Re: [Jack-Devel] JACK in Chrome !!

On Wed, 16 Jan 2013 15:56:06 +0100, Patrick Shirkey  
<[hidden]> wrote:

>
> On Wed, January 16, 2013 9:03 pm, Kaj Ailomaa wrote:
>> On Wed, 16 Jan 2013 08:00:35 +0100, Patrick Shirkey
>> <[hidden]> wrote:
>>
>>
>>> This is the process for enabling jack and pulse on Debian systems.
>>>
>>> 1: Add the user to the audio group
>>
>> On Debian, the user is already a member of the audio group, but not on
>> Ubuntu.
>>
>
> I recall that on a clean install of Debian Wheezy 64 bit that I had to  
> add
> the user to the audio group.
>  It might be me but I only had to do it a couple of months back so its  
> all
> still pretty fresh.
>
>>> 2: logout of the desktop to apply the changes to the user priviledges
>>> 3: install "pulseaudio-module-jack"
>>> 4: disable autospawn by manually editing the ~/pulse/client.conf file
>>>
>>>     autospawn = no
>>>
>>> 5: restart pulseaudio
>>>
>>>     pulseaudio -k
>>>     pulseaudio -D
>>>
>>
>> If this is required for you to get jackdbus grab the card from  
>> pulseaudio
>> (by killing pulseaudio before starting jack), then you'd have to do the
>> same on any system with the same version of pulseaudio and jack. The
>> reason to that is what you explain at 7:, namely, pulseaudio suffering
>>  from a bug making it not release the card at certain circumstances.
>>
>
> Not necessarily. it only kicks in if you originally start PA with
> autospawn enabled.

What kicks in? The only reason why jackdbus is not grabbing the card  
sometimes, is because of a bug in pulseaudio. Has nothing to do with  
packaging, or the pulse config for autospawn. If jackdbus is failing to  
grab the card, you can also use pasuspender to temporarily have PA ungrab  
the card.


> I don't know how PA is configured differently on Fedora but I didn't have
> the same problems using it with that OS.
>

I might look into that. However, the only two things I can think of what  
makes their way work (if that is actually so) is either:
1. they've pathched the pulse code and not shared upstream (which sounds  
utterly implausible)
2. They are using some kind of wrapper with jack (which then would  
circumvent the bug)

Also, you may have had more than one audio device in your system, and  
started jack with another card.

In any way, it's not a packaging problem per se. The pulseaudio bug has  
seen some attention, so it aught to get fixed in time.

>> So far, I'm not seeing any distro specific problem. Only buggy code. And
>>  from what I've read, there may be more than one bug causing problems in
>> pulse and jack integration.
>>
>
> The other thing that causes severe headaches is the naming of the jack1/2
> packages on Debian.
>
>

I also fail to see the problem here. There are two different jacks after  
all.

1358351161.11870_0.ltw:2,a <op.wq0gr8p82rg3it at melete dot mousike dot me>