Re: [osg-users] .osg model converted to flt would not load in 3rd party software
Hi Per, Macher -- When I was last in that code, the flt exporter supported v15.7, v15.8, and v16.1, controllable with the "version=" option, with v16.1 being the default. All these versions support the use of Mesh Records. The export options are fully documented. See the ExportOptions.cpp file, or run: osgconv --format flt Or see the Support/KnowledgeBase page: http://www.openscenegraph.org/projects/osg/wiki/Support/KnowledgeBase/OpenFlight Rather than add an option to not use Mesh Records, the right way to do this would be to add support for OpenFlight v15.6. (If, indeed, Mesh Records are the problem.) -Paul Per Fahlberg wrote: Hi Macher, The model loads fine for me in Remo 3D 2.0.1 and the only issue is that we by default flip the dds textures, to correct this there is a checkbox in the texture palette, under the Edit menu, that can be unchecked. By the way if you would like to convert your model to openflight and avoid the mesh problem you can import the osg model directly in Remo 3D and will then get a model only constructed of faces, that will most likely work better in the other tools as well. I don't think there is an option to get the openflight plugin in osg to only produce faces instead of meshes, please correct me if I am wrong. Regards, Per Fahlberg Remograph On 07/29/2010 11:54 AM, Mach Bhai wrote: I have downloaded and installed it today. version 2.0.1 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] setUseVertexBufferObjects(true) has no effect
Hi Fred -- If you want some basic shape support combined with use of buffer objects, you should consider osgworks.googlecode.com, which contains convenience functions to create some basic shapes using osg::Geometry objects (which supports use of buffer objects). If you need shape metadata (sphere radius, box extents, etc) you can store that information as UserData on the Geometry objects. -Paul Fred Smith wrote: Hi, setUseVertexBufferObjects(true) has no effect for me with the code below: Code: osg::Geode* geode = new osg::Geode(); geode->addDrawable(new osg::ShapeDrawable(new osg::Cylinder(osg::Vec3(4.4f,0.0f,0.0f),1.0f,1.4f))); geode->getDrawable(0)->setUseVertexBufferObjects(true); When I run this code with an OpenGL API tracer, I can see display lists are still being used (glNewList to build the display list, then glCallList to use it). This test is easy to reproduce with the osgShaders sample, in GL2Scene.cpp, CreateModel() can be changed to reflect the above code. I guess I must be doing something wrong to make use of VBOs. Any idea? Cheers, Fred [/code] -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=30275#30275 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] VS 2010 and OSG v2.8.3?
Hi all -- Has anyone successfully built OSG v2.8.3 with VS 2010? Back when we did 2.8.3 release testing, I vaguely recall some people reporting (successful) build results using VS 2010, but just wanted to make sure this actually works. Thanks. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] What is setRenderBinDetails ................
Nitin Rangari wrote: Hi All, i have very simple question i want to know What is setRenderBinDetails when its use.. I recommend you search the archives for "render bin" or "setRenderBinDetails". -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Transform Feedback
Ethan Kerzner wrote: Hi, I guess my first question was too vague. Sorry about that. I am still trying to use transform feedback on an entire scene inside osg. I have created a CustomDrawable class with a drawImplementation that can perform transform feedback on anything drawn inside it. I am now trying to draw the entire scene inside my drawImplementation using a node visitor to traverse the graph and locate and draw all of the drawables with transform feedback enabled. I do not understand the parameters of "void draw(osg::RenderInfo &renderInfo) const" function. Can you please explain what information RenderInfo contains? Is the RenderInfo different for each drawable or global for the scene? Is it possible to determine the RenderInfo of a drawable? I think you're making this more complex than it needs to be. If you had a plain OpenGL program, you'd bind a transform feedback buffer before drawing, do the drawing, then process the buffer's contents after drawing. In OSG, the way to do pre- and post-draw operations is via Camera callbacks. You should be able to attach a pre-draw callback that binds a transform feedback buffer (and creates it if it doesn't already exist, of course), and a post-draw callback that unbinds the buffer and processes the contents. Any children of the camera will then "draw" into the transform feedback buffer. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] cross axes
Gianni Ambrosio wrote:
Looking around I found a topic that could help but I will ask some help about
that later.
Now, can anybody explain me please why I don't see the axes object in the
bottom left corner of the viewer using the following code? I can see the cessna
plane only in the viewer.
...
osg::Camera* createHUD()
{
osg::Camera* camera = new osg::Camera;
camera->setProjectionMatrix(osg::Matrix::ortho2D(0,1280,0,1024));
camera->setReferenceFrame(osg::Transform::ABSOLUTE_RF);
camera->setViewMatrix(osg::Matrix::identity());
camera->setClearMask(GL_DEPTH_BUFFER_BIT);
camera->setRenderOrder(osg::Camera::POST_RENDER);
camera->setAllowEventFocus(false);
osg::Node* axes = osgDB::readNodeFile("axes.osg");
camera->addChild(axes);
return camera;
}
The axes.osg model has a bounding volume that extends from approximately (0, 0,
0) to (1, 1, 1). With the transformation you've specified, the model will
subtend less that one pixel of screen space.
--
-Paul Martz Skew Matrix Software
http://www.skew-matrix.com/
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] [build] osg-2.8.3 and ffmpeg-0.6
Taurnil Oronar wrote:
/usr/include/libavutil/common.h:154: error: 'UINT64_C' was not declared in this
scope
make[2]: ***
[src/osgPlugins/ffmpeg/CMakeFiles/osgdb_ffmpeg.dir/FFmpegImageStream.o] Error 1
make[1]: *** [src/osgPlugins/ffmpeg/CMakeFiles/osgdb_ffmpeg.dir/all] Error 2
make: *** [all] Error 2
After searching the only thing I could really find ffmpeg related was a post
about changing ;
${FFMPEG_STDINT_INCLUDE_DIR}
to
(FFMPEG_STDINT_INCLUDE_DIR)
in the OSG FindFFmpeg.cmake which had no effect. Any pointers?
There were no changes to the ffmpeg plugin source for 2.8.3, so I'd think you'd
get the same build results as with 2.8.2, presuming you're using the same
version of the dependency as you use for 2.8.2. Is this the case? Is it broke
the same way for 2.8.2?
--
-Paul Martz Skew Matrix Software
http://www.skew-matrix.com/
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] How to know if a pre render pass has finished rendering
Prakhar Jain wrote: I wish to do a Render to Texture in which I have set the render target for the camera to be an osg image. Is there a way by which I can know if the prerender pass has finished or not ?. Attach a post-draw callback to the Camera. See the Camera header file. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Using point sprites for particles - blending
Jean-Sébastien Guay wrote: what I want. And I can't disable depth test or anything like that because my particles are supposed to interact with other objects, for example be contained in a shovel. Another idea: draw your point sprites after all the opaque geometry, with depth write disabled and depth test enabled. Then they will interact in z correctly with other objects in your scene. This will result in some visual artifacts due to incorrect ordering of the point sprites, of course, but perhaps your users will tolerate that. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Using point sprites for particles - blending
Jean-Sébastien Guay wrote: Does anyone have any tips about how to handle this case? Seems to me like it would be a pretty common thing to do. I've been toying around with integrating point sprites into a depth peeling algorithm. This presents some pretty interesting issues, as points don't rasterize like polygons -- no easy way to interpolate data over the surface of the point sprite. This makes sampling the depth map rather complex. I'm about ready to switch over to billboarded triangle pairs for this task. Assuming I do that, depth peeling should solve the sorting issue. There was an interesting blog on opengl.org headline news recently: "Fast and Accurate Single-Pass A-Buffer using OpenGL 4.0+" (you might need to scroll down a bit to find it). I don't know how well-suited it is for point sprites. The same sorting issue is present with fields of billboard grass. Try a search for grass rendering algorithms. They all seem to have their own back-to-front sorting solution. Might be good to mine them for ideas. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] how to animate a scene
Gianni Ambrosio wrote: Dear All, I have an object (osg::Node) in the scene and I would like to move it changing its position. I set the osg::Node as child of an osg::PositionAttitudeTransform to change its posiotion in time. Now the problem. I have a set of coordinates that represent the object position every 1/100 of a second. What should I implement to refresh the object position correctly? I mean, the synchronization is my basic problem. You can use AnimationPath and AnimationPathCallback for this task. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] finding adjacent vertices
Thomas Klemmer wrote: I have the vertexArray of a geometry and need to finde the curvatures of the mesh on each vertex. Thus I'm in need of all adjacent vertices of a vertex I'm trying to compute the curvature! The osgUtil::Simplifier has similar requirements so you might look at its code and see how it performs the same task. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Small test program gets segmentation fault when doing own viewer loop in Qt environment
LightModel::apply dereferences a pointer when it calls glLightModelfv to set the
ambient scene color. Check under a debugger to see if that's a NULL pointer. If
so, the way to resolve this would be to look at how it gets initialized in the
rest of OSG, then duplicate that in your viewer.
My guess is that LightModel is part of default OSG state, and you're not setting
it up the same way as OSG's viewers. But this is just a guess. You'd need to dig
in the code.
-Paul
Zachary Hilbun wrote:
Hi,
I'm writing a small test program (source below) that is doing its own viewer loop.
This program runs fine if I run it using viewerWindow->show. If I run it with
my own viewer loop it crashes the fist time it calls frame with a Segmentation
fault in osg::LightModel::apply (see stack trace below).
OSG v 2.8.1
Qt 4 (Eclipse integration version 1.6.1) environment
Linux using Eclipse in debug mode
I got ViewerQT from the Qt integration example AdapterWidget.cpp.
I got the viewer loop code from the Quick Start Guide. I noticed some examples had viewerWindow->realize in them so I tried it with and without that. It didn't make any difference.
Any ideas of what the problem is?
ViewerQT*viewerWindow = new ViewerQT;
viewerWindow->setCameraManipulator (new osgGA::TrackballManipulator);
viewerWindow->setSceneData
(osgDB::readNodeFile("/home/zacharyh/OpenSceneGraph-Data-2.8.0/cow.osg"));
// display the scene graph /
#define loop
#ifndef loop
viewerWindow->show ();
#else
viewerWindow->getCamera ()->setProjectionMatrixAsPerspective (40., 1., 1., 60.
* 36.);
// Create a matrix to specify a distance from the viewpoint.
osg::Matrix trans;
trans.makeTranslate (0., 0., 0.);
double angle (0.); // Rotation angle (in radians)
viewerWindow->realize ();// crashes with and without this
while ( ! viewerWindow->done ())
{ // do view loop
// Create the rotation matrix.
osg::Matrix rot;
rot.makeRotate (angle, osg::Vec3 (0., 0., 1.));
// angle += 0.01;
// Set the view matrix (the concatenation of the rotation and
// translation matrices).
viewerWindow->getCamera ()->setViewMatrix (rot * trans);
// Draw the next frame.
viewerWindow->frame ();
} // do view loop
#endif
a.connect (&a, SIGNAL (lastWindowClosed ()), &a, SLOT (quit ()));
return a.exec ();
gdb Debugger (6/9/10 10:15 AM) (Suspended)
Thread [1] (Suspended: Signal 'SIGSEGV' received. Description:
Segmentation fault.)
18 osg::LightModel::apply(osg::State&) const() 0x0027778a
17 osg::State::applyAttribute() 0x002d33b0
16 osg::State::applyAttributeList() 0x002d3613
15 osg::State::apply() 0x002ce9b1
14 osgUtil::RenderLeaf::render() 0x016f6b1e
13 osgUtil::RenderBin::drawImplementation() 0x016f0d13
12 osgUtil::RenderStage::drawImplementation() 0x016f93c4
11 osgUtil::RenderBin::draw() 0x016f09a0
10 osgUtil::RenderStage::drawInner() 0x016f9776
9 osgUtil::RenderStage::draw() 0x016fd51d
8 osgUtil::SceneView::draw() 0x017065af
7 osgViewer::Renderer::cull_draw() 0x00480cb0
6 osgViewer::Renderer::operator()() 0x0047e208
5 osg::GraphicsContext::runOperations() 0x0024b928
4 osgViewer::ViewerBase::renderingTraversals() 0x004bfb42
3 osgViewer::ViewerBase::frame() 0x004bd8dd
2 shotTrail()
/home/zacharyh/workspace/osgViewerQt/ShotTrail.cpp:86 0x0805454d
1 main() /home/zacharyh/workspace/osgViewerQt/main.cpp:40
0x0805edf5
Thank you!
