Re: jackd-1.9.5 and jackd2 transition

2010-02-16 Thread Adrian Knoth
On Tue, Feb 16, 2010 at 04:16:48AM +0100, Jonas Smedegaard wrote:

[jackd-1.9.5]
 From brief testing:

 It seems jackd v2 is less flexible in referencing ALSA devices.

Confirmed by upstream. I also saw this.


 This works:

   jackd -d alsa --device=hw:0,3

Yep.

 Also, killing jac_netsource spews the following to the console (unlike  
 v1 which just in a single line informed that it was killed):

 *** glibc detected *** jack_netsource: double free or corruption  
 (!prev): 0x08580810 ***
 === Backtrace: =
 /lib/i686/cmov/libc.so.6[0xb7dac824]

Can't confirm that. Over here, it looks like this:

a...@hex:~$ jack_netsource -H localhost
Connected :-)
netjack: at frame 93 - total netxruns 1  (1%) queue time= 42669
^c...@hex:~$ 


Tested with jackd -d netone.


-- 
mail: a...@thur.de  http://adi.thur.de  PGP/GPG: key via keyserver

___
pkg-multimedia-maintainers mailing list
pkg-multimedia-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/mailman/listinfo/pkg-multimedia-maintainers


jackd-1.9.5 and jackd2 transition

2010-02-15 Thread Adrian Knoth
Hi!


I just merged the jackd-1.9.5 release into our repo. I also enabled the
jackdbus feature which is required for ladish.

If you like, please give it a whirl:

   
http://git.debian.org/?p=pkg-multimedia/jack-audio-connection-kit.git;a=summary


To me, it gives better results than jackd-0.118.x. Better here means:
pulseaudio runs fine on top of jackd2, way more stable than with jackd1.
It supports glitch-free graph updates (start with -S), this means, the
audio stream isn't interrupted when you add a new jack client or track
in ardour.

To my knowledge, there's only one drawback:

   http://subversion.ffado.org/ticket/264

jackd2 might cause lots of error messages when there's a buffer underrun
in FFADO, that is, when the firewire ISO streaming interrupts. With
jackd1, you get exactly one underrun message, with jackd2, your terminal
might get flooded, which in turn could make the system unresponsive.

Since I'm also affected by this behaviour, I'll ping upstream to fix it.

What's missing? Manpages. jackd2 doesn't ship them, so I'll copy them
from jackd1. I'll also propose to include them in the official jackd2
release. I guess we shouldn't let jackd2 enter unstable without
manpages.


Remaining question: Do you think jackd-1.9.5 should be included in
squeeze? It surely has a lot of benefits, and given the lifespan of a
release, having jackdbus in squeeze would easy backporting ladish. OTOH,
we currently don't have user feedback, IOW, it lacks testing.

Possible solution: upload jackd2 to unstable instead of experimental
(which means no way back to jackd1), file a RC bug against the package
to prevent it from entering testing, and once we see that users are fine
with it, let it slip into squeeze/testing before the freeze.



Cheerio

-- 
mail: a...@thur.de  http://adi.thur.de  PGP/GPG: key via keyserver

___
pkg-multimedia-maintainers mailing list
pkg-multimedia-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/mailman/listinfo/pkg-multimedia-maintainers


Re: jackd-1.9.5 and jackd2 transition

2010-02-15 Thread Jonas Smedegaard

On Mon, Feb 15, 2010 at 12:54:37PM +0100, Adrian Knoth wrote:
I just merged the jackd-1.9.5 release into our repo. I also enabled the 
jackdbus feature which is required for ladish.


If you like, please give it a whirl:

  
http://git.debian.org/?p=pkg-multimedia/jack-audio-connection-kit.git;a=summary


To me, it gives better results than jackd-0.118.x. Better here means:
pulseaudio runs fine on top of jackd2, way more stable than with jackd1.
It supports glitch-free graph updates (start with -S), this means, the
audio stream isn't interrupted when you add a new jack client or track
in ardour.

To my knowledge, there's only one drawback:

  http://subversion.ffado.org/ticket/264


From brief testing:

It seems jackd v2 is less flexible in referencing ALSA devices.

According to aplay -L S/PDIF of my VIA EPIA board is 
iec958:CARD=V8237,DEV=0.


This works with jackd v1:

  jackd -d alsa --device=iec958:CARD=V8237,DEV=0

With jackd v2 it fails.  These also fails (untested if they would work 
with jackd v1):


  jackd -d alsa --device=plug:iec958

  jackd -d alsa --device=iec958:V8237

This works:

  jackd -d alsa --device=hw:0,3



Also, killing jac_netsource spews the following to the console (unlike 
v1 which just in a single line informed that it was killed):


