On Wed, 2016-04-13 at 10:48 +0900, Sangchul Lee wrote: > Now, trigger_roles, ducking_roles and volume can be divided into several groups by slash. > That means each group can be affected by its own volume policy. > > If we need to apply ducking volume level differently that is triggered from > each trigger role(s), this feature would be useful for this purpose. > > For example, let's assume that tts should take music and video's volume down to 40% > whereas voice_recognition should take those and tts's volume down to 20%. > In this case, the configuration can be written as below. >  trigger_roles=tts/voice_recognition ducking_roles=music,video/music,video,tts volume=40%/20% > > If one of ducking role is affected by more than two trigger roles simultaneously, > volume of the ducking role will be applied by method of multiplication. > And it works in the same way as before without any slash. > > Signed-off-by: Sangchul Lee <sc11.lee at samsung.com> > --- > > Here's updated patch for v3. > Thanks for the review, tanu. > > Notes: >  v2 changelog: revise codes according to Tanu's comments >  - commit message enhancement >  - definition of group structure >  - rename variable and revise logs >  - handle error case of empty parsed string >  - fix codes to follow coding convention for private struct >  - pass pointer directly instead of group index > >  v3 changelog: it is integrated with the latest upstream codes >  - remove a duplicate member of userdata >  - revise an inappropriate log message and comparison for indicating a failure >  - fix invalid logic regarding removing hashmap for interaction_state > >  src/modules/module-role-ducking.c |   6 +- >  src/modules/stream-interaction.c  | 293 +++++++++++++++++++++++++++----------- >  2 files changed, 212 insertions(+), 87 deletions(-) Thanks! Applied. -- Tanu
