🐛 修复 @include 脚本代码被注入到所有页面 (#1803)

docs/references/architecture-execution.md

@@ -35,7 +35,14 @@ Key points:
 
 Normal userscripts run in the page. The SW builds a `RegisteredUserScript` from the script's `@match`/`@include`
 patterns and registers the compiled payload (the `scripting` bundle) with `chrome.userScripts.register`, in the
-`MAIN` or `USER_SCRIPT` world as required. At document time the content/inject pair
+`MAIN` or `USER_SCRIPT` world as required. The browser injects where a URL hits `matches` **or** `includeGlobs`
+(a union, not a filter), so `getApiMatchesAndGlobs` ([`url_matcher.ts`](../../src/pkg/utils/url_matcher.ts)) keeps
+`matches` to the script's real match patterns and never widens it to `*://*/*` to make room for globs. A regex
+`@include` that cannot be narrowed to a host-level glob is the exception: `getUserScriptRegister` marks it
+`deliverOnPageLoad` and the script is not registered. Its injection code instead travels in the `pageLoad` reply
+(`TScriptInfo.code`) to the content script, which mounts it from the `USER_SCRIPT` world, where an inserted
+`<script>` is not subject to the page CSP. `@early-start` and `@unwrap` scripts, and Firefox, keep the site-wide
+registration. At document time the content/inject pair
 ([`script_runtime.ts`](../../src/app/service/content/script_runtime.ts),
 [`exec_script.ts`](../../src/app/service/content/exec_script.ts)) evaluates the compiled function with the GM
 context.

e2e/include-injection-scope.spec.ts

@@ -0,0 +1,82 @@
+import { test, expect, startMockServer, type MockServer } from "./server-fixtures";
+import { installScriptByCode } from "./utils";
+import type { Page } from "@playwright/test";
+
+// #1803：浏览器对 matches 与 includeGlobs 取并集，@include 脚本曾被注册成全站，代码挂到每个页面的 window 上。
+// 这里在真实 Chrome 里验证三件事：glob 精确注册后仍在目标页（含 #hash 网址）执行；正则收不窄的脚本改由 pageLoad
+// 下发代码后，在严格 CSP 的目标页照常执行；两者都不会把代码留在不相关的页面上。
+
+function markerScript(name: string, rule: string, marker: string) {
+  return `// ==UserScript==
+// @name         ${name}
+// @namespace    issue-1803-e2e
+// @version      1.0.0
+// @include      ${rule}
+// @grant        none
+// ==/UserScript==
+document.documentElement.dataset.${marker} = "ran";`;
+}
+
+const GLOB_SCRIPT = markerScript("Issue 1803 glob", "http://*siteb.test*/page", "scGlob");
+// 没有 scheme 的局部正则抽不出网域，只能不预注册、随 pageLoad 下发
+const REGEX_SCRIPT = markerScript("Issue 1803 regex", "/content-security-policy\\.test:\\d+\\/page/", "scRegex");
+const CONTROL_SCRIPT = `// ==UserScript==
+// @name         Issue 1803 control
+// @namespace    issue-1803-e2e
+// @version      1.0.0
+// @match        http://sitea.test/*
+// @grant        none
+// ==/UserScript==
+document.documentElement.dataset.scControl = "ran";`;
+
+const waitForMarker = (page: Page, marker: string) =>
+  page.waitForFunction((key) => document.documentElement.dataset[key] === "ran", marker, { timeout: 10_000 });
+
+test.describe("Issue 1803: @include 脚本只注入到匹配的页面", () => {
+  let server: MockServer;
+
+  test.beforeEach(async () => {
+    server = await startMockServer();
+  });
+
+  test.afterEach(async () => {
+    await server.close();
+  });
+
+  test("glob 与收不窄的正则 @include 都在各自的目标页执行，严格 CSP 也不例外", async ({ context, extensionId }) => {
+    await installScriptByCode(context, extensionId, GLOB_SCRIPT);
+    await installScriptByCode(context, extensionId, REGEX_SCRIPT);
+    const page = await context.newPage();
+    try {
+      // glob 不以 * 结尾，浏览器拿带 hash 的完整网址匹配，靠补上的 #* 变体命中
+      await page.goto(server.url("siteb.test", "/page#section"), { waitUntil: "domcontentloaded" });
+      await waitForMarker(page, "scGlob");
+
+      // 该主机下发 script-src 'none'，代码由 content 环境插入页面
+      await page.goto(server.url("content-security-policy.test", "/page"), { waitUntil: "domcontentloaded" });
+      await waitForMarker(page, "scRegex");
+    } finally {
+      await page.close();
+    }
+  });
+
+  test("不匹配的页面上不留下任何脚本代码", async ({ context, extensionId }) => {
+    await installScriptByCode(context, extensionId, GLOB_SCRIPT);
+    await installScriptByCode(context, extensionId, REGEX_SCRIPT);
+    await installScriptByCode(context, extensionId, CONTROL_SCRIPT);
+    const page = await context.newPage();
+    try {
+      await page.goto(server.url("sitea.test", "/page"), { waitUntil: "domcontentloaded" });
+      // 对照脚本执行说明本页的 pageLoad 已经完成，随 pageLoad 下发的代码此时也已注入（如果有的话）
+      await waitForMarker(page, "scControl");
+
+      const leaked = await page.evaluate(() => ({
+        flags: Object.keys(window).filter((key) => key.startsWith("#-")),
+        markers: { ...document.documentElement.dataset },
+      }));
+      expect(leaked).toEqual({ flags: [], markers: { scControl: "ran" } });
+    } finally {
+      await page.close();
+    }
+  });
+});