Cheers,
Zachary
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=28752#28752
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
--
-Paul Martz Skew Matrix Software
http://www.skew-matrix.com/
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Books on OpenSceneGraph
Sanat Talmaki wrote: So I was wanting to know if there are some books out there that someone like me who is not very proficient with Computer Graphics (OpenGL in this case) can look at to understand the basic concepts. You'll definitely need to get a firm grasp of 3D graphics and OpenGL before you can use OSG wisely. The OSG Quick Start Guide has a "recommended reading" list near the front, which should be a good start. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] How to delete osg graphs properly?
Ku Krapox wrote:
@ Paul : I've read your QSG, and found lots of useful information, thanks
again! In particular I've learnt by heart the part on memory management. :P
But the problem is, I'm doing everything right... All my objects are now in
ref_ptr and I still have memory leaks.
So, you're clearing a ref_ptr, and stepping through it in the debugger, and
you're watching the reference count go to zero, but delete isn't being called?
Please post a reproducer code that shows this behavior.
I had an idea : might inheritance interfere with the osg::Referenced system? I
mean, if I create my own class derived from PAT, then put it in a ref_ptr, will
it be handled correctly?
Any class derived from Referenced will work with OSG's memory management system.
'cause I've found out the following behaviour :
class CustomPAT : public osg::PositionAttitudeTransform {...}
int main()
{
PositionAttitudeTransform* myPAT1 = new PositionAttitudeTransform();
delete myPAT1; // Compilation error -> normal
CustomPAT* myPAT2 = new CustomPAT();
delete myPAT2; // Compilation OK !!! -> shouldn't be
}
Looks like you failed to declare your destructor protected.
--
-Paul Martz Skew Matrix Software
http://www.skew-matrix.com/
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] How to delete osg graphs properly?
Ku Krapox wrote: Hi all, It's me again, with another practical problem : how to delete properly my osg objects when they contain lots of stuff? Basically I have a PAT with lots of groups and nodes inside it. At some point I need to delete it, then reinstanciate it right away. But I found out that doing this produces one big memory leak... My object is in an osg::ref_ptr, so I tried to .release() it first, it didn't quite change anything... I also tried to mess with the destructor, deleting manually some of the groups and nodes, with no more success... I have to mention that I use several NodeCallbacks inside my PAT, which may be active when I delete it. I don't know how to deal with them, or if I have to... So my question is : what is the proper method to delete an osg scene or part of scene (in order to avoid memory leaks)? Do I have to access and delete every leaf of my graph? Thanks for any advice! :) I'd advise you to read the OSG Quick Start Guide. http://stores.lulu.com/store.php?fGroupID=2076 -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] A challenge with diamond-shaped graphs...
Frank Sullivan wrote: But I want to make the wheels look as though they are turning as well. I accomplish this with a simple texture matrix. I simply find the node that represents each wheel, and I add a texture matrix to each of them that will slide the s/t coordinates of the wheels' textures so as to give the illusion that the wheels are turning. This seems to work well also. Well enough for my purposes, anyway. The problem is when I have two cars in the scene, and the graph looks like it does above, I can't turn the wheels on one car without turning them on the other as well. Is the best solution to this to do a deep copy of the car's sub-graph rather than reusing the same sub-graph with two different matrix transforms? If the texture matrix is in the shared part of the subgraph, then the same texture matrix will be used for both instances -- that is sort of the definition of sharing. You have correctly observed that making a hard copy will eliminate both the sharing and the issue you've encountered. Another solution would be to use shaders to render the wheels. Each matrix transform would specify a uniform to transform the texture that's applied to the wheel texture coordinates. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] OBJ export of Shapes, any progress or workarounds?
Dženan Zukić wrote: Can anyone provide pointers? You could try posting to osgWorks google group, or look at osgBullet (which uses osgWorks)... -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] [osgPlugins] Exporting OpenFlight from OSG and DDS textures
Luca Colombari wrote: Hi Paul, Features like: - osgVolume::Volume - osgVolume::Image Layer - are not recognized. Right, currently not recognized or used by the FLT plugin. So one way could be to try mapping osgVolume object to FLT records by knowing the OpenFlight format and how write their information into the .flt file. Right. You would need to figure out how to support these OSG features when creating the FLT file, using OpenFlight's (relatively speaking) limited features. By the way: I've never seen a FLT file used to store volumetric data. But, if I recall correctly, FLT does support 3D textures. If my intention is to generate a VOLUME of dicom images seen as overlapping dicom images I'm wondering if it is possible to obtain with OpenSceneGraph an osg file that uses only osg features (for example osg::Group instead of osgVolume::Volume and so on). That's an interesting approach. You wouldn't need to modify the FLT plugin if you could do that. I'm sure the osgVolume nodekit has many useful features, but honestly I don't think it's required to create an OSG scene graph to render volumetric data. In fact, I remember seeing an OSG app running at SIGGRAPH 2004 that was displaying volumetric data. This was long before the osgVolume nodekit existed. So, certainly this is possible. I'm not sure exactly what your scene graph needs to display. But the general approach for rendering volumetric data is to use a 3D texture with several screen-oriented triangle pairs. You'd orient those to the screen with a shader. This can all be done with core OSG and no need for the custom node types present in the osgVolume nodekit. It should export to FLT. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Does OSG have a vertex-reducer?
Alberto Luaces wrote:
"Sasha Koruga" writes:
Hey,
I was wondering if OSG has the built-in functionality to reduce the vertex
count of a mesh to a programmer-specified amount. For example, if I have a
125,000 vertex model of an airplane, can I call some function like
reduce(airplane, 10), that will reduce the vertex count of that mesh to
10 while doing its best to maintain the mesh's general shape?
Hi Sasha,
take a look at the osgsimplifier example.
In addition to the osgUtil::Simplifier in OSG, you might want to look at a
couple external projects.
osgWorks (osgworks.googlecode.com) has two additional vertex reduction
techniques. One ("Decimator") is based on the Simplifier and attempts to
compensate for the Simplifier's inability to remove vertices that fall on an
edge. The other ("Reducer") employs a different algorithm entirely. It produces
good results and is very fast, but currently has limitations regarding which
primitive types it works with. osgWorks comes with a script that lets you run
these reduction techniques on various models and compare the results
side-by-side visually.
osgBullet (osgbullet.googlecode.com) also has a vertex reduction algorithm
called "VertexAgg". It aggregates nearby vertices together, then uses Bullet to
create a convex hull around the vertices. The result is very approximate, and
doesn't contain normals or tex coords, but it's good enough for creating physics
collision shapes for complex objects (which is what it was intended for).
--
-Paul Martz Skew Matrix Software
http://www.skew-matrix.com/
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Limitation of NodeMask
Andrew Cunningham wrote: Hi, May I am not seeing something here, but using the NodeMask seems to be not able to handle 'multiple' states. For example, I am using a mask ,0x01 to control the 'visibility' of the nodes. This works great. I have another mask to control 'pickable' state of the node, say 0x02 when used with IntersectionVisitor. Obviously a node that is not visible is not pickable. That is (pickable && !visible) should be culled during a IntersectionVisitor traversal Therein lies the problem that I see. A simple "bitwise AND" is performed between the traversal mask and the node mask and the result compared to 0. There is no mask that can be passed to IntersectionVisitor that will cull (pickable && !visible). As long as the pickable bit is on, the node will always be traversed. Maybe I'm stating the obvious but, you appear to be overriding the visible bit to mean both visible and pickable. But you already have a pickable bit. So there's no reason to override the visible bit for pickability. Use the visible bit for visibility, and the pickable bit for pickability, and it should work fine. Or am I missing something? -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Collision Detection
John Galt wrote: Hi Paul, I only plan on detecting intersections between a PATNode and other Nodes. I do not intend to write code for a "reaction to a collision", I want the intersection to happen and only want to know which objects are intersecting. Do you suggest I still use osgbullet for that? If you're just doing intersection testing, then you should look at using osgUtil's facilities for that. You could transform your pyramid using the local to world matrix and use the result to configure a polytope intersector. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Collision Detection
John Galt wrote: Hi, I am trying to implement a code that detects collision between a PAT Node with the geometry of a pyramid and multiple other fixed Nodes that are loaded with ac3d models. How do you determine which node (if any) falls within the geometry of the PAT Node of the pyramid? I would use a library designed for collision detection (and rigid body dynamics) such as Bullet. I maintain a toolkit for Bullet and OSG apps: osgbullet.googlecode.com. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] [osgPlugins] Exporting OpenFlight from OSG and DDS textures
Luca Colombari wrote: Attachments: http://forum.openscenegraph.org//files/osg_output1_127.osg The FLT plugin doesn't have a dependency on the osgVolume nodekit. You'll have to figure out how to map osgVolume objects to FLT records. Good luck with that. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] [osgPlugins] Exporting OpenFlight from OSG and DDS textures
Luca Colombari wrote: So I supposed that the FLT_WRITER code of flt_plugin could be rewritten by me to allow a correct conversion from OSG or IVE to FLT format. OSG supports many more features than FLT. The FLT importer doesn't exercise all OSG features. Similarly, the exporter doesn't (can't) support converting all OSG features to FLT. The FLT exporter was designed primarily to support exporting the same subset of OSG features used by the FLT importer. Thus you should be able to load a FLT file into OSG, then export it back to FLT, and get something close to the original FLT file. It was not a goal of the exporter (nor is it possible) to support exporting arbitrary OSG scene graphs. Please post here with a list of the OSG features in your scene graph that are not being exported to FLT, and what OpenFlight record types you plan to use to support these features in the exporter. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] [osgPlugins] Exporting OpenFlight from OSG and DDS textures
Luca Colombari wrote: Hi, Thanks for the answer... Ok, Currently I don't know very much the logic about OpenSceneGraph and I'm studing it by going into the code. So now my problem is to understand how can I generate an OpenFlight file by handling also all the related images. I need to convert DICOM files into OpenFlight model to use another tool that manages .FLT files. With OpenFlight Plugin code it is possible to convert (command 'osgconv') directly DICOM files /or intermediate generated .osg file (with external automatic generated DDS texture file) into a new .FLT file where is loaded the volume (nodes): now if I use the command osgviewer to visualize .flt file I am not able to see nothing, only an empty volume, because no textures are supported. I have downloaded an example of .flt file together with .rgb and rgb.attr files and in this case I am able to visualize it with osgviewer. So by using RGB plugin I have generate .rgb files from images read from DICOM files but...how could I link the converted OpenFlight file with the external RGB textures? Sorry but I'm confused and I'm trying to understand how implement code with the use of OpenFlight plugin to generate an FLT file with handled textures and visualize a 3D image with osgviewer or other tools that support OpenFlight format. Well, there's quite a few things that could be going wrong here. If you can view the DICOM file with DDS textures, then it sounds like your DDS plugin works. So I guess the question is: Why is it that when you view a FLT file, the DDS textures don't show up? First thing to check would be: Is it loading the DDS files? To find out, set OSG_NOTIFY_LEVEL=DEBUG and then run osgviewer on your FLT file. Look through the output to see if it is finding and loading your DDS files correctly. If it is loading them correctly, then next thing is to compare your first OSG file with your second OSG file to see what's wrong. For example, in your first post, you said you used osgvolume to create an OSG file called output.osg. Then you convert it to FLT and can't view it. So the question is why? The way to find out: convert that FLT file back to OSG, call it "output2.osg". Then open both OSG files in a text editor and compare them side by side and look for differences to explain why the first file works and the second doesn't. It could be the case that your first OSG file, output.osg, uses features of OSG that aren't supported in FLT. I don't have enough information to say for sure. You just need to investigate and debug. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] [osgPlugins] Exporting OpenFlight from OSG and DDS textures
Luca Colombari wrote: Hi, I'm new with OpeneSceneGraph 2.9.6: i'm analysing plugin code to understand if there is the possibility to generate an "OpenFlight (.FLT) file together with textures" by starting directly from a list of DICOM files or by starting from the .OSG and .DDS files generated with the command: < osgvolume --images *.dcm -o OutputFile.osg >. At the moment the .flt file obtained by the command < osgconv Output.osg Output.flt > contains only the volume and not textures model. I need some suggestion to understand how generate also textures/image information to be able to visualize DICOM files from a generated OpenFlight format. Thanks in advance Are you saying that the .flt file doesn't store the textures, instead they exist as separate files? If so, that's correct: OpenFlight doesn't support storing texture image date within the .flt file. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] dimension oh a model
Alberto Luaces wrote: how could I have missed this! Thank you for the tips :) OSG is big. :-) -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] dimension oh a model
Jean-Sébastien Guay wrote: Hi Daniele, Alberto, Assuming that no transformation nodes exist in the graph describing your node, you can write a node visitor that joins all the bounding boxes of every geode. Else you'll need to modify every bounding box by its parent transforms in order to join them. Actually you don't need to write this, it's so common that it exists in OSG already: osg::ComputeBoundsVisitor. And it does the right thing with transforms. osg::ComputeBoundsVisitor cbbv; node->accept(cbbv); osg::BoundingBox bb = cbbv.getBoundingBox(); osg::Vec3 size = bb._max - bb._min; // Then size contains the size in X, Y and Z. For purposes of analyzing models, it's often useful to have access to this information from the command line. The osgWorks project (osgworks.googlecode.com) contains the osgwbvv application to display bounding volume information, either as a sphere or (using ComputeBoundsVisitor) as a box. >osgwbvv cow.osg Sphere: Center 0.776125 -0.43866 0 Radius 6.35558 >osgwbvv --box cow.osg Box: Center 0.776125 -0.43866 0 Radius 6.35558 Extents 10.4439 3.40282 6.39676 The bounding volume (sphere or box) is also displayed graphically as line geometry around the model. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] depth test and Deferred Shading
Cedric Pinson wrote: I am doing a deferred shading and i have a problem to use the depth buffer as i would like. Here the order of drawing: consider using a scene A 1 Draw only in the zbuffer 2 setup a camera with mtr and draw normal and color into two render target. Dont touch to the depth buffer - for this step i enable the zbuffer in read only and i dont attach a depth buffer to the camera, i want the z to be tested from the viewport done on pass 1. It's like the zbuffer is lost between 1 and 2 You said you DON'T attach the depth buffer to the camera in step 2? I believe you need to attach it for the z test. Leave the depth test enabled with the same depth function, but set the depth write mask to false. Then it will test against the depth buffer. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Two Pass Rendering?
