Simplify and split audio configs across frameworks

Yes, latest AI (making pulseaudio an option) can be postponed. Not an issue ATM. But we should still anticipate with the change to have pulseaudio being optional: that will make transition easier.

See similar comment on https://lf-automotivelinux.atlassian.net/browse/SPEC-1601 .

more globally, make pulseaudio an optional component in AGL: this is achieved by removing direct and/or useless dependencies and add a DISTRO_FEATURE which can be used to conditionally build with pulseaudio support.

• radio recipe

• 4a-mixer recipe

• homescreen recipe homescreen code

• conditionnal pull into images

• add pulseaudio for agl-audio-soundmanager

This is a too fundamental change for FF. Do the necessary conditionals and research and then realize that in GG.

The immediate  goals for FF are:

• Make 4a the default alsa sound path

• Clean-up the bitrotted pulse config files

• get the pulse-plugin working ( owns the necessary additions to the config files)

As seen during SAT meeting on 20180719, the following AIs are taken:

• make 4A the default option for agl-demo (review here)

• when 4A is enabled, remove the pulseaudio alsa hook that may produce issues with audio clients, like gstreamer (review here)

• more globally, make pulseaudio an optional component in AGL: this is achieved by removing direct and/or useless dependencies and add a DISTRO_FEATURE which can be used to conditionally build with pulseaudio support.

  • radio recipe

  • 4a-mixer recipe

  • homescreen recipe homescreen code

  • conditionnal pull into images

  • add pulseaudio for agl-audio-soundmanager

Does this still apply as long as:

• Audio config is splitted in HighLevel and HALs.

• We are getting rid of pulseaudio (so that if some apps still have an issue, then we should move them out of pulseaudio, not fixing pulseaudio config)