src/app/repo/resource.ts

@@ -83,14 +83,16 @@ export type CompiledResource = {
   uuid: string;
   require: string[]; // 仅存储url，节省空间
   matches: string[]; // primary
-  includeGlobs: string[]; // includeGlobs applied after matches
+  includeGlobs: string[]; // 浏览器与 matches 取并集，见 getApiMatchesAndGlobs
   excludeMatches: string[];
   excludeGlobs: string[];
   allFrames: boolean;
   world: string;
   runAt: string;
   scriptUrlPatterns: URLRuleEntry[];
   originalUrlPatterns: URLRuleEntry[] | null;
+  // 不向浏览器预注册，注入代码由 pageLoad 按匹配结果下发（见 getUserScriptRegister）
+  deliverOnPageLoad: boolean;
 };
 
 export const ResourceNamespace = "76f45084-91b1-42c1-8be8-cbcc54b171f0";
@@ -110,7 +112,7 @@ export class ResourceDAO extends Repo<Resource> {
 }
 
 // CompiledResource结构或 matches 计算规则变更时，建议修改 CompiledResourceNamespace 以删除旧Cache
-export const CompiledResourceNamespace = "39a5eaa9-1cd2-446b-82b8-dd7bd3719f52";
+export const CompiledResourceNamespace = "8f0d6c1e-3b4a-4f55-9d2e-6a7c1b9e5f13";
 
 export class CompiledResourceDAO extends Repo<CompiledResource> {
   constructor() {

src/app/repo/scripts.ts

@@ -136,6 +136,7 @@ export type TScriptInfo = Override<
     originalMetadata?: Partial<Record<string, string[]>>;
     resource: PageScriptResource;
     requireCssResource?: PageScriptResource;
+    // 只有不预注册（CompiledResource.deliverOnPageLoad）的脚本才非空：挂载 window[flag] 的注入代码，由 content 负责注入
     code: "" | string;
     sort?: number;
     flag: string;

src/app/service/content/script_runtime.ts

@@ -52,6 +52,23 @@ export class ScriptRuntime {
       const nodeId = msg.sendRelatedTarget(el);
       return nodeId;
     });
+
+    // 未预注册的脚本（见 getUserScriptRegister 的 deliverOnPageLoad），注入代码随 pageLoad 下发到这里挂载 window[flag]，
+    // 执行环境照常通过 flag 接手。content 运行在 USER_SCRIPT world，其 CSP 放行内联脚本与 eval，
+    // 插入页面的 <script> 不受页面 CSP 限制
+    this.server.on("loadScriptCode", (data: { pageCodes: string[]; contentCodes: string[] }) => {
+      for (const code of data.contentCodes) {
+        new Function(code)();
+      }
+      for (const code of data.pageCodes) {
+        // createElementNS 确保 XML 文档里也能执行（#211）
+        const el = document.createElementNS("http://www.w3.org/1999/xhtml", "script");
+        el.setAttribute("type", "text/javascript");
+        el.textContent = code;
+        document.documentElement.appendChild(el);
+        el.remove();
+      }
+    });
   }
 
   init() {

src/app/service/content/scripting.ts

@@ -169,9 +169,26 @@ export default class ScriptingRuntime {
       }
       this.activeStorageNames = new Map(Object.entries(pairs));
 
