Re: [osg-users] mac -install_name
Andy, I use install_name_tool -change . and install_name_tool -id ... in a script as a Run Script Build phase. I don't know if you can do it in XCode directly. Seems to work. Let me know if you need more info and I'll dig up the exact script. Martins Andy Skinner wrote: I’ll admit to being a little past the limit of my knowledge here ... We would like to specify the –install_name arg with which we link the shared libraries. I think it is because we don’t want the paths where these things were built to be in the files. We’re trying to replace the value with something like: [EMAIL PROTECTED]/filename”. Is there a way to specify the –install_name argument on the Mac? I tried setting it in our linker flags, then saw it was in there already, and I think it is added by the OSG build. I have looked at the gmane archive, and found some mention of –install_name, but nothing I can see how to use. thanks andy ___ 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] mac -install_name
Thanks, but that's what I'm trying to avoid. Don't you have to do the -id version for every file (changing its own name) and the -change version for every dependency in every file? I'm hoping I can do something in the link commands and avoid having to patch up the files. andy -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Martins Innus Sent: Tuesday, May 27, 2008 1:03 PM To: OpenSceneGraph Users Subject: Re: [osg-users] mac -install_name Andy, I use install_name_tool -change . and install_name_tool -id ... in a script as a Run Script Build phase. I don't know if you can do it in XCode directly. Seems to work. Let me know if you need more info and I'll dig up the exact script. Martins Andy Skinner wrote: I'll admit to being a little past the limit of my knowledge here ... We would like to specify the -install_name arg with which we link the shared libraries. I think it is because we don't want the paths where these things were built to be in the files. We're trying to replace the value with something like: @load_path/filename. Is there a way to specify the -install_name argument on the Mac? I tried setting it in our linker flags, then saw it was in there already, and I think it is added by the OSG build. I have looked at the gmane archive, and found some mention of -install_name, but nothing I can see how to use. thanks andy ___ 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] Getting world space normal
Well, not so much a requirement as OpenGL convention. GL defines a normal as a 3-component vector (there are only the glNormal3*() calls, unlike glVertex which has many sizes) And the GLSL builtins : gl_NormalMatrix is a mat3 and gl_Normal is a vec3. But sure, you can make the math with larger number of components. -- mew On Tue, May 27, 2008 at 9:41 AM, Paul Melis [EMAIL PROTECTED] wrote: Mike Weiblen schreef: Normals must be transformed by a normal matrix (which is a mat3). It is not correct to promote a normal to a vec4 and try to use the transformation matrix (which is a mat4). Just curious: the mat3 requirement comes from the OpenGL shading language? As mathematically there's nothing wrong with representing a normal with (x, y, z, 1) as long as the 4x4 matrix being multiplied with is the inverse transponse of the point transformation (as you remarked) and has a zero translation part. Paul You'll need to generate a new normal matrix, which is defined as the upper-left mat3 of the inverse transpose of the transformation matrix. -- mew On Tue, May 27, 2008 at 8:30 AM, zhangguilian [EMAIL PROTECTED] wrote: Hi, I want to get the normal in world space, I can not use gl_NormalMatrix because it will translate the normal from local space to eye space, from you hint I do in this way: mat4 ToWorldMatrix = osg_ViewMatrixInverse* gl_ModelViewMatrix; WorldPos= ToWorldMatrix * gl_Vertex; vec4 normal=vec4(gl_Normal,0.0); WorldNrm= (ToWorldMatrix * normal).xyz; WorldNrm=normalize(WorldNrm); the worldpos seems right, but worldNrm seems has problem, Can I get the world space normal in this way? Thanks very much for your consideration! zhangguilian 2008-05-27 ___ 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 -- Mike Weiblen -- Austin Texas USA -- http://mew.cx/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] mac -install_name
