[pulseaudio-discuss] An raop2 support

[Hajime Fujita]

Hi Anton,

I think I fixed the crash issue which you reported.

>> E: [lt-pulseaudio] rtsp_client.c: Assertion 'c->url' failed at
modules/rtp/rtsp_client.c:400, function rtsp_exec(). Aborting.

Please try the latest code.
https://github.com/hfujita/pulseaudio-raop2/commits/raop2-for-merge1


Thanks,
Hajime

Hajime Fujita wrote:
> Hi Anton,
> 
> Thank you for testing and reporting several issues.
> 
> Anton Lundin wrote:
>> On 06 September, 2013 - Hajime Fujita wrote:
>>
>>> Hi Anton,
>>>
>>> Thank you for your interest.
>>>
>>> I have switched to a new branch and several new features/fixes are
>>> integrated. Could you test this one, if you'd like?
>>> https://github.com/hfujita/pulseaudio-raop2/commits/raop2-for-merge
>>>
>>> The following command should allow you to switch to the latest branch.
>>> $ cd pulseaudio-raop2
>>> $ git fetch origin
>>> $ git checkout -b raop2-for-merge origin/raop2-for-merge
>>>
>>> Or you can simply clone the whole tree again (assuming that your
>>> internet connection to github is fast enough.)
>>
>> I tested the new branch, and stuff worked loads better!
>>
>> I also tested the patches you posted here on the list on top of current
>> master and it worked the same.
>>
>> I have quite a few assertions:
>> raop_packet_buffer.c: Assertion 'packet->data[4 + 3] == (seq_num & 0xff)' failed at modules/raop/raop_packet_buffer.c:154, function pa_raop_pb_read_packet(). Aborting.
>>
>> This should be fully reproducible if you test the raop module against
>> XBMC.
> 
> Hmm... I thought this should not happen (that's why I inserted assertion
> instead of warning).
> I found one clear bug in the current implementation.
> 
> I've pushed two patches to the `raop2-for-merge1` branch with aiming to
> solve this issue (again, sorry for the terrible branch naming!)
> https://github.com/hfujita/pulseaudio-raop2/commits/raop2-for-merge1
> 
>> Against the Denon receiver i se some errors like:
>> E: [lt-pulseaudio] rtsp_client.c: Assertion 'c->url' failed at modules/rtp/rtsp_client.c:400, function rtsp_exec(). Aborting.
>>
>> Thats related to when i move the volume slider. I could get you a
>> network trace if that helps?
> 
> I'm aware of this issue and found out the cause.
> I'm now trying to figure out what would be the best way to fix it...
> 
>>
>> I also se some:
>> D: [lt-pulseaudio] rtsp_client.c: Sending command: RECORD
>> E: [raop-sink] module-raop-sink.c: Failed to send UDP packet: Resource temporarily unavailable
>>
>> When setting up a stream.
>>
>> While i tested the patches posted to the list, i also tested the new
>> tunnel sink and it worked to tunnel to a raop sink with the new module
>> without the "stuttering", yey =)
>>
>>
>> //Anton
>>
>>
>>>
>>> Anton Lundin wrote:
>>>> On 28 August, 2013 - Hajime Fujita wrote:
>>>>
>>>>> Hello,
>>>>>
>>>>> Recently I found that the current PulseAudio did not support the latest
>>>>> AirPlay audio streaming, so I started to brush up the raop2
>>>>> implementation based on Christophe Fergeau and Martin Blanchard's
>>>>> implementation.
>>>>>
>>>>> Here is my ongoing work.
>>>>> https://github.com/hfujita/pulseaudio-raop2 (repo)
>>>>> http://hfujita.github.io/pulseaudio-raop2/ (brief document)
>>>>>
>>>>> My primary goal is to enjoy musics in my AV receiver from Linux, but a
>>>>> vague ambitious is to get the code merged into upstream so that anyone
>>>>> can use the feature by just using distro packages.
>>>>>
>>>>> Current status:
>>>>> * Plays music with Pioneer VSX-43.
>>>>> * Sound test (launched from the GNOME sound setting dialog) does not
>>>>> work (no sound heard).
>>>>> * When stop playing music and then start again, PulseAudio may go into
>>>>> an abnormal state (no sound/hang up/crash).
>>>>> * Works reasonably fine with VLC/totem but when playing music from
>>>>> Rhythmbox, quite rough sound is heard.
>>>>> * Packet retransmission is not supported yet.
>>>>>
>>>>> I'm very happy to hear any sort of suggestions.
>>>>> In the near future I may ask several technical questions about
>>>>> PulseAudio internals.
>>>>
>>>> Hi!
>>>>
>>>> I thought i take it for a spinn and try it against my XBMC machine and my
>>>> Denon-1912, and got some mixed results.
>>>>
>>>> First i thought i test it by running another pulseaudio server on my
>>>> machine and use tunnel-sink to connect my "regular" server to the server
>>>> with the new raop-module in it, but that didn't really work. Something
>>>> with the raop-module doesn't really behave when a tunnel sink is feeding
>>>> it.
>>>
>>> Could you point me out any document that I can follow to reproduce the
>>> above situation?
>>>
>>>> Later i tested to run a real client against it, and that works quite
>>>> well!
>>>> Occasional stutter, maybe once a minute, when running against XBMC, but
>>>> stutter every 10 seconds or so when running against the Denon receiver.
>>>
>>> This might be fixed by the latest patch I showed before, as they
>>> introduce packet retransmission and more precise timing calculation.
>>>
>>>> Looks like XBMC uses libshairport and it might be a interesting project
>>>> to write a module using that to receive raop to pulseaudio?
>>>
>>> Could be interesting, but my current interest is to use Linux machine as
>>> an audio source, not the sink. Also it's beyond my capability right now.
>>>
>>>>
>>>>
>>>> //Anton
>>>>
>>>>
>>>
>>> Thanks,
>>> Hajime
>>>
>>>
>>> _______________________________________________
>>> pulseaudio-discuss mailing list
>>> pulseaudio-discuss at lists.freedesktop.org
>>> http://lists.freedesktop.org/mailman/listinfo/pulseaudio-discuss
>>
> 
> 
> Thanks,
> Hajime
> 
> _______________________________________________
> pulseaudio-discuss mailing list
> pulseaudio-discuss at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/pulseaudio-discuss
>