Bruce Wheaton wrote: What's the preferred way to do this? I need to separate a shader into an H and V pass. I suppose I could double the scene in the graph, but since the first pass needs to be to a texture that wouldn't do it, would it. So I suppose I need another camera with an FBO using the same scene - render to that with my H shader, then use that plus all the new FBO as the source for the final pass? That seems odd to me, since a lot of the objects might be partially obscured in the first pass. Do I need to use an interim texture/camera for each object? Is there an example doing this? I understand it's a favorite way to do glows. Most articles on glow describe rendering the scene using MRT to create a rendered scene image and a separate glow map. Then blur the glow map first in x then in y. Finally, composite the two textures together to create the final image. The tool that (unmodified) OSG offers for this is the Camera. You would use one camera to create the rendered scene texture and the glow texture. Then use a second camera to blur the glow texture in x. Use a third camera to blur it in y. Finally, use a fourth camera to combine the rendered scene texture with the blurred glow map. All these Cameras adds a lot of overhead on the CPU side, but the only way to reduce that would be to derive custom RenderStage classes, something most OSG developers choose not to do. osgPPU offers an alternative and might be worth considering. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Instanced Rendering Example in OSG
Nicholas Vidovich wrote: Hi Paul -- I didn't mean to imply that the technique wasn't viable -- due to my errors and lack of experience I was unable to correctly implement instanced rendering by passing the data in a floating point texture. OK. Yeah, it's definitely possible. Sounds like you just need to debug it. I'd suggest starting with just a couple instances, and once you have that working to your satisfaction, then move up to larger numbers. Trying to debug position issues with 100k instances is well-nigh impossible; start with a smaller test case and debug that first. Do you have an example you would be willing to post? I have an example, but it's proprietary. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Instanced Rendering Example in OSG
Nicholas Vidovich wrote: Thanks, Christian -- I have tried modifying the osgdrawinstanced example as well. Do you have an example that sets the positions of the instances? I tried passing positions in as a uniform sampler2D texture, but that didn't work out. I have done instanced rendering where position and orientation are passed as elements of a floating point texture. This works fine. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] [osgPPU] Can't compile svn version of osgPPU against osg 2.8.3
Hi Art -- I'm a little confused. Maybe we're not communicating clearly. Art Tevs wrote: As for osgPPU it is not only the getOrCreateGLBufferObject() which don't work, but also much more things. I tried yesterday to bring combine current trunk version of osgPPU with osg 2.8 and that didn't really worked out without introducing additional bugs. OK. I understand you choose not to try to make your svn trunk head compatible with older versions of OSG, and instead you are drawing a compatibility line going forward. That's fine, and I understand this. So, I decided to submit several patches to the 0.4 release but do not include new features, only patches. This will be a 0.4.1 version, which will be compatible with up-to osg 2.8.3. This part I do not understand. I just downloaded the osgPPU 0.4 tar.gz, and it builds fine with OSG 2.8.3. Please tell me what you think you need to change in osgPPU 0.4 to make it compatible with OSG 2.8.3. Thanks. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] OT: Fast software rendering
Interesting. http://www.wired.com/gadgetlab/2010/04/unlimited-detail-3-d-graphics/?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+wired%2Findex+%28Wired%3A+Index+3+%28Top+Stories+2%29%29&utm_content=Google+Feedfetcher -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] [osgPPU] Can't compile svn version of osgPPU against osg 2.8.3
Art Tevs wrote: However if some time ago there was a split into another branch it might became painful to have an external NodeKit compatible with all official osg branches that exists. I mean 2.8.3 is based on 2.8, so between 2.8 and 2.9 there was an API change, this is ok. However both versions 2.8.3 and 2.9.x were maintained in parallel. So for NodeKit developers this would mean we have to take care also on osg branches in our trunks, which might became painful. Therefor I just asked, which kind of solutions there might exists. I, for one, would not expect any version of osgPPU to be compatible with OSG 2.8.3 other than the same version of osgPPU that was compatible with 2.8.2. I think users should be able to use the same version of osgPPU and simply upgrade OSG from 2.8.2 to 2.8.3 to get the new plugins in 2.8.3. (But they will probably need to rebuild osgPPU because 2.8.3 is not binary compatible with 2.8.2.) But, ok, never mind, I will try to come up now with a solution to have the best of both worlds, new features and compatible API to the official stable release. If you want to maintain compatibility with 2.8.x on osgPPU svn trunk, then that is certainly possible, and a worthy goal, but (in my opinion) it is entirely optional and not required for you to do so, because you already have a tagged osgPPU release that is compatible with OSG 2.8.x. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] [osgPPU] Can't compile svn version of osgPPU against osg 2.8.3
Allen Saucier wrote: http://forum.openscenegraph.org//files/screenshot2_903.png The statement in this image implies that svn trunk of osgPPU is compatible with the latest stable release of OSG. This is not correct. It looks like the osgPPU 0.4 tag should be compatible with OSG 2.8.x. Maybe Art can comment on version compatibility? (A compatibility matrix might be useful, sort of like what I have done with osgBullet: http://code.google.com/p/osgbullet/wiki/CompatibilityMatrix ) -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] [osgPPU] Can't compile svn version of osgPPU against osg 2.8.3
Art Tevs wrote: Ok, then it explains why it does not compile. Yes, svn trunk of osgPPU tracks the svn trunk of OSG. Hmm, I wonder how to handle different osg branches in future. I mean if again something like this happens, then I had either to create additional branch in osgPPU repository to track with the according osg branch (which I didn't really like) or to detect the version and write preprocessor or cmake workarounds for that (which is also not a good solution). I'm not sure what new problem OSG 2.8.3 has created for osgPPU that didn't already exist with the OSG 2.8.2 release. Can you explain the problem? The same version of osgPPU that works with 2.8.2 should also work with 2.8.3. osgPPU trunk is broke with OSG 2.8.3, but it was also broke with OSG 2.8.2. Suppose the 2.8.3 release never happened, and Allen was trying to build svn trunk of osgPPU with the OSG 2.8.2 release. That would also not work, right? What would you tell him to do in that case? The same answer should apply to 2.8.3. For this release I will go for a preprocessor solution, I think. My external projects osgWorks and osgBullet are compatible with OSG 2.6.1 and later. osgWorks contains version-specific code in only two places (to handle the OSG API changes in Registry and Traits that occurred in OSG 2.6.0 and 2.8.0 respectively). And osgBullet contains no version-specific code at all; all osgBullet code is 100% compatible with OSG 2.6.1 and later. So, my use of the C preprocessor for OSG compatibility is very infrequent and is not a problem in itself. However, testing on different versions can be a burden to external project managers, and that's a decision you have to make. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] About osg::OcclusionQueryNode
Jean-Sébastien Guay wrote: The osgocclusionquery example has a visitor that will only insert OQNs when the vertex count of children exceeds a threshold (5000 by default) but I wanted the control to place it where I want, because if I have a graph with 2 children, one with 5010 vertices and one with 4000, then with the visitor I would get an OQN above the first child but not the second, whereas I want an OQN above both (above their parent). Yes, that visitor is pretty crude. I have not found an easy way to place OQNs programmatically in an arbitrary scene graph and get optimal performance results. The best approach, as you are doing, is to have a priori knowledge of the model, and educated modelers that know the most effective locations for placing OQNs. (The problem is identical to placing OccluderNode, OSG's other occlusion mechanism.) Good solution, using a configuration file. Do I assume correctly that you have a configuration file per model? -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] [osgPPU] Can't compile svn version of osgPPU against osg 2.8.3
Allen Saucier wrote: [al...@nlosdemobox:]$ make Scanning dependencies of target osgPPU [ 2%] Building CXX object src/osgPPU/CMakeFiles/osgPPU.dir/Unit.o In file included from /home/allen/projects/osgPPU/src/osgPPU/Unit.cpp:17: /home/allen/projects/osgPPU/include/osgPPU/Unit.h: In member function ‘void osgPPU::Unit::popFrameBufferObject(osg::State&)’: /home/allen/projects/osgPPU/include/osgPPU/Unit.h:311: error: ‘class osg::FBOExtensions’ has no member named ‘glBindFramebuffer’ Names of the member functions in the FBOExtensions class changed on trunk after the 2.8.2 release. OSG 2.8.3 was released off the 2.8 branch and is 99.9% API compatible with OSG 2.8.2. The FBOExtensions name changes are not in 2.8.3. It looks to me like trunk of osgPPU is tracking trunk of OSG (which I would expect). Art and I discussed osgPPU and the 2.8.3 release previously. If you check the archives, I believe you'll find a post from me stating that the same version of osgPPU that worked with 2.8.2 should also work with 2.8.3. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] About osg::OcclusionQueryNode
Hi J-S -- Thanks for taking the new OcclusionQueryNode for a test drive.