Yeah, thats what I do, but I only have a couple files in my case I need to worry about. Although I just tried adding: -install_name @executable_path/../Frameworks/file.dylib to LDFLAGS and that seemed to work. I do this in a Makefile project and use the resulting dylib in a larger XCode project. Then i didn't have to do the install_name_tool voodoo. Martins Andy Skinner wrote: Thanks, but that's what I'm trying to avoid. Don't you have to do the -id version for every file (changing its own name) and the -change version for every dependency in every file? I'm hoping I can do something in the link commands and avoid having to patch up the files. andy -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Martins Innus Sent: Tuesday, May 27, 2008 1:03 PM To: OpenSceneGraph Users Subject: Re: [osg-users] mac -install_name Andy, I use install_name_tool -change . and install_name_tool -id ... in a script as a Run Script Build phase. I don't know if you can do it in XCode directly. Seems to work. Let me know if you need more info and I'll dig up the exact script. Martins Andy Skinner wrote: I'll admit to being a little past the limit of my knowledge here ... We would like to specify the -install_name arg with which we link the shared libraries. I think it is because we don't want the paths where these things were built to be in the files. We're trying to replace the value with something like: @load_path/filename. Is there a way to specify the -install_name argument on the Mac? I tried setting it in our linker flags, then saw it was in there already, and I think it is added by the OSG build. I have looked at the gmane archive, and found some mention of -install_name, but nothing I can see how to use. thanks andy ___ 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] mac -install_name
When I added -install_name something to our LDFLAGS, it complained that the option was in there twice. So I assumed CMAKE was putting it in. Maybe I should look to see if we're doing it somewhere. andy -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Martins Innus Sent: Tuesday, May 27, 2008 2:24 PM To: OpenSceneGraph Users Subject: Re: [osg-users] mac -install_name Yeah, thats what I do, but I only have a couple files in my case I need to worry about. Although I just tried adding: -install_name @executable_path/../Frameworks/file.dylib to LDFLAGS and that seemed to work. I do this in a Makefile project and use the resulting dylib in a larger XCode project. Then i didn't have to do the install_name_tool voodoo. Martins Andy Skinner wrote: Thanks, but that's what I'm trying to avoid. Don't you have to do the -id version for every file (changing its own name) and the -change version for every dependency in every file? I'm hoping I can do something in the link commands and avoid having to patch up the files. andy -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Martins Innus Sent: Tuesday, May 27, 2008 1:03 PM To: OpenSceneGraph Users Subject: Re: [osg-users] mac -install_name Andy, I use install_name_tool -change . and install_name_tool -id ... in a script as a Run Script Build phase. I don't know if you can do it in XCode directly. Seems to work. Let me know if you need more info and I'll dig up the exact script. Martins Andy Skinner wrote: I'll admit to being a little past the limit of my knowledge here ... We would like to specify the -install_name arg with which we link the shared libraries. I think it is because we don't want the paths where these things were built to be in the files. We're trying to replace the value with something like: @load_path/filename. Is there a way to specify the -install_name argument on the Mac? I tried setting it in our linker flags, then saw it was in there already, and I think it is added by the OSG build. I have looked at the gmane archive, and found some mention of -install_name, but nothing I can see how to use. thanks andy ___ 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] Parallel Split Shadow Map (Update)
Hi Adrian, I have run out of day, in my day long merge spree, so I'll need to start up again tomorrow, at which point I'll check out your latest changes. If you have any further updates to the ones sent of the 23rd May please just send them through, Cheers, Robert. On Fri, May 23, 2008 at 8:11 AM, Adrian Egli OpenSceneGraph (3D) [EMAIL PROTECTED] wrote: Latest Version to Test, i tested this on: ATI X1600 M : Windows Vista osgshadow --pssm --SingleThreaded --screen 0 -3 2008/5/21 Adrian Egli OpenSceneGraph (3D) [EMAIL PROTECTED]: Hi all again, now i came from ATI X1600M test, same artifact as you described, small shadowed line. i will try to figure out where it comes from. i guess we have to set another polygonoffset. 0.5,0.5 is may not as good as it seams to be. may -0.5,-0.5 or -1.0,-1.0 will work for ATI and NVidia. please check this with the option for polygonoffset and tell me what is a good offset. thanks adrian 2008/5/21 Ralph Kern [EMAIL PROTECTED]: Adrian Egli OpenSceneGraph (3D) schrieb: Hi all, i still come around with some changes. :-) now the near - far clip plane are more optimized, but i hope it's shows still correct shadows for all of your test cases. Its updated with latest PSSM version (from this morning,MEZ) http://webaddon3d.assoftware.ch/AE/PSSM_SHADOW_TEST_OpenSceneGraph_WIN32.rar (use the demo.bat) and in the attachement you will find the latest source. test osgshadow --pssm --SingleThreaded --no-base-texture just checked again. I removed the other two REM's on PATH and LIBRARY_PATH to get the program running. Works on WinXP SP2 GeForce FX5700, but slow frame rate (2.3 fps). On WinXP SP2 ATI X1600 dual monitor the park scenes run fine using screen 0 with 70 fps. There is a small lit line ( 1 pix wide) around the buildings like a self shadow offset which is to large. Chess scene is still entirely black. regards Ralph ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org -- Adrian Egli -- Adrian Egli ___ 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] Debugging example hangs in Cygwin and need some insight into cygwin_osgdb_osg.dll
