Switch Audio: Added Audio Out usage and touchscreen code rework

[Wohlstand]

Replacement for #18

Audio Render is a second layer that converts the audio stream to output. Audio Out is the direct output of the audio stream without conversions.

Description

• Added the second module SDL_switchaudioout.c which uses Audio Out instead of Audio Render
• Added SWITCHAUDIOOUT_bootstrap entry with the highest priority
• Renamed SWITCHAUDIO_bootstrap into SWITCHAUDIOREN_bootstrap
• Fixed some warnings by adding of (void)this; line into empty calls
• Fixed the touchscreen work:
  • The ID must be bigger than 0. If ID is 0, that means it's an invalid device, and this is confusion for such API calls as SDL_GetTouchDevice()
  • I rewrote the touchscreen code with a better one

Existing Issue(s)

1) Audio Render API plays with less stability, and in some cases, may turn into silence because of overloads. Audio Out is the direct output of the audio stream without any conversions. 2) Touchscreen API worked incorrectly

Demo

This is a simple program that draws a white screen and plays the music. When you put fingers, it should draw a tiny red dot at the place of a finger.

• Source code: music-demo.zip
• Pre-built demo: DecoderUZK.nro.zip

[Wohlstand]

I see. Anyway, AUDREN has an unique ability to supply ANY type of audio stream: different sample rates, different sample formats, etc.

• AUDOUT is strictly limited on 48000 Hz and 2 channels number, but, it's the fastest and most stable API here.

So, if somebody want's to submit the sound stream without an ability to convert it to the stream format, the AUDREN will be useful.

[Wohlstand]

Anyway, I'll get rid of AudioRender soon. Right now I'm out of home, will get back soon.

[Wohlstand]

Okay, I completely removed the AudioRender support as you had been requested.

[Wohlstand]

When you will merge this into upstream?

[WinterMute]

When you will merge this into upstream?

4 days ago.

[Wohlstand]

When you will merge this into upstream?

4 days ago.

Ah! Now I see. Usually it's need to merge via "merge" button at GitHub itself to gracefully confirm the fact of the merge. Manual merges by local pushes were suggested to be done if you going to apply some own tweaks on the fly.

[WinterMute]

Normally I'd prefer that the audio & touchscreen changes be done in separate PRs but I just squashed the audio changes manually this time.

[Wohlstand]

In next time, I'll do the separate PR for each feature. Thanks for the notice.

[Wohlstand]

Tested out the thing after recent dkp-pacman sync: now it works just fine!