I maintain an addon that seems to be having issues with Firefox 22. There is a JavaScript module that uses loadFrameScript, which in turn injects some libraries using mozIJSSubScriptLoader. The file brought in by loadFrameScript looks similar to below:
// Create a JS sub-script loader.
var loader = Components.classes["#mozilla.org/moz/jssubscript-loader;1"]
.getService(Components.interfaces.mozIJSSubScriptLoader);
// Create a context object.
var executionContext = Object.create(content);
// Load the libraries.
loader.loadSubScript("chrome://my-package/content/libs/jquery.js", executionContext);
loader.loadSubScript("chrome://my-package/content/logic.js", executionContext);
However, the act of loading jQuery throws an exception:
Error: NS_ERROR_XPC_BAD_OP_ON_WN_PROTO: Illegal operation on WrappedNative prototype object
Source File: chrome://my-package/content/libs/jquery.js
Line: 829
It does not look like jQuery is doing anything crazy on that line, just calling setTimeout. Googling around for this message, I found a similar situation in the Scriptish extension, but no resolution. I am at a loss as to what I should be doing differently, or what changes broke the way I load jQuery in Firefox 22. Is there a better way to bring in jQuery?
Update
This really is the most aggravating problem. I dropped using the executionContext object, because I don't even remember why I used it in the first place, and jQuery loads into the content just dandy.
loader.loadSubScript("chrome://my-package/content/libs/jquery.js", content);
loader.loadSubScript("chrome://my-package/content/logic.js", content);
Now, however, other scripts that also get loaded into content cannot use sendAsyncMessage. I suppose this makes sense, as it's a whole new scope that does not have the addon API's, but now I am not sure how to read the page DOM. How do I load my logic and jQuery into content and still retain the ability to sendAsyncMessage results?
Just my two cents -
I'm also maintaining an extension that runs into the problem. For me, the solution is really the same as indicated in scriptish - use window.xxxx instead of directly referencing that method.
For example, previously one of the lines calls setTimeout() directly, after I changed it to window.setTimeout(), the code works.
Since you said that line is not doing anything other than calling setTimeout, I suppose it's the same issue. Try adding window. before that call.
Good luck!
The likely reason you did use executionContext in the first place was that otherwise stuff would be directly defined on content, which could conflict with the website, other add-ons and/or leak to the website. So better have a wrapper around window where you load your stuff.
I just coded up a minimal "content script" loader based on frame scripts. Nothing much, but should get the job done. I verified on FX 24 that jquery will work in it, and the stuff does not leak into the content window.
// Frame scripts share a scope, so better not mess them up ;)
(function() {
"use strict";
const {classes: Cc, interfaces: Ci, utils: Cu} = Components;
const utils = {};
try {
throw new Error();
}
catch (ex) {
let url = ex.fileName.replace(/\/[^\/]*?$/, "/");
const ssm = Cc["#mozilla.org/scriptsecuritymanager;1"].getService(Ci.nsIScriptSecurityManager);
Object.defineProperties(utils, {
"url": {
enumerable: true,
value: function(fn) {
return url + fn;
}
},
"mayLoad": {
enumerable: true,
value: function(o) {
let node = (o.document || o);
let window = (o.ownerDocument || o).defaultView || o;
try {
return window.location != "about:blank" &&
!ssm.isSystemPrincipal(node.nodePrincipal);
}
catch (ex) {
Cu.reportError(ex);
return false;
}
}
},
});
Object.freeze(utils);
}
try {
const loader = Cc["#mozilla.org/moz/jssubscript-loader;1"]
.getService(Ci.mozIJSSubScriptLoader);
// Create a context object for each window that get's loaded.
// Or use DOMWindowCreated, like the add-on manager does to install
// the InstallTrigger.
addEventListener("DOMContentLoaded", function(e) {
let window = e.target.defaultView;
if (!utils.mayLoad(window)) {
// Refuse to load in chrome (system) pages.
return;
}
// Need to create our context in the window scope (compartment).
// The reason to create a wrapper/context it in the first place
// is to avoid clashes with other add-ons, the website itself,
// etc.
let executionContext = Cu.createObjectIn(window);
// Wire up the window to be the prototype.
executionContext.__proto__ = window;
// Add some useful stuff you want the "content scripts" to have
// access to.
