Re: [osg-users] vertex_array_object branch merged with master
Hi Curtis, Thanks for testing things out. Potentially we could de-instate the osg::State::dirtyTexCoordPointer() as there osg::VertexArrayState takes over this functionality, you can get the current VertexArrayState via state.getCurrentVertexArrayState(). I removed the functions like dirtyTexCoordPointer() to help make sure the rest of the OSG doesn't rely on deprecated methods. I wasn't expect it to be used widely by users. Is there code you had to modify open sourced? If so could you point me at it so I can work why it's being used and the way it's being used. This might help inform how to adapt things. For OSG-3.6 it would be ideal to smooth out glitches that end uses are seeing so any assistance on homing on what the issues are will really help in this. Robert. On 13 October 2016 at 16:46, Curtis Rubel wrote: > Hi Robert, > > I did some very basic testing of this for you yesterday. There is > no need to respond back to me with any possible fixes or anything, > I am just passing this along to you for your information right now. > We have no plans on updating our OSG version in the near future. > > I was able to successfully compile the release on OpenSuSE 13.2 64bit Linux. > > Next I recompiled our Opensource OpenIG application against this release and > everything compiled OK, with exception of one minor change to the source code > for a missing function: > > osg::State* state = renderInfo.getState(); > #if OSG_VERSION_LESS_OR_EQUAL(3,5,3) > state->dirtyTexCoordPointer(1); > #endif > > I used 3.5.3 above only because that is what I know for sure as I do not > have a copy of 3.5.4 to check against. > > After this change our OpenIG application runs with no segv's, however > there are a number of visual issues noted with the 3.5.6 master: > > 1. The one FBX model in the scene is not positioned properly anymore. > Although the shadow for it seems to still be in the correct location and > even shows the animation of that model, even though visually the > model itself is not in that location anymore. > 2.. There appears to be something going on with our terrain shaders too > as when I change the TOD to 10:00, the terrain all turns bright white > from its normal textures. In fact just moving around in the scene > also results in the terrain textures disappearing totally. > 3. Seems that our F+ lighting implementation will also need some > adjustments as it appears not to be rendering properly anymore. > > I know very little about how the shaders were implemented in > this application as most of this was done by our consultant, just passing > on some of the basic results as I seem them right now for you. > > All of this is working properly in the 3.5.3 version we have been using > so at some point between this and whats on the master branch now these > visual changes have shown up. > > Anyhow I know its just some very basic info, but I figured I would pass > it along for you in case this is not what you might have expected with > this latest release. > > We will be curious to see what if anything anyone else reports back to you. > > ... > > Thank you! > > Cheers, > Curtis > > -- > Read this topic online here: > http://forum.openscenegraph.org/viewtopic.php?p=68993#68993 > > > > > > ___ > 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] osgversion command fails with symbol lookup error on new clone of OpenSceneGraph 3.5.5 on Linux
On 13 October 2016 at 16:23, Alberto Luaces wrote: > Robert, > > by the way, due to those recent changes in the OT API, I think it would > be a good move to bump the soname to 21. Totally overlooked the need to do this - I updated the OSG's SO_VERSION but totally forgot about OpenThreads having it's own. I have now bumped the version and so version of OpenTheads and checked it into git master. Thanks for the vigilance, Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] vertex_array_object branch merged with master
Hi Robert, I did some very basic testing of this for you yesterday. There is no need to respond back to me with any possible fixes or anything, I am just passing this along to you for your information right now. We have no plans on updating our OSG version in the near future. I was able to successfully compile the release on OpenSuSE 13.2 64bit Linux. Next I recompiled our Opensource OpenIG application against this release and everything compiled OK, with exception of one minor change to the source code for a missing function: osg::State* state = renderInfo.getState(); #if OSG_VERSION_LESS_OR_EQUAL(3,5,3) state->dirtyTexCoordPointer(1); #endif I used 3.5.3 above only because that is what I know for sure as I do not have a copy of 3.5.4 to check against. After this change our OpenIG application runs with no segv's, however there are a number of visual issues noted with the 3.5.6 master: 1. The one FBX model in the scene is not positioned properly anymore. Although the shadow for it seems to still be in the correct location and even shows the animation of that model, even though visually the model itself is not in that location anymore. 2.. There appears to be something going on with our terrain shaders too as when I change the TOD to 10:00, the terrain all turns bright white from its normal textures. In fact just moving around in the scene also results in the terrain textures disappearing totally. 3. Seems that our F+ lighting implementation will also need some adjustments as it appears not to be rendering properly anymore. I know very little about how the shaders were implemented in this application as most of this was done by our consultant, just passing on some of the basic results as I seem them right now for you. All of this is working properly in the 3.5.3 version we have been using so at some point between this and whats on the master branch now these visual changes have shown up. Anyhow I know its just some very basic info, but I figured I would pass it along for you in case this is not what you might have expected with this latest release. We will be curious to see what if anything anyone else reports back to you. ... Thank you! Cheers, Curtis -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=68993#68993 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] osgversion command fails with symbol lookup error on new clone of OpenSceneGraph 3.5.5 on Linux
