[pulseaudio-discuss] [PATCH RFC 3/4] sink: Inherit the volume from the active port.

[Damir Jelić]

On Fri, Jan 11, 2013 at 04:13:18PM +0100, David Henningsson wrote:
> On 01/05/2013 12:19 AM, poljar (Damir Jelić) wrote:
> >Since the port now holds the volume information we can simply inherit
> >the volume from the active port if the volume isn't set while creating
> >the sink.
> 
> Here's where it gets tricky: sinks are having a massive amount of
> volumes, so I have difficulties telling all them apart. But I hope
> you're doing the right thing :-)
> 

More eyes would be welcome here, I'm not sure either. :P

> >
> >BugLink: https://bugs.freedesktop.org/show_bug.cgi?id=55262
> >---
> >  src/pulsecore/sink.c | 14 +++++++++++---
> >  1 file changed, 11 insertions(+), 3 deletions(-)
> >
> >diff --git a/src/pulsecore/sink.c b/src/pulsecore/sink.c
> >index a8a91d6..37cabb6 100644
> >--- a/src/pulsecore/sink.c
> >+++ b/src/pulsecore/sink.c
> >@@ -276,7 +276,6 @@ pa_sink* pa_sink_new(
> >      s->n_corked = 0;
> >      s->input_to_master = NULL;
> >
> >-    s->reference_volume = s->real_volume = data->volume;
> >      pa_cvolume_reset(&s->soft_volume, s->sample_spec.channels);
> >      s->base_volume = PA_VOLUME_NORM;
> >      s->n_volume_steps = PA_VOLUME_NORM+1;
> >@@ -309,10 +308,18 @@ pa_sink* pa_sink_new(
> >                  s->active_port = p;
> >      }
> >
> >-    if (s->active_port)
> >+    if (s->active_port) {
> >          s->latency_offset = s->active_port->latency_offset;
> >-    else
> >+
> >+        if (pa_cvolume_valid(&s->active_port->volume) && !data->volume_is_set)
> >+            s->reference_volume = s->real_volume = s->active_port->volume;
> >+        else
> >+            s->reference_volume = s->real_volume = data->volume;
> >+
> >+    } else {
> >          s->latency_offset = 0;
> >+        s->reference_volume = s->real_volume = data->volume;
> >+    }
> >
> >      s->save_volume = data->save_volume;
> >      s->save_muted = data->save_muted;
> >@@ -1976,6 +1983,7 @@ void pa_sink_set_volume(
> >          }
> >
> >          pa_cvolume_remap(&new_reference_volume, &s->channel_map, &root_sink->channel_map);
> >+        pa_device_port_set_volume(s->active_port, new_reference_volume);
> 
> Hmm, are we really guaranteed that (s->active_port != NULL) here? I
> don't think so.
> 

Yes that's a good catch. Will fix it.