Discussion:
just fixed regression in coreaudio non-interleaved callback i/o
(too old to reply)
Ross Bencina
2011-06-30 12:50:53 UTC
Permalink
Hi Guys

Just a heads up: I just found and fixed regression in r1612 that broke
coreaudio non-interleaved callback i/o. It would cause Pa_OpenStream to
return paFormatNotSupported when trying to open a non-interleaved callback
stream.

Presumably Phil's change was just supposed to return an not-supported error
for non-interleaved blocking i/o (which is currently broken) but in fact it
returned the error for callback streams too:

http://www.portaudio.com/trac/changeset?reponame=&new=1612%40portaudio%2Ftrunk%2Fsrc%2Fhostapi%2Fcoreaudio%2Fpa_mac_core.c&old=1544%40portaudio%2Ftrunk%2Fsrc%2Fhostapi%2Fcoreaudio%2Fpa_mac_core.c

I think we should probably update the snapshot soon. That's a pretty bad
regression.

Thanks

Ross
Bjorn Roche
2011-06-30 14:34:10 UTC
Permalink
THanks Ross. I'm not sure I see how this is a regression, but I'm also unclear what change by Phil you are referring to. But don't feel the need to explain unless it seems important for some reason.

bjorn
Post by Ross Bencina
Hi Guys
Just a heads up: I just found and fixed regression in r1612 that broke coreaudio non-interleaved callback i/o. It would cause Pa_OpenStream to return paFormatNotSupported when trying to open a non-interleaved callback stream.
http://www.portaudio.com/trac/changeset?reponame=&new=1612%40portaudio%2Ftrunk%2Fsrc%2Fhostapi%2Fcoreaudio%2Fpa_mac_core.c&old=1544%40portaudio%2Ftrunk%2Fsrc%2Fhostapi%2Fcoreaudio%2Fpa_mac_core.c
I think we should probably update the snapshot soon. That's a pretty bad regression.
Thanks
Ross
_______________________________________________
Portaudio mailing list
Portaudio at music.columbia.edu
http://music.columbia.edu/mailman/listinfo/portaudio
-----------------------------
Bjorn Roche
http://www.xonami.com
Audio Collaboration
Ross Bencina
2011-06-30 16:33:39 UTC
Permalink
Hi Bjorn

It's a regression because prior to r1612 non-interleaved buffering with
callbacks worked fine in PA/CoreAudio. Between r1612 and today the streams
would fail to open and report paFormatNotSupported.

The link below shows the diff of Phil's change.

Ross.


----- Original Message -----
From: "Bjorn Roche" <bjorn at xowave.com>
To: "Portaudio Mailing List" <portaudio at music.columbia.edu>
Sent: Friday, July 01, 2011 12:34 AM
Subject: Re: [Portaudio] just fixed regression in coreaudio
non-interleavedcallback i/o
Post by Bjorn Roche
THanks Ross. I'm not sure I see how this is a regression, but I'm also
unclear what change by Phil you are referring to. But don't feel the need
to explain unless it seems important for some reason.
bjorn
Post by Ross Bencina
Hi Guys
Just a heads up: I just found and fixed regression in r1612 that broke
coreaudio non-interleaved callback i/o. It would cause Pa_OpenStream to
return paFormatNotSupported when trying to open a non-interleaved
callback stream.
Presumably Phil's change was just supposed to return an not-supported
error for non-interleaved blocking i/o (which is currently broken) but in
http://www.portaudio.com/trac/changeset?reponame=&new=1612%40portaudio%2Ftrunk%2Fsrc%2Fhostapi%2Fcoreaudio%2Fpa_mac_core.c&old=1544%40portaudio%2Ftrunk%2Fsrc%2Fhostapi%2Fcoreaudio%2Fpa_mac_core.c
I think we should probably update the snapshot soon. That's a pretty bad regression.
Thanks
Ross
_______________________________________________
Portaudio mailing list
Portaudio at music.columbia.edu
http://music.columbia.edu/mailman/listinfo/portaudio
-----------------------------
Bjorn Roche
http://www.xonami.com
Audio Collaboration
_______________________________________________
Portaudio mailing list
Portaudio at music.columbia.edu
http://music.columbia.edu/mailman/listinfo/portaudio
Phil Burk
2011-06-30 18:22:09 UTC
Permalink
Hi Ross,
Post by Ross Bencina
Just a heads up: I just found and fixed regression in r1612 that
broke coreaudio non-interleaved callback i/o. It would cause
Pa_OpenStream to return paFormatNotSupported when trying to open a
non-interleaved callback stream.
Thanks for catching that. That patch was in response to Roman Shulakov's
email on 2/28/11 about broken nonInterleaved audio. I did not read it
closely enough. Apparently nonInterleaved works OK for callback. Your
change looks good.