Object.defineProperties(executionContext, {
"sendAsyncMessage": {
enumerable: true,
value: sendAsyncMessage.bind(null)
},
"reportError": {
enumerable: true,
value: Cu.reportError.bind(Cu)
},
"doSomething": {
enumerable: true,
value: function(arg) {
Cu.reportError("did something " + arg);
}
},
"loadScript": {
enumerable: true,
value: function(fn) {
loader.loadSubScript(utils.url(fn), executionContext);
}
}
});
// Normalize the properties, i.e. move them over to the correct
// window scope (compartment);
Cu.makeObjectPropsNormal(executionContext);
// Load initial scripts
executionContext.loadScript("test.js");
});
}
catch (ex) {
content.console.error(ex);
}
})();
The key points are:
Use Cu.createObjectIn(window) to get the scope (compartment in Spidermonkey) right and avoid the NS_ERROR_XPC_BAD_OP_ON_WN_PROTO exceptions.
Use Cu.makeObjectPropsNormal(), if you define additional stuff on your context.
Don't try to inject stuff into chrome-privileged windows (utils.mayLoad).
The throw new Error() try-catch is just a reliable hack to get the current URI (ex.fileName) to later allow specifying relative paths when loading scripts.
Related
I'm trying to convert our requirejs calls to use SystemJS, but I'm not exactly sure what I'm doing wrong.
Our original calls look like this:
return function(callback) {
requirejs(["/app/shared.js"], function(result){
callbackFunction = callback;
callback(dashboard);
main();
})
}
And what I'm trying instead is:
return function(callback) {
console.log(callback.toString())
SystemJS.import('app/shared.js').then(function(result){
callbackFunction = callback;
callback(dashboard);
main();
});
}
I've had to remove some leading / to get things to load properly, which is fine, but I've now ran into an issue where variables that were defined at the top of shared.js aren't visible in my local main.js file. In my browser console I get:
Potentially unhandled rejection [1] ReferenceError: dashboard is not defined
shared.js defines dashboard:
var dashboard = { rows: [], }
// Other definitions...
define(["/app/custom-config.js", /* etc */]);
I guess I have two questions:
is this the correct way to replace requirejs calls?
if so, why aren't my variables from shared.js accessible?
For a fuller picture, main() just sets up the dashboard object, and then calls callbackFunction(dashboard) on it.
Your problem can be reduced to the following case where you have two AMD modules, with one that leaks into the global space, and the 2nd one that tries to use what the first one leaked. Like the two following modules.
src/a.js requires the module that leaks and depends on what that module leaks:
define(["./b"], function () {
console.log("a loaded");
callback();
});
src/b.js leaks into the global space:
// This leaks `callback` into the global space.
var callback = function () {
console.log("callback called");
}
define(["./b"], function () {
console.log("b loaded");
});
With RequireJS, the code above will work. Oh, it is badly designed because b.js should not leak into the global space, but it will work. You'll see callback called on the console.
With SystemJS, the code above won't work. Why? RequireJS loads modules by adding a script element to the header and lets script execute the module's code so callback does end up in the global space in exactly the same way it would if you had written your own script element with an src attribute that points to your script. (You'd get an "Mismatched anonymous define" error, but that's a separate issue that need not detain us here.) SystemJS, by default, uses eval rather than create script elements, and this changes how the code is evaluated. Usually, it does not matter, but sometimes it does. In the case at hand here callback does not end up in the global space, and module a fails.
Ultimately, your AMD modules should be written so that they don't use the global space to pass information from one another.
However, there is another solution which may be useful as a stepping-stone towards a final solution. You can use scriptLoad: true to tell SystemJS to use script elements like RequirejS does. (See the documentation on meta for details and caveats.) Here is a configuration that does that:
System.config({
baseURL: "src",
meta: {
"*": {
scriptLoad: true, // This is what fixes the issue.
}
},
packages: {
// Yes, this empty package does something. It makes `.js` the
// default extension for modules.
"": {}
},
});
// We have to put `define` in the global space to
// so that our modules can find it.
window.define = System.amdDefine;
If I run the example code I've given here without scriptLoad: true, then module a cannot call the callback. With scriptLoad: true, it can call the callback and I get on the console:
b loaded
a loaded
callback called
I'm trying to load a JS file from a bookmarklet. The JS file has this JS that wraps the module:
(function (root, factory) {
if (typeof module === 'object' && module.exports) {
// Node/CommonJS
module.exports = factory();
} else if (typeof define === 'function' && define.amd) {
// AMD. Register as an anonymous module.
define(factory);
} else {
// Browser globals
root.moduleGlobal = factory();
}
}(this, function factory() {
// module script is in here
return moduleGlobal;
}));
Because of this, if the webpage uses RequireJS, the script will not export a global when it loads. To get around this I temporarily set define to null, load the script, then reset define to its original value:
function loadScript(url, cb) {
var s = document.createElement('script');
s.src = url;
s.defer = true;
var avoidRequireJS = typeof define === 'function' && define.amd;
if (avoidRequireJS) {
var defineTmp = define;
define = null;
}
s.onload = function() {
if (avoidRequireJS) define = defineTmp;
cb();
};
document.body.appendChild(s);
}
This works, but it seems to me like it could be problematic to change a global variable when other parts of the application could depend on it. Is there a better way to go about this?
