24 Jan
2005
24 Jan
'05
6:46 p.m.
I have been trying to get dssi up and running on my machine this
afternoon and I have run into a snag. It appears to compile correctly
(both CVS and tarballs). But whenever I try to do the tests I get a
segfault. I know its some user error, but its difficult to track down
with only a segfault.
Is there a common mistake I am doing?
thanks
m.
25 Jan
25 Jan
4:52 a.m.
On Mon, Jan 24, 2005 at 02:46:08 -0800, Matthew Allen wrote:
I have been trying to get dssi up and running on my
machine this
afternoon and I have run into a snag. It appears to compile correctly
(both CVS and tarballs). But whenever I try to do the tests I get a
segfault. I know its some user error, but its difficult to track down
with only a segfault.
Is there a common mistake I am doing?
Segfaults are (almost) never due to user errors. Unless you did
something truely wierd or forced an install its the developers fault :)
It could be that your system networking config is a bit odd. Verisons of
liblo < 0.16, which the DSSI test code depends on, segfaulted under some
condititons.
Unfortunatly I cant remember the fix, but hopefully someone here can...
You could just upgrade liblo, but it would be good to get your netowrking
config sorted out for the future, assuming you have a network.
If upgrading liblo doesnt fix it then please send a gdb backtrace, mail me
offlist if you dont know how to do that.
- Steve
26 Jan
26 Jan
4:46 a.m.
Steve Harris wrote:
On Mon, Jan 24, 2005 at 02:46:08 -0800, Matthew Allen
wrote:
IIRC, make sure that the name reported by hostname(1) is in the
/etc/hosts file. Otherwise liblo would segfault when trying to launch a
new server.
Cheers,
Walco
I have been trying to get dssi up and running on
my machine this
afternoon and I have run into a snag. It appears to compile correctly
(both CVS and tarballs). But whenever I try to do the tests I get a
segfault. I know its some user error, but its difficult to track down
with only a segfault.
Is there a common mistake I am doing?
Segfaults are (almost) never due to user errors. Unless you did
something truely wierd or forced an install its the developers fault :)
It could be that your system networking config is a bit odd. Verisons of
liblo < 0.16, which the DSSI test code depends on, segfaulted under some
condititons.
Unfortunatly I cant remember the fix, but hopefully someone here can...
12:56 p.m.
On Wed, 26 Jan 2005 09:46:17 +0100, Walco <walco-linux(a)n--tree.net> wrote:
IIRC, make sure that the name reported by hostname(1) is in the
/etc/hosts file. Otherwise liblo would segfault when trying to launch a
new server.
Cheers,
Walco
And remeber correctly you did! My hostname in /etc/hosts was different
than the actual hostname reported. Edit file, reboot, and Wooo Hooo
dssi land has arrived. Thasnk for all the help guys.
2:02 p.m.
On Wed, 26 Jan 2005 08:56:33 -0800, Matthew Allen <mtallen(a)gmail.com> wrote:
And remeber correctly you did! My hostname in
/etc/hosts was different
than the actual hostname reported. Edit file, reboot, and Wooo Hooo
dssi land has arrived. Thasnk for all the help guys.
I was having problems with the sooperlooper gui recently and placing
my hostname in /etc/hosts fixed that.
Fluidsynth-dssi was segfaulting rosegarden for me. I could select
fluidsynth-dssi as a soft synth but when I clicked on the editor it
segfaulted. Now with the liblo update I can click on the editor button
but nothing happens.
I'm still thinking it might be a hostname routing issue. What should
be in /etc/hosts? Currently all that's there is my static IP address
for my LAN, hostname and alias (same as hostname). Should the
loopback interface or something else be there for dssi to work?
--Jason
2:30 p.m.
actually the only thing I have in my /etc/hosts is my loopback
127.0.0.1 soundbox
and a bunch of IPv6 stuff I don't have the slightest clue about.
I havent tried rosegarden yet, but all of the synths work well from
the command line. I will try Rosegarden tonight.
m.
On Wed, 26 Jan 2005 13:01:43 -0500, jason.ronallo(a)gmail.com
<jason.ronallo(a)gmail.com> wrote:
On Wed, 26 Jan 2005 08:56:33 -0800, Matthew Allen
<mtallen(a)gmail.com> wrote:
And remeber correctly you did! My hostname in
/etc/hosts was different
than the actual hostname reported. Edit file, reboot, and Wooo Hooo
dssi land has arrived. Thasnk for all the help guys.
I was having problems with the sooperlooper gui recently and placing
my hostname in /etc/hosts fixed that.
Fluidsynth-dssi was segfaulting rosegarden for me. I could select
fluidsynth-dssi as a soft synth but when I clicked on the editor it
segfaulted. Now with the liblo update I can click on the editor button
but nothing happens.
I'm still thinking it might be a hostname routing issue. What should
be in /etc/hosts? Currently all that's there is my static IP address
for my LAN, hostname and alias (same as hostname). Should the
loopback interface or something else be there for dssi to work?
