As always, if you have steps to reproduce we would love to hear them. Startup time is still the #1 performance metric we optimize for.
Isn't it a bit wrong direction now that people keep the browser open all day long (compared to years ago that they close and open it frequently)?
Don't get me wrong, though. I would like to have Firefox startup fast. I just think maybe the focus at this stage of browser development should be something like faster tab creation and less memory leaks?
Firefox 4 is pretty amazing, but I just encountered a problem today: on the Mac, it doesn't even try to use the system's language priority, and by default it displays Japanese characters before Chinese (thus inconsistent rendering of characters), even though the system preference is configured to have Chinese over Japanese. Eventually I found there is a Firefox-only language setting in the browser. Maybe this is just design-philosophy, but still it feels not very native system-friendly.
So far, Firefox 4 hasn't yet won my heart. Most of geek friends I know are using Chrome, and giving the current situation, are unlikely to revert to Firefox any time soon. I'm interested to see how the Firefox could regain its older crown.
Here's a simple example I botched together. Firefox has the worst performance when it comes to rendering a simple css transforms. Firefox hits > 80% CPU time when the effect is in play compared to < 30% for the competition and < 7% for Safari.
Copy / paste and mouse hover like a mad man.
Firefox and Performance still don't belong in the same sentence for me as it stands.