Slashdot Mirror
Mozilla Firefox 3.6 Released
Shining Celebi writes "Mozilla has released Firefox 3.6 today, which adds support for Personas, lightweight themes that can be installed without restarting the browser, and adds further performance improvements to the new Tracemonkey Javascript engine. One of the major goals of the release was to improve startup time and general UI responsiveness, especially the Awesomebar. You can read the full set of release notes here."
If you have the Switch Proxy Tool, I strongly suggest you disable it. Caused all sort of issues when upgrading. If you've already upgraded, right click on the shortcut and run in safe mode, there you can disable it. YMMV.
=================
Unix is very user friendly, it's just picky about who its friends are.
Microsoft's patch vs. Mozilla's release. I can't wait. The Excitement is almost too much.
It's an improvement. That's what counts, some of us don't want to trade our lovely open-source browser for a product from Google or Apple, or MS for that matter. I can wait on javascript performance, TYVM.
Proof that Firefox is heading for doom. Stop wasting time on making the browser look different than the fucking OS you idiots.
about:config
set mousewheel.withnokey.sysnumlines to false
set mousewheel.withnokey.numlines to 3
As good as new... Wait a second.
That delay is nothing to do with your browser - that's Slashdot scanning of a bunch of ports on your IP address. I spotted this a few weeks back when I made a post to Slashdot while running a "tail -f" on my firewall logs, although I've been aware of the lag a lot longer than that. It seems that if your firewall just DROPs the packets you get a delay while it retries a couple of times, whereas if you REJECT then it's a good deal quicker. There's some caching going on as well, once you've gone through this the lag disappears for a day or two, then re-starts. As it says in my .sig - WTF?
UNIX? They're not even circumcised! Savages!
These tests are mostly pointless anyway. They measure raw JS performance, which would matter if you'd be doing, say, number crunching. In practice, the most heavyweight operation that is likely to be done by scripts in a browser is DOM manipulation, and that's an entirely different thing. What does it matter if your super-efficient JS AOT compiler based on quantum branch prediction can call a method on a DOM object as fast as a plain native JMP, if the implementation of said method causes reflow and redraw of most of the page?
Coincidentally, it's why Opera feels so fast for actual browsing while still using an interpreter for JS (and consequently sucking in any synthetic JS perf tests) - its interpreter is an order of magnitude slower than e.g. Chrome, yes, but it's got an extremely fast layout engine and renderer, so DOM updates are instantaneous.
to
"A witty saying proves nothing." - Voltaire