+      const contentClient = new Client(this.senderToContent, "content");
+      // 未预注册脚本的注入代码交给 content 环境注入：本环境是 ISOLATED world，插入的 <script> 会被页面 CSP 拦下
+      const pageCodes: string[] = [];
+      const contentCodes: string[] = [];
+      for (const [list, codes] of [
+        [injectScriptList, pageCodes],
+        [contentScriptList, contentCodes],
+      ] as const) {
+        for (const script of list) {
+          if (!script.code) continue;
+          codes.push(script.code);
+          script.code = "";
+        }
+      }
+      if (pageCodes.length || contentCodes.length) {
+        contentClient.do("loadScriptCode", { pageCodes, contentCodes });
+      }
+
       // 向页面 发送脚本列表及环境信息
       if (contentScriptList.length) {
-        const contentClient = new Client(this.senderToContent, "content");
         // 根据@inject-into content过滤脚本
         contentClient.do("pageLoad", { scripts: contentScriptList, envInfo });
       }

src/app/service/service_worker/runtime.test.ts

@@ -347,6 +347,31 @@ describe.concurrent("RuntimeService - getPageScriptMatchingResultByUrl 脚本匹
     expect(unregisterSpy).toHaveBeenCalledWith([script.uuid]);
   });
 
+  it.concurrent(
+    "正则 @include 收不窄的脚本改随 pageLoad 下发：记入 CompiledResource 并注销旧注册，不再注册",
+    async () => {
+      const { runtime, mockScriptService } = createRuntimeTestContext();
+      (runtime as any).resource = {
+        getScriptResourceValueByType: vi.fn().mockResolvedValue({ require: {}, "require-css": {}, resource: {} }),
+      };
+      runtime.isUserScriptsAvailable = true;
+      const script = createMockScript({
+        metadata: { include: ["/live\\.bilibili\\.com/"] },
+        status: SCRIPT_STATUS_ENABLE,
+      });
+      mockScriptService.buildScriptRunResource.mockResolvedValue(createScriptRunResource(script));
+      const saveSpy = vi.spyOn(runtime.compiledResourceDAO, "save").mockResolvedValue(undefined as never);
+      const unregisterSpy = vi.spyOn(runtime, "unregistryPageScripts").mockResolvedValue(undefined);
+      const loadSpy = vi.spyOn(runtime, "loadPageScript").mockResolvedValue(undefined);
+
+      await runtime.updateResourceOnScriptChange(script);
+
+      expect(saveSpy).toHaveBeenCalledWith(expect.objectContaining({ uuid: script.uuid, deliverOnPageLoad: true }));
+      expect(unregisterSpy).toHaveBeenCalledWith([script.uuid]);
+      expect(loadSpy).not.toHaveBeenCalled();
+    }
+  );
+
   describe.concurrent("includeDisabled 选项", () => {
     it.concurrent("当 includeDisabled=false 时不返回禁用脚本；当 includeDisabled=true 时返回禁用脚本", async () => {
       // Arrange
@@ -523,6 +548,7 @@ describe.concurrent("RuntimeService - getPageScriptMatchingResultByUrl 脚本匹
         runAt: "document-idle",
         scriptUrlPatterns: patterns.scriptUrlPatterns,
         originalUrlPatterns: null,
+        deliverOnPageLoad: false,
       };
 
       const mockCompiledResourceDAO = {
@@ -983,6 +1009,7 @@ describe("getScriptsForTab 附加边界场景", () => {
       runAt: "document-idle",
       scriptUrlPatterns: patterns.scriptUrlPatterns,
       originalUrlPatterns: null,
+      deliverOnPageLoad: false,
     };
 
     const mockCompiledResourceDAO = {
@@ -1140,6 +1167,41 @@ describe("getScriptsForTab 附加边界场景", () => {
     expect(buildSpy).toHaveBeenCalledTimes(1);
     expect(result).not.toBeNull();
   });
+
+  it("未预注册（deliverOnPageLoad）的脚本随 pageLoad 下发挂载 window[flag] 的注入代码", async () => {
+    const { runtime, scriptRes, compiledResource } = createFullContext();
+    compiledResource.deliverOnPageLoad = true;
+    await runtime.applyScriptMatchInfo(scriptRes);
+
+    const result = await runtime.getScriptsForTab({ url: pageUrl, tabId: undefined, frameId: undefined });
+
+    const [info] = result!.injectScriptList;
+    expect(info.code.startsWith(`window['${info.flag}'] = function(){`)).toBe(true);
+    expect(info.code).toContain("// test");
+  });
+
+  it("已预注册的脚本 pageLoad 不携带代码", async () => {
+    const { runtime, scriptRes } = createFullContext();
+    await runtime.applyScriptMatchInfo(scriptRes);
+
+    const result = await runtime.getScriptsForTab({ url: pageUrl, tabId: undefined, frameId: undefined });
+
+    expect(result!.injectScriptList[0].code).toBe("");
+  });
+
+  it.each([
+    [false, 1],
+    [true, 0],
+  ])("批量注册时 deliverOnPageLoad=%s 的脚本注册条数为 %i", async (deliverOnPageLoad, count) => {
+    const { runtime, script, compiledResource, mockScriptDAO } = createFullContext();
+    compiledResource.deliverOnPageLoad = deliverOnPageLoad;
+    (mockScriptDAO as any).all = vi.fn().mockResolvedValue([script]);
+    vi.spyOn(runtime, "restoreJSCodeFromCompiledResource").mockResolvedValue("window['flag'] = function(){}");
+
+    const list = await runtime.getParticularScriptList({ excludeMatches: [], excludeGlobs: [] });
+
+    expect(list).toHaveLength(count);
+  });
 });
 
