version 0.2

1. 更换了html处理库，之前的有bug
2. 兼容了窗口跨域有多个线程栈的情况
3. 扩大关键词，能够搜到更多结果

README.md

@@ -67,7 +67,7 @@ npm install
 ```text
 src/proxy-server/proxy-server.js
 ```
-要用anyproxy抓取https请求需要信任它的证书，访问它的web管理界面：
+要用anyproxy抓取https请求需要信任它的证书，在运行这个文件之前，先用`anyproxy ca`选项启动，访问它的web管理界面：
 ```text
 http://localhost:8002/
 ```

package.json

@@ -14,6 +14,7 @@
     "@babel/types": "^7.12.12",
     "anyproxy": "^4.1.3",
     "body-parser": "^1.19.0",
+    "cheerio": "^1.0.0-rc.5",
     "express": "^4.17.1",
     "jssoup": "0.0.12",
     "shelljs": "^0.8.4"

src/components/global-assign-hook-component/core/global-assign-hook-component-main.js

@@ -1,7 +1,7 @@
 const fs = require("fs");
 const shell = require("shelljs");
 const crypto = require("crypto");
-const JSSoup = require("jssoup").default;
+const cheerio = require("cheerio");
 
 const {loadPluginsAsStringWithCache} = require("./plugins-manager");
 const {injectHook} = require("./inject-hook");
@@ -75,6 +75,7 @@ function isHtmlResponse(responseDetail) {
 // HTML中可能会夹带script标签，里面的JS代码也要能Hook到
 function processHtmlResponse(requestDetail, responseDetail) {
     // 使用了这个库来解析HTML  https://github.com/chishui/JSSoup
+    // 上面那个库有bug，替换为这个库： https://github.com/cheeriojs/cheerio/
 
     // 对所有的内嵌JavaScript内容注入Hook
     const url = requestDetail.url;
@@ -84,35 +85,45 @@ function processHtmlResponse(requestDetail, responseDetail) {
         return;
     }
 
-    const soup = new JSSoup(body);
-    const scriptArray = soup.findAll("script");
+    const $ = cheerio.load(body);
+    const scriptArray = $("script");
     if (!scriptArray?.length) {
         return;
     }
     let alreadyInjectHookContext = false;
     for (let script of scriptArray) {
+
         // 对于是src引用的外部，其标签内容都会被忽略
-        if (script.attrs.src) {
+        if (script.attribs.src) {
             continue;
         }
-        const jsCode = script.contents.join("\n");
-        if (!script.contents.length || !jsCode) {
+
+        // 空script
+        if (!script.children.length) {
             continue;
         }
-        // 要保留住原来的属性
-        let oldAttrs = "";
-        for(let key in script.attrs) {
-            oldAttrs += ` "${key.replace("\"", "\\\"")}"="${script.attrs[key].replace("\"", "\\\"")}" `
+
+        // script的内容
+        let jsCode = "";
+        for(let child of script.children) {
+            jsCode += child.data;
         }
+        if (!jsCode) {
+            return;
+        }
+
         let newJsCode = injectHook(jsCode);
         // 随着script替换时注入，不创建新的script标签
         if (!alreadyInjectHookContext) {
             newJsCode = loadPluginsAsStringWithCache() + newJsCode;
             alreadyInjectHookContext = true;
         }
-        script.replaceWith("<script "+ oldAttrs +">" + newJsCode + "</script>")
+
+        const newScript = cheerio.load("<script>" + newJsCode + "</script>")("script");
+        newScript.attribs = script.attribs;
+        $(script).replaceWith(newScript);
     }
-    responseDetail.response.body = soup.prettify();
+    responseDetail.response.body = $.html();
 }
 
 /**

src/components/global-assign-hook-component/core/plugins-manager.js

@@ -25,12 +25,15 @@ function loadPluginsAsStringWithCache() {
 function loadPluginsAsString() {
 
     // 用来保证Hook代码只被加载一次
+    // TODO 妥善处理Worker环境
     const loadOnce = "\n" +
-        // 在Worker环境下会出现window未定义
-        "if (!window || window.cc11001100_hook_done) {\n" +
-        "    return;\n" +
-        "}\n" +
-        "window.cc11001100_hook_done = true;\n\n";
+        "   if (!window) {\n" +
+        "       return; \n" +
+        "   } \n" +
+        "   if (window.cc11001100_hook_done) {\n" +
+        "       return;\n" +
+        "   }\n" +
+        "   window.cc11001100_hook_done = true;\n\n";
 
     const hookJsCode = fs.readFileSync("../components/global-assign-hook-component/core/hook.js").toString();
 
@@ -42,7 +45,7 @@ function loadPluginsAsString() {
         pluginsJsContentArray.push(pluginJsContent);
     }
 
-    return "// ----------------------------------------- Hook代码开始 ----------------------------------------------------- \n" +
+    return "\n// ----------------------------------------- Hook代码开始 ----------------------------------------------------- \n" +
         "\n(() => {\n" +
 
         loadOnce +