*** glibc detected *** jack_netsource: double free or corruption 
(!prev): 0x08580810 ***

=== Backtrace: =
/lib/i686/cmov/libc.so.6[0xb7dac824]
/lib/i686/cmov/libc.so.6[0xb7dae0b3]
/lib/i686/cmov/libc.so.6(cfree+0x6d)[0xb7db10ad]
jack_netsource[0x804bcc4]
jack_netsource[0x804ad4a]
/lib/i686/cmov/libc.so.6(__libc_start_main+0xe5)[0xb7d57b55]
jack_netsource[0x80493f1]
=== Memory map: 
08048000-0804f000 r-xp  fe:01 9502   /usr/bin/jack_netsource
0804f000-0805 rw-p 6000 fe:01 9502   /usr/bin/jack_netsource
08572000-08593000 rw-p  00:00 0  [heap]
b2be4000-b2be5000 rw-p  00:00 0 
b2be5000-b2be6000 ---p  00:00 0 
b2be6000-b2c66000 rw-p  00:00 0 
b7ace000-b7ae rw-s  00:10 20486  
/dev/shm/jack-shm-registry
b7ae-b7ae1000 ---p  00:00 0 
b7ae1000-b7b61000 rw-p  00:00 0 
b7b61000-b7b62000 ---p  00:00 0 
b7b62000-b7be2000 rw-p  00:00 0 
b7be2000-b7bf3000 r-xp  fe:00 26396  
/lib/i686/cmov/libresolv-2.10.2.so
b7bf3000-b7bf4000 r--p 0001 fe:00 26396  
/lib/i686/cmov/libresolv-2.10.2.so
b7bf4000-b7bf5000 rw-p 00011000 fe:00 26396  
/lib/i686/cmov/libresolv-2.10.2.so
b7bf5000-b7bf7000 rw-p  00:00 0 
b7bf7000-b7bfc000 r-xp  fe:00 26384  
/lib/i686/cmov/libnss_dns-2.10.2.so
b7bfc000-b7bfd000 r--p 4000 fe:00 26384  
/lib/i686/cmov/libnss_dns-2.10.2.so
b7bfd000-b7bfe000 rw-p 5000 fe:00 26384  
/lib/i686/cmov/libnss_dns-2.10.2.so
b7bfe000-b7c08000 r-xp  fe:00 26367  
/lib/i686/cmov/libnss_files-2.10.2.so
b7c08000-b7c09000 r--p 9000 fe:00 26367  
/lib/i686/cmov/libnss_files-2.10.2.so
b7c09000-b7c0a000 rw-p a000 fe:00 26367  
/lib/i686/cmov/libnss_files-2.10.2.so
b7c0a000-b7c0c000 rw-p  00:00 0 
b7c0c000-b7c28000 r-xp  fe:00 10837  /lib/libgcc_s.so.1

b7c28000-b7c29000 rw-p 0001c000 fe:00 10837  /lib/libgcc_s.so.1
b7c29000-b7d0e000 r-xp  fe:01 775
/usr/lib/libstdc++.so.6.0.13
b7d0e000-b7d12000 r--p 000e5000 fe:01 775
/usr/lib/libstdc++.so.6.0.13
b7d12000-b7d13000 rw-p 000e9000 fe:01 775
/usr/lib/libstdc++.so.6.0.13
b7d13000-b7d1a000 rw-p  00:00 0 
b7d1a000-b7d3e000 r-xp  fe:00 26379  
/lib/i686/cmov/libm-2.10.2.so
b7d3e000-b7d3f000 r--p 00023000 fe:00 26379  
/lib/i686/cmov/libm-2.10.2.so
b7d3f000-b7d4 rw-p 00024000 fe:00 26379  
/lib/i686/cmov/libm-2.10.2.so
b7d4-b7d41000 rw-p  00:00 0 
b7d41000-b7e82000 r-xp  fe:00 26382  
/lib/i686/cmov/libc-2.10.2.so
b7e82000-b7e84000 r--p 00141000 fe:00 26382  
/lib/i686/cmov/libc-2.10.2.so
b7e84000-b7e85000 rw-p 00143000 fe:00 26382  
/lib/i686/cmov/libc-2.10.2.so
b7e85000-b7e88000 rw-p  00:00 0 
b7e88000-b7ff2000 r-xp  fe:01 10989  
/usr/lib/libsamplerate.so.0.1.7
b7ff2000-b7ff3000 rw-p 00169000 fe:01 10989  
/usr/lib/libsamplerate.so.0.1.7
b7ff3000-b8002000 r-xp  fe:01 10925  
/usr/lib/libcelt.so.0.0.0
b8002000-b8003000 rw-p e000 fe:01 10925  
/usr/lib/libcelt.so.0.0.0
b8003000-b8018000 r-xp  fe:00 26377  
/lib/i686/cmov/libpthread-2.10.2.so
b8018000-b8019000 r--p 00014000 fe:00 26377  
/lib/i686/cmov/libpthread-2.10.2.so
b8019000-b801a000 rw-p 00015000 fe:00 26377  
/lib/i686/cmov/libpthread-2.10.2.so
b801a000-b801c000 rw-p  00:00 0 
b801c000-b8023000 r-xp  fe:00 26378  
/lib/i686/cmov/librt-2.10.2.so
b8023000-b8024000 r--p 6000 fe:00 26378  
/lib/i686/cmov/librt-2.10.2.so
b8024000-b8025000 rw-p 7000 fe:00 26378  
/lib/i686/cmov/librt-2.10.2.so
b8025000-b8026000 rw-p  00:00 0 

Re: jackd-1.9.5 and jackd2 transition

2010-02-15 Thread Jonas Smedegaard

On Mon, Feb 15, 2010 at 12:54:37PM +0100, Adrian Knoth wrote:
I just merged the jackd-1.9.5 release into our repo. I also enabled the 
jackdbus feature which is required for ladish.


If you like, please give it a whirl:

  
http://git.debian.org/?p=pkg-multimedia/jack-audio-connection-kit.git;a=summary


For those interested in testing but lazy/inexperienced compiling from 
source, I have made available unofficial packages for i386 and amd64, 
Sid and Squeeze, using one of the following in your 
/etc/apt/sources.list:


deb http://debian.jones.dk/ sid audio
deb http://debian.jones.dk/ squeeze audio


Regards,

 - Jonas

--
* Jonas Smedegaard - idealist  Internet-arkitekt
* Tlf.: +45 40843136  Website: http://dr.jones.dk/

 [x] quote me freely  [ ] ask before reusing  [ ] keep private


signature.asc
Description: Digital signature
___
pkg-multimedia-maintainers mailing list
pkg-multimedia-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/mailman/listinfo/pkg-multimedia-maintainers