feat:实现正则表达式编译缓存功能 #3130 (#3131)

hengyunabc · ban-xiu · web-flow · commit 878f8844e47b · 2026-01-12T14:43:23.000+08:00
Co-authored-by: lbs &lt;2322701154@qq.com&gt;
diff --git a/core/src/main/java/com/taobao/arthas/core/util/RegexCacheManager.java b/core/src/main/java/com/taobao/arthas/core/util/RegexCacheManager.java
@@ -0,0 +1,65 @@
+package com.taobao.arthas.core.util;
+
+import com.taobao.arthas.core.shell.term.impl.http.session.LRUCache;
+import java.util.regex.Pattern;
+
+/**
+ * 正则表达式缓存管理器
+ * 用于缓存编译后的正则表达式对象，避免重复编译的开销
+ */
+public class RegexCacheManager {
+    private static final RegexCacheManager INSTANCE = new RegexCacheManager();
+    
+    // 使用LRUCache缓存编译后的正则表达式
+    private final LRUCache<String, Pattern> regexCache;
+    
+    // 缓存大小限制
+    private static final int MAX_CACHE_SIZE = 100;
+    
+    private RegexCacheManager() {
+        // 初始化LRUCache，设置最大缓存大小
+        this.regexCache = new LRUCache<>(MAX_CACHE_SIZE);
+    }
+
+    public static RegexCacheManager getInstance() {
+        return INSTANCE;
+    }
+
+    /**
+     * 获取正则表达式Pattern对象，优先从缓存获取，缓存未命中则编译并缓存
+     */
+    public Pattern getPattern(String regex) {
+        if (regex == null) {
+            return null;
+        }
+        
+        // 从LRUCache获取
+        Pattern pattern = regexCache.get(regex);
+        if (pattern != null) {
+            return pattern;
+        }
+
+        // 缓存未命中，编译正则表达式
+        // 不捕获PatternSyntaxException，让异常向上抛出，以便及时发现无效的正则表达式
+        pattern = Pattern.compile(regex);
+        // 缓存编译结果
+        regexCache.put(regex, pattern);
+        
+        return pattern;
+    }
+    
+    /**
+     * 清理缓存
+     */
+    public void clearCache() {
+        regexCache.clear();
+    }
+    
+    /**
+     * 获取缓存大小
+     */
+    public int getCacheSize() {
+        return regexCache.usedEntries();
+    }
+
+}
diff --git a/core/src/main/java/com/taobao/arthas/core/util/matcher/RegexMatcher.java b/core/src/main/java/com/taobao/arthas/core/util/matcher/RegexMatcher.java
@@ -1,21 +1,32 @@
 package com.taobao.arthas.core.util.matcher;
 
+import com.taobao.arthas.core.util.RegexCacheManager;
+import java.util.regex.Pattern;
+
 /**
  * regex matcher
  * @author ralf0131 2017-01-06 13:16.
  */
 public class RegexMatcher implements Matcher<String> {
 
     private final String pattern;
+    private volatile Pattern compiledPattern;
 
