Performance Testing

This week I've been doing performance testing for nightly builds of Firefox and chrome.
We (my OSD600 class) have been using Chrome Experiments for these tests.
These experiments were meant to show off how well Chrome does HTML5.
Our goal is this: to identify and profile the performance issues that Firefox has in running these experiments.

I was assigned 10 experiments to handle. The experiments ranged from graphics demonstrations to games to audio/video demontrations.  While Firefox performed just as good as chrome in several experiments, there were many problems that I noticed.

The biggest issue I found was in a game called SandTrap.  When running in Firefox 4 beta, any time sand particles fall out of the "maze" they all freeze and the game cannot continue until reload.  I brought up the issue with Dave on irc and he threw together some debug versions of the game (surprisingly quickly). he narrowed down the issue to a line of Javascript code that seemed to be accessing parts of an array that didn't exist.  The problem with that is that Chrome uses the exact same code and has no trouble there.  I filed a bug for it (My first bug ever!) here and it was marked as a duplicate of another bug just one minute later :'( but at least the issue is being dealt with.

Another bug I posted for Sandtrap can be found here.  It is about how sand particles in Firefox were painted as large squares whereas in chrome they were fine dots.  It is still open.

I'm putting together information to file another bug in the Javascript Engine that makes typed letters in the drawing tool for The Wilderness Downtown not appear properly. I posted a screenshot on Flickr here. Seems like a pretty intense bug if you ask me.