Robert Osfield wrote: The bug certainly seems to be a cygwin bug, unless the OSG's handle is being corrupted for some reason. This code works fine on other platforms so I wouldn't expect the later to be an issue. Perhaps you could write a small example that just loadeds the osg plugin, then unloads it all using dlopen/dlclose calls directly rather than using osgDB/osgconv, this would isolate the problem to just standard unix calls. I was going to try that but here is another interesting catch - if I do: osgconvd cow.osg brian.osg (or the equivalent in the debugger) I get the hang as I documented but If I do: osgconvd cow.osg brian.jpg or osgconvd test.jpg brian.cow I simply get a message: Error no data loaded. This almost seems as if iot has something to do with the actually writing of the osg file when it writes the data and then something not terminating as it should. bk ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Counting visible vertex
osgOQ is no longer an extension -- OcclusionQueryNode is part of OSG v2.4. However, as currently implemented, it does not return to you a count of the number of visible vertices. OSG, like many rendering doesn't provide direct support to return the number of visible vertices. However, you should be able to use any general algorithms that exist to perform this task. To determine the pixel size of a bounding sphere, you can use the distance and radius along with the combined modelview and projection matrices. OcclusionQueryNode could probably be modified easily to return a count of the number of visible pixels of the bounding box, but it doesn't currently do this, as that was not a design goal. -Paul _ From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of a.lagnier Sent: Tuesday, May 27, 2008 10:28 AM To: osg-users@lists.openscenegraph.org Subject: [osg-users] Counting visible vertex Hi all, I want to calculate number of no occlusing vertex. But, I don't know how determinate visibility. Must I use osgOQ extension? I have an other question : how can I access to pixel size of an object on screen? I think I must use Cullstack::pixelSize(boundingsphere). I want it for determinate what LOD is visible (with RangeMode PIXEL_SIZE_ON_SCREEN). Maybe there is an other way to determinate it... Any idea ? Thanks, Aude Créez votre adresse électronique [EMAIL PROTECTED] http://www.laposte.net 1 Go d'espace de stockage, anti-spam et anti-virus intégrés. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Please test SVN of OpenSceneGraph and VirtualPlanetBuilder in prep for dev releases
Hi All, Today I've begun clearing the osg-submissions back log. I've merged the majority of the backlog, and will continue tomorrow morning and then put out a OSG-2.5.1 and VPB-0.9.9 dev releases. There has been lots of build changes checked in, which, fingers crossed will solve build problems such as when using CMake 2.6, and without breaking build on prior versions of CMake. However, all this needs a serious work out across all platforms so I'd appreciate testing on SVN before I go a tag a dev release. Thanks in advance for you time and effort testing, Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] mac -install_name
On 5/27/08, Andy Skinner [EMAIL PROTECTED] wrote: I'll admit to being a little past the limit of my knowledge here ... We would like to specify the -install_name arg with which we link the shared libraries. I think it is because we don't want the paths where these things were built to be in the files. We're trying to replace the value with something like: @load_path/filename. Is there a way to specify the -install_name argument on the Mac? I tried setting it in our linker flags, then saw it was in there already, and I think it is added by the OSG build. I have looked at the gmane archive, and found some mention of -install_name, but nothing I can see how to use. thanks andy The old, now mostly defunc Xcode project set the install_name via the normal Xcode mechanisms. We set the path to @executable_path/../Frameworks In addition to manually adding the flags to the generic CMake flags, CMake has new direct support for this property. We're adding this to the framework build in CMake. (dylibs are less clear about what the default convention should be.) Also, regardless, CMake also builds with the max path padding so you can use the install_name_tool to redefine the install_name after the fact. I'm struggling with time constraints to finish the CMake/OS X integration. If anybody wants to help, please let me know. I currently have put up a Mercurial repository at: http://www.assembla.com/wiki/show/OpenSceneGraphHg There is an example of the install_name stuff in CMake in CMakeModules/OsgMacroUtils.cmake. (Grep for 'FRAMEWORK'). Thanks, Eric ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Please test SVN of OpenSceneGraph and VirtualPlanetBuilder in prep for dev releases