Phil
Phil Burk
2011-06-30 18:37:31 UTC
Permalink
Post by Ross Bencina
I think we should probably update the snapshot soon.
The old SS2 was based on rev 1647. We are now up to rev 1697. So there
have been lots of changes.

Do you feel we are at a fairly stable point?

Is there anything critical we need for V19-SS3?

BTW, I notice Trac said we are 5 months late for V19-SS2 milestone. So I
marked SS2 complete on 3/26/11 and added V19-SS3 for the next stable
snapshot.

Phil Burk
Chris
2011-06-30 21:31:49 UTC
Permalink
I don't know if Viktor's problems are directly related to it, or
whether he is still having problems, but I think ticket #146 can be
closed. I have been unable to get the reported problem to surface since
adding wav headers to my test files. Viktor?


-----Original Message-----
From: Phil Burk <pburk at mobileer.com>
To: portaudio <portaudio at music.columbia.edu>
Sent: Thu, Jun 30, 2011 11:37 am
Subject: [Portaudio] another snapshot
Post by Ross Bencina
I think we should probably update the snapshot soon.
The old SS2 was based on rev 1647. We are now up to rev 1697. So there
have been lots of changes.

Do you feel we are at a fairly stable point?

Is there anything critical we need for V19-SS3?

BTW, I notice Trac said we are 5 months late for V19-SS2 milestone. So
I
marked SS2 complete on 3/26/11 and added V19-SS3 for the next stable
snapshot.

Phil Burk
Viktor Ivanov
2011-07-01 09:20:27 UTC
Permalink
I don't know if Viktor's problems are directly related to it, or whether
he is still having problems, but I think ticket #146 can be closed. I
have been unable to get the reported problem to surface since adding wav
headers to my test files. Viktor?
I'll run a check in a few days (the project was put on hold). IIRC we
agreed that the problems are unrelated. There is a little update for my
case: the issue is only for WASAPI input streams running in exclusive
event mode.

When we're back on that project I'll do a detailed report and maybe it
will be better to open a new ticket then?


Cheers
Chris
2011-07-01 10:09:12 UTC
Permalink
<< When we're back on that project I'll do a detailed report and maybe
it
will be better to open a new ticket then? >>

If it is conclusively determined that the problem is with PortAudio and
not your code or a device driver. I haven't experienced any problems
using WASAPI in exclusive and callback modes.

-----Original Message-----
From: Viktor Ivanov <viktor.ivanov at securax.org>
To: portaudio <portaudio at music.columbia.edu>
Sent: Fri, Jul 1, 2011 2:20 am
Subject: Re: [Portaudio] another snapshot
I don't know if Viktor's problems are directly related to it, or whether
he is still having problems, but I think ticket #146 can be closed. I
have been unable to get the reported problem to surface since adding wav
headers to my test files. Viktor?
I'll run a check in a few days (the project was put on hold). IIRC we
agreed that the problems are unrelated. There is a little update for my
case: the issue is only for WASAPI input streams running in exclusive
event mode.

When we're back on that project I'll do a detailed report and maybe it
will be better to open a new ticket then?


Cheers
Phil Burk
2011-07-01 16:21:40 UTC
Permalink
I closed ticket #146.

Thanks Chris.

Phil
Post by Chris
Post by Chris
I don't know if Viktor's problems are directly related to it, or
whether
Post by Chris
he is still having problems, but I think ticket #146 can be closed. I
have been unable to get the reported problem to surface since adding
wav
Post by Chris
headers to my test files. Viktor?
Loading...