Develop Reference

Jint+JSfuck - 'Index was outside the bounds of the array'

I'm trying to run the following code in jint:
Jint.Engine engine = new Jint.Engine();
var result = engine.SetValue("data", data).Execute("(/\\n(.+)/.exec(eval(data.replace(/\\s+/, \"\").slice(0, -2)))[1]);").GetCompletionValue();
Which, when unescaped, is executing the following javascript:
(/\n(.+)/.exec(eval(data.replace(/\s+/, "").slice(0, -2)))[1]);
the data variable corresponds to a JSfuck string, similar to this: https://pastebin.com/vmGAebW5
The problem is that I always get a 'Index was outside the bounds of the array' exception, even though the javascript works fine when run in a browser. Any ideas as to what is causing the issue?

windows.location is not available in Jint as the window object is provided by the browser. If you wanted to be able to run this kind of code then you would need to mock all the browser exposed objects, which is a huge task.
Also you are using SetValue('window.location', "") which I assume is not a valid call. The property name should not contain dots.

Find text in source code of page and create hyperlink

I need to use JavaScript to search the source code of the current page for a string, e.g data-userId="2008", then extracts the id number (2008 in this case) and creates a hyperlink to include it, e.g. http://www.google.com?2008.
I've been attempting to use indexOf and document.documentElement.innerHTML approaches but not getting anywhere. I've got closer with the help of this post but no success yet.
Here's what i have so far:
<script type="text/javascript">
function getVals() {
code = document.getElementsByTagName("html")[0].innerHTML;
results = code.match(/data-userId=(\d*)&/g);
for (i=0;i<results.length;i++) {
value = results[i].match(/data-userId=(\d*)&/);
}
}
onload = getVals;
document.write(code);
</script>
Due to restrictions on our network the solution needs to be JavaScript.

Use the getAttribute() method of the Element object http://www.w3schools.com/jsref/met_element_getattribute.asp.
I'm not sure why exactly you'd query the html element and then look at innerHTML. Do you not know which tags/selectors will contain the data attribute you're looking for? If you don't know which selectors you're looking for you can use a recursive function like this to walk the DOM and store the values in a data structure of your choice - I'll use an array of objects in this example (not sure how you need this data formatted).
Also, I'm not sure how you're choosing to visit these pages, but the below code could be easily modified to create hyperlinks when/if the attributes you're looking for are found.
var storeElements = [];
function walkTheDOM(node){
if(node.getAttribute('data-userId') !== null){
storeElements.push({"element": node, "attrValue": node.getAttribute('data-userId'});
}
node = node.firstChild;
while(node){
walkTheDOM(node);
node = node.nextSibling;
}
}
Then you can call the function like this:
walkTheDOM(document.querySelectorAll('html')[0]);
If this isn't what you're looking for let me know and I can change my answer. For those of you who've read Douglas Crockford's "Javascript: The Good Parts" this function may look very familiar :).

How to develop a custom protractor locator?

I have an angular-application where I put custom html-attributes ("data-testid='saveButton'") for identification on e.g. buttons, inputs etc.
Protractor has no built-in locator for this: the xpath-way does not work as this will always search from the document root (to overcome this limitation I would have to do quite a lot of string-magic to build the xpath-string, so I discarded this variant).
So I wanted to write my own locator, and consulted the official Protractor Documentation.
The documentation leaves me with an important question:
What methods are available on the parent-element-object? In their example they use using.querySelectorAll('button'); but how do they know they can use this?
I searched for 'querySelectorAll' in the protractor-documentation to see if it is maybe a method of ElementFinder or WebElement, but I didn't find it.
Also I found no way to debug the locator (only stop the protractor-test by browser.pause()), so I can't look at runtime to see what I can do with the given parent-element.
So does anybody know how I can get information about what methods the parent-element has?

The parent-element, which is essentially the context of the selector is a DOM element that gets handled by the browser - it can use whatever DOM function there is, in this case Element.querySelectorAll. As the addLocator's example says about the function you pass to it:
This function will be serialized as a string and will execute in the browser.
Also, keep in mind that there is a way to provide context to the xpath selectors thus avoiding the search from root:
var context = element(by.css('.some-parent-element-class'));
var divDirectDescendants = context.element(by.xpath('div'));
var divAllDescendants = context.element(by.xpath('.//div'));
With CSS these would be:
var divDirectDescendants = element(by.css('.some-parent-element-class > div'));
var divAllDescendants = element(by.css('.some-parent-element-class div'));
Also, there's a way to achieve what you want with a CSS selector (which is what querySelectorAll is, after all):
var everyElementWithMyAttr = element.all(by.css('[data-testid="saveButton"]'));

Getting the prototype of an object as a string with javascript

So I have a situation where I am trying out some DI based approached in JS, now lets not discuss the worth or validity of DI in JS etc let us just focus on the question.
So I get a targetConstructor, which would be something like:
function SomeClass(){};
doSomethingWithTargetConstructor(SomeClass);
So the above would pass the constructor of SomeClass into the method and that all works fine, however there is a test case which I cannot find a way to solve, as the targetConstructor will be a named function so you can extract the name from it. However the below situation does not work:
var someRootNamespace = {};
someRootNamespace.someSubNamespace = {};
someRootNamespace.someSubNamespace.someNamespacedClass = function(someArg){};
doSomethingWithTargetConstructor(someRootNamespace.someSubNamespace.someNamespacedClass);
So in the above example you cannot extract the name of the target constructor as it is a nameless function, HOWEVER in the debugger (as seen below) you can see that the targetConstructor prototype does contain the name information needed.
As basically I want the string representation of someRootNamespace.someSubNamespace.someNamespacedClass
However I cannot for the life of me find a way to get at the information shown in the debugger, so is it even possible to get at? and if so how do you go about it? as the information is there, I just need a way to get access to it.

Dynamically Setting a JavaScript Object's Method's Source

I've recently been working on a nice little JavaScript game engine that works a lot like Game Maker, but lets people create basic JavaScript games within a browser. Every instance of every object will have it's own preset methods, which the runner will iterate through and execute. I'm trying to find a way to let the user / creator dynamically edit any of the methods source code. When I say 'preset methods', I mean blank methods stored under specific preset names within the objects / object instances. Here's a basic example:
var newObject = object_add("object_name"); // Adds a new object 'blueprint' and returns the reference.
The function object_add(); creates a JavaScript object, and adds a number of preset methods to it, such as:
create
destroy
step
draw
.. and many more
Each of these methods will have no code in them to start with. I need to let the creator dynamically change any of the methods source code. I could simply overwrite the variable that points towards the method, with a new method, but how can you set method's source code using a string?
I know that something like:
newObject.create = function(){textbox.innerHTML};
definitely wouldn't work. Any ideas?
Many thanks,
Dan.

Looks like you want to use eval function, but it's generally a bad idea.

The answer was found at: Creating functions dynamically in JS
Here's the answer (copied from the other page).
Well, you could use Function, like in this example:
var f = new Function('name', 'return alert("hello, " + name + "!");');
f('erick');
//This way you're defining a new function with arguments and body and assigning it to a variable f. You could use a hashset and store many functions:
var fs = [];
var fs['f1'] = new Function('name', 'return alert("hello, " + name + "!");');
fs['f1']('erick');
//Loading xml depends if it is running on browser or server.
Thanks, #CBroe https://stackoverflow.com/users/1427878/cbroe