⚡ Improve find performance

Signed-off-by: Daniel <845765@qq.com>

kernel/search/find.go

@@ -17,8 +17,8 @@
 package search
 
 import (
-	"bufio"
 	"bytes"
+	"io"
 	"os"
 	"path/filepath"
 	"runtime"
@@ -47,13 +47,20 @@ func FindAllMatches(root string, targets []string) []Match {
 		return nil
 	}
 
-	var searchBytes [][]byte
+	// 构建基于首字节的模式索引，并计算最长模式长度
+	patternIndex := make(map[byte][][]byte)
+	var maxLen int
 	for _, t := range targets {
-		if t != "" {
-			searchBytes = append(searchBytes, []byte(t))
+		if t == "" {
+			continue
 		}
+		b := []byte(t)
+		if len(b) > maxLen {
+			maxLen = len(b)
+		}
+		patternIndex[b[0]] = append(patternIndex[b[0]], b)
 	}
-	if len(searchBytes) == 0 {
+	if len(patternIndex) == 0 {
 		return nil
 	}
 
@@ -78,7 +85,7 @@ func FindAllMatches(root string, targets []string) []Match {
 		go func() {
 			defer wg.Done()
 			for p := range jobs {
-				hits := scanFileForTargets(p, searchBytes)
+				hits := scanFileForTargets(p, patternIndex, maxLen)
 				if len(hits) > 0 {
 					for _, t := range hits {
 						results <- Match{Path: p, Target: t}
@@ -102,47 +109,73 @@ func FindAllMatches(root string, targets []string) []Match {
 	return matches
 }
 
-// scanFileForTargets 在文件中搜索所有目标，返回去重后的命中目标字符串列表
-func scanFileForTargets(path string, targets [][]byte) []string {
+// scanFileForTargets 在文件中流式搜索所有目标（基于首字节索引），返回去重后的命中目标字符串列表
+func scanFileForTargets(path string, patternIndex map[byte][][]byte, maxLen int) []string {
 	f, err := os.Open(path)
 	if err != nil {
 		return nil
 	}
 	defer f.Close()
 
-	reader := bufio.NewReaderSize(f, 1024*1024) // 1MB 缓冲
+	// 构建字节位图，加速首字节检测
+	var bitmap [256]bool
+	for b := range patternIndex {
+		bitmap[b] = true
+	}
+
 	found := make(map[string]struct{})
-	remaining := len(targets)
+	buf := make([]byte, 64<<10) // 64KB
+	// 保留上一次块末尾的重叠数据以支持跨块匹配
+	var tail []byte
 
 	for {
-		line, err := reader.ReadSlice('\n')
-		if len(line) > 0 {
-			for _, t := range targets {
-				ts := string(t)
-				if _, ok := found[ts]; ok {
-					continue
+		n, err := f.Read(buf)
+		if n > 0 {
+			// data = tail + buf[:n]
+			data := make([]byte, len(tail)+n)
+			copy(data, tail)
+			copy(data[len(tail):], buf[:n])
+
+			// 扫描 data，查找任意候选首字节位置
+			i := 0
+			for i < len(data) {
+				// 快速跳过非候选字节
+				for i < len(data) && !bitmap[data[i]] {
+					i++
 				}
-				if bytes.Contains(line, t) {
-					found[ts] = struct{}{}
-					remaining--
-					if remaining == 0 {
-						// 找到所有目标，提前返回
-						res := make([]string, 0, len(found))
-						for k := range found {
-							res = append(res, k)
+				if i >= len(data) {
+					break
+				}
+				b := data[i]
+				// 对应首字节的所有模式进行校验
+				for _, pat := range patternIndex[b] {
+					pl := len(pat)
+					// 如果剩余字节不足以完全匹配，则交由下一轮（通过 tail 保证）
+					if i+pl <= len(data) {
+						if bytes.Equal(pat, data[i:i+pl]) {
+							found[string(pat)] = struct{}{}
 						}
-						return res
 					}
 				}
+				i++
+			}
+
+			// 保留最后 maxLen-1 字节作为下一块的 tail（避免超长内存分配）
+			if maxLen <= 1 {
+				tail = nil
+			} else {
+				if len(data) >= maxLen-1 {
+					tail = append(tail[:0], data[len(data)-(maxLen-1):]...)
+				} else {
+					tail = append(tail[:0], data...)
+				}
 			}
 		}
 		if err != nil {
-			if err == bufio.ErrBufferFull {
-				for err == bufio.ErrBufferFull {
-					_, err = reader.ReadSlice('\n')
-				}
-				continue
+			if err == io.EOF {
+				break
 			}
+			// 读取出错，返回已有结果
 			break
 		}
 	}