PulseAudio branch to test BlueZ master

PulseAudio branch to test BlueZ master

[Mikel Astiz]

Hi all,

It's been recently quite hard to test the audio part of BlueZ without
having an updated PulseAudio which keeps up with the new D-Bus API in
BlueZ.

I've been doing some work to solve this and here you have the result:
git://github.com/mastiz/pulseaudio-bluez5.git

It's a big patchset on top of Tanu's next branch for PulseAudio. Only
the last patches are related to BlueZ 5, and the rest should slowly
get merged into PA master, in some way or another (first patches
already submitted upstream).

I warn you that it's not going to be stable in general, and in
particular I haven't done any testing with oFono. So I'd expect quite
a few issues.

Consider this work-in-progress and note that it will get rebased
periodically on top of Tanu's tree. If you have any feedback or code
contributions, I'd suggest you contact me directly.

Cheers,
Mikel

Re: PulseAudio branch to test BlueZ master

[Mikel Astiz]

Hi all,

On Thu, Dec 6, 2012 at 5:03 PM, Mikel Astiz <mikel.astiz.oss@gmail.com> wrote:
> Hi all,
>
> It's been recently quite hard to test the audio part of BlueZ without
> having an updated PulseAudio which keeps up with the new D-Bus API in
> BlueZ.
>
> I've been doing some work to solve this and here you have the result:
> git://github.com/mastiz/pulseaudio-bluez5.git

I just updated the PulseAudio branch to test the BlueZ 5 API and the
patches for oFono. It's once again in:
git://github.com/mastiz/pulseaudio-bluez5.git branch bluez5-head

The main changes include the rebase on top of a more recent PA "next"
branch and some minor fixes. In addition, the feedback from Vinicious
has been integrated, affecting the PropertiesChanged signals from
oFono to PA.

> It's a big patchset on top of Tanu's next branch for PulseAudio. Only
> the last patches are related to BlueZ 5, and the rest should slowly
> get merged into PA master, in some way or another (first patches
> already submitted upstream).

Many of these patches are now merged in the "next" PA branch, so the
remaining patches are BlueZ 5-specific (except the first one).

> I warn you that it's not going to be stable in general, and in
> particular I haven't done any testing with oFono. So I'd expect quite
> a few issues.

For the first time, I got HFP working with BlueZ+oFono+PA. Still
unstable though.

On the other hand, A2DP is not working for me any more. I think this
might be an unrelated issue, probably a regression in BlueZ, but I
have to do further research.

> Consider this work-in-progress and note that it will get rebased
> periodically on top of Tanu's tree. If you have any feedback or code
> contributions, I'd suggest you contact me directly.

I will post the patches as RFC to PA mailing list once Tanu's "next"
branch gets merged into master.

Cheers,
Mikel

Re: PulseAudio branch to test BlueZ master

[Vinicius Costa Gomes]

Hi Mikel,

On 11:19 Tue 18 Dec, Mikel Astiz wrote:
> On the other hand, A2DP is not working for me any more. I think this
> might be an unrelated issue, probably a regression in BlueZ, but I
> have to do further research.
> 

This problem may have been caused by me :-) The commit "bluetooth: Add
HFP 1.6 codec ID" was buggy, and was registering the A2DP endpoint with
the wrong codec. Sorry about that. I pushed an updated version to my
fork of pulseaudio[1].


Cheers,
-- 
Vinicius

[1] git://github.com/vcgomes/pulseaudio-bluez5.git 

Re: PulseAudio branch to test BlueZ master

[Mikel Astiz]

Hi Vinicius,

On Tue, Dec 18, 2012 at 9:24 PM, Vinicius Costa Gomes
<vinicius.gomes@openbossa.org> wrote:
> Hi Mikel,
>
> On 11:19 Tue 18 Dec, Mikel Astiz wrote:
>> On the other hand, A2DP is not working for me any more. I think this
>> might be an unrelated issue, probably a regression in BlueZ, but I
>> have to do further research.
>>
>
> This problem may have been caused by me :-) The commit "bluetooth: Add
> HFP 1.6 codec ID" was buggy, and was registering the A2DP endpoint with
> the wrong codec. Sorry about that. I pushed an updated version to my
> fork of pulseaudio[1].

You're right, it was caused by this patch. Thanks for testing.

I updated now and it seems to work both with A2DP and HFP.

Btw, now that the PulseAudio "next" branch has been merged to the
master branch, I'll send the patches as RFC to the PulseAudio mailing
list.

Cheers,
Mikel