Jean-Sébastien Guay wrote:
First of all, running the osgocclusionquery example, I get erratic
results (though this is the one from the svn trunk, not 2.8.3, but my
understanding is that the version in 2.8.3 is basically up to date with
the trunk). For example, if I turn the camera so I see the back of the
box straight-on (perpendicular to the view vector) I get the same stats
as if I view it from the front (i.e. seeing all the lines). If I just
turn the camera slightly so the back is not exactly perpendicular, then
the GPU time goes down considerably (as I'd have expected even when
looking at the back). Is this normal?
Here's the problem: What to do when the eye is so close to the query geometry
that the eye is actually inside the query geometry itself? It would be
inefficient to actually do a query in this case, and makes more sense to just
assume that the geometry is visible.
One of the recent changes (in 2.8.3 and svn trunk) fixes some previous problems
by changing how to test for this condition. The current code checks to see if
the near plane penetrates the bounding sphere of the query geometry. If so, the
geometry is considered visible. Unfortunately, in the simple osgocclusionquery
example, the autocomputation of near/far causes the near plane to be inside the
query geometry bound almost all the time. When the query geometry is part of a
larger scene and surrounded by a larger view volume, with the eye at some
distance, the current algorithm performs correctly.
So I think the osgocclusionquery test case is contrived, but if you really need
it to work, then one workaround would be to change the example to disable
autocomputation of near/far. But to "fix" OcclusionQueryNode so that it handles
this case "better", we'd need to do a full query in this situation, and I guess
it's debatable as to whether this would actually be an improvement or not.
The second thing I wonder is that looking at the code, I don't see
anything that lets the OcclusionQueryNode work if the required GL
extensions are not present. I don't see any explicit test for the
required extensions. Perhaps I just missed it, could you point it out to
me if I did? Or is it up to the app to not use OcclusionQueryNodes if
they're not supported on the hardware it's running on?
In particular, in RetrieveQueriesCallback, I see this line of code:
ext->glGetQueryObjectiv( tr->_id, GL_QUERY_RESULT, &(tr->_numPixels) );
glGetQueryObjectiv is defined as this:
void Drawable::Extensions::glGetQueryObjectiv(GLuint id, GLenum pname,
GLint *params) const
{
if (_gl_get_query_objectiv_arb)
_gl_get_query_objectiv_arb(id, pname, params);
else
osg::notify(osg::WARN) << "Error: glGetQueryObjectiv not supported
by OpenGL driver" << std::endl;
}
which means that if it's not supported, then the call will return,
printing a warning, but not modifying tr->_numPixels. But the default
constructor of TestResult sets this to 0. So since there's no test to
see if glGetQueryObjectiv is supported or not before the call, that
means that the geometry under the OQN will always be culled as if it had
0 visible pixels if it's not supported? Surely that's not the intended
result.
Perhaps the TestResult default constructor should set _numPixels to -1,
and then if it stays -1 even after the call to glGetQueryObjectiv, the
OQN should be disabled? Or just explicitly test if glGetQueryObjectiv is
supported...
I admit I didn't design this code to work on older hardware, nor have I tested
it on hardware without this capability. Please feel free to make any necessary
changes to that it at least behaves like a regular Group node when the feature
isn't present.
--
-Paul Martz Skew Matrix Software
http://www.skew-matrix.com/
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] [build] Install problem with OSG 2.8.3
Nick Schultz wrote: Not sure why the path is being set as : "E:/OpenSceneGraph-dev/bin/Release/../../bin/osg66-osg.dll" when it is actually located at : "E:/OpenSceneGraph-dev/bin/Release/osg66-osg.dll" why is this happening in the first place? Are you sure you're installing 2.8.3? The so version is 65 for 2.8.3. 66 is the so version number for 2.9.8. Did you try to do an svn switch that went awry? I suggest you start from a clean source tree and delete your CMake cache. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] [build] Install problem with OSG 2.8.3
Tony Vasile wrote: When I try to install OSG 2.8.3 on a linux box I get the following message when it attempts the install part of the process: - Install configuration: "" -- Up-to-date: /project/uws/Development/avasile/lib/pkgconfig/openscenegraph.pc -- Up-to-date: /project/uws/Development/avasile/lib/pkgconfig/openthreads.pc -- Up-to-date: /project/uws/Development/avasile/lib/libOpenThreads.so.2.4.0 -- Installing: /project/uws/Development/avasile/lib/libOpenThreads.so.11 CMake Error at src/OpenThreads/pthreads/cmake_install.cmake:41 (FILE): file INSTALL cannot duplicate symlink "/home/tony/src/OpenSceneGraph-2.8.3/lib/libOpenThreads.so.11" at "/project/uws/Development/avasile/lib/libOpenThreads.so.11". Call Stack (most recent call first): src/OpenThreads/cmake_install.cmake:41 (INCLUDE) src/cmake_install.cmake:37 (INCLUDE) cmake_install.cmake:41 (INCLUDE) Any ideas? I'd suggest you try to do the exact same kind of build and install with 2.8.2 or svn trunk and see if the issue is specific to 2.8.3 or not. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Absolute position of an object
daniele argiolas wrote: Hi, I create a simple scene with osg where I load an osg model. Then I try to position it with PositionAttitudeTransform and setPosition but it remains always in the same position (center of the scene). How can I position objects in absolute way? The viewer always looks directly at your model. So unless you've loaded something else so show your model's relative position, such as axes.osg, then it will appear to be in the same place. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Fwd: Re: Color array optimization
Vincent Bourdier wrote: Just one last question ... Why does BIND_PER_PRIMITIVE is wrong ? what about BIND_PER_PRIMITIVESET ? It all depends on what performs well in OpenGL. Look at the implementations of these and examine what OpenGL commands they issue. There's no way to implement BIND_PER_PRIMITIVE in OpenGL without using glBegin/glEnd, which is widely recognized as OpenGL's oldest and least efficient mechanism for specifying geometric data. BIND_PER_PRIMITIVESET should perform well because there's a drawing command per PrimitiveSet. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] HUD camera and depth
Don Leich wrote: Hi all, I needed to change an HUD camera so that it would write into the depth buffer. To get it to work I had to do something that I don't understand. I thought that a depth range of (0.,0.) and function of ALWAYS should cause all subsequent rendering to write at the minimum depth value. Doing so would render lines OK but all filled polys and text went missing. I found that I had to open up the range a bit in order to see these too. Hi Don -- I've found that a more direct way to write to the min z value is to use a vertex shader that outputs gl_Position with z = -1 and w = 1. Regardless, there's no reason why your technique shouldn't work in plain OpenGL: http://www.opengl.org/sdk/docs/man/xhtml/glDepthRange.xml Not sure why this isn't working in OSG. First thing to do would be set a breakpoint in your Drawable's draw() to see whether or not drawing commands are being sent to OpenGL. If your breakpoint doesn't get hit, then you've narrowed it down to an OSG issue. If the OpenGL drawing commands are getting issued, then something about the OpenGL state machine is being misconfigured. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] OSGExplore
Biron, Patrick D. (LARC-E402)[ANALYTICAL MECHANICS ASSOCIATES INC] wrote: Hello, I haven’t heard any input from anyone interested in checking out OSGExplore. I realized that I have not provided an executable of the application, so I added a windows executable (and the necessary DLL’s for the application to run) to the OSGExplore sourceforge site (rather than leaving the user to compile the application). I recommend opening a small file, at first (such as fountain.osg) to see how the application operates (playing with osg::Material, specifically, with that example – focusing on color properties, will help with understanding the application). I also included an example document to help show how OSGExplore works. Thank you, Patrick *From:* Biron, Patrick D. (LARC-E402)[ANALYTICAL MECHANICS ASSOCIATES INC] *Sent:* Tuesday, February 02, 2010 10:35 AM *To:* osg-users@lists.openscenegraph.org *Subject:* RE: OSGExplore Also, the application is hosted on sourceforge. *From:* osg-users-boun...@lists.openscenegraph.org [mailto:osg-users-boun...@lists.openscenegraph.org] *On Behalf Of *Biron, Patrick D. (LARC-E402)[ANALYTICAL MECHANICS ASSOCIATES INC] *Sent:* Tuesday, February 02, 2010 10:32 AM *To:* osg-users@lists.openscenegraph.org *Subject:* [osg-users] OSGExplore Hello All, I’m a graduate student looking for, for any who are interested, input on an application I’ve developed - if it’s useful, things that it does poorly, decently…and, importantly: inputs on the concepts behind the application, etc. (warning – I am a tad rusty on C++ and I am new to QT, among other things…so, though the application mostly works, the development behind it might not be as efficient/proper as it should be). The application is called OSGExplore (based/borrowed on OSGEdit…and borrowing from basic examples from OSG and QT). Essentially, OSGExplore uses data visualization techniques to help users find osg types (such as osg::Material) on a visualization screen, rather than with a tree view (though one is provided) so that it can be found and edited quickly. I’ve developed the application using visual studio, QT 4.5.2, Boost 1.40.0 (cmake 2), osg 2.8.0, and qt property browser 2.5.1. Environment variables: $BOOST_DIR (location of boost – example “C:\Boost\boost-1.40.0.cmake2”) $QTDIR (location of QT – example “C:\QT\4.5.2”) $QTPropertyBrowser (Location of the qt property browser – example “C:\QT\qtpropertybrowser-2.5_1-opensource”) $OSG_ROOT (location of OSG – example “C:\Program Files\OpenSceneGraph”) As for how the application is used, here is an example on how I hope it would help/work. Using the fountain.osg example, there are four osg::Materials…finding each material might take some work using a tree view, given that there are multiple copies of each osg::Material being used. But, using the filter of osg::Explore, the user can choose osg::Material as the filter class type, then choose _specularFront as the filter property. The four osg::Material nodes will now be colored with the color used for “specular front.” From there, we can “attract” osg::Materials to each other on the visualization screen (by clicking the attract check box), and change their shape to “Big”, so that they’re easier to spot on our visualization screen. From there, we could click on each node until we find our proper osg::Material (or just look at the colors of each material “node”), and then (from the property view on the left hand side of the application) change the material properties. Warning – if you have a large number of nodes of a certain type, choosing to turn on “class links” may slow down the application drastically (this is also the case for parent/child links in general). Here are some application commands: “R”: randomly places the visualization nodes “-“: Zooms out “+”: Zooms in Mouse Wheel: Zoom out/in Clicking nodes drags them “Esc”: Lose focus of the current node “Space”: Pause node movement Some other options to play with: Node attraction and repulsion strength (the forces between each node globally and the forces between filtered nodes to global nodes). Just to let you know you are not being ignored. :-) I have kept this post from you in my inbox with the intent of trying it. I haven't had time due to other obligations. Hopefully I'll get to it in the near future. It sounds like an interesting tool. Keep up the good work. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 released!
Robert Osfield wrote: I've update the SpreadingTheNews wiki page with your entries for opengl.org and khronos.org having been done, there are lots more to do. Community now is your time to dive in an post the news to your favourite sites, just add (done) to each site you pass the news on to: http://www.openscenegraph.org/projects/osg/wiki/Community/Tasks/SpreadingTheNews Sheesh, forgot about that. Yes, I'll echo Robert's plea for community help in pushing this out, and thanks to everyone in advance. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Add a color node in the scenegraph?
Vincent Bourdier wrote: Hello, You should try just adding a material on each PAT. If I'm right, material will overpass the color of the geode... If anyone can confirm :-[ That will work if lighting is on. If lighting is off, set a BlendColor at each PAT and then configure blending on the Drawable to reference the blend color. Le 07/04/2010 11:33, Benjamin GODIN a écrit : Hello, I have a scene with one object but represented many times. My graph is something like that : [Image: http://img576.imageshack.us/img576/1522/osg.png ] I want to give different colors to the different objects. The method I've found is setColor on a ShapeDrawable but that means I need to create many instances of this object. Is there any way I can add the color in the scene graph (as a node)? Thanks. :)[/img] -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 released!
