User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.3a2pre) Gecko/20100211 Minefield/3.7a2pre (.NET CLR 3.5.30729) Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.3a2pre) Gecko/20100211 Minefield/3.7a2pre (.NET CLR 3.5.30729) Panning and zooming on the http://seadragon.com/ site's image viewer is much slower in Fx 3.7 than in Fx 3.6. Reproducible: Always Steps to Reproduce: 1. Go to http://seadragon.com/ 2. Choose an image and switch the image viewer to full-window mode 3. Pan and zoom. Note the fluidity of the motion & the frame rate. Compare the performance in Firefox 3.6 and 3.7. Actual Results: Pan & zoom is not fluid in 3.7, as it is in 3.6 Expected Results: Image panning should be as fast in 3.7 as it is in 3.6. On my computer (1.7 GHz Pentium M), panning and zooming images is completely fluid in Fx 3.6 (i.e. the frame rate is high enough that it is not perceptible that the animation consists of frames.) In the Firefox 3.7 nightly, the motion is very choppy: the frame rate is so low that the frames of the animation can easily be distinguished. Before posting something like "it works for me", please understand that of course the performance difference will not be visually perceptible if your computer is fast enough that the motion is fluid in 3.7 too!
Does the performance get better if you turn OOPP off? (set dom.ipc.plugins.enabled to false in about:config)
Why is this in plug-ins? I don't see any plugins on seadragon.com
It seems that on 3.6 seadragon.com is using silverlight, and on 3.7 nightlies it's not, using HTML instead.
And flipping the pref doesn't have any effect on that behavior. I wonder if the site is doing improper useragent sniffing of some sort?
Switched the useragent to the Firefox 3.6 useragent and it loads silverlight. Moving to tech-evangelism, since this is a website issue not a plugins issue.
OK, that explains the difference. I didn't notice that it wasn't using Silverlight on 3.7 until you mentioned it ... It doesn't use it on Chrome either.
Is it worth filing a bug on the fact that the HTML performs poorly?
Boris, originally I filed the bug because I thought that 3.7 was using Silverlight too, and it was Silverlight that performed differently in the versions of Firefox ... Since this is not the case, I guess someone could mark this bug as invalid?
bz, I didn't see significant perf issues with the HTML impl, but that was on a very fast machine. Szabolcs, if you can find a contact at Microsoft/Seadragon and report the bad UA detection to them, that would be great. In any case, INCOMPLETE here.