从Chrome扩展程序读取页面的链接脚本

Question

I am building a Chrome Extension with the ability to detect custom javascript code from a Browser Action.

So far, I managed to detect when the custom code is in the page itself, but I can't detect it when it is loaded from an external .js file (and that's the purpose of this question, I want to detect the content of a <script src="http://www.externaldomain.com/file.js"></script>

I know this could be achieved since Google did it in its "Tag Assistant" extension but checking their source code, everything is minified and it's hard to figure out how they did it. eg. Tag Assistant

Do you have any idea of which API/trick they used ?

Answer 1

When a website loads up, it runs all the scripts kind of like the console on chrome. if you run alert("hello") using the console,it will run once, and you will have no way to get it back using javascript. Similarily, embedded js files are all run once, and you can't just look into the html or anywhere else to get the code back since the js file is unloaded after being run. All the functions are stored in variables, events linked to those functions, etcetera.

What you can do, and probably will have to do is re-fetch those files using ajax.

I think that is exactly what tag assistant does. After looking for a while I found this:

//for each document.script, get the src, and load it with ajax using the function gf(src,return function), and send the response.responseText to the text identification functions
cf.prototype.qa = function(a, b) {
    var c = {};
    Ea(document.scripts, function(d) {//for each
        !d.src || "" == d.src || c[d.src] || ef(this, d.src) || d.src == a || (c[d.src] = !0, gf(d.src, function(a) {
            cf.M().W.forEach(function(c) {
                c.u() && (c = c.qa(a.responseText, d.src), c.length && b(c, d.src))
            })
        }))
    }, this)
};

function gf is a simple ajax call with a callback function

gf = function(a, b) {
    try {
        var c = new XMLHttpRequest;
        c.open("GET", a, !0);
        var d = !1;
        c.onreadystatechange = function() {
            d || 4 != c.readyState || 200 != c.status || (d = !0, b(c))
        };
        c.send()
    } catch (e) {}
}

source from chrome-extension://kejbdjndbnbjgmefkgdddjlbokphdefk/tag_assistant.js (the content script)

Seeing what tag assistant does, I suggest you get every script element, retrieve its src,and in your content scripts, run a simple ajax call like the one on this page: http://code.tutsplus.com/articles/how-to-make-ajax-requests-with-raw-javascript--net-4855

//for each script, get its src,and do whatever you want with it
for(i=0;i<document.scripts.length;i++){
    load(document.scripts[i].src, function(xhr){
       myCodeAnalyzingFunction(xhr.responseText)
    })
}