{ with } 1 \leq i < j < k \leq N \text{ such that } h'_i > h'_j < h'_k
$$
This is equivalent to making the array **valley-free**, which occurs if and only if the array is **bitonic**: i.e., non-decreasing up to some peak, then non-increasing, or more generally, has no local minima strictly lower than both neighbors.
But note: the minimal way to eliminate all valleys is to raise every local minimum to the height of the **minimum of its two neighboring peaks**.
Thus, the minimal number of animals required is the **sum over all local minima** $ j $ of:
$$
\max\left(0, \min(h_{j-1}, h_{j+1}) - h_j\right)
$$
**Formal Output per Query:**
After updating $ h_{p_q} \leftarrow h_{p_q} + 1 $, compute:
$$
\sum_{j=2}^{N-1} \max\left(0, \min(h_{j-1}, h_{j+1}) - h_j\right)
$$
**Constraints:**
- Updates are persistent.
- $ x_i = 1 $ for all queries.
- Only local minima (interior points satisfying $ h_{j-1} > h_j < h_{j+1} $) contribute to the sum.
---
**Final Formal Statement:**
Let $ h^{(0)} = [h_1, h_2, \dots, h_N] $ be the initial elevation array.
For $ q = 1 $ to $ Q $:
1. $ h^{(q)}_{p_q} \leftarrow h^{(q-1)}_{p_q} + 1 $
2. $ h^{(q)}_i = h^{(q-1)}_i $ for $ i \ne p_q $
3. Compute:
$$
A_q = \sum_{j=2}^{N-1} \max\left(0, \min(h^{(q)}_{j-1}, h^{(q)}_{j+1}) - h^{(q)}_j\right)
$$
4. Output $ A_q $
**Note:** The sum is over all interior indices $ j \in \{2, 3, \dots, N-1\} $, and only contributes when $ h_j $ is strictly less than both neighbors.