Re: [Om-synth] Om problems

[Ross Clement]

On Sun, 2006-01-29 at 18:10 -0500, Dave Robillard wrote:
> On Sun, 2006-29-01 at 02:09 +0000, Ross Clement wrote:
> > Well, having said that I have never managed to compile OM, I then manage
> > to compile it.
> 
> Forget to install the gtk development packages did we?  Ask for help
> next time. :)

It was a combination of my yum configuration being a bit funny. I added
a whole lot of repos before I read that not all repos are compatible,
and things have been a bit funny since then. I had to install gtkmm2.4
twice before it took effect. I also had to see PKG_CONFIG_PATH by hand.
However, there is still a problem. I have libflowcanvas
in /usr/local/lib, but om_gtk can't find it. I had to create symbolic
links in /usr/lib before om_gtk would start. I do have /usr/local/lib
entered in my /etc/ld.so.conf. Then I managed to make om-synth compile.
So I don't think it was a no-brainer.

> > However, as soon as I run it, I get a crash. Does anyone
> > know what might be going wrong?
> > 
> > I run the following 
> > 
> > jackd -d alsa
> > om
> > om_gtk
> > 
> > Om starts and seems happy until I start om_gtk, then I get:
> > 
> > om
> > [AlsaMidiDriver] Successfully opened ALSA sequencer.
> > [OSC] Started OSC server at osc.udp://clarissa:16180/
> > [PreProcessor] Launching thread.
> > [PostProcessor] Starting.
> > [JackAudioDriver] Activated Jack client.
> > [AlsaMidiDriver] Started realtime MIDI thread (SCHED_FIFO, priority 10)
> > [OSC] Registered client osc.udp://clarissa:17904/ (1 clients)
> > [JackAudioDriver] Enabling.
> > *** glibc detected *** om: double free or corruption (out): 0xb7d396d8
> > ***
> > ======= Backtrace: =========
> > /lib/libc.so.6[0x4e7424]
> > /lib/libc.so.6(__libc_free+0x77)[0x4e795f]
> > /lib/libc.so.6(freeaddrinfo+0x35)[0x53583e]
> > /usr/local/lib/liblo.so.0[0xe78359]
> > /usr/local/lib/liblo.so.0(lo_send_message_from+0x45)[0xe784a1]
> > /usr/local/lib/liblo.so.0(lo_send_message+0x21)[0xe784e1]
> > /usr/local/lib/liblo.so.0(lo_send_internal+0x77)[0xe7855f]
> > om[0x805756d]
> > om[0x805768f]
> > om[0x807243d]
> > /lib/libpthread.so.0[0x6ccb80]
> > /lib/libc.so.6(__clone+0x5e)[0x54edee]
> 
> Hm.. that' interesting.  lo_send_message_from is a new function in the
> latest version of liblo (which has a new maintainer) which I havn't
> tried yet.
> 

I have liblo-0.23.tar.gz. I built it from source. My distro is Fedora
Core 4, and liblo isn't (as far as I can see) available through any of
the yum repos I'm using.

> What version of liblo are you running and where did you get it? (ie
> build it yourself, or came with your distro, etc)
> 
> > Also, even if you configure om without ladspa so that it reports that
> > you won't have ladspa support, you still need ladspa.h to compile.
> 
> Not too crucial, since compiling Om without LADSPA support is something
> noone is going to want to do anyway. :)

I'm not sure what I need to do to install ladspa. At the moment I have
copied ladspa.h to /usr/include/. Should I do more? Could this be the
cause of my problem?

Cheers,

Ross-c