Robert Osfield wrote: Hi Paul, Congratulations on this release - the most advanced and stable OSG release yet :-) I understand how much effort goes into corralling everyone, reviewing and merging fixes, doing testing - kudos. It's a lot of work. I've been tracking my hours on this and it's like two full 40 hour weeks of effort. I should be thanking you, though, as I know you put in 10 times this much. So, thanks, and keep up the good work. On Mon, Apr 5, 2010 at 8:39 PM, Paul Martz wrote: I don't have write access to the downloads page, so I've made a source ZIP available here, until it can be moved to the OSG server: http://www.skew-matrix.com/osg283/OpenSceneGraph-2.8.3.zip Robert, if you could move that, I'd appreciate it. I've now done this, and update the downloads page. I've checked the various links and they all seem to be OK. http://www.openscenegraph.org/projects/osg/wiki/Downloads Excellent, thanks. Looks great! I'll push the press release out to opengl.org and khronos.org ASAP. I think the webmaster is in the U.K. so I want to get it there before he signs off for the night. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 press release
Robert Osfield wrote: Hi Paul, Sorry for the slow reply, it's holiday in the Osfield household right now :-) I hope you and your family are having a good time. On Mon, Apr 5, 2010 at 6:51 PM, Paul Martz wrote: Thanks for the suggestions and changes. The Wiki version is now up: http://www.openscenegraph.org/projects/osg/wiki/News/Press/OSG2.8.3 Thanks for putting this up, I've now added a link to this from the News page: http://www.openscenegraph.org/projects/osg/wiki/News Excellent, thanks for updating that. I just changed the wiki front page to have a news bullet point for 2.8.3. I also just sent an email to Jose Luis asking if I could get write permission to the Downloads page so that I can add my source zip. I've done a quick review of the the OSG2.8.3 wiki page and it looks OK to me, and nice to see another company announcing the release - a healthy sign for the community ;-) Thanks for your help with this release! -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] 2.8.3 released!
Hi all -- The OSG v2.8.3 final release has just been tagged. Congratulations and many thanks to everyone in the community who pitched in and helped me get this out the door. The community effort really caught a lot of issues that I would never have caught on my own, so again, thanks! It's hard to believe it's been 6 months since 2.8.2. The 2.8.3 release takes new functionality, long available on trunk and in developer releases, and makes those features available in a stable release for developers to use in production releases of their OSG-based apps. I don't have write access to the downloads page, so I've made a source ZIP available here, until it can be moved to the OSG server: http://www.skew-matrix.com/osg283/OpenSceneGraph-2.8.3.zip Robert, if you could move that, I'd appreciate it. Does anyone want to volunteer to build binaries? Preferably, someone who has done it in the past, so that the packages remain consistent from release to release. After the downloads page is in shape, I'll push the press release out to opengl.org and khronos.org. Use this URL to svn checkout the 2.8.3 final release tag: http://www.openscenegraph.org/svn/osg/OpenSceneGraph/tags/OpenSceneGraph-2.8.3 The release status page serves as the release notes: http://www.openscenegraph.org/projects/osg/wiki/Support/283release And the press release is still located here: http://www.openscenegraph.org/projects/osg/wiki/News/Press/OSG2.8.3 The only issue I'm aware of with 2.8.3 at this point is a problem I discovered with the 3DS exporter handling certain MatrixTransforms, possibly an issue with how the matrix decomposition handles corner-case matrices. The behavior is the same on svn trunk and not a show stopper in my opinion. I've exchanged offline emails with Sukender and supplied him with a reproducer .osg model. Again, thanks to all for the assistance with v2.8.3. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 press release
Jean-Sébastien Guay wrote: Hi Paul, You changed the sentence I pointed out to: "numerous commercial and open source academic application software projects" I would have rather said: "numerous commercial, open source and academic application software projects" Easily fixed, thanks. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 press release
Thanks for the suggestions and changes. The Wiki version is now up: http://www.openscenegraph.org/projects/osg/wiki/News/Press/OSG2.8.3 -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 rc3 tagged -- hopefully the last rc
Michael Rohn wrote: Hi, thanks for the changes in the PNGPlugin. It builds and works fine for me. But I have a new problem: If I select the option 'BuildQtExamples' in cmake then I get an error. Two examples, namely 'osgviewerQtWidget' and 'osgQtBrowser' are missing. I think there are two solutions: 1. Add the examples or 2. Remove them from the CMake-Files Maybe the second version is better one. Agreed. Good catch. I've removed the references to these examples from the examples/CMakeLists.txt on the 2.8 branch. This will go into the final release (or the next release candidate... ) -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 rc3 tagged -- hopefully the last rc
Hi again -- I just realized we can't do this release over a holiday weekend. I would like to release at the end of Monday, though, so please help out by doing a bit of testing if your schedule allows. In particular, exercise the PNG plugin (which was the only code change in the most recent rc). Almost a final release! Thanks everyone! -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3rc2 Mac Xcode project - Windowing system
Bruce Wheaton wrote: The Xcode project seems better - executables can now create the correct sort of windows. Thanks Stephan! Good to hear. I think we're really close to a release now with the new rc3, which has the updated XCode file. I also got to learn something more about OSG, so that was nice. BTW, Paul - thanks for the OSGQSG - very nice, good introduction. You're welcome. It needs updating, starting to be irrelevant in spots... -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] 2.8.3 rc3 tagged -- hopefully the last rc
Hi all -- I've just tagged 2.8.3 rc3. You can download it from here: http://www.openscenegraph.org/svn/osg/OpenSceneGraph/tags/OpenSceneGraph-2.8.3-rc3 More info on the 2.8.3 release is here: http://www.openscenegraph.org/projects/osg/wiki/Support/283release Changes in rc3 include: * Fixes for mingw builds. * Fix for static linking of 3DS, OBJ, and FBX plugins. * Fix XCode problems with the osgviewer and osgmovie executables. * Add the OSG_CPP_EXCEPTIONS_AVAILABLE CMake option. * Update PNG plugin for compatibility with multiple versions of libPNG. * CMake build enhancements for the COLLADA plugin. Really the only code change was the update of the PNG plugin, and the other changes are all build-related. As a result, I'm optimistic I can rename the rc3 as the final release on 3 April, late Saturday. I appreciate any testing you can do between now and then. Thanks to everyone who contributed suggestions, changes, fixes, and all the testing! I think we're getting really close to the final release now. I'm changing my name to Paul "Merge Monkey" Martz. :-) -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 release imminent
Rafa Gaitan wrote: Hi Paul, I'm not sure if we arrive on time, but maybe it will be interesting to add lastest mods of CMake for Collada plugin. What do you think? Hi Rafa -- As the release was held back due to other changes anyway, and the COLLADA build changes were non-invasive and didn't change any actual code, I have added these CMake changes. I've tested on Windows w/ VS9. With this change, setting up and building the COLLADA plugin on 2.8.3 is just as easy as on trunk. This is in 2.8.3 rc3 (just tagged). Thanks for the input. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] libPNG API changes
Bruce Wheaton wrote:
LibPNG changed their API in January. I have a current Arch Linux install
which uses the newer version, but we can presume it will crop up more.
The only two current issues are addressed by:
/OpenSceneGraph/src/osgPlugins/png/ReaderWriterPNG.cpp
176d175
< #if PNG_LIBPNG_VER < 10004
178,180d176
< #else
< if (fin.gcount() == 8 && png_sig_cmp(header, 0, 8) == 0)
< #endif
231,232d226
< {
< #if PNG_LIBPNG_VER < 10004
234,237d227
< #else
<png_set_expand_gray_1_2_4_to_8(png);
< #endif
< }
As per the libPNG recommendations, to replace png_check_sig
and png_set_gray_1_2_4_to_8()
http://www.libpng.org/pub/png/src/libpng-1.2.x-to-1.4.x-summary.txt
If this isn't the best/place to report this, someone please let me know.
Hi Bruce -- This appears to be fixed on trunk with r11121. It's a non-invasive
change, except for I also need to add r10763 (essentially just a change to CMake
except for the PNG plugin code changes). But I have now merged r10763 and r11121
onto the 2.8 branch, so now we should have support for many versions of libPNG
(I'm using an older version) and also adds the OSG_CPP_EXCEPTIONS_AVAILABLE
CMake option.
I'll tag a release candidate shortly.
--
-Paul Martz Skew Matrix Software
http://www.skew-matrix.com/
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3rc2 Mac Xcode project - Windowing system
Bruce Wheaton wrote: I certainly defer to Stephan on this... but I presume the fix he found is similar, or addresses the core problem. If time becomes a factor Stephan, or you need testing of a changed version, let me know. Stephan sent me a new XCode project file and I have checked it in to the 2.8 branch but I have not yet made another release candidate. I want to consider updating the PNG plugin before I tag another rc. (If the changes to PNG are non-intrusive, and I'm suspecting they are, it might be a good idea to update it.) In the meantime, if you could test the XCode project files on the head revision of the 2.8 branch, I'd appreciate it. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3rc2 Mac Xcode project - Windowing system
Bruce Wheaton wrote: OK, it seems that when I set: -DUSE_DARWIN_COCOA_IMPLEMENTATION for the osgViewer framework, then recompile, the problem goes away. I'm not 100% if this is the right place, but it seems appropriate - the targets seem to be broken down by carbon and cocoa here in the viewer more than anywhere. I also tried -DUSE_DARWIN_CARBON_IMPLEMENTATION in the carbon target configurations with apparent success. Hopefully that's it - a fairly minor change. I don't see a reason why it wouldn't always be set, since without the flag the correct window implementation doesn't get compiled. If Stephan doesn't get back before you need it, I can modify a clean Xcode project and send it back? Where would that be too? Bruce This is excellent news! Thanks! Stephan's follow-up email indicates we still lack consensus. If you folks decide on a change, you can zip up the project files and just email them directly to me to avoid sending megabytes through the mail server. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] libPNG API changes
Bruce Wheaton wrote:
LibPNG changed their API in January. I have a current Arch Linux install
which uses the newer version, but we can presume it will crop up more.
The only two current issues are addressed by:
/OpenSceneGraph/src/osgPlugins/png/ReaderWriterPNG.cpp
176d175
< #if PNG_LIBPNG_VER < 10004
178,180d176
< #else
< if (fin.gcount() == 8 && png_sig_cmp(header, 0, 8) == 0)
< #endif
231,232d226
< {
< #if PNG_LIBPNG_VER < 10004
234,237d227
< #else
<png_set_expand_gray_1_2_4_to_8(png);
< #endif
< }
As per the libPNG recommendations, to replace png_check_sig
and png_set_gray_1_2_4_to_8()
http://www.libpng.org/pub/png/src/libpng-1.2.x-to-1.4.x-summary.txt
If this isn't the best/place to report this, someone please let me know.
Am I correct in assuming that this is a change request for the upcoming 2.8.3
release? I believe someone else has mentioned this too.
--
-Paul Martz Skew Matrix Software
http://www.skew-matrix.com/
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] How to set an ID to a node?