You may fetch the script using XMLHttpRequest, jQuery.ajax or the new Fetch API.
This will allow you to manipulate the script and reassign define before executing it. Two options:
Have the module export a global by wrapping the script with:
(function(define){ ... })(null);
Handle the module exports yourself by wrapping the script with:
(function(define, module){ ... })((function() {
function define(factory) {
var exports = factory();
}
define.amd = true;
return define;
})());
You can then load it using a new <script> element or eval 😲.
Note that when using XHR, you may have to address CORS issues.
If you can use the AJAX method above, that will be best. But as stated, you will need to deal with CORS issues, which is not always trivial - even impossible if you do not control the origin server.
Here is a technique which uses an iframe to load the script in an isolated context, allowing the script to export its global object. We then grab the global object and copy it to the parent. This technique does not suffer from CORS restrictions.
(fiddle: https://jsfiddle.net/qu0pxesd/)
function loadScript (url, exportName) {
var iframe = document.createElement('iframe');
Object.assign(iframe.style, {
position: 'fixed',
top: '-9999em',
width: '0px'
});
var script = document.createElement('script');
script.onload = function () {
window[exportName] = iframe.contentWindow[exportName];
document.body.removeChild(iframe);
}
script.src = url;
document.body.appendChild(iframe);
iframe.contentWindow.document.open();
iframe.contentWindow.document.appendChild(script);
iframe.contentWindow.document.close();
}
loadScript('https://ajax.googleapis.com/ajax/libs/jquery/3.2.1/jquery.min.js', 'jQuery');
I ran a quick test to see if a memory leak would happen from deleting the iframe, and it appears to be memory safe. Here's the snapshot of loading a script 100 times, resulting in 100 different iframes and 100 different instances of jQuery loading.
The parent window's jQuery variable is continuously overwritten, meaning only the last one prevails and all previous references are cleaned up. This is not entirely scientific and you will need to do your own testing, but this should be safe enough to get you started.
Update: The above code requires that you know the name of the exported object, which is not always known. Some modules may export multiple variables too. For example, jQuery exports both $ and jQuery. The following fiddle illustrates a technique for solving this issue by copying any global objects which did not exist before the script was loaded:
https://jsfiddle.net/qu0pxesd/3/
Which approach would work best really depends on the specific needs of the project. Context would determine which one I'd use.
Undefining define Temporarily
I'm mentioning it because you tried it.
DON'T DO THIS!
The approach of undefining define before you load your script and restoring it after is not safe. In the general case, it is possible for other code on the page to perform a require call that will resolve after you've undefined define and before you've defined it again. After you do document.body.appendChild(s); you're handing back control to the JavaScript engine, which is free to immediately execute scripts that were required earlier. If the scripts are AMD module, they'll either bomb or install themselves incorrectly.
Wrapping the Script
As Dheeraj V.S. suggests, you can wrap the script to make define locally undefined:
(function(define) { /* original module code */ }())
can work for trivial cases like the one you show in your question. However, cases where the script you try to load actually has dependencies on other libraries can cause issues when it comes to dealing with the dependencies. Some examples:
The page loads jQuery 2.x but the script you are trying to load depends on a feature added in jQuery 3.x. Or the page loads Lodash 2 but the script needs Lodash 4, or vice-versa. (There are huge differences between Lodash 2 and 4.)
The script needs a library that is not otherwise loaded by something else. So now you are responsible for producing the machinery that will load the library.
Using RequireJS Contexts
RequireJS is capable of isolating multiple configurations from one another by defining a new context. Your bookmarklet could define a new context that configures enough paths for the script you are trying to load to load itself and its dependencies:
var myRequire = require.config({
// Within the scope of the page, the context name must be unique to
// your bookmarklet.
context: "Web Designer's Awesome Bookmarklet",
paths: {
myScript: "https://...",
jquery: "https://code.jquery.com/jquery-3.2.1.min.js",
},
map: {...},
// Whatever else you may want.
});
myRequire(["myScript"]);
When you use contexts like this, you want to save the return value of require.config because it is a require call that uses your context.
Creating a Bundle with Webpack
(Or you could use Browserify or some other bundler. I'm more familiar with Webpack.)