Hi Robert, There has been lots of build changes checked in, which, fingers crossed will solve build problems such as when using CMake 2.6, and without breaking build on prior versions of CMake. However, all this needs a serious work out across all platforms so I'd appreciate testing on SVN before I go a tag a dev release. Windows Vista, MS Visual Studio 2005, both CMake 2.4.7 and CMake 2.6 On both CMake versions, OSG built fine. VPB will need some of the same changed made to OSG to support CMake 2.6 correctly, but it builds fine with CMake 2.4.7. Good work, J-S -- __ Jean-Sebastien Guay[EMAIL PROTECTED] http://www.cm-labs.com/ http://whitestar02.webhop.org/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Please test SVN of OpenSceneGraph and VirtualPlanetBuilder in prep for dev releases
Hi Robert, OSG built fine on linux (suse 10.3) with cmake 2.4.7. Checked examples also they are working properly cheers RJ Robert Osfield wrote: Hi All, Today I've begun clearing the osg-submissions back log. I've merged the majority of the backlog, and will continue tomorrow morning and then put out a OSG-2.5.1 and VPB-0.9.9 dev releases. There has been lots of build changes checked in, which, fingers crossed will solve build problems such as when using CMake 2.6, and without breaking build on prior versions of CMake. However, all this needs a serious work out across all platforms so I'd appreciate testing on SVN before I go a tag a dev release. Thanks in advance for you time and effort testing, 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] threading trouble with slave cameras
Hi, I just tried to replace SceneView with osgViewer::Viewer to take advantage of threading. My scene has 3 views of the same scene with cameras set up like this: Viewer camera / | \ far_cam1far_cam2 far_cam3 | | | near_cam1 near_cam2 near_cam3 (POST_RENDER) (POST_RENDER)(POST_RENDER) \ | / scene The viewer camera is only used for clearing. The near and far cameras are used to render different depth ranges of my scene to prevent a lot of z-fighting; each pair of near/far cameras draws into the same viewport. With no threading, I use 1 core at 100%. With CullThreadPerCameraDrawThreadPerContext I use 1 core at 100% and another core at about 25%. There are still 2 cores that are nearly unused. The other threading modes do not appear to help. I tried to trace through the threading code in ViewerBase a little. In cameras (the vector of osg::Cameras) there is only 1 camera, but I expected to find 7. Is it impossible to have 7 cull threads in this situation, or am I just setting something up wrong? - Terry ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Please test SVN of OpenSceneGraph and VirtualPlanetBuilder in prep for dev releases
Bravo, Robert! Works great: * Mac OSX 10.5, cmake 2.4.8 producing gcc makefiles; * WinXP SP3, cmake 2.6 producing VS2005 sln/vcproj files. -Paul ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Openthreads problem on 64 bit windows system.
I'm having a problem with an existing application that runs well on a 32-bit windows XP (VS 2003) system, but does not seem to work when ran on a 64-bit windows XP (VS 2003) system. I narrowed down the problem to Openthreads, seems that my threaded while loop never (or rarely) gets called. I was able to recreate the problem with this simple program below. It runs two threads, both printing a message to the screen. Only one message is displayed on a 64bit machine. I turned on OSG notification to DEBUG, but no extra information is displayed. Has anyone else experience this problem? Am I doing something wrong? Thanks in advance. /// //Header file MutexTest.h /// #ifndef __MUTEXTEST_H_ #define __MUTEXTEST_H_ #include OpenThreads/Thread #include OpenThreads/Mutex #include OpenThreads/Barrier #include OpenThreads/Block #include OpenThreads/ScopedLock #include iostream class MutexTest : public OpenThreads::Thread { public: MutexTest(); ~MutexTest(); //Functions virtual void run(); void quit(); void printHello(); void setMutex(OpenThreads::Mutex* newValue); OpenThreads::Mutex* getMutex(); private: mutable OpenThreads::Mutex* _mutex; }; #endif /// //Source file MutexTest.cpp /// #include MutexTest.h MutexTest::MutexTest(){ _mutex = new OpenThreads::Mutex(); } MutexTest::~MutexTest(){ std::coutMutexTest Test Shuting down.std::endl; } void MutexTest::setMutex(OpenThreads::Mutex* newValue){ _mutex = newValue; } OpenThreads::Mutex* MutexTest::getMutex(){ return _mutex; } void MutexTest::quit() { } void MutexTest::printHello(){ OpenThreads::ScopedLockOpenThreads::Mutex lock(*_mutex); std::cout-- Mutex Class: Hello --std::endl; } void MutexTest::run() { while(true){ // //This only gets called once when running on 64bit system / printHello(); //sleep OpenThreads::Thread::microSleep(5000); }//while }//run // --- // // main // // --- int main(int argC, char* argV[]) { //Create a mutex we can share OpenThreads::Mutex* mutable my_mutex = new OpenThreads::Mutex; MutexTest* testMutex = new MutexTest(); testMutex-setMutex(my_mutex); //Start thread testMutex-start(); while(true) { OpenThreads::ScopedLockOpenThreads::Mutex lock(*my_mutex); std::cout Main: Hello std::endl; //sleep OpenThreads::Thread::microSleep(2000); }//while return 0; }//main ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Getting world space normal