--Jason
27 Jan
27 Jan
5:22 a.m.
On Wed, Jan 26, 2005 at 10:30:21 -0800, Matthew Allen wrote:
actually the only thing I have in my /etc/hosts is my
loopback
127.0.0.1 soundbox
and a bunch of IPv6 stuff I don't have the slightest clue about.
liblo does tend to pick ipv6 addresses if it can (some artifact of the way
the new name resoltuion api works I guess).
I only have 127.0.0.1 and my hostname though, and that works.
- Steve
5:19 a.m.
On Wed, Jan 26, 2005 at 08:56:33 -0800, Matthew Allen wrote:
On Wed, 26 Jan 2005 09:46:17 +0100, Walco
<walco-linux(a)n--tree.net> wrote:
OK, I had hoped that the fix in the latest version would actually make
this case work for the local case. Maybe we should ship some testcode to
spot this case and warn people?
- Steve
IIRC, make sure that the name reported by hostname(1) is in the
/etc/hosts file. Otherwise liblo would segfault when trying to launch a
new server.
Cheers,
Walco
And remeber correctly you did! My hostname in /etc/hosts was different
than the actual hostname reported. Edit file, reboot, and Wooo Hooo
dssi land has arrived. Thasnk for all the help guys.
7:41 a.m.
Hi Steve,
I was assuming he was still using liblo 0.15 - the 'local case' should
work in 0.16. In the scenario that the name resolution setup is *really*
screwed (empty hosts file, no/misconfigured dns), liblo's fallback on
hostname 'localhost' (which then can also not be found) will only work
if liblo clients connect explicitly using the loopback IP address
instead of the default 'localhost'. In that case a warning might be
appropriate.
Cheers,
Walco
And remeber
correctly you did! My hostname in /etc/hosts was different
than the actual hostname reported. Edit file, reboot, and Wooo Hooo
dssi land has arrived. Thasnk for all the help guys.
OK, I had hoped that the fix in the latest version would actually make
this case work for the local case. Maybe we should ship some testcode to
spot this case and warn people?
10:29 a.m.
On Thu, Jan 27, 2005 at 12:41:05 +0100, Walco wrote:
Hi Steve,
I was assuming he was still using liblo 0.15 - the 'local case' should
work in 0.16. In the scenario that the name resolution setup is *really*
screwed (empty hosts file, no/misconfigured dns), liblo's fallback on
hostname 'localhost' (which then can also not be found) will only work
if liblo clients connect explicitly using the loopback IP address
instead of the default 'localhost'. In that case a warning might be
appropriate.
I have considered that, but the problem is that there is more than one
loopback address. I'm taking the poistion that liblo wont work if your
networking setup is too deeplytruely mangled.
- Steve
And
remeber correctly you did! My hostname in /etc/hosts was different
than the actual hostname reported. Edit file, reboot, and Wooo Hooo
dssi land has arrived. Thasnk for all the help guys.
OK, I had hoped that the fix in the latest version would actually make
this case work for the local case. Maybe we should ship some testcode to
spot this case and warn people?
11:16 a.m.
On Thu, 27 Jan 2005 14:28:42 +0000, Steve Harris
<S.W.Harris(a)ecs.soton.ac.uk> wrote:
On Thu, Jan 27, 2005 at 12:41:05 +0100, Walco wrote:
Just for the record. The original segfault was using 0.15 (downloaded
from tarball on saturday). The segfaults went away with 0.16 (out of
CVS), but I had no gui until I fixed my hosts file.
Also I dont really thinkg my networking was really in that bad of a
state. (at least from my persepctive). Both wireless and wired
connections work great with a dhcp server and other apps that really
heavily on a network connection (Pure data being the one I use most)
also work well.
m.
Hi Steve,
I was assuming he was still using liblo 0.15 - the 'local case' should
work in 0.16. In the scenario that the name resolution setup is *really*
screwed (empty hosts file, no/misconfigured dns), liblo's fallback on
hostname 'localhost' (which then can also not be found) will only work
if liblo clients connect explicitly using the loopback IP address
instead of the default 'localhost'. In that case a warning might be
appropriate.
I have considered that, but the problem is that there is more than one
loopback address. I'm taking the poistion that liblo wont work if your
networking setup is too deeplytruely mangled. And
remeber correctly you did! My hostname in /etc/hosts was different
than the actual hostname reported. Edit file, reboot, and Wooo Hooo
dssi land has arrived. Thasnk for all the help guys.
OK, I had hoped that the fix in the latest version would actually make
this case work for the local case. Maybe we should ship some testcode to
spot this case and warn people?
7269
days inactive
7272
days old
linux-audio-user@lists.linuxaudio.org
10 comments
4 participants
participants (4)
-
jason.ronalloï¼ gmail.com -
Matthew Allen -
Steve Harris -
Walco