You could use Webpack to consume all the AMD modules necessary for the script you are trying to load to produce a bundle that exports its "module" as a global. At a minimum, you'll need something like this in your configuration:
// Tell Webpack what module constitutes the entry into the bundle.
entry: "./MyScript.js",
output: {
// This is the name under which it will be available.
library: "MyLibrary",
// Tell Webpack to make it globally available.
libraryTarget: "global",
// The final bundle will be ./some_directory/MyLibrary.js
path: "./some_directory/",
filename: "MyLibrary.js",
}
Once this is done, the bookmarklet only needs to insert a new script element that points to the produced bundle and no longer has to worry about wrapping anything or dealing with dependencies.
If it were me, I would have the url provide the hint as to how to load the module. Instead of having just a "scripts/" directory -> I would make "scripts/amd/", "scripts/require/", etc. Then query the url for "amd", "require", etc. within your loadScript method... using, e.g.,
if (url.includes('amd')) {
// do something
} else if (url.includes('require')) {
// do something different
}
That should let you avoid the global var entirely. It might also provide a better structure for your app in general.
You could also return an object with a script property and loadType property that specifies amd, require, etc... but imho the first option would be the quickest and save you some additional typing.
Cheers
I am working on project that uses quite a few js libraries and one of them is outputting awful lot into console, it is polluting the airwaves so bad that it makes it hard to debug....
I know how to disable logging completely by overriding console.log with this,
(function (original) {
console.enableLogging = function () {
console.log = original;
};
console.disableLogging = function () {
console.log = function () {};
};
})(console.log);
but how do it do that per source(file/url) of where message originated?
Preamble
The beginning discusses how stuff works in general. If you just care for the code, skip Introduction and scroll to the Solution heading.
Introduction
Problem:
there is a lot of console noise in a web application. A significant amount of that noise is coming from third party code which we do not have access to. Some of the log noise might be coming from our code, as well.
Requirement:
reduce the noise by stopping the log. Some logs should still be kept and the decision about those should be decoupled from the code that is doing the logging. The granularity needed is "per-file". We should be able to choose which files do or do not add log messages. Finally, this will not be used in production code.
Assumption: this will be ran in a developer controlled browser. In that case, I will not focus on backwards compatibility.
Prior work:
First off logging can be enabled/disabled globally using this
(function (original) {
console.enableLogging = function () {
console.log = original;
};
console.disableLogging = function () {
console.log = function () {};
};
})(console.log);
(code posted in the question but also here for reference)
However, that does not allow for any granularity.
This could be modified to work on only specific modules but that cannot be done for third party code.
A mixed approach would be to disable logging globally but enable it in each of our modules. Problem there is that we have to modify each of our files and we will not get some potentially useful external messages.
A logging framework can be used but it might be an overkill. Although, to be honest, that's what I'd go for, I think, but it may need some integration into the product.
So, we need something light-weight-ish that has some configuration and does not need to be pretty.
Proposal:
The Loginator (title subject to change)
Let's start with the basics - we already know we can override the global log function. We'll take that and work with it. But first, let's recognise that the console object supports more than just .log. There could be various logging functions used. So-o-o, let's disable all of them.
Silence everything
//shorthand for further code.
function noop() {}
const savedFunctions = Object.keys(console)
.reduce((memo, key) => {
if(typeof console[key] == "function") {
//keep a copy just in case we need it
memo[key] = console[key];
//de-fang any functions
console[key] = noop;
}
return memo;
},
{});
console.log("Hello?");
console.info("Hello-o-o-o?");
console.warn("Can anybody hear me?");
console.error("I guess there is nobody there...");
savedFunctions.log("MUAHAHAHA!")
This can obviously be improved but it showcases how any and ll logging can be stopped. In reality, console.error should probably be left and console.warn might be also useful. But this is not the be-all-and-end-all solution.
Next, since we can override console functionality...why not supply our own?
Custom logging
const originalLog = console.log;
console.log = function selectiveHearing() {
if (arguments[0].indexOf("die") !== -1) {
arguments[0] = "Have a nice day!";
}
return originalLog.apply(console, arguments)
}
console.log("Hello.");
console.log("My name is Inigo Montoya.");
console.log("You killed my father.");
console.log("Prepare to die.");
That is all the tools we need to roll our own mini-logging framework.
How to do selective logging
The only thing missing is to determine which file something is coming from. We just need a stack trace.
// The magic
console.log(new Error().stack);
/* SAMPLE:
Error
at Object.module.exports.request (/home/vagrant/src/kumascript/lib/kumascript/caching.js:366:17)
at attempt (/home/vagrant/src/kumascript/lib/kumascript/loaders.js:180:24)
at ks_utils.Class.get (/home/vagrant/src/kumascript/lib/kumascript/loaders.js:194:9)
at /home/vagrant/src/kumascript/lib/kumascript/macros.js:282:24
at /home/vagrant/src/kumascript/node_modules/async/lib/async.js:118:13
at Array.forEach (native)
at _each (/home/vagrant/src/kumascript/node_modules/async/lib/async.js:39:24)
at Object.async.each (/home/vagrant/src/kumascript/node_modules/async/lib/async.js:117:9)
at ks_utils.Class.reloadTemplates (/home/vagrant/src/kumascript/lib/kumascript/macros.js:281:19)
at ks_utils.Class.process (/home/vagrant/src/kumascript/lib/kumascript/macros.js:217:15)
*/
(Relevant bit copied here.)
