Re: [Jack-Devel] port disconnection callbacks (JackPortConnectCallback)

PrevNext  Index
DateFri, 10 Oct 2014 09:39:43 +1100
From Matt Flax <[hidden] at flatmax dot org>
ToPaul Davis <[hidden] at linuxaudiosystems dot com>, Stéphane Letz <[hidden] at grame dot fr>
CcJACK <[hidden] at lists dot jackaudio dot org>
In-Reply-ToPaul Davis Re: [Jack-Devel] port disconnection callbacks (JackPortConnectCallback)
On 10/10/14 00:12, Paul Davis wrote:
>
>
> On Thu, Oct 9, 2014 at 9:04 AM, Stéphane Letz <[hidden] 
> <mailto:[hidden]>> wrote:
>
>
>     Le 9 oct. 2014 à 14:57, Paul Davis <[hidden]
>     <mailto:[hidden]>> a écrit :
>
>     > return 0 on success, otherwise a non-zero error code, causing JACK
>
>
>     When this is the documentation of the
>     "jack_set_thread_init_callback" …
>
>     The documentation of jack_set_process_thread is " * @return 0 on
>     success, otherwise a non-zero error code."
>
>     ==> so?
>
>
> this is why i said the documentation was ambiguous because it doesn't 
> make it clear that the "@return" text describes the behaviour of the 
> process callback, *not* jack_set_process_callback().
>

Let me try to understand a little of what's going on here ....

Is the original problem that the real time thread (which processes 
audio) is receiving a non-zero return from a client - which indicates to 
stop processing that client. At this point, the client's ports are 
disconnected by the real time thread.

The non-realtime thread is the one which executes port (dis)connection 
callbacks ... this non-realtime thread has no way of knowing that the 
real time thread has had a client change.

Matt
PrevNext  Index

1412894394.24561_0.ltw:2, <54370EAF.8070609 at flatmax dot org>