Benjamin GODIN wrote: In Fact, I have different objects in my scene and each geode has his own PositionAttitudeTransform as his parent. At t+1 time, each PositionAttitudeMatrix gets a new translation vector. To do this, I'm using the NodeVisitor. I have a table containing each translation vector for each geode for each timestamp. But in the NodeVisitor, I don't know which geode refers to which line in my table. So I needed a reference between the two. :) You could just store the address of the PAT in the table, then you could just iterate over the table and avoid the overhead of a NodeVisitor. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3rc2 Mac Xcode project - Windowing system
FYI, I'm sort of waiting to hear what's up with this issue, will it require a change or not. The change I just put in for static linking on the 2.8 branch really is so insignificant that it hardly merits a new release candidate. But if there's a change for XCode too, then I'd like to cook another release candidate and let people taste the cooking before the final release. On the other hand, I don't want to hold this up forever. So any progress you folks can make on the XCode issue will be greatly appreciated. At the time I write this, the XCode issue is the only item holding up the release. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] duplicate symbols in 3DS and OBJ plugins
Robert Osfield wrote: If the code is C++ then I would be inclined towards using unique namespaces for the different plugins. Some already do this. If the the code is C or pulled into from an external project like lib3ds then use of namespaces is problematic so making sure names are unique would be appropriate. Hi Robert -- I added namespaces to the FBX and 3DS plugins, but left the OBJ plugin alone. This has resolved the issue for the 2.8.3 release. This is on the 2.8 branch as revision 11303. If the merge back to trunk turns difficult, let me know, and I'll do the change there and post full changed files to osg-submissions as usual. (Now that I've already committed the change, I'm thinking we should also use the same convention in OBJ, for consistency, so you might want to consider that for trunk.) -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3rc2 Mac Xcode project - Windowing system
Bruce Wheaton wrote: On the Mac Xcode project, the Cocoa viewer works, but on the osgviewer and osgmovie targets, they don't run. I've tried the Carbon 32-bit and Cocoa 32-bit versions. I linked to Carbon, Cocoa, OpenGl and Quicktime Frameworks. osgViewer, OpenThreads, osg, osgDB, osgText, osgUtil, osgGA Frameworks are copied into the bundle's Frameworks folder All plugins are copied into the bundle's Plugins folder But when the apps try to make a window, they fail with: View::setUpViewAcrossAllScreens() : Error, no WindowSystemInterface available, cannot create windows. As I recall, this used to be enough. Did the introduction of the Cocoa viewer change something? Is there an environment variable that needs to be set? Since I intend to make a Mac bundle, I'd rather not have any user installation steps. Thanks for the issue report. I'm not encountering any such issues when I build with CMake / makefiles / gcc on OS X 10.5, and I have offline emails indicating this same build configuration works for 10.6. It seems to be XCode-only. Stephan Huber contributed the XCode project files, perhaps you two can work together to resolve the issue? It'd be nice to have this resolved prior to release, so anything you guys can come up with would be greatly appreciated. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] duplicate symbols in 3DS and OBJ plugins
Robert Osfield wrote: Hi Paul, On Wed, Mar 31, 2010 at 3:57 PM, Paul Martz wrote: Does anyone have any other thoughts or ideas on this? Seems like we need a policy that we can use in OSG going forward, so that this same issue doesn't keep creeping back into the code (only to be discovered the day before the scheduled release :-( ...) If the code is C++ then I would be inclined towards using unique namespaces for the different plugins. Some already do this. If the the code is C or pulled into from an external project like lib3ds then use of namespaces is problematic so making sure names are unique would be appropriate. Thanks for the feedback. I am able to reproduce this issue on Windows, and there is also a symbol collision, as suspected, with the FBX plugin. I'll work on the namespace solution and see how far I get. Whee! -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 release imminent
Robert Osfield wrote: I'm easily swayed either way. Paul has already put everything behind a 2.8.3 release is almost ready to push the button, I believe it should be his final call as to whether he feels bumping the version 2.10.0 is OK. What ever Paul decides I'll run with and do what I can to assist. Thanks, I'm inclined to stick with 2.8.3 at this point. OSG 2.8.0/1/2 apps should require only a recompile/relink to build with 2.8.3, except if they use osgAnimation (which is technically not what I'd call "core OSG"). Art, your osgPPU tagged for 2.8 should compile and link with 2.8.3 just fine, so no need for you to do a new tag. Let me know if that's not the case. The bottom line is that I think the "2.8.3" moniker is appropriate. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 release imminent
Art Tevs wrote: Hi Paul, just a question, maybe I missed it: why we actully started now with the numbering 2.8.x Why not to call new release 2.10? I mean, I am waiting for major release in order to tag also new version of osgPPU. However due to the used patch numbers, I thought that these are just small updates. Hi Art -- 2.8.x releases are developed and tagged off of the 2.8 branch. So will we continue with 2.8.x numbering or is there a plan to go further with the old numbering, 2.10, 2.12, ... That's up to you and the rest of the community. If someone wants to do another release on the 2.8 branch, that would be "2.8.4". As for 2.10 and 2.12, they seem to be off the radar, although they were announced at the SIGGRAPH OSG BOF last August in the slides from Robert that I presented. I don't recall the reasoning for canceling 2.10 and 2.12, maybe Robert can comment on that. But my understanding is that he's intending "3.0" to be the next stable release off of trunk. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 release imminent
Jean-Sébastien Guay wrote: Hi Rafa, I'm not sure if we arrive on time, but maybe it will be interesting to add lastest mods of CMake for Collada plugin. What do you think? I had already requested the recent changes to the Collada plugin and Paul said it would be too big a change to include in 2.8.3, given how much is already there, so perhaps in a possible 2.8.4 after that... Thanks, J-S. Yes, this was discussed before. The Collada changes have a pretty invasive revision dependency chain, and those are just the revisions I know about, so it seemed too risky to attempt it. And, as it turns out, I now have my hands full with the static linking issue, so I'll be lucky if I can resolve that and get the release out the door. I'm not the only person who can run "svn merge". :-) Anyone in the community is welcome to merge the collada changes onto their local copy of the 2.8 branch and, once you get something viable, post it to osg-submissions like any other change (clearly stating it's a change to the 2.8 branch). If it turns out to be totally trivial, we can reconsider adding it to 2.8.3, or it can go into a 2.8.4 release. -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] duplicate symbols in 3DS and OBJ plugins
Hi all -- Stefan uncovered an issue when building 2.8.3 with static libraries. The OBJ and 3DS plugins have duplicate symbol names, notably PrimitiveIndexWriter::drawArrays, which causes a link error. The FBX plugin might also have this issue. See the thread "2.8.3-rc2 tagged" for full details. I wanted to open a new discussion thread for this issue, as it's almost certainly a problem on trunk and therefore applicable to anyone in the community building OSG as static libraries. Any fix I put in to 2.8.3 will probably need to be merged back to trunk. One way to resolve this would be to wrap the symbols in a plugin-specific namespace. Alternatively, we could hit it with a hammer and just change the symbol names so that they are unique across all plugins. Does anyone have any other thoughts or ideas on this? Seems like we need a policy that we can use in OSG going forward, so that this same issue doesn't keep creeping back into the code (only to be discovered the day before the scheduled release :-( ...) -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] How to set an ID to a node?
Benjamin GODIN wrote: Hello, I was wondering if there was a way to set an ID to a node. Then, I could use that ID in my NodeVisitor to refer to an external table which contains the values to use. Is it possible ? In fact, the only thing I have seen in the doc is the possibility to give the node a name (a String). What functionality do you need beyond a string? You can attach anything that derives from Referenced as UserData on anything that derives from Object, including Node. The osgWorks project defines a RefID class, which is essentially a reference counted string, for just such a purpose. (See osgworks.googlecode.com.) -- -Paul Martz Skew Matrix Software http://www.skew-matrix.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3-rc2 tagged
Thanks, Stefan. Good catch. I'm pretty sure this will be a problem on svn trunk too, if anyone were to try a static link there on a Linux/Darwin box. What I _think_ is happening is this: By default, Linux/Darwin exports symbols from libraries. In a non-static build, shared object symbols are contained, so that's not going to be a problem either. But for static builds on Linux/Darwin, two plugins that have symbols with the same name will cause a conflict at link time. Does this sound right? If this is an accurate description of the problem, then I can think of three ways to resolve the issue: 1) If I remember right, there is a way on Linux/Darwin to tell the linker to hide symbols rather than export them. It's been a while since I've done *nix development; this might just apply to shared objects. 2) We could use namespaces in the plugin. You might try this, 3) The ugly, but direct method would be to change the class names so that they are unique within the plugin. But honestly I'd prefer to use a namespace. I don't think you're seeing a problem with FBX at this time because the XCode project files don't include the FBX plugin. But this might be an issue on other platforms that try to do a static build, so we should probably address it in some way as well. Stefan, would you like to take a crack at option 2 above, and use a namespace qualifier in the 3DS plugin to see if it resolves this issue? If so, please submit your change. I'm willing to hold up the release for this fix. Paul Martz Skew Matrix Software LLC _http://www.skew-matrix.com_ <http://www.skew-matrix.com/> +1 303 859 9466 stefan nortd wrote: Hi Paul, I have a linker problem with the latest rc2 that I did not have with 2.8.2. Code: ld: duplicate symbol PrimitiveIndexWriter::drawArrays(unsigned int, int, int)in /Users/noema/Development/git/VirtualAwesome/libs/osg-osx/lib/libosgdb_obj.a(OBJWriterNodeVisitor.o) and /Users/noema/Development/git/VirtualAwesome/libs/osg-osx/lib/libosgdb_3ds.a(WriterNodeVisitor.o) collect2: ld returned 1 exit status Command /Developer/usr/bin/g++-4.2 failed with exit code 1 (xcode, snow leopard) It looks to me that there a three occasions with an identical PrimitiveIndexWriter::drawArrays(...) function. It's in the obj, 3ds, and fbx plugin. I am using obj and 3ds together. I also link statically via the USE_OSGPLUGIN(...) call but I would assume the same error occurs when using the osg dynamically. Any ideas? Cheers, stefan -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=26272#26272 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 release imminent
Michael Rohn wrote: I tested it under Win7 with MinGW and had some problems during the build. Thanks for the testing and the input! TXPNode.cpp -> needs #include Thanks, that's r10965. osgPlugins PNG -> problem with new libpng. Some functions in libpng are renamed Sorry, but this amounts to a new feature request, and with the release just 1 day away we are well past considering new features, so I'll have to propose you include this in 2.8.4, or wait for the next stable release off trunk. osgviewerSDL -> need to use the CMakeLists.txt from trunk Thanks, this is r10963, except it looks like this revision contains changes to PNG along the lines of upgrading to a new version of libPNG, so I'll include all the CMakeLists.txt changes in the revision, but not the change to the PNG plugin source. I'm merging in these changes now, they will be in the 2.8 branch head shortly. -Paul ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 release imminent
Robert Osfield wrote: Hi Paul, I've tested 2.8.3-rc2 out at my end and so far everything compiles and works well under Kubuntu 9.04. I don't have 3rd party apps to test against so I've only tested against the core OSG apps. Unfortunately VPB requires the svn/trunk version of the OSG so I haven't been able to test against this. Thanks for the update. I'm digging in to one issue I've uncovered today, but it has to do with a custom node so I doubt it's a show stopper for the release. Other than that I think we're clear for a release. -Paul ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Make text ignore polygonmode
