Optimize Text rendering / SharedBuffers

[cart]

This one isn't rocket science. Creating buffers every frame is really slow. We now actually share buffers in SharedBuffers. This was the intended design all along, I'm just now getting around to implementing it 😄

I used the text_debug example to benchmark.

Before:

Diagnostics:
---------------------------------------------------------------------------------------------
fps                                                              : 360.538804  (avg 362.820610)
frame_time                                                       : 0.002983    (avg 0.002756)

After

Diagnostics:
---------------------------------------------------------------------------------------------
fps                                                              : 1185.222059  (avg 1157.447952)
frame_time                                                       : 0.000818    (avg 0.000871)

[cart]

Minor bug with apps that don't use text. Easy fix.

[cart]

Fixed!