Hello Mike and Paul, I can easily get the point transformation matrix in this way: mat4 ToWorldMatrix = osg_ViewMatrixInverse* gl_ModelViewMatrix; WorldPos= ToWorldMatrix * gl_Vertex; How can I get the inverse transpose of the ToWorldMatrix and set its translation part to zero? There seems to be no inverse and transpose funtions in glsl. Thank you very much! guilianzhang 2008-05-28 发件人: Mike Weiblen 发送时间: 2008-05-28 01:36:52 收件人: OpenSceneGraph Users 抄送: 主题: Re: [osg-users] Getting world space normal Well, not so much a requirement as OpenGL convention. GL defines a normal as a 3-component vector (there are only the glNormal3*() calls, unlike glVertex which has many sizes) And the GLSL builtins : gl_NormalMatrix is a mat3 and gl_Normal is a vec3. But sure, you can make the math with larger number of components. -- mew On Tue, May 27, 2008 at 9:41 AM, Paul Melis [EMAIL PROTECTED] wrote: Mike Weiblen schreef: Normals must be transformed by a normal matrix (which is a mat3). It is not correct to promote a normal to a vec4 and try to use the transformation matrix (which is a mat4). Just curious: the mat3 requirement comes from the OpenGL shading language? As mathematically there's nothing wrong with representing a normal with (x, y, z, 1) as long as the 4x4 matrix being multiplied with is the inverse transponse of the point transformation (as you remarked) and has a zero translation part. Paul You'll need to generate a new normal matrix, which is defined as the upper-left mat3 of the inverse transpose of the transformation matrix. -- mew On Tue, May 27, 2008 at 8:30 AM, zhangguilian [EMAIL PROTECTED] wrote: Hi, I want to get the normal in world space, I can not use gl_NormalMatrix because it will translate the normal from local space to eye space, from you hint I do in this way: mat4 ToWorldMatrix = osg_ViewMatrixInverse* gl_ModelViewMatrix; WorldPos= ToWorldMatrix * gl_Vertex; vec4 normal=vec4(gl_Normal,0.0); WorldNrm= (ToWorldMatrix * normal).xyz; WorldNrm=normalize(WorldNrm); the worldpos seems right, but worldNrm seems has problem, Can I get the world space normal in this way? Thanks very much for your consideration! zhangguilian 2008-05-27 ___ 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 -- Mike Weiblen -- Austin Texas USA -- http://mew.cx/ ___ 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] issues with geometry and 2.4.0 multi-threaded
i have just freshly build 2.4.0 from source acquired from the zip file on the downloads page for VC8 on WinXP. i have ran the osgviewer app using both a custom obj file and the cow.osga from the new osg data archive. it seems that in all instances of multithreaded applications the loaded geometry becomes disorganized into a horrible mess. if set the threading model to single threaded for the viewer, all is well. i can only imagine im missing something rather obvious thanks ~jason coposky ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] issues with geometry and 2.4.0 multi-threaded
Are you on a multimonitor NVIDIA system by any chance? If so, this is a known issue in the NVIDIA device driver. Not sure what the workaround was -- I think it had something to do with additional MakeCurrent calls. You should check the archives; this was discussed within the past month. You might try the current SVN head to see if it resolves this issue. Paul Martz Skew Matrix Software LLC http://www.skew-matrix.com +1 303 859 9466 -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Jason Coposky Sent: Tuesday, May 27, 2008 8:22 PM To: osg-users Subject: [osg-users] issues with geometry and 2.4.0 multi-threaded i have just freshly build 2.4.0 from source acquired from the zip file on the downloads page for VC8 on WinXP. i have ran the osgviewer app using both a custom obj file and the cow.osga from the new osg data archive. it seems that in all instances of multithreaded applications the loaded geometry becomes disorganized into a horrible mess. if set the threading model to single threaded for the viewer, all is well. i can only imagine im missing something rather obvious thanks ~jason coposky ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-opensce negraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] can't remove all PrimitiveSet
Paul , thank you. Yes, the right way is : *geometry-removePrimitiveSet( 0 , geometry-getNumPrimitiveSets() );* The problem is caused by my another code. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org