memory error when using geomip

C++ coding using Panda3D
#1

I’m trying to get a modified version of the second tutorial in the manual to work (in c++ of course) using the following code:

#include "pandaFramework.h"
#include "pandaSystem.h"
#include "geoMipTerrain.h"
 
#include <iostream>
 
PandaFramework framework;
 
int main(int argc, char *argv[]) {
    //open a new window framework
  framework.open_framework(argc, argv);
    //set the window title to My Panda3D Window
  framework.set_window_title("terrain tester");
    //open the window
  WindowFramework *window = framework.open_window();
 
  //here is room for your own code
  GeoMipTerrain *terr= new GeoMipTerrain("");
  std::cerr<<"loading ";
  terr->set_heightfield((string)"part_0.png");
  std::cerr<<"done \n";
  NodePath terrRoot = terr->get_root();
  std::cerr<<"1\n";
  terr->generate();
  std::cerr<<"2\n";
  terrRoot.reparent_to( window->get_render() );
  std::cerr<<"3\n";
  terrRoot.set_scale( 5,5,5);
  std::cerr<<"4\n";
  terrRoot.set_pos( 0,0,-10);
  std::cerr<<"5\n";
 
    //do the main loop, equal to run() in python
  framework.main_loop();
    //close the window framework
  framework.close_framework();
  return (0);
}

when I compile will g++ on fedora17/linux and run it I get the following output:

Known pipe types:
  glxGraphicsPipe
(all display modules loaded.)
:display:glxdisplay(warning): No suitable FBConfig contexts available; using XVisual only.
depth_bits=24 color_bits=24 alpha_bits=8 stencil_bits=8 back_buffers=1 force_hardware=1
loading done
1
base: dtool/src/dtoolbase/typeHandle.cxx:99: void TypeHandle::dec_memory_usage(TypeHandle::MemoryClass, int): Assertion `rnode->_memory_usage[memory_class] >= 0' failed.

I’ve tried moving the get_root after the generate. That does not change the output at all. If the geoMip code is replaced with the tutorials model loader everything works as expected.

Any help would be appreciated.

#2

Hi

I was able to compile and run your code with my own png file using Visual C++. Maybe you are not seeing the Config.prc located in your Panda’s etc folder

#3

As I change the options in the /etc/Config.prc file the window seems responsive before it dies. Specificity the starting window position changes as expected. I’ve lost the

base: dtool/src/dtoolbase/typeHandle.cxx:99: void TypeHandle::dec_memory_usage(TypeHandle::MemoryClass, int): Assertion `rnode->_memory_usage[memory_class] >= 0' failed.

error. It now reads

Known pipe types:
  glxGraphicsPipe
(all display modules loaded.)
:display:glxdisplay(warning): No suitable FBConfig contexts available; using XVisual only.
depth_bits=24 color_bits=24 alpha_bits=8 stencil_bits=8 back_buffers=1 force_hardware=1 
loading done 
1
make: *** [run] Aborted (core dumped)

All I’ve done is change windowOrigin in the /etc/Config.prc and reboot a couple of times.

#4

What is the size (width, height) of part_0.png?

#5

257 by 257 pixels, I’ve gotten it to compile and run on another computer running ubuntu. I’m now convinced it’s an error with how my fedora’s panda3d is compiled. How would you go about chasing down this kind of error?

EDIT: error line that vanished is back again.

#6

I haven’t much to offer but your code runs just fine on my linux box. Here’s my setup:
Ubuntu 10.10
g++ (Ubuntu/Linaro 4.4.4-14ubuntu5.1) 4.4.5
panda3d1.7_1.7.2~maverick_i386.deb

$g++ -c test.cpp -o test.o -fPIC -O2 -I/usr/include/python2.6/ -I/usr/include/panda3d/
$g++ test.o -o test -fPIC -L/usr/lib/panda3d/ -lp3framework -lpanda -lpandafx -lpandaexpress -lp3dtoolconfig -lp3dtool -lp3pystub -lp3direct
$./test
Known pipe types:
  glxGraphicsPipe
(all display modules loaded.)
loading done 
1
2
3
4
5
#7

After much frustration and trial and error I’ve switched from the cvs head to panda 1.8. The problem vanished after switching.

Before giving up on the cvs I traced to crash back to ~pvector() with the signal being thrown from a non panda3d library.