OSMF SVN

OSMFPlayer: adding font info to readme.txt, extending configuration params with 'autoSwitchQuality', and having media errors forwarded to JavaScript using the external interface.

Updating MediaContainer-, NestedMediaContainer- and ExamplePlayer samples to work with the layout changes.

Pre-PARB'ing Layout API:

- Removal of ILayoutContext (merged with ILayoutTarget).
- Layout targets no longer get their width and height set directly. Instead, an 'updateMediaDisplay()' method is invoked, passing the target the width and height that it can use for displaying its media.
- 'intrinsicWidth' and height properties have been renamed 'mediaWidth' and height in order to map to the DisplayObjectTrait more closely.
- A context can no longer be assigned a layout renderer. Coupling of a renderer and a context now takes place exclusively by setting the renderer's context property. The renderer will emit a LayoutRendererChange event that the target can use to collect the reference for its read-only layoutRenderer property.
- Introducing ExternalProperty that isolates the boiler plate code for classes that hold getters to values that are externally controlled by means of a value change event.
- RegistrationPoint now defines String constants.
- DefaultLayoutRenderer now sets default width, height and scaling mode properties when targets get added to the renderer (unless some layout data has already been set).

Updating unit tests accordingly.

Renaming lingering 'gateway' namings to 'container'.

Fixing CompositeMetadata from nullifying set parent values on encountering a null synthesized child value originating from a NullFacetSynthesizer.

Fixing setElementId to add a NullFacetSynthesizer to the created facet, so the ID doesn't get inherited on composites.

Created tag sprint9-stable.

Created branch sprint9-candidate.

Changed the plugin metadata media factory key to an ObjectIdentifier per Brian's request.

Fix bug ZE-16 - Playhead slider doesn't set to the beginning for restart. The problem was that the F4F file handler never dispatched a notifyTimeBias event upon the seek(0) that happens on autoRewind, so HTTPNetStream continued to believe that it's time property was non-zero. The solution is to dispatch the event upon the seek(0) request.