liblo-devel Mailing List for LibLO

Hi,

I have just migrated the LibLO CVS repository to SourceForge's new  
Subversion service.

You can browse the code repository here:
http://svn.sourceforge.net/liblo/

And check-out the latest version of the code out with this command:
svn co https://svn.sourceforge.net/svnroot/liblo/trunk liblo


nick.

Whoops, nevermind guys. It was subtle and took forcing a core dump to
find - apparently I was passing the OSC address to connect to as the
address to bind to.

Thanks for getting my gears turning. I've been sitting on this one for
over a month.

--=20

World wasn't remembered / World in my world / This hasn't been
remembered --Front 242, Waste

Public Key: http://sparked.zadzmo.org/

On Wed, 10 May 2006 09:34:38 +0100
Steve Harris <S.W...@ec...> wrote:

> That error will be coming from libc. The Servname refers to service names
> from /etc/services I think, and I don't think there's code that uses that,
> so it implies something is seriously messed up. the -8 is probably not a
> good sign either. Have you tried running it through valgrind incase
> theres a stack smash or pointer walk?

Actually, yes, valgrind is part of my regular stress test that I run on
just about every new feature I put in. Currently my program is squeaky
clean, with the only error currently being this:

=3D=3D6311=3D=3D Syscall param timer_create(evp) points to uninitialised by=
te(s)
=3D=3D6311=3D=3D    at 0x1BA16752: timer_create (in /lib/tls/librt-2.3.3.so)

As far as I can understand, that bit is harmless.

The only other thing from Valgrind are memleaks. One might be
interesting to look at:

=3D=3D6311=3D=3D 39 bytes in 1 blocks are definitely lost in loss record 6 =
of 20
=3D=3D6311=3D=3D    at 0x1B8FE8A2: malloc
(in /usr/lib/valgrind/vgpreload_memcheck.so) =3D=3D6311=3D=3D    by 0x1B91A=
73:
lo_server_get_url (server.c:1027)

=20
> Heh, one of my favourite songs, though Motion is better.

No one with 15 miles of my city has even heard of them. Nice :)

--=20

World wasn't remembered / World in my world / This hasn't been
remembered --Front 242, Waste

Public Key: http://sparked.zadzmo.org/

On Tue, May 09, 2006 at 10:45:33PM -0400, Aaron Bader wrote:
> 
> Hello all:
> 
> Been developing my application which uses a split architecture - the
> graphical interface does a standard fork/exec of a seperate daemon
> program, to keep the engine and ui code seperate.
> 
> I'm getting the OSC error 'Servname not supported for ai_socktype',
> number -8, immediately on startup of the backend - but ONLY if it is
> forked from the ui process. If I start the backend manually from the
> command line, everything is peachy.

That error will be coming from libc. The Servname refers to service names
from /etc/services I think, and I don't think there's code that uses that,
so it implies something is seriously messed up. the -8 is probably not a
good sign either. Have you tried running it through valgrind incase
theres a stack smash or pointer walk?
 
> World wasn't remembered / World in my world / This hasn't been
> remembered --Front 242, Waste

Heh, one of my favourite songs, though Motion is better.

- Steve

Hello all:

Been developing my application which uses a split architecture - the
graphical interface does a standard fork/exec of a seperate daemon
program, to keep the engine and ui code seperate.

I'm getting the OSC error 'Servname not supported for ai_socktype',
number -8, immediately on startup of the backend - but ONLY if it is
forked from the ui process. If I start the backend manually from the
command line, everything is peachy.

I first found this on 0.22, but it's still manifesting in version 0.23.
I'm not even certain where to start tracking this down after a quick
grep didn't even find that error string in the osc source code.

--=20

World wasn't remembered / World in my world / This hasn't been
remembered --Front 242, Waste

Public Key: http://sparked.zadzmo.org/