String.fromCharCode(charCodeMap(e.keyCode, e.shiftKey));

This command would take the ascii value return by charCodeMap and return a string of the character. This would be great, but key should hold a char not string. When using the example program, int(key) no long returns the ascii value because it cannot cast the string to and interger. Next I tried this:

new Char(String.fromCharCode(charCodeMap(e.keyCode, e.shiftKey)).charAt(0));

Lets take that string value and create a processing.js Char type. This would work great but turns out that processing.js Char type returns the ascii value and not the character. To retrieve a character, toString() must be used. I can modify processing.js Char to work the way I want and return a character, but then I have broken Char in other places. Seems to be no perfect answer to this bug, functionality compared to p5 will be sacrficed somewhere. I guess I’m off to irc to see what other people think.

First we were shown how to use Mozilla cross reference effectively. The cross reference can be used to search through firefox source code. Using key terms, one can locate the necessary lines of code and then makes changes in your local copy. In class we searched through code were tab creation may be handled. Eventually we found some code, in tabbrowser.xml, where we believed that our changes can be made.

// Check if we're opening a tab related to the current tab and
// move it to after the current tab.
// aReferrerURI is null or undefined if the tab is opened from
// an external application or bookmark, i.e. somewhere other
// than the current tab.
if ((aRelatedToCurrent == null ? aReferrerURI : aRelatedToCurrent) &&
 Services.prefs.getBoolPref("browser.tabs.insertRelatedAfterCurrent")) {

The code we found had comments explaining that during the creation of a tab, it will check to see if the tab is related, if so, it will create the tab next to the tab in focus. This was the functionality that we wanted, but we wanted it with every tab. So by commenting out the the check, it should create the tab next to the tab in focus every time. Using the command below we created a patch and then re-compiled minefield.
hg diff -p -u 8 > tab.patch

That simple changed seemed to work. As you can see in the image below, the new tab is create next to the tab in focus and not at the end. Minefield is imitating the functionality of being a related tab for every tab created. The patch can be found here on my wiki.

In dps909 this week we were asked to help a group project to convert all the processing.js examples in the seneca directory into ref tests. The examples are sketches that were created in the early stages of processing.js. Most sketches have an image of how the sketch looks in p5 and then it renders it using pjs. After the sketches have been loaded, it is up to the user to compare the sketches. As pjs grows, these tests have multiplied to the point where all the tests cannot be performed manually. Ref tests takes a series of pixel information from a passed test so it can be used to check the same sketch in the future. With this pixel information, tests can be performed and checked easily and quickly.

Its very simple to create ref tests. Go to the seneca directory of pjs and run an example. Take the sketch code and run it in p5. Compare it to the result in p5.

Sketch Code:
void setup() {
size();
background(153);
}

void draw() {
line(0, 0, width, height);
}

p5	pjs

Being that the sketches look the same, we can create a ref test using the pjs helper. The result is a ref test that can be saved as a pde file and placed in the ref test directory. Information on the tests I completed can be found on my wiki.

} else if (arguments.length > 0 && typeof arguments[0] !== ‘number’) {
array = arguments[0];

The code above accepts arguments such as array, objects, array-like objects. I found this site helpful when trying to understand objects, arrays and array-like objects. By changing the code to array.arguments[0].toArray(), the values get copied properly so they can be used. The problem that this creates is that, toArray is a arraylist function, therefore if the constructor receives an array, the toArray function will not be available.
As a result I thought to check for an arraylist and an array.

if(arguments[0].toArray){
array = arguments[0].toArray();
}else if (arguments[0].slice){
array = arguments[0].slice(0);
}

After review, I was told that we should limit the types of acceptable parameters to only Arraylist and number. So I removed the section for array and I had to change the clone() function because it uses an array. I changed it from:

return new ArrayList(array.slice(0));
to
return new ArrayList(this);

Now everything should work, arraylist now accepts an arraylist as a parameter.

ArrayList second = new ArrayList(first);

Because the constructor was expecting an int, it would take the object and create and array of the same size without taking the values. Using the “typeof” command, I could test the parameters and if the parameter was an object, create a copy of the array. 

if(typeof size === “object”){
var array = size.slice(0);
}else{
var array = new Array(0 | size);
}
Now what was left was to create some tests, I created an ArrayList test file because I could not find one and inputted some tests.

ArrayList first = new ArrayList();
first.add(“1″);
first.add(“2″);
first.add(“3″);
first.add(“4″);
first.add(“5″);

ArrayList second = new ArrayList();
second.add(“1″);
second.add(“2″);
second.add(“3″);
second.add(“4″);
second.add(“5″);

ArrayList third = new ArrayList(first);

_checkEqual(first, second);
_checkEqual(first, third);
_checkEqual(third, second);

Then I ran all the tests.

Solving this bug was quite simple, espcially now because we have covered javascript heavily in class since our first release. I moved the colour variable, which is a long list of colours being assigned their hex colour codes, out of the processing instance. This will result in saving memory because the list of colours does not have to be compiled during every processing instance.
So before pushing to my remote repo, I performed the necessary tests. All tests passed and I pushed to my repo, which can be found here.