What metrics? Almost every page loading benchmark I've seen are based on
the time when 'onload' event fires. That cannot reflect user's
perception on browser speed and will give misleading result when used to
compare different browsers -- different browser fires that event on
different phase of the page loading (IIRC, chrome cheats by doing that
before layout[1], firefox does that after layout but before rendering),
so that's not comparing apple-to-apple.
/Roger
[1] http://trac.webkit.org/browser/trunk/WebCore/dom/Document.cpp#L1748
1748 // Make sure both the initial layout and reflow happen afterthe onload
1749 // fires. This will improve onload scores, and other browsers do it.
1750 // If they wanna cheat, we can too. -dwh
The updated link is
http://trac.webkit.org/browser/trunk/WebCore/dom/Document.cpp#L1828
>
> 1748 // Make sure both the initial layout and reflow happen afterthe onload
> 1749 // fires. This will improve onload scores, and other browsers do it.
> 1750 // If they wanna cheat, we can too. -dwh
>
>>> Is Firefox's trunk gaining ground on Chrome's trunk and roughly by
>>> how much? Has Firefox overtaken Chrome again at all?
The main metrics I was thinking of was javascript. I was ambiguous
because I didn't know if there were any others. I've never used Chrome
so I've never experienced any difference. I'm just curious because of of
some market share data I've seen lately showing Chrome steal Firefox's
market share for the last couple months.
It really depends on what you measure. Some tests show Chrome up to 5x
faster than Gecko. Other tests show Gecko up to 5x faster than chrome.
The ratios change around all the time on various tests as both work on
speeding up their code.
There is certainly active work going on to address the areas where we
are currently slower. See https://wiki.mozilla.org/JaegerMonkey for one.
-Boris