Hi Luiz, On Tue, 1 Sep 2020 at 14:30, Archie Pusaka <apusaka@xxxxxxxxxx> wrote: > > On Tue, 1 Sep 2020 at 13:42, Luiz Augusto von Dentz > <luiz.dentz@xxxxxxxxx> wrote: > > > > Hi Archie, > > > > On Mon, Aug 31, 2020 at 10:33 PM Archie Pusaka <apusaka@xxxxxxxxxx> wrote: > > > > > > Hi Luiz, > > > > > > On Tue, 1 Sep 2020 at 13:20, Luiz Augusto von Dentz > > > <luiz.dentz@xxxxxxxxx> wrote: > > > > > > > > Hi Archie, > > > > > > > > On Mon, Aug 31, 2020 at 8:19 PM Archie Pusaka <apusaka@xxxxxxxxxx> wrote: > > > > > > > > > > Hi Luiz, > > > > > > > > > > The reason we introduced this method is because we need to send some > > > > > of AVRCP category 2 passthroughs in order to be able to qualify the > > > > > AVRCP absolute volume control as an audio source, despite we don't > > > > > actually plan to send any passthrough in the practice. > > > > > > > > I wonder what those are, are they volume up/down? > > > Yes, volume up/down and mute are the ones that are unique to category > > > 2, but there are also some other buttons which are included in > > > category 2 as well but not unique to it, for example button 0 - 9, > > > help, and power. > > > > Weird, there are dedicated commands for setting absolut volume, not > > sure why one would use volume up/down to qualify? Or is that testing > > any button that belongs to category 2? Anyway it would be nice to know > > what test that is and what is the expected behavior the test case > > expects from us. > > Weird, yes. Sending absolute volume command capability is declared in > the AVRCP ICS 2/50, but it can only be selected if 2/8 is also > selected, which is sending passthrough command for category 2. > Therefore, we also need to be able to send a category 2 passthrough. > > The test itself is AVRCP/CT/PTT/BV-02-I. This is the test to see > whether we can send a category 2 passthrough, not the one to see > whether we can send the absolute volume command. The test asks us to > send every category 2 passthrough we declared in the ICS file (table > 4). There should be at least one passthrough declared. > > > > > > > > > > > > > > > Therefore, which passthrough we send is actually not important (to > > > > > us), as long as we can do that to satisfy the requirement. > > > > > However, I see that there are over 50 passthroughs across category 1 - > > > > > 4, and thought that it would be cumbersome to have a different method > > > > > for each key if one day we decided to support them all, hence the > > > > > generic method. > > > > > > > > Yep, but there are button presses so things gets a little more > > > > complicated if we need to emulate press/hold/release semantics. > > > For now, I don't think having a separate method for each key will make > > > it easier if we need to emulate hold and release semantics. > > > > > > Could we have 3 methods as follows, all have avc_key as a param: > > > (a) press. This means BlueZ will auto-release it within a short amount of time > > > (b) hold. Bluez wouldn't release it unless it received (c). > > > (c) release. Bluez will release the previously held button. > > > What do you think? > > > > I'm fine with that, we might have to start the status of each request > > though, also the holding should not last longer than 2 seconds if I > > recall correctly. > > Then I shall try to have this implemented. > Btw I only took a quick glance, but I did not find the 2 seconds > limitation in the spec. I found it in the AV/C Panel Subunit specification from 1394 trade association. It says we want to support "hold", then we need to frequently resend the key with "pressed" state, because if 2 seconds passed without the TG receiving it, then the TG will assume a "release" state is sent (but is lost). Also, the specification doesn't support multiple simultaneous button presses, because we are supposed to send "release" before sending a new "pressed" key, so I think we don't need to track the status for each request. > > > > > > > > > > > > May I have some questions? > > > > > (1) Why do you prefer a dedicated method per key? > > > > > (2) Let's say I need to add a dedicated method for a single category 2 > > > > > passthrough key, do you have a key in mind or I can just pick by > > > > > myself? > > > > > > > > > > Thanks, > > > > > Archie > > > > > > > > > > > > > > > On Tue, 1 Sep 2020 at 01:16, Luiz Augusto von Dentz > > > > > <luiz.dentz@xxxxxxxxx> wrote: > > > > > > > > > > > > Hi Archie, > > > > > > > > > > > > On Mon, Aug 31, 2020 at 12:37 AM Archie Pusaka <apusaka@xxxxxxxxxx> wrote: > > > > > > > > > > > > > > From: Archie Pusaka <apusaka@xxxxxxxxxxxx> > > > > > > > > > > > > > > This allows us to send any passthrough command. > > > > > > > > > > > > > > Reviewed-by: Michael Sun <michaelfsun@xxxxxxxxxx> > > > > > > > --- > > > > > > > > > > > > > > doc/media-api.txt | 8 ++++++++ > > > > > > > 1 file changed, 8 insertions(+) > > > > > > > > > > > > > > diff --git a/doc/media-api.txt b/doc/media-api.txt > > > > > > > index dabc69936..fe7222eef 100644 > > > > > > > --- a/doc/media-api.txt > > > > > > > +++ b/doc/media-api.txt > > > > > > > @@ -199,6 +199,14 @@ Methods void Play() > > > > > > > Possible Errors: org.bluez.Error.NotSupported > > > > > > > org.bluez.Error.Failed > > > > > > > > > > > > > > + void Press(byte avc_key) > > > > > > > + > > > > > > > + Press a specific key to send as passthrough command. > > > > > > > + > > > > > > > + Possible Errors: org.bluez.Error.InvalidArguments > > > > > > > + org.bluez.Error.NotSupported > > > > > > > + org.bluez.Error.Failed > > > > > > > + > > > > > > > Properties string Equalizer [readwrite] > > > > > > > > > > > > > > Possible values: "off" or "on" > > > > > > > -- > > > > > > > 2.28.0.402.g5ffc5be6b7-goog > > > > > > > > > > > > For now I would prefer to use dedicated method per key or are there > > > > > > many keys you would like to use from the upper layer? > > > > > > > > > > > > > > > > > > -- > > > > > > Luiz Augusto von Dentz > > > > > > > > > > > > > > > > -- > > > > Luiz Augusto von Dentz > > > > > > Thanks, > > > Archie > > > > > > > > -- > > Luiz Augusto von Dentz Thanks, Archie