Robert, by the way, due to those recent changes in the OT API, I think it would be a good move to bump the soname to 21. Thanks! Alberto Robert Osfield writes: > Hi Curtis, > > Curious, what Linux variant are you using? > > FYI, I develop under Kubuntu 16.04 at present, so all the new Affinity > work written in Linux and I've run osgversion a number of times and > not seen any issues with missing linkage. For dev work I don't > usually install the OSG though, instead setting my LD_LIBRARY_PATH to > pick up the OSG libs where they were built. This means I don't have > to do a make install each time I modify the OSG codebase, I wouldn't > expect it to change behaviour w.r.t linkgage, but perhaps there is > some relationship. > > I just checked the applications/osgversion/CMakeLists.txt and it > includes OpenThreads on the link line so it should work. > > Is there any chance you've only install the updated OSG but not the > updated OpenThreads? > > Robert. > > > On 11 October 2016 at 18:31, Curtis Rubel wrote: >> Hi, >> >> After cloning out a copy of OpenSceneGraph 3.5.5 from GitHub, >> and running cmake with default options, then enabled building >> of examples and documentation on Linux only. >> >> After clean build with no errors, running make install;sudo ldconfig >> >> osgversion command fails with the following results: >> >> osgversion >> osgversion: symbol lookup error: /usr/local/lib64/libosgDB.so.145: undefined >> symbol: _ZN11OpenThreads6Thread20setProcessorAffinityERKNS_8AffinityE >> >> Actually seems all the default built binaries fail with the same error. >> >> Are there some other cmake flags, defines or compile options that need to be >> specified before compiling 3.5.5 and have it run properly on 64bit Linux? >> >> If more data is needed, just let me know what/how to get it for you >> and I will do my best to get it. >> >> Cheers, >> Curtis >> >> -- >> Read this topic online here: >> http://forum.openscenegraph.org/viewtopic.php?p=68963#68963 >> >> >> >> >> >> ___ >> osg-users mailing list >> osg-users@lists.openscenegraph.org >> http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org -- Alberto ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] LOD for drawing lines of a co-ordinate system
Hi, I had the idea to use an osg::LOD for drawing a co-ordinate system in a geometry program. The closer you get, the more fine-grained lines you see. This works fine as long as you zoom in to (0,0,0). If you zoom to a different point, it doesn´t work, as the distance from eye to center is still too big. I tried to use PIXEL_SIZE_ON_SCREEN, but then I see nothing all the time (probably because I use glLines for the coordinate-system which are very fine). I could solve this by breaking the coordinate-system into several nodes, thus the center would be more accurate. But before doing that: Is there an easier way? Thanks for your comments, Andreas PS: Is it allowed to post pictures here, small ones of course? ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Vector script out of range
Hai Robert, Yeah I am using same project setting. I am trying this simple code in windows 8 system. robertosfield wrote: > Hi Uma, > > I have load the h.obj and trail.3ds files under linux without problems > using OSG git master. > > As Laurens suggestions it may be making sure that you are building > your own application with the same setting settings a the OSG version > you are linking to. When setting paths to the plugins it's also worth > setting the path to the parent directory of the plugins other than the > plugin directory itself as the OSG appends a version number to the > plugin directory so when you link to different versions of the OSG it > can choose the correct plugins directory to avoid problems with mixing > dll's. > > Robert. > > On 7 October 2016 at 05:21, Uma Devi Selvaraj <> wrote: > > > Hi all, > > > > Thanks for the reply.I am facing error not only for .ply models. For other > > models like .obj, .dxf and .3ds. I have attached both the call stack of my > > program and one of the model I used.(call stack for all models are same) > > > > > > > > ... > > > > Thank you! > > > > Cheers, > > Uma > > > > -- > > Read this topic online here: > > http://forum.openscenegraph.org/viewtopic.php?p=68898#68898 > > > > > > > > > > ___ > > osg-users mailing list > > > > http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org > > > > > ___ > osg-users mailing list > > http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org > > -- > Post generated by Mail2Forum -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=68990#68990 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Pragmatic shader composition - bug
Success, it works (tested on the 3.4 branch). Cheers. Glenn Waldron On Wed, Oct 12, 2016 at 12:26 PM, Robert Osfield wrote: > Hi Glenn, > > >> Is actually wrong and should be: > >> > >> if (versionLine[versionLine.versionLine.size()-1]!='\n') > >> versionLine.push_back('\n'); > >> > > I have checked this change into OSG master and OpenSceneGraph-3.4 branch. > > Could you let me know if that works fine. > > 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
Re: [osg-users] Vector script out of range
Hi , I have cloned osg repo and checked out to the branch 3.4.0. I am facing same problem for bunny.ply model (Smooth Triangle).Hope Laurens changes are merged. ... Thank you! Cheers, Uma -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=68988#68988 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Off scale in StackedMatrixElement
Hi, I have osgAnimation::Animation, which uses StackeMatrixElement as transform. Problem is that I get scale animation instead of rotation, even when I set only rotate and translate for matrix. Are there some ways to off scale animations in StackedMatrixElement transforms? Thank you! Cheers, Lyubov -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=68987#68987 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org