True, there are some better ways to do it but not a lot. It would either require a framework or it's browser specific - error stacks are not officially supported but they work in Chrome, Edge, and Firefox. Also, come on - it's literally one line - we want simple and don't mind dirty, so I'm happy for the tradeoff.
Solution
Putting it all together. Warning: Do NOT use this in production
(function(whitelist = [], functionsToPreserve = ["error"]) {
function noop() {}
//ensure we KNOW that there is a log function here, just in case
const savedFunctions = { log: console.log }
//proceed with nuking the rest of the chattiness away
Object.keys(console)
.reduce((memo, key) => {
if(typeof console[key] == "function" && functionsToPreserve.indexOf(key) != -1 ) {
memo[key] = console[key];
console[key] = noop;
}
return memo;
},
savedFunctions); //<- it's a const so we can't re-assign it. Besides, we don't need to, if we use it as a seed for reduce()
console.log = function customLog() {
//index 0 - the error message
//index 1 - this function
//index 2 - the calling function, i.e., the actual one that did console.log()
const callingFile = new Error().stack.split("\n")[2];
if (whitelist.some(entry => callingFile.includes(entry))) {
savedFunctions.log.apply(console, arguments)
}
}
})(["myFile.js"]) //hey, it's SOMEWHAT configurable
Or a blacklist
(function(blacklist = [], functionsToPreserve = ["error"]) {
function noop() {}
//ensure we KNOW that there is a log function here, just in case
const savedFunctions = {
log: console.log
}
//proceed with nuking the rest of the chattiness away
Object.keys(console)
.reduce((memo, key) => {
if (typeof console[key] == "function" && functionsToPreserve.indexOf(key) != -1) {
memo[key] = console[key];
console[key] = noop;
}
return memo;
},
savedFunctions); //<- it's a const so we can't re-assign it. Besides, we don't need to, if we use it as a seed for reduce()
console.log = function customLog() {
//index 0 - the error message
//index 1 - this function
//index 2 - the calling function, i.e., the actual one that did console.log()
const callingFile = new Error().stack.split("\n")[2];
if (blacklist.some(entry => callingFile.includes(entry))) {
return;
} else {
savedFunctions.log.apply(console, arguments);
}
}
})(["myFile.js"])
So, this is a custom logger. Sure, it's not perfect but it will do the job. And, hey, since the whitelisting is a bit loose, it could be turned to an advantage:
to whitelist a bunch of files that share a substring, say, all myApp can include myApp1.js, myApp2.js, and myApp3.js.
although if you want specific files, you can just pass the full name, including extension. I doubt there would be a bunch of duplicate filenames.
finally, the stack trace will include the name of the calling function, if any, so you can actually just pass that and that will whitelist on per-function basis. However, it relies on the function having a name and it's more likely for function names to clash, so use with care
Other than that, there can certainly be improvements but that is the basis of it. The info/warn methods can also be overriden, for example.
So, this, if used, should only be in dev builds. There are a lot of ways to make it not go into production, so I won't discuss them but here is one thing I can mention: you can also use this anywhere if you save it as a bookmarklet
javascript:!function(){function c(){}var a=arguments.length<=0||void 0===arguments[0]?[]:arguments[0],b=arguments.length<=1||void 0===arguments[1]?["error"]:arguments[1],d={log:console.log};Object.keys(console).reduce(function(a,d){return"function"==typeof console[d]&&b.indexOf(d)!=-1&&(a[d]=console[d],console[d]=c),a},d),console.log=function(){var c=(new Error).stack.split("\n")[2];a.some(function(a){return c.includes(a)})&&d.log.apply(console,arguments)}}(["myFile.js"]);
This is it minified (although I passed it through Babel first, to use ES5 minification) and still configurable, to an extent, as you can change the very end where you can pass the whitelist. But other than that, it will work the same and is completely decoupled from the codebase. It will not run at pageload but if that's needed you can either use this as a userscript (still decoupled) or include it before other JS files in dev/debug builds only.
A note here - this will work in Chrome, Edge and Firefox. It's all the latest browsers, so I assume a developer will use at least one of them. The question is tagged as Chrome but I decided to widen the support. A Chrome only solution could work slightly better but it's not really a big loss of functionality.
