GitLab

This CL introduces two main updates to constraints processing for audio:
1. The SelectSettings() algorithm now considers audio-processing
   properties as belonging to each device instead of being considered
   independent of the device. The reason is that, while the
   audio-processing module can be applied to any source, once a source
   has processing applied, all tracks for that source must use the same
   audio-processing settings.
2. The getUserMedia() implementation leverages (1) and makes sure that
   candidate settings for audio sources already in use are limited to
   their current configuration.

Bug: 802198
Change-Id: I3554206e852ff41116a1fe02be45a7a8dff4760d
Reviewed-on: https://chromium-review.googlesource.com/881183Reviewed-by: Henrik Boström <hbos@chromium.org>
Commit-Queue: Guido Urdaneta <guidou@chromium.org>
Cr-Commit-Position: refs/heads/master@{#533288}