Tag Archives: Firefox

Google Chrome is the winner, Javascript speed-wise

Did some benchmark testing on the Javascript engines of several Mac OS X web browsers by using SunSpider and Google’s V8 benchmark suite. Although Google Chrome is still a development release on OS X, it’s already the winner in this test. Safari 4 comes second, it shows similar results in the SunSpider test but loses more in the V8 test. Sadly Firefox 3.5 loses to both of them by quite a significant margin, despite the fact that Firefox 3.5 introduces the new TraceMonkey JavaScript engine. Opera 10.00 beta 2 is just not a viable competitor to all of the others, I hope this is just because of the beta thing.

Here are some charts that show the results:

sunspider javascript benchmark of firefox, safari, opera and chrome

OS X web browser Javascript speed comparison among firefox, safari, opera and google chrome

V8 benchmark total score of Firefox, Safari, Opera and Google Chrome

I helped Firefox today, I’m feeling great

It’s not easy to find a reproducible bug in a stable open source software that has the exposure as large as Mozilla Firefox, which has more than 20% market share of the web browser market in Europe. But I did, and I think I’m very lucky. Not only that, this bug has been hiding in Firefox since 2003 (was it the Firefox 1.0 days?), so this bug has been in existence for 5 years, wow, I think I’m really very lucky to have spotted it.

This bug is serious, but very difficult to manifest itself. Under rare conditions, a form post will be sent to the web server twice without your notice. Say, if you’re shopping in an online e-shop, and you’re so “lucky” that you encountered this bug, you would be buying the same goods twice, or if you’re making a payment, you would be paying the amount double.

Thanks Alfred Peng for confirming the bug so that Firefox developers began taking a look at the bug. From the perspective of open source development, identifying and testing thoroughly for the existence of the bug and filing a detail bug report with reproducible steps is the task I should do. The fix for the bug is just less than 10 lines long (the test case is much longer), but it may take me days if I have to come up with a patch. Filing a bug report is something anyone who speaks English can do, so if you find a bug in Firefox, please go to their bug tracking system and submit a problem report. The more people contribute to it, the better the software will become, and by doing so, we, in turn, are helping ourselves to create a more pleasant web experience.