Martin Beckett wrote: I have some text labels in my scene, but when I switch the polygon mode (eg with the 'w' key in osgviewer) the text is rendered broken (in line mode) and dissapears (in point mode). How do I force osg::PolygonMode::FILL, or force text to ignore it? I tried: Code: label->getOrCreateStateSet()->setMode(GL_POLYGON_MODE, osg::StateAttribute::ON); I think you'll want to OR in the PROTECTED flag, right? -Paul ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] 2.8.3 release imminent
Hi all -- Just a note to say that I'm planning to tag the 2.8.3 release late Wednesday 31 March. I'm not seeing a lot of posts about 2.8.3 experiences, so if you've been using it, please post and let me know how it's working, good or bad. I am using it myself for many external projects, and have received offline emails about testing on various platforms. At this point I'm not aware of any show stoppers. If that continues to be the case, then I'll rename the rc2 tag as the final tag. I'll try to put together a press release as time allows. If the authors of these modules could send me a summary of the changes that have gone into 2.8.3 (I took them from trunk-head), I'd appreciate it: * FBX plugin * ply plugin * ffmpeg plugin * 3ds plugin * osgAnimation * ImageIO plugin * QuickTime plugin * Inventor plugin Thanks, -- Paul Martz Skew Matrix Software LLC _http://www.skew-matrix.com_ <http://www.skew-matrix.com/> +1 303 859 9466 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] osg::Switch , UpdateCallback not being dissabled
Ted Morris wrote: OK, so I turn on one vehicle as in: _vehiclesw->setSingleChildOn(1); The scene just renders the single object and runs through the callback. fine. But then it also is still running through object "0" callback (although it is not being rendered in the scene). Interestingly, the other children > 1 don't have their callbacks invoked under this circumstance. The UpdateVisitor constructor shows that it uses the TRAVERSE_ALL_CHILDREN traversal mode by default, so it doesn't necessarily care which child is active. If a subgraph contains update callbacks, the UpdateVisitor will traverse. -Paul ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] 2.8.3-rc2 tagged
Hi all -- v2.8.3 release candidate 2 was just tagged, with the following additions over rc1: * Updated XCode project file (includes ply plugin). * r11286 fixes OpenThreads pthreads build on FreeBSD. * r11289 adds pixel format getter to GraphicsWindowCocoa. You can check out rc2 from here: http://www.openscenegraph.org/svn/osg/OpenSceneGraph/tags/OpenSceneGraph-2.8.3-rc2 Full v2.8.3 release info is on the wiki: http://www.openscenegraph.org/projects/osg/wiki/Support/283release Please continue to test and let me know of any issues. Thanks! -Paul ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3-rc1 tagged
Stephan Huber wrote: FBX and PLY are not part of the xcode project file, because of the external dependencies. If PLY has no external dependencies I can add to the xcode-project-file. FBX does have an external dependency, and I understand that dependency is available on OS X. PLY doesn't appear to have any external dependencies. Whether you add these to the XCode project file is your call. I build on OS X with makefiles and gcc, so I'm already happy. :-) -Paul ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3-rc1 tagged
Bruce Wheaton wrote: Things are going pretty well, until I tried osgMovie - seems like a --framework OpenThreads flag is still in the Target. Oh, wait, it's in a few places - osgViewer too. Am I right that it can be safely removed now? The target seems to be gone from the Xcode file. Huh, not sure. I think this is up to you XCode folks, you will need to decide what you want in the XCode project files. If there are any changes that need to be folded in, let me know and I'll commit them. -Paul osgViewerCocoa was working, I guess the other targets need some work? Bruce On Mar 23, 2010, at 1:27 PM, Paul Martz wrote: Bruce Wheaton wrote: Sorry, Paul, did the Mac OS X frameworks and Xcode stuff get updated properly? I saw you were looking for help (afraid I'm too new to the project to offer). XCode project files were added and are in rc1 (thanks, Stephan). I have a Mac, but primarily use it as a *nix box (gcc and a shell prompt), so I really have no experience with XCode. I was able to load the project into XCode v3.1.4, but didn't do any further testing with it. So right now Stephan is really the only XCode tester. If you have XCode experience and your schedule allows, please try out the XCode project files, and let us know how things work for you. Did the ply and fbx plugins make it into the XCode project file, for example? That would be one thing to check... -Paul ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3-rc1 tagges
Robert Osfield wrote: Hi Paul, I've done a check out of 2.8.3-rc1 and it everything (including wrappers, examples and package targets) builds just fine under Kubuntu 9.04. So thumbs up so far :-) I've run out of daytime here so will have to come back to doing runtime testing tomorrow morning. Great news, thanks! Are you building any of ffmpeg, fbx, or the Inventor plugin, by any chance? -Paul ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3-rc1 tagged
Bruce Wheaton wrote: Sorry, Paul, did the Mac OS X frameworks and Xcode stuff get updated properly? I saw you were looking for help (afraid I'm too new to the project to offer). XCode project files were added and are in rc1 (thanks, Stephan). I have a Mac, but primarily use it as a *nix box (gcc and a shell prompt), so I really have no experience with XCode. I was able to load the project into XCode v3.1.4, but didn't do any further testing with it. So right now Stephan is really the only XCode tester. If you have XCode experience and your schedule allows, please try out the XCode project files, and let us know how things work for you. Did the ply and fbx plugins make it into the XCode project file, for example? That would be one thing to check... -Paul ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] 2.8.3-rc1 tagges
Hi all -- Things are looking pretty stable so far, so I've gone ahead and tagged release candidate 1. You can check it out from here: http://www.openscenegraph.org/svn/osg/OpenSceneGraph/tags/OpenSceneGraph-2.8.3-rc1/ Please test as soon as your schedule allows and report any problems. See the wiki page for major features, and guide your testing accordingly: http://www.openscenegraph.org/projects/osg/wiki/Support/283release The only potential issue I'm aware of at this time is the ongoing discussion concerning FreeBSD 8 pthreads. If this requires a change, or if other issues are found requiring a change, I'll tag another release candidate. Thanks for everyone who has helped so far. I'm overwhelmed by the responsiveness of people here, directing me to specific fixes and trunk changes in order to fix bugs and add new features to the 2.8 branch. Outstanding work, everyone! -Paul ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 testing
Jim Brooks wrote: Had compile errors on FreeBSD 8. Fixed with these two lines: src/OpenThreads/pthreads/CMakLists.txt IF(HAVE_PTHREAD_SETAFFINITY_NP) IF(NOT CMAKE_SYSTEM_NAME MATCHES "FreeBSD") #---fix--- ADD_DEFINITIONS(-DHAVE_PTHREAD_SETAFFINITY_NP) ENDIF(NOT CMAKE_SYSTEM_NAME MATCHES "FreeBSD") #---fix--- Thanks; let's discuss this in osg-submissions where Robert has already replied. -Paul ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 testing
Jean-Sébastien Guay wrote: Hi Paul, Building now, Windows 7 64 bit, VC9 sp1, 32 bit build. After building I'll test what I can. Builds fine. I've also set up a nightly build (called "Whitestar_vc9sp1 (2.8 branch)"). Thanks. I'm also running daily CDash builds for Windows and Darwin/OSX, though we still have the reporting issue with Darwin that I've mentioned before. (The Darwin build fails to report any errors when the build fails, and always reports that the build took less than a minute even though it clearly takes about 40 minutes. Not sure what's going wrong here.) As for testing, I've tested a few models in osgviewer, as well as the fbx plugin and the dae plugin. The fbx plugin seems to indeed have all the latest fixes, and the dae plugin is missing several fixes (as expected). I have to apologize about the dae plugin. I was just too chicken to go through with the merge, after a couple attempts. :-( Other than that, could you make a list of what 2.8.3 is supposed to fix over 2.8.2, and from that list maybe we can devise a list of tests to run to confirm that those things, at least, are really fixed? The wiki page has a bullet point summary of major changes, plus also contains a nearly complete list of revisions merged from trunk: http://www.openscenegraph.org/projects/osg/wiki/Support/283release Are you after more info than that? (Perhaps a more human-readable list of changes, instead of just revision numbers?) Tomorrow I'll test 2.8.3 in our software which will make for a more complete test. Thanks. I'm also testing with my external projects. -Paul ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] trouble in MSFBO
Kai Xia wrote: but the texture output seem not more smooth than before, and which is worse is that there is some obvise black segment on the edge of the objects in 3d scene, I dont know what is wrong about this? even I change the texture format to GL_RGB / GL_RGBA, the problems remained, could any one tell me how to implement MultiSample in FBO, thanks very much! Hm. It works fine for me with a GL_RGBA texture, and the camera configured as follows: rootCamera->attach( osg::Camera::COLOR_BUFFER0, tex, 0, 0, false, 8, 8 ); rootCamera->setRenderTargetImplementation( osg::Camera::FRAME_BUFFER_OBJECT, osg::Camera::FRAME_BUFFER ); rootCamera->setImplicitBufferAttachmentMask( osg::Camera::IMPLICIT_COLOR_BUFFER_ATTACHMENT|osg::Camera::IMPLICIT_DEPTH_BUFFER_ATTACHMENT, osg::Camera::IMPLICIT_COLOR_BUFFER_ATTACHMENT ); -Paul ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 testing
Thanks, Cedric. I found r10362 had this change, and it's now merged into the 2.8 branch. I definitely get different results, but haven't had a chance to compare to trunk, so if you could please take a look and let me know if it's now correct, I'd appreciate it. -Paul Cedric Pinson wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 03/22/10 22:03, Cedric Pinson wrote: On 03/22/10 20:15, Robert Osfield wrote: Hi Cedric, On Mon, Mar 22, 2010 at 7:08 PM, Cedric Pinson wrote: I tried osgAnimation with this release, and there is difference with the trunk behavior. The test is to use osganimationhardware with bignathan.osg the one with the trunk works as expected and the one in this release seems to blend multiple animations. It looks like the clone problem... because i instanciate multiple item from the input. I will try to check it this evening Didn't you provide updated .osg files for the changes in osgAnimation? Could there be a coupling between other changes in OpenSceneGraph-Data and which version of osgAnimation you have? I presume 2.8.3 will need to have a new OpenSceneGraph-Data made from it's own svn/trunk. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org I dont know yet the format should be the same if osgAnimation is synced to trunk, so i am doing some diff to check where it could have a problem Cheers, Cedric I think i get it comes from node.cpp: diff Node.cpp ~/dev/osg-trunk-svn/src/osg/Node.cpp 77c77 < _updateCallback(node._updateCallback), - --- _updateCallback(copyop(node._updateCallback.get())), 80c80 < _cullCallback(node._cullCallback), - --- _cullCallback(copyop(node._cullCallback.get())), copying Node.cpp into the release 2.8.3 works as expected Cheers, Cedric - -- Provide OpenGL services around OpenSceneGraph and more +33 659 598 614 Cedric Pinson mailto:cedric.pin...@plopbyte.net http://www.plopbyte.net -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.14 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkun4HMACgkQs6ZHzVQN0Ij7fACeIapSBIod7TLbJMg4zLvyIpVQ ppwAoIYKZD1H1bCecbdZ2TpBTNghxmw/ =o67H -END PGP SIGNATURE- ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 testing
Chuck Seberino wrote: Andy and Paul - I submitted a fix for this (r10927) which reorders some of the ifdef checking so that it works properly under OSX. Thanks, r10927 is now merged onto the 2.8 branch. Andy, let me know how that works for you. (I wasn't seeing this problem on 32bit OS X 10.5; I do a ccmake to generate makefiles and build with gcc on the command line, apparently a different configuration from you?) And, sorry I wasn't able to get to this sooner. We had a temporary internet outage in the area. Back online now (obviously). -Paul ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 testing