 describe("pageLoad 按消息发送方标签页区分隐身上下文", () => {
@@ -1422,6 +1484,7 @@ describe("restoreJSCodeFromCompiledResource 还原代码时的生效 metadata",
       runAt: "document_idle",
       scriptUrlPatterns: scriptURLPatternResults(scriptRes)!.scriptUrlPatterns,
       originalUrlPatterns: null,
+      deliverOnPageLoad: false,
     };
     mockScriptService.buildScriptRunResource.mockResolvedValue(scriptRes);
     (runtime as any).script = {

src/app/service/service_worker/runtime.ts

@@ -105,6 +105,7 @@ type TPageLoadScriptCache = {
   userConfig: UserConfig | undefined;
   resourceByType: TRuntimeResourceByType;
   localResources: TLocalResourceCache[];
+  deliverOnPageLoad: boolean;
 };
 
 const runtimeGlobal = {
@@ -471,6 +472,11 @@ export class RuntimeService {
       await this.unregistryPageScripts([script.uuid]);
       return;
     }
+    if (ret.compiledResource.deliverOnPageLoad) {
+      // 规则改成收不窄的正则后，之前的预注册必须撤掉，否则代码仍挂在所有页面
+      await this.unregistryPageScripts([script.uuid]);
+      return;
+    }
     const { apiScript } = ret;
     await this.loadPageScript(script, apiScript!);
   }
@@ -869,11 +875,10 @@ export class RuntimeService {
     const scriptMatchInfo = await this.applyScriptMatchInfo(scriptRes);
     if (!scriptMatchInfo) return undefined;
     // 生效规则一条 inclusion 都不剩（用户把当前站点从匹配中移除后可能如此）时不能注册：
-    // getApiMatchesAndGlobs 对没有 match pattern 的规则集会退回 *://*/*，注册出去等于全站运行。
+    // 这种规则集转换不出任何 matches，提前返回，免得后面把它当成「只匹配不支持的 scheme」记警告。
     if (!scriptMatchInfo.scriptUrlPatterns.some((rule) => rule.ruleType & RuleTypeBit.INCLUSION)) return undefined;
 
-    const res = getUserScriptRegister(scriptMatchInfo);
-    const registerScript = res.registerScript;
+    const { registerScript, deliverOnPageLoad } = getUserScriptRegister(scriptMatchInfo);
 
     let jsCode = "";
     if (withCode) {
@@ -906,6 +911,7 @@ export class RuntimeService {
       runAt: registerScript.runAt || "",
       scriptUrlPatterns: scriptUrlPatterns,
       originalUrlPatterns: scriptUrlPatterns === originalUrlPatterns ? null : originalUrlPatterns,
+      deliverOnPageLoad,
     } as CompiledResource;
 
     this.compiledResourceDAO.save(result);
@@ -976,10 +982,10 @@ export class RuntimeService {
           if (!ret) return undefined;
           result = ret.compiledResource;
           resultCode = ret.jsCode;
-        } else {
+        } else if (!result.deliverOnPageLoad) {
           resultCode = await this.restoreJSCodeFromCompiledResource(script, result);
         }
-        if (!resultCode) return undefined;
+        if (result.deliverOnPageLoad || !resultCode) return undefined;
         const registerScript = {
           id: result.uuid,
           js: [{ code: resultCode }],
@@ -1450,6 +1456,7 @@ export class RuntimeService {
       userConfig: codeInfo.userConfig,
       resourceByType: this.cloneRuntimeResourceByType(resourceByType),
       localResources: this.getLocalResourceCacheList(resourceByType),
+      deliverOnPageLoad: compiledResource.deliverOnPageLoad,
     };
   }
 
@@ -1676,7 +1683,11 @@ export class RuntimeService {
     const contentScriptList: TScriptInfo[] = [];
     for (const script of enableScriptList) {
       const list = isInjectIntoContent(script.metadata) ? contentScriptList : injectScriptList;
-      list.push(trimScriptInfo({ ...script }));
+      const info = trimScriptInfo({ ...script });
+      if (this.pageLoadCaches.get(script.uuid)?.deliverOnPageLoad) {
+        info.code = compileInjectionCode(script, script.code, script.scriptUrlPatterns);
+      }
+      list.push(info);
     }
 
     // 发布给 inject 和 content 的信息