The layout management in OSMF prior to 1.6 sprint 3 was somehow inconsistant in composite elements.
That(I guess) is a problem of layout measurement.
Our player relies on osmf layouting greatly so I have posted a lot here already.
May be it can help you:
This topic contains referencies to other and to jira:
There is a "duct tape" proxy attached to jira bug it may help you if you put your composition to that proxy BEFORE settting it to media player.
Since 1.6 sprint 3 the whole layout thing has changed and has become consistent - it does not work as expected at all :O)
But as soon as you are on 1.0 the duct tape proxy may help
Thanks for the quick reply, I will have a look at that tomorrow. That's a really weird bug (and really hard to debug, maybe there could be room for improvement there, such as a function that trace the structure and the layout of each children?) , but fortunately I've got time to work on it. Is there any major release planed in the next month or so? If yes, I think I should probably wait for that and upgrade from 1.0 to 1.7... Will try the duck tape thingy as well meanwhile, thanks!!
Oh. I've not said that for a duct tape to work you should also put composition to media container before media player.
So the correct sequence would be:
1) Create composition
2) Put inside proxy
3) Put to MediaContainer
4) Put to media player
Jira page explains why the sequence should be like that.
And, once again, it helps only before 1.6 Sprint3 and I'm not that sure about serial compositions (we use a makeshift PlaylistElement instead mostly)
As for the bug itself - I personally doubt it will be fixed because it is almost a year old without any activity :O)
As a result we stick with 1.5 - can't invest in player redesign due to the whole layout framework has become unusible since Sprint3...
Managed to fix that problem just by specifying a width and height in the <media:content> tag in the mediaRSS feed Glad that problem is fixed, and I don't know if it has anything to deal with the layout issues in OSMF 1.0. Might be useful to run some test with nested elements with/without specified size in pixels in the visual elements ...
Thanks for your support and quick replies, even if it didn't fix my problem (but might have as I haven't tested it?) - that's how a community gets stronger ;-)
Glad that you've made it!
You are right - settting pixels explicitly works.
However we mostly use relative positioning to make fullscreen toggle available without explicit resize of each element in collection. and in that way things are getting worse :)