import org.objectweb.asm.ClassReader;
import org.objectweb.asm.ClassWriter;
public class MethodCallCountTransformer implements ClassFileTransformer {
@Override
public byte[] transform(ClassLoader loader, String className, Class classBeingRedefined,
ProtectionDomain protectionDomain, byte[] classfileBuffer) throws IllegalClassFormatException {
try {
if (!“java/util/regex/Pattern$CharProperty”.equals(className)
&& !“java/util/regex/Pattern$BmpCharProperty”.equals(className)) {
return classfileBuffer;
}
ClassWriter writer = new ClassWriter(ClassWriter.COMPUTE_MAXS);
ClassAdapter adapter = new MethodCallClassAdapter(writer, className);
ClassReader reader = new ClassReader(classfileBuffer);
reader.accept(adapter, 0);
// 生成新类字节码
return writer.toByteArray();
} catch (Exception e) {
e.printStackTrace();
// 返回旧类字节码
return classfileBuffer;
}
}
}
package com.alibaba.taobao.tinyprofiler;
import java.util.concurrent.atomic.AtomicInteger;
public class Counter {
private static AtomicInteger methodCallCount = new AtomicInteger(0);
public static void printAndIncCount(String className, String methodName) {
System.out.println(className + “.” + methodName + “ called, total times ” + methodCallCount.incrementAndGet());
}
}
OK,现在我们输入:
String regex = “.*.*.*D”;
String target = “22asdvasdx”;
Pattern.compile(regex).matcher(target).matches();
System.out.println(“Xuanyin’s estimated count: ” + findTotalWays(4, target.length()) + “; depth: ” + recursionDepth);
输出结果(注:这里的 depth 并不是递归深度,而是递归次数,当时搞错了):
肿么样,分毫不差~OK,那么我们现在回到最开始的问题,输入 .*.*.*.*.*.*D 去匹配 com.taobao.binary.bogda.query.service.RulesInfoQueryService:1.0.0.daily
结果显示需要 5 亿 匹配, 还要进出栈近 2.5 亿次哦
这里我的机器是i7-2600K 超4.5G,结果显示需要5秒,这还是不是最差的情况哦~