I'll look at this shortly. Andy, if you're really aggressive, you could merge it onto your local copy and let me know if it fixes things. Kind of "pre-test" it for me. But I'll get to it soon. -Paul Chuck Seberino wrote: Andy, The fix is in trunk, but hasn't been merged into the 2.8 branch yet. Chuck On Mar 22, 2010, at 10:44 AM, Andy Skinner wrote: Does that mean I need to wait for it to get onto 2.8? Or that it is there? thanks andy -Original Message- From: osg-users-boun...@lists.openscenegraph.org [mailto:osg-users- boun...@lists.openscenegraph.org] On Behalf Of Chuck Seberino Sent: Monday, March 22, 2010 1:00 PM To: OpenSceneGraph Users Subject: Re: [osg-users] 2.8.3 testing Andy and Paul - I submitted a fix for this (r10927) which reorders some of the ifdef checking so that it works properly under OSX. Chuck On Mar 22, 2010, at 9:47 AM, Andy Skinner wrote: I just got OpenSceneGraph-2.8, and tried to build on OS X. I'm getting an error in OpenThreads/Atomic, complaining that int32_t does not name a type. I'm following the unix instructions, doing configure and then make. I had just build the trunk (at least the first part) and got past this. Have I got it set up incorrectly, or is there a problem on 2.8? thanks andy -Original Message- From: osg-users-boun...@lists.openscenegraph.org [mailto:osg-users- boun...@lists.openscenegraph.org] On Behalf Of Paul Martz Sent: Monday, March 22, 2010 12:31 PM To: OpenSceneGraph Users Subject: [osg-users] 2.8.3 testing Hi all -- Here's an update on the 2.8.3 release, and a plea for testing. The 2.8.3 release is slightly behind my original schedule, but nonetheless getting close to a release candidate. ISSUES: There are two open issues at this point: 1. XCode project files need to be updated. (Stephan, will you have any time to contribute this in the near future, or do we need another volunteer?) 2. Updates to the DAE plugin and osgManipulators nodekit. These are late feature requests. I'm concerned about updating these two modules. The changes for both these modules are not limited to the modules themselves, and require changes to core OSG and elsewhere. Many changes on trunk often depend on previous changes, and chasing all of them down with a high level of confidence is difficult. It would certainly introduce issues if I missed any required changes. Let me make this suggestion: Let's leave DAE and osgManipulators out of the 2.8.3 release. If someone really wants them, and can't wait for v2.10 (or v3.0), then we could do a 2.8.4 release. The benefit of this approach is that it doesn't prevent 2.8.3 from getting out the door in a timely fashion, should merging these changes produce instability that would otherwise delay the release. Does this make sense? TESTING: Please test the 2.8 branch if you have time. For convenience, here's the 2.8 branch URL: http://www.openscenegraph.org/svn/osg/OpenSceneGraph/branches/OpenScene Graph-2.8 I'd appreciate build/run testing for the following platforms and features: Platforms: Mac 10.5/10.6 32/64bit Windows XP/Vista/7, 32/64bit *nix, 32/64bit Features (latest stuff from trunk): osgAnimation nodekit FBX plugin 3ds plugin (esp writer capability) ImageIO plugin QuickTime plugin Inventor plugin ply plugin ffmpeg plugin There were also minor bug fixes to the 3dc, ac, dds, and flt plugins. When you send a test report, please mention which plugins you are building, as not all plugins build by default of course. Once we get the XCode project files updated, I'll tag a release candidate, and further candidates will follow as needed. I'm still hoping for a 31 March release date. If we start testing now, this should give us plenty of time for testing and resolving issues. Thanks! -Paul ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users- openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users- openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users- openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 release likely, need community assistance
Tony Horrobin wrote: Hi Paul, It seems happy now. gcc version 4.4.1 (Ubuntu 4.4.1-4ubuntu9) Thanks. The fact that this was broken at all is a source of concern for me. I wonder how many other changes I've merged where the changes have simply vanished into thin air. :-( -Paul ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] 2.8.3 testing
Hi all -- Here's an update on the 2.8.3 release, and a plea for testing. The 2.8.3 release is slightly behind my original schedule, but nonetheless getting close to a release candidate. ISSUES: There are two open issues at this point: 1. XCode project files need to be updated. (Stephan, will you have any time to contribute this in the near future, or do we need another volunteer?) 2. Updates to the DAE plugin and osgManipulators nodekit. These are late feature requests. I'm concerned about updating these two modules. The changes for both these modules are not limited to the modules themselves, and require changes to core OSG and elsewhere. Many changes on trunk often depend on previous changes, and chasing all of them down with a high level of confidence is difficult. It would certainly introduce issues if I missed any required changes. Let me make this suggestion: Let's leave DAE and osgManipulators out of the 2.8.3 release. If someone really wants them, and can't wait for v2.10 (or v3.0), then we could do a 2.8.4 release. The benefit of this approach is that it doesn't prevent 2.8.3 from getting out the door in a timely fashion, should merging these changes produce instability that would otherwise delay the release. Does this make sense? TESTING: Please test the 2.8 branch if you have time. For convenience, here's the 2.8 branch URL: http://www.openscenegraph.org/svn/osg/OpenSceneGraph/branches/OpenSceneGraph-2.8 I'd appreciate build/run testing for the following platforms and features: Platforms: Mac 10.5/10.6 32/64bit Windows XP/Vista/7, 32/64bit *nix, 32/64bit Features (latest stuff from trunk): osgAnimation nodekit FBX plugin 3ds plugin (esp writer capability) ImageIO plugin QuickTime plugin Inventor plugin ply plugin ffmpeg plugin There were also minor bug fixes to the 3dc, ac, dds, and flt plugins. When you send a test report, please mention which plugins you are building, as not all plugins build by default of course. Once we get the XCode project files updated, I'll tag a release candidate, and further candidates will follow as needed. I'm still hoping for a 31 March release date. If we start testing now, this should give us plenty of time for testing and resolving issues. Thanks! -Paul ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 release likely, need community assistance
Andy Skinner wrote: I'm building the trunk again now on OS X, to see if Robert has fixed the problem you mention. If he has, maybe you could bring this change back in. Yes, it does look like trunk has a fix, plus I was missing another change. I have put your fix back in the 2.8 branch, plus also merged in r9904 and r11273, and this is building for me without errors on OS X 10.5 (and trunk is also building fine). Please test the latest 2.8 branch to make sure this resolves your static initializer issue. Thanks. -Paul ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 release likely, need community assistance
Tony Horrobin wrote: Hi Paul, To clarify, this behaviour was only on the 2.8 branch. It turns out that r10149 on the trunk had the essentially same fix as I attached in my last post. Thanks for the additional details. I had already merged all ply changes (including r10149) onto the 2.8 branch, so I was confused as to why this would build on trunk but be broke on the branch. Mysteriously, many of the changes I merged onto 2.8 for ply are missing, even though svn thinks they have been merged (and therefore does nothing if I attempt to merge the same changes again). Not sure how this happened. I have resolved the issue by manually copying the ply plugin source files from svn trunk to the 2.8 branch. If you don't mind, please try it again and report back. Thanks for testing, otherwise this might not have been caught until after the release. -Paul ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 release likely, need community assistance
Tony Horrobin wrote: Hi Paul, Here are a couple of fixes if it's not too late: Revision 10601 on the trunk fixed 'Matrices' to 'Materials' in src/osgViewer/StatsHandler.cpp Thanks -- I did not merge r10601; it was quite extensive. But I did modify StatsHandler.cpp to fix this type. Under Ubuntu 9.10, gcc 4.4.1 the ply plugin fails to build because uint8_t is unknown. The solution would be to include stdint.h, except that under windows with Visual Studio 2005 and 2008, stdint.h is missing and the ply plugin uses its own typedefs instead. As a workaround, I've replaced uint8_t with unsigned char. If I understand you correctly, the ply plugin doesn't build on this platform, both on the 2.8 branch as well as svn trunk? Has there been a submission to fix this? If not, could someone make a submission please? I don't have enough information to do this myself. -Paul ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 release likely, need community assistance
Paul Martz wrote: Hi Andy -- This is now in the 2.8 branch (r11207, and also yesterday's r11263). It would be great if you could do some testing and verify this resolves the issue. Thanks in advance. Well, not so fast... It turns out that r11207 breaks the build on OS X 10.5. I get the same build errors on trunk as I do on the 2.8 branch with this change. Errors below. I'll back this out in 2.8. -Paul /Users/pmartz/Projects/OSG/trunk/src/osgViewer/GraphicsWindowCarbon.cpp: In member function Ôvirtual void osgViewer::CarbonWindowingSystemInterface::_init()Õ: /Users/pmartz/Projects/OSG/trunk/src/osgViewer/DarwinUtils.h:123: error: Ôbool osgDarwin::DarwinWindowingSystemInterface::_initializedÕ is private /Users/pmartz/Projects/OSG/trunk/src/osgViewer/GraphicsWindowCarbon.cpp:1069: error: within this context /Users/pmartz/Projects/OSG/trunk/src/osgViewer/GraphicsWindowCarbon.cpp:1071: error: ÔinitÕ is not a member of ÔosgDarwin::DarwinWindowingSystemInterfaceÕ make[2]: *** [src/osgViewer/CMakeFiles/osgViewer.dir/GraphicsWindowCarbon.cpp.o] Error 1 make[1]: *** [src/osgViewer/CMakeFiles/osgViewer.dir/all] Error 2 make: *** [all] Error 2 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 release likely, need community assistance
Hi Andy -- This is now in the 2.8 branch (r11207, and also yesterday's r11263). It would be great if you could do some testing and verify this resolves the issue. Thanks in advance. On a slightly different topic, I've also merged in the change from trunk that adds the version convenience macros to the Version header, so that your OSG v2.8.3-based application can now contain code like this: #if OSG_VERSION_GREATER_THAN(2,8,3) // version specific code here #endif This is a handy feature on trunk, and I'm glad I remembered to grab this for 2.8.3. -Paul Paul Martz wrote: Thanks for the updated info, Robert. I see your change as r11207. It's good to know that this issue is limited to just a small handful of OSX-related files. I have already updates the 2.8 branch to the latest (except for r11207) GraphicsWindowCocoa and DarwinUtils; this was required for OS X 10.6 support. For now, I'll wait on this, and revisit the issue as 2.8.3 rc1 approaches. -Paul Robert Osfield wrote: HI Paul, On Sun, Mar 14, 2010 at 10:36 PM, Paul Martz wrote: Hi guys -- Given this work is still in progress, I agree with Robert and I don't think it's right for the 2.8.3 release. Sorry. I've done a little work on the static initialization side and it looks like the particular one that was an issue for Andy was the way that the automatic initialization of osgViewer's under OSX was making windowing system calls, this isn't normally a problem but when working on a headless machine you'd get error even if you didn't open a window. The solution for this particular issue was making moving the windowing system initialization calls form the constructor to be lazily initialized on demand. To make this work I modified : src/osgViewer/DawinUtils.h src/osgViewer/DawinUtils.mm src/osgViewer/GraphicsWindowCocoa.mm src/osgViewer/GraphicsWindowCarbon.cpp I checked these changes into svn/trunk over the weekend, so they will still need testing by the community. Andy reported success with the first iteration of the changes I made, but made a few more changes since then to clean them up so there is still a bit of risk associated with these changes. Once the community tests them out a bit further we should be confident enough to run with them. Save for the above changes for static it could be that none of the other changes would be required for Andy as this is the only one that his reported as a show stopper so far. There is a but though... I've just checked the 2.8 branch at it doesn't have any Cocoa support at all, so my changes wouldn't be able to merge directly unless you just added Cocoa support to the OSG-2.8 branch. For the OSX community faced with enforced Cocoa usage under 64bit build this would be no bad thing, but it's more work for yourself to merge. Without these we'll just have to keep the 2.8 series as 32bit under OSX, and ear mark the 3.x series for 64bit support under OSX. The sooner I can get 3.0 out the door the less of an issue this will be, as it'll be an easy sell to the community - just use 3.x if you want OSX 64bit support ;-) Cheers, Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] OS X 10.6, is OSG building OK here?
Hi all -- I've got an offline report that the 2.8 branch isn't building on OS X 10.6 with the 10.6 SDK, so I wanted to post here and ask how others are getting along with OS X 10.6, either on the 2.8 branch head, or on trunk head. If there are issues, let's weed them out. Thanks for any info... -Paul ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 2.8.3 release likely, need community assistance
Jean-Sébastien Guay wrote: Well, here's one such request - could you include Michael Platings' recent changes to the FBX plugin, as well as the change he posted to osg-submissions for the Collada plugin (whenever Robert reviews and merges it into the trunk)? OK, the FBX plugin on the 2.8 branch is up to date with trunk r11262. (I haven't updated the wiki yet as I'm experiencing issues with the trac server.) For the DAE back port, r11129 is holding things up. In addition to modifying the DAE plugin proper, it appears to add matrix array support to core OSG, which is causing a merge conflict as the core OSG files are missing some preceding revisions. The addition of matrix arrays seems isolated enough that I ought to be able to add this feature manually. But not on a Saturday night. :-) I'll take another look at this tomorrow. -Paul ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