I was as troubled as you. This is my approach. https://github.com/jchnxu/guard-with-debug
Simple usage:
localStorage.debug = [
'enable/console/log/in/this/file.ts',
'enable/console/log/in/this/folder/*',
'-disable/console/log/in/this/file.ts',
'-disable/console/log/in/this/folder/*',
// enable all
'*',
].join(',');
The benefit: it's zero-runtime.
Disclaimer: I am the author of this tiny utility
It work in chrome:
...index.html
<html>
<body>
<script>
(function(){
var original = console.log;
console.log = function(){
var script = document.currentScript;
alert(script.src);
if(script.src === 'file:///C:/Users/degr/Desktop/script.js') {
original.apply(console, arguments)
}
}
})();
console.log('this will be hidden');
</script>
<script src="script.js"></script>
</body>
</html>
...script.js
console.log('this will work');
Console.log does not work from index.html, but work from script.js. Both files situated on my desctop.
I've found these settings in the latest (July 2020) Chrome DevTools console to be helpful:
DevTools | Console | (sidebar icon) | user messages
DevTools | Console | (gear icon) | Select context only
DevTools | Console | (gear icon) | Hide network
I like (1) most, I only see the messages from "my" code. (2) hides messages from my iframe.
If it's an option to modify file, you can set a flag at top of file for disabling logs for that:
var DEBUG = false;
DEBUG && console.log("cyberpunk 2077");
To disable logs for all js files, put it once at top of any js file:
var DEBUG = false;
if (!DEBUG) {
console.log = () => {};
}
This is not pretty but will work.
Put something like this in your file before the <script> tag of the "bad" library :
<script>function GetFile(JSFile) {
var MReq = new XMLHttpRequest();
MReq.open('GET', JSFile, false);
MReq.send();
eval(MReq.responseText.replace(/console.log\(/g,"(function(){})("));
}</script>
Then replace the tag
<script src="badLib.js">
With:
GetFile("badLib.js")
Only for short time debugging.
I wrote modules for both browser and nodejs using requirejs.
Everything works fine, but I want to include a module just for the browser, not for node as I don't need it and it would not work in node. (It's just a fancy design library for the browser).
My code looks like this:
define([
'requirement',
'libs/fancy'
], function(Requirement, fancy) {
// do stuff
});
fancy is the lib that I don't want in node. So I could write a workaround like this:
if (typeof window !== 'undefined') { // cheap detection of browser/node
define([
'requirement',
'libs/fancy'
], start);
} else {
define([
'requirement'
], start);
}
function start(Requirement, Fancy) {
// do stuff
}
But obviously this is ugly. Does anyone know a better way to do it?
-- EDIT 1:
var requirements = ['requirement'];
if (typeof window !== 'undefined') {
requirement.push('libs/fancy');
}
define(requirements, function(Requirement, Fancy) {
// do stuff
}
Still not perfect
I've sometimes used the second method you show of creating an array of dependencies on which I push depending on what I need.
There's another method, however, that I've used when I don't want to modify the list of dependencies. Presumably the code inside your module will have to work with an undefined value for Fancy. So you could use something like what follows. The idea is to configure RequireJS to load a module that returns an undefined value when loaded. This way you don't need to modify your dependency list. It just needs to be able to handle a case where Fancy is undefined.
var requirejs = require("requirejs");
// Create a fake module that we name immediately as "undefined".
requirejs.define("undefined", [], function () { return undefined; });
var req = requirejs.config({
map: {
// Make it so that all requests for `foo` load `undefined` instead.
"*": {
foo: "undefined"
}
}
});
req(["foo"], function (foo) {
console.log(foo);
});
The example above maps foo to undefined so when console.log executes, the value on the console is undefined. In your own code you'd map libs/fancy to undefined.
A variation on this method would be to have the undefined module return an object which shows the same interface as the real library but does nothing. This would avoid having to test whether Fancy is defined or not inside your module. I'd call the fake module something else than undefined though. Maybe something like fake-fancy.
I'm wondering if it's possible to sandbox JavaScript running in the browser to prevent access to features that are normally available to JavaScript code running in an HTML page.
For example, let's say I want to provide a JavaScript API for end users to let them define event handlers to be run when "interesting events" happen, but I don't want those users to access the properties and functions of the window object. Am I able to do this?
In the simplest case, let's say I want to prevent users calling alert. A couple of approaches I can think of are:
Redefine window.alert globally. I don't think this would be a valid approach because other code running in the page (i.e., stuff not authored by users in their event handlers) might want to use alert.
Send the event handler code to the server to process. I'm not sure that sending the code to the server to process is the right approach, because the event handlers need to run in the context of the page.
