On Tue, Jul 14, 2009 at 3:05 AM, Robert O'Callahan<rob...@ocallahan.org> wrote: > On Tue, Jul 14, 2009 at 9:46 PM, Boris Zbarsky <bzbar...@mit.edu> wrote: >> >> For the current model, note that all the text says is "should not show >> this content to the user". While this is not defined anywhere, it doesn't >> seem to indicate that the content's DOM should not exist, for example. In >> Gecko, at least, the image in your example will be loaded and hence its >> onload will fire. > > > There's actually a fairly major related problem here. We hide the fallback > content by treating it as display:none. Currently Gecko has a huge bug where > a display:none plugin does not load/run. This works out well for the video > fallback case. If we fix that bug, then unless we do some special magic, > plugin-based video fallback will run and play audio while the <video> > element plays --- very bad.
It shouldn't be that hard to for plugins walk the parent chain to see if there are any "active" video elements before starting the plugin. > People have already discovered that <video src="hellokitty.ogv"><object > data="hellokitty.ogv"></object></video> plays the audio track twice in > Firefox, and aren't happy about it, although it's what the spec seems to > suggest. I'm not sure how to modify the spec in a sane way to fix this case, > though. We can do what's described above for videos and audios too (i.e. walk parent chain etc). / Jonas