谷歌浏览器插件MV3报错"Uncaught ReferenceError: window is not defined"

  • 谷歌浏览器插件MV3报错"Uncaught ReferenceError: window is not defined"已关闭评论
  • 14 次浏览
  • A+
所属分类:Web前端
摘要

配置mv3后,在后台代码background.js使用DOMPurify发现无法访问window,会一直报错
Uncaught ReferenceError: window is not defined
查看后台,globalThis变成了一个叫ServiceWorkerGlobalScope的玩意


出错

配置mv3后,在后台代码background.js使用DOMPurify发现无法访问window,会一直报错
Uncaught ReferenceError: window is not defined
查看后台,globalThis变成了一个叫ServiceWorkerGlobalScope的玩意

原因

mv3使用了一个叫Service workers的东西替代原来的background页面,无法访问完整的dom API,所以不管是window还是document、HTMLElement……都会xx is not defined
chrome官方介绍:

Manifest V3 replaces background pages with service workers.
Like their web page counterparts, extension service workers listen for and respond to events in order to enhance the end user's experience. For web service workers this typically means managing cache, preloading resources, and enabling offline web pages. While extension service workers can still do all of this, the extension package already contains a bundle of resources that can be accessed offline. As such, extension service workers tend to focus on reacting to relevant browser events exposed by Chrome's extensions APIs.

大意就是在后台访问dom API不好,用chrome插件API好
另附mv2和mv3的区别对比表:

MV2 - Background page MV3 - Service worker
Can use a persistent page. Terminates when not in use.
Has access to the DOM. Doesn't have access to the DOM.
Can use XMLHttpRequest(). Must use fetch() to make requests.

可见官方明确了禁止直接访问dom,只能通过一些方式间接实现了

解决

放弃在后台页直接调用dom API

chrome官方态度还是比较明确的,所以最好不要花太大力气去搞。能用chrome API的全都用chrome API实现,特别是标签和窗口的操作。另外chrome也提供了一些mv2到mv3迁移的建议migrating_to_service_workers

使用undom模拟

undom是比jsdom轻量的Document接口实现,基本满足dom操作的需求,简单易用。
安装

pnpm install --save undom 

简单使用

import undom from 'undom'; let document = undom();  let foo = document.createElement('foo'); foo.appendChild(document.createTextNode('Hello, World!')); document.body.appendChild(foo);  // 驱动第三方库DOMPurify const purify = DOMPurify(document); purify.sanitize(html..) 

另外undom本身不支持直接输出HTML的,它只实现了Document的骨架,不能直接用innerHTML、querySelector等,但提供了手动解析的方法

function serialize(el) { 	if (el.nodeType===3) return el.textContent; 	var name = String(el.nodeName).toLowerCase(), 		str = '<'+name, 		c, i; 	for (i=0; i<el.attributes.length; i++) { 		str += ' '+el.attributes[i].name+'="'+el.attributes[i].value+'"'; 	} 	str += '>'; 	for (i=0; i<el.childNodes.length; i++) { 		c = serialize(el.childNodes[i]); 		if (c) str += 'nt'+c.replace(/n/g,'nt'); 	} 	return str + (c?'n':'') + '</'+name+'>'; }  function enc(s) { 	return s.replace(/[&'"<>]/g, function(a){ return `&#${a};` }); }  // 输出完整html console.log(serialize(document.childNodes[0])); // 转义html console.log(enc(serialize(document.childNodes[0])));