     public RegexMatcher(String pattern) {
         this.pattern = pattern;
     }
 
     @Override
     public boolean matching(String target) {
-        return null != target
-                && null != pattern
-                && target.matches(pattern);
+        if (null == target || null == pattern) {
+            return false;
+        }
+
+        // 在第一次matching时才编译正则表达式
+        if (compiledPattern == null) {
+            compiledPattern = RegexCacheManager.getInstance().getPattern(pattern);
+        }
+        
+        return compiledPattern != null && compiledPattern.matcher(target).matches();
     }
 }
diff --git a/core/src/test/java/com/taobao/arthas/core/util/RegexCacheManagerTest.java b/core/src/test/java/com/taobao/arthas/core/util/RegexCacheManagerTest.java
@@ -0,0 +1,138 @@
+package com.taobao.arthas.core.util;
+
+import org.junit.Assert;
+import org.junit.Before;
+import org.junit.Test;
+
+import java.util.regex.Pattern;
+import java.util.regex.PatternSyntaxException;
+
+/**
+ * RegexCacheManager测试类
+ */
+public class RegexCacheManagerTest {
+    private RegexCacheManager cacheManager;
+
+    @Before
+    public void setUp() {
+        // 获取单例实例
+        cacheManager = RegexCacheManager.getInstance();
+        // 清理缓存，确保测试环境干净
+        cacheManager.clearCache();
+    }
+
+    /**
+     * 测试基本缓存功能
+     */
+    @Test
+    public void testBasicCacheFunctionality() {
+        // 测试缓存未命中的情况
+        String regex1 = ".*Test.*";
+        Pattern pattern1 = cacheManager.getPattern(regex1);
+        Assert.assertNotNull(pattern1);
+        Assert.assertEquals(1, cacheManager.getCacheSize());
+
+        // 测试缓存命中的情况
+        Pattern pattern1Cached = cacheManager.getPattern(regex1);
+        Assert.assertNotNull(pattern1Cached);
+        Assert.assertSame(pattern1, pattern1Cached); // 应该是同一个对象
+        Assert.assertEquals(1, cacheManager.getCacheSize()); // 缓存大小应该保持不变
+
+        // 测试多个正则表达式
+        String regex2 = "^Test.*";
+        Pattern pattern2 = cacheManager.getPattern(regex2);
+        Assert.assertNotNull(pattern2);
+        Assert.assertEquals(2, cacheManager.getCacheSize());
+
+        // 测试空正则表达式
+        Pattern nullPattern = cacheManager.getPattern(null);
+        Assert.assertNull(nullPattern);
+
+        Pattern emptyPattern = cacheManager.getPattern("");
+        Assert.assertNotNull(emptyPattern);
+        Assert.assertTrue(emptyPattern.matcher("").matches());
+        Assert.assertFalse(emptyPattern.matcher("non-empty").matches());
+        Assert.assertEquals(3, cacheManager.getCacheSize());
+    }
+
+    /**
+     * 测试LRU淘汰策略
+     */
+    @Test
+    public void testLRUEvictionPolicy() {
+        // 生成多个正则表达式，超过最大缓存大小
+        int maxCacheSize = 100;
+        for (int i = 0; i < maxCacheSize + 5; i++) {
+            String regex = "TestRegex" + i;
+            Pattern pattern = cacheManager.getPattern(regex);
+            Assert.assertNotNull(pattern);
+        }
+
+        // 缓存大小应该等于最大缓存大小
+        Assert.assertEquals(maxCacheSize, cacheManager.getCacheSize()); // 100 是实际的最大缓存大小
+
+        // 测试访问顺序，确保LRU策略生效
+        String firstRegex = "TestRegex0";
+
+        // 再次访问第一个正则表达式，使其成为最近使用的
+        Pattern firstPattern = cacheManager.getPattern(firstRegex);
+        Assert.assertNotNull(firstPattern);
+
+        // 再添加一个新的正则表达式，应该淘汰最久未使用的
+        String newRegex = "NewTestRegex";
+        Pattern newPattern = cacheManager.getPattern(newRegex);
+        Assert.assertNotNull(newPattern);
+
+        // 第一个正则表达式应该仍然在缓存中（因为刚被访问过）
+        Pattern firstPatternAgain = cacheManager.getPattern(firstRegex);
+        Assert.assertNotNull(firstPatternAgain);
+    }
+
+    /**
+     * 测试缓存清理功能
+     */
+    @Test
+    public void testCacheClear() {
+        // 添加一些缓存项
+        cacheManager.getPattern(".*Test1");
+        cacheManager.getPattern(".*Test2");
+        Assert.assertTrue(cacheManager.getCacheSize() > 0);
+
+        // 清理缓存
+        cacheManager.clearCache();
+        Assert.assertEquals(0, cacheManager.getCacheSize());
+
+        // 清理后应该可以重新添加缓存项
+        Pattern pattern = cacheManager.getPattern(".*Test3");
+        Assert.assertNotNull(pattern);
+        Assert.assertEquals(1, cacheManager.getCacheSize());
+    }
+
+    /**
+     * 测试无效正则表达式处理
+     */
+    @Test
+    public void testInvalidRegexHandling() {
+        // 测试无效的正则表达式，应该抛出PatternSyntaxException
+        String invalidRegex = "[a-z";
+        try {
+            cacheManager.getPattern(invalidRegex);
+        } catch (Exception e) {
+            // 验证抛出的是PatternSyntaxException
+            Assert.assertTrue("Expected PatternSyntaxException but got " + e.getClass().getName(), e instanceof PatternSyntaxException);
+        }
+        
+        // 测试另一个无效的正则表达式，应该抛出PatternSyntaxException
+        String anotherInvalidRegex = "(a-z";
+        try {
+            cacheManager.getPattern(anotherInvalidRegex);
+        } catch (Exception e) {
+            // 验证抛出的是PatternSyntaxException
+            Assert.assertTrue("Expected PatternSyntaxException but got " + e.getClass().getName(), e instanceof PatternSyntaxException);
+        }
+        
+        // 确保缓存大小没有增加
+        Assert.assertEquals("无效正则表达式不应该被缓存", 0, cacheManager.getCacheSize());
+    }
+
+}
diff --git a/core/src/test/java/com/taobao/arthas/core/util/matcher/RegexMatcherTest.java b/core/src/test/java/com/taobao/arthas/core/util/matcher/RegexMatcherTest.java
@@ -16,6 +16,12 @@ public void testMatchingWithNullInputs(){
         Assert.assertTrue(new RegexMatcher("foobar").matching("foobar"));
     }
 
+    @Test
+    public void testMatchingWithEmptyPattern() {
+        Assert.assertTrue(new RegexMatcher("").matching(""));
+        Assert.assertFalse(new RegexMatcher("").matching("foobar"));
+    }
+
     /**
      * test regux with . | * + ? \s \S \w \W and so on...
      */
