[Gnash-dev] sound_handler redesign status

[strk]

If you haven't followed the commit list, here's a summary:

SDL sound handler is being cleaned up to generalize management
interfaces. The idea is that most functionalities are really
part of the core, and not of a modular handler.

Core functionalities are:
        - management of defined (embedded) sounds
        - management of a set of playing sounds

Possibly modular functionalities are:
        - mixing of playing sounds (input streams or channels)
        - communication with system mixer

In current trunk, SDL sound_handler needs some more cleanup, then
the core functionalities could be moved up to the base class
and modified as needed to take possibly threaded implementation of
module-specific features.
Already, playing sounds (InputStream) use is generalized, as is
management of defined sounds, only still not in base class.

The deliverable for the final cleanup would be a NullSoundHandler
providing manual samples fetching functionality, to allow fully
predictable data flow for the testing framework.

--strk;