Collected and moved existing camera infrastructure from game.cpp to camera.cpp and camera.h. Introduced configuration settings 'fov' which chooses the camera's (vertical) field of view and 'view_bobbing' which currently does nothing. Other code refactored to not expect the FOV to be a build time constant.

Removed unused camera_position and camera_direction fields from Client. Moved ClientEnvironment::drawPostFx to ClientMap::renderPostFx -- this will make the camera management classes easier to write, as ClientMap already knows the camera position but ClientEnvironment doesn't and has to be told about it. This also eliminates the need for Client::getEnv(). Made the post effect color a content feature defined in content_mapnode.h.

And some simple animation when trying to dig...

This is done with the new TOSERVER_PLAYERITEM message, that informs the
server on the index of the wielded item.

Collect the player's eye position coding in a Player method, and
use it in client to pass the eye position information to the game.

--HG--
extra : rebase_source : 7260636295d9068fbeeddf4143c89f2b8a91446c

--HG--
extra : rebase_source : e8ec407f60711d42d33be4811b2880088f617b5b

Made a proper queued thread to client for handling some block mesh updates. Also made client mutex-free to allow easier adding of new stuff.