Perhaps a solution where the server processes the user defined function and then generates a callback to be executed on the client would work? Even if that approach works, are there better ways to solve this problem?
Google Caja is a source-to-source translator that "allows you to put untrusted third-party HTML and JavaScript inline in your page and still be secure."
Have a look at Douglas Crockford's ADsafe:
ADsafe makes it safe to put guest code (such as third party scripted advertising or widgets) on any web page. ADsafe defines a subset of JavaScript that is powerful enough to allow guest code to perform valuable interactions, while at the same time preventing malicious or accidental damage or intrusion. The ADsafe subset can be verified mechanically by tools like JSLint so that no human inspection is necessary to review guest code for safety. The ADsafe subset also enforces good coding practices, increasing the likelihood that guest code will run correctly.
You can see an example of how to use ADsafe by looking at the template.html and template.js files in the project's GitHub repository.
I created a sandboxing library called jsandbox that uses web workers to sandbox evaluated code. It also has an input method for explicitly giving sandboxed code data it wouldn't otherwise be able to get.
The following is an example of the API:
jsandbox
.eval({
code : "x=1;Math.round(Math.pow(input, ++x))",
input : 36.565010597564445,
callback: function(n) {
console.log("number: ", n); // number: 1337
}
}).eval({
code : "][];.]\\ (*# ($(! ~",
onerror: function(ex) {
console.log("syntax error: ", ex); // syntax error: [error object]
}
}).eval({
code : '"foo"+input',
input : "bar",
callback: function(str) {
console.log("string: ", str); // string: foobar
}
}).eval({
code : "({q:1, w:2})",
callback: function(obj) {
console.log("object: ", obj); // object: object q=1 w=2
}
}).eval({
code : "[1, 2, 3].concat(input)",
input : [4, 5, 6],
callback: function(arr) {
console.log("array: ", arr); // array: [1, 2, 3, 4, 5, 6]
}
}).eval({
code : "function x(z){this.y=z;};new x(input)",
input : 4,
callback: function(x) {
console.log("new x: ", x); // new x: object y=4
}
});
An improved version of RyanOHara's web workers sandbox code, in a single file (no extra eval.js file is necessary).
function safeEval(untrustedCode)
{
return new Promise(function (resolve, reject)
{
var blobURL = URL.createObjectURL(new Blob([
"(",
function ()
{
var _postMessage = postMessage;
var _addEventListener = addEventListener;
(function (obj)
{
"use strict";
var current = obj;
var keepProperties =
[
// Required
'Object', 'Function', 'Infinity', 'NaN', 'undefined', 'caches', 'TEMPORARY', 'PERSISTENT',
// Optional, but trivial to get back
'Array', 'Boolean', 'Number', 'String', 'Symbol',
// Optional
'Map', 'Math', 'Set',
];
do
{
Object.getOwnPropertyNames(current).forEach(function (name)
{
if (keepProperties.indexOf(name) === -1)
{
delete current[name];
}
});
current = Object.getPrototypeOf(current);
}
while (current !== Object.prototype)
;
})(this);
_addEventListener("message", function (e)
{
var f = new Function("", "return (" + e.data + "\n);");
_postMessage(f());
});
}.toString(),
")()"],
{type: "application/javascript"}));
var worker = new Worker(blobURL);
URL.revokeObjectURL(blobURL);
worker.onmessage = function (evt)
{
worker.terminate();
resolve(evt.data);
};
worker.onerror = function (evt)
{
reject(new Error(evt.message));
};
worker.postMessage(untrustedCode);
setTimeout(function ()
{
worker.terminate();
reject(new Error('The worker timed out.'));
}, 1000);
});
}
Test it:
https://jsfiddle.net/kp0cq6yw/
var promise = safeEval("1+2+3");
promise.then(function (result) {
alert(result);
});
It should output 6 (tested in Chrome and Firefox).
As mentioned in other responces, it's enough to jail the code in a sandboxed iframe (without sending it to the server-side) and communicate with messages.
I would suggest to take a look at a small library I created mostly because of the need to providing some API to the untrusted code, just like as described in the question: there's an opportunity to export the particular set of functions right into the sandbox where the untrusted code runs. And there's also a demo which executes the code submitted by a user in a sandbox:
http://asvd.github.io/jailed/demos/web/console/
I think that js.js is worth mentioning here. It's a JavaScript interpreter written in JavaScript.
It's about 200 times slower than native JavaScript, but its nature makes it a perfect sandbox environment. Another drawback is its size – almost 600 KB, which may be acceptable for desktops in some cases, but not for mobile devices.
All the browser vendors and the HTML5 specification are working towards an actual sandbox property to allow sandboxed iframes -- but it's still limited to iframe granularity.
