FF3 RC1 and Firebug 1.2 beta

[jack.slocum]

For those who haven't tried the upgrade yet, it is awesome! The "peg the CPU on page load" issue is gone. :D

[Animal]

Yes, I've upgraded to FF3 now that FB works with it.

I'm still having a few rendering issues with complex pages though.

Thes pages work fine in FF2, Opera and IE (though rendering fine on IE is no guarantee that a page is a valid, conforming page)

My application is a single page app which loads Components though Ajax, and the callback adds a new Panel encapsulating the Component to the Viewport's center region. The Component may have subcomponents which it then Ajax-loads into it's own layout (And most use the standard "north Grid, south detail" border layout to represent a query and the focussed row)

What's happening is that the Component's Panel (the Panel that is added to the Viewport's center) is positioned correctly. Its north and south Panels have the correct positioning attributes, but FF is displaying them 12 pixels to the left of where they should be.

I can poke them with Firebug, and see their correct attributes and style. But then I can hover over their elements in FB's HTML tab, and the light blue highlighting is 12 pixels to the left of where the styling says they should be. There is a "gap" of 12 pixels on the right edge of the Component's Panel.

If I set the left to "12px", they look OK, but on layout which corrects the position values, they jump back.

I think it's something to do with timing, and the fact that FF3 renders much quicker than FF2. At some point, some measurements are taken by FF which fix some positioning wrong.

[jgarcia@tdg-i.com]

The "peg the CPU on page load" issue is gone. :D

http://buzzlife.com/forums/images/smilies/smoothcriminal.gif
:D

[mystix]

http://buzzlife.com/forums/images/smilies/smoothcriminal.gif

OT: i used to like that game!

[jack.slocum]

I am not sure what would cause that Animal. I know one thing that is a big help with performance in FF3 is they have implemented IEs getBoundingClientRect. It's nice because Ext detects it and uses for XY positioning. I did some initial testing and it seems to report identical values as the old way of manually calculating it. Perhaps somehow the problem is related?