In general, no degree of regular expressions, etc. can safely sanitise arbitrary user provided JavaScript as it degenerates to the halting problem :-/
An ugly way, but maybe this works for you:
I took all the globals and redefined them in the sandbox scope, as well I added the strict mode so they can't get the global object using an anonymous function.
function construct(constructor, args) {
function F() {
return constructor.apply(this, args);
}
F.prototype = constructor.prototype;
return new F();
}
// Sanboxer
function sandboxcode(string, inject) {
"use strict";
var globals = [];
for (var i in window) {
// <--REMOVE THIS CONDITION
if (i != "console")
// REMOVE THIS CONDITION -->
globals.push(i);
}
globals.push('"use strict";\n'+string);
return construct(Function, globals).apply(inject ? inject : {});
}
sandboxcode('console.log( this, window, top , self, parent, this["jQuery"], (function(){return this;}()));');
// => Object {} undefined undefined undefined undefined undefined undefined
console.log("return of this", sandboxcode('return this;', {window:"sanboxed code"}));
// => Object {window: "sanboxed code"}
https://gist.github.com/alejandrolechuga/9381781
An independent JavaScript interpreter is more likely to yield a robust sandbox than a caged version of the built-in browser implementation.
Ryan has already mentioned js.js, but a more up-to-date project is JS-Interpreter. The documentation covers how to expose various functions to the interpreter, but its scope is otherwise very limited.
As of 2019, vm2 looks like the most popular and most regularly-updated solution to running JavaScript in Node.js. I'm not aware of a front-end solution.
With NISP you'll be able to do sandboxed evaluation.
Though the expression you write is not exactly JavaScript code, instead you'll write S-expressions. It is ideal for simple DSLs that doesn't demand extensive programming.
Suppose you have code to execute:
var sCode = "alert(document)";
Now, suppose you want to execute it in a sandbox:
new Function("window", "with(window){" + sCode + "}")({});
These two lines when executed will fail, because "alert" function is not available from the "sandbox"
And now you want to expose a member of window object with your functionality:
new Function("window", "with(window){" + sCode + "}")({
'alert':function(sString){document.title = sString}
});
Indeed you can add quotes escaping and make other polishing, but I guess the idea is clear.
Where is this user JavaScript code coming from?
There is not much you can do about a user embedding code into your page and then calling it from their browser (see Greasemonkey). It's just something browsers do.
However, if you store the script in a database, then retrieve it and eval() it, then you can clean up the script before it is run.
Examples of code that removes all window. and document. references:
eval(
unsafeUserScript
.replace(/\/\/.+\n|\/\*.*\*\/, '') // Clear all comments
.replace(/\s(window|document)\s*[\;\)\.]/, '') // Removes window. Or window; or window)
)
This tries to prevent the following from being executed (not tested):
window.location = 'http://example.com';
var w = window;
There are a lot of limitations you would have to apply to the unsafe user script. Unfortunately, there isn't any 'sandbox container' available for JavaScript.
I've been working on a simplistic JavaScript sandbox for letting users build applets for my site. Although I still face some challenges with allowing DOM access (parentNode just won't let me keep things secure =/), my approach was just to redefine the window object with some of its useful/harmless members, and then eval() the user code with this redefined window as the default scope.
My "core" code goes like this... (I'm not showing it entirely ;)
function Sandbox(parent){
this.scope = {
window: {
alert: function(str){
alert("Overriden Alert: " + str);
},
prompt: function(message, defaultValue){
return prompt("Overriden Prompt:" + message, defaultValue);
},
document: null,
.
.
.
.
}
};
this.execute = function(codestring){
// Here some code sanitizing, please
with (this.scope) {
with (window) {
eval(codestring);
}
}
};
}
So, I can instantiate a Sandbox and use its execute() function to get code running. Also, all new declared variables within eval'd code will ultimately bound to the execute() scope, so there will not be clashing names or messing with existing code.
Although global objects will still be accessible, those which should remain unknown to the sandboxed code must be defined as proxies in the Sandbox::scope object.
You can wrap the user's code in a function that redefines forbidden objects as parameters -- these would then be undefined when called:
(function (alert) {
alert ("uh oh!"); // User code
}) ();
Of course, clever attackers can get around this by inspecting the JavaScript DOM and finding a non-overridden object that contains a reference to the window.
Another idea is scanning the user's code using a tool like JSLint. Make sure it's set to have no preset variables (or: only variables you want), and then if any globals are set or accessed do not let the user's script be used. Again, it might be vulnerable to walking the DOM -- objects that the user can construct using literals might have implicit references to the window object that could be accessed to escape the sandbox.