[osg-users] IVE referenced textures...
Hi All, Sorry to bother you but I have an FLT file that references a number of textures via relative paths, and I wish to save this scenegraph as an IVE file, maintaining the relative path references to the textures so that the IVE doesn't become too large. When I save the scene graph as an IVE all my textures disappear. Could anyone suggest what I might be doing wrong please ? Thanks for any help. Kind regards Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 3DS loader revisited?
Hi Sukender Robert, Just to emphasise what has already been said really, but we have literally 10's of thousands of 3DS models for various manufacturer's products, so whilst I accept there are significant limitations with 3DS structure now compared to more modern formats, the sheer size of the back catalogue means it will be around for a long time to come. Hence I think support for 3DS is important within OSG. Robert's comments about the 3DS reader supporting streams is vital for us now with our web based catalogue viewing system. In regards to the ability to write a 3DS file, this would be fantastic. If help/assistance/testing is wanted I'm sure I'd be able to find some time. If you need sample models these could probably be provided as well. In regards to Collada, my recent excursions into conversions from 3DS to collada via osg has had mixed results. Still investigating (when time permits) but there is certainly an issue in validity of the collada file produced if the node name of the scene graph has spaces, also issues encountered with materials. However, these are for another thread. But I tend to agree that even with the teething problems, Collada does look to be very powerful. Is anyone looking at the new Autocad format - FBX? Just wondering whether anyone considering a Reader/Writer for it? Looked at trying to take an OSG collada file and convert it to FBX via Autocad's converter, but it failed. End of memory dump :-) Kind regards Neil. Robert Osfield robert.osfi...@gmail.com wrote: Hi Sukender, On Fri, Aug 21, 2009 at 8:35 AM, Sukendersuky0...@free.fr wrote: Of course, my goal is to work on the 3DS reader/writer without loosing functionality. Actually, I simply tried to replace the old lib3DS with the new one... and it was very straightforward: some calls were renamed, some others changed a bit... One areas of significant deviation was for the support of reading from istreams, in your work can you still read from istreams? The ability to read from istream enables reading from http via the curl plugin. Anyway I could load my 3DS models exactly the same way I did before. I could even remove a tiny limitation (meshes were taversed by name, which can be dangerous if two identical names exist. That should not, but now it's okay). Everything worked like a charm! I guess some testing would be a good idea, but it seems there was no functionality loss. Side note: I had a few compilation errors using VC9 in the lib3DS, because of implicit conversions from void* to SomethingElse*. I made those conversions explicit and gave the fix to the lib3DS patches tracker. I had to make a number of changes to fix the pedantic warnings emitted by g++ 4.x. This might need to be rolled in as well. About linking, I guess putting lib3DS as a static library could be a good idea, even if my testing was with direct inclusion of source files. I'll work on it and submit VC (8? 9?) binaries I guess... or may I keep the files as they are in a first time? We could possibly just include the source to lib3ds as a subdirectory of the 3ds plugin, as CMake allows us to manage the source directories better than the old days of trying to support both VS6 and Makefiles. About the writer, well, we have some customers with old software... Moreover they're absolutely not 3D/vis/sim-oriented, so there are few supported formats. I agree 3DS is very poor, but it is a non-flexible format (as some XML-based formats could be); and as a non-flexible format it can be loaded exactly the same way on various (old) software. What kind of output format would you use else? Collada? A major problem with 3ds is that because there is no official/public specification and lots of different attempts at implementations the files generated can be very patchy in the their consistency. 3ds is not likely to map a general purpose scene graph too well either, but this issue applies to most writers - a general purpose scene graph is far to flexible to be easily mapped by the constraints of most formats. Collada has to be one of the best for being able to map things, but even it is a long way from perfect. 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] Ask for osgviewerMFC memory consumption explanations
Hi Alex, There could be any number of reasons why this could be happening, and obviously some down to coding, however if you are using caching at all, then it maybe down to how long you've told the cache to keep the models, even when not used. I can't remember what the default is, but if set to never get rid of the models, then the only time their memory will be released is when the application shuts down - not merely closing and destroying the window. This leads on to another reason why you maybe seeing the memory usage you report in that, depending on the design of your app architecture, if you leave the management of dll's to the O.S., then dll's are only loaded on an as-required basis, and are only unloaded when the app shuts down. OSG has a number of static memory constructs singletons that are created at dll load time, and are not released until dll unload time (app shut down). One way around this is to manage the dll load and unload yourself, however this can be very tiresome ;-( I would suggest that, unless you are witnessing a continual increase in the memory allocated to your app, even when you are loading and unloading the same model, then I wouldn't be too concerned. Obviously if you are witnessing unrestricted memory consumption then something is up, but I suspect not at the OSG end. Finally, in my experience, beware of the false positive memory leaks reported by Visual Studio. Plenty of back talk on this in the archives, but suffice it to say VS unloads MFC too early in the app shut down process, which triggers a dump of unreleased memory warnings. VS crystal ball not working too well as when OSG dll's unload - usually after MFC due to VS problem with scheduling library loads/unloads - it releases its own memory. Where real leaks actually exist, more often than not it can be traced back to app coding errors, or inappropriate use of non ref_ptr references to scenegraph entries. I've wasted days in the past on this. You can decrease some of the false positives by forcing the load of the OSG dll's yourself, but sometimes life is just too short to be bothered ! Hope this helps. Kind regards Neil. Alexandre Amalric alex.pix...@gmail.com wrote: Hi osg-users, I'm developping an application based on the example osgviewerMFC but I don't understand the memory consumption from this example. 1) when I launch osgviewerMFC.exe the memory used is about 10Mo 2) then I open a model, memory rise to let's say 50Mo (depends on the model) 3) then I close the window and memory do not get back to 10Mo like desired but stay at let say 30Mo (in my case). Why osgviewerMFC.exe use much memory after opening and closing a model than before doing nothing ?? I am asking this because I have the same behaviour with my own application, I suppose something isn't well destroyed when closing a window (removing a viewer). Kind regards, -- Alexandre AMALRIC Ingénieur RD === PIXXIM S.A. 73E, rue Perrin-Solliers 13006 Marseille http://www.pixxim.fr ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Problem with Collada plugin...
Hi All, Has anyone else had a problem with the collada plugin whereby if the name of an osg node has a space in it, then if this is assigned to a url, source or target attribute - as is the case for geometries and materials - the spaces are replaced with %20. This would be ok, but the actual definition for the corresponding collada item doesn't get its ID attribute changed to have %20. The collada file therefore references instances of a geometry as #A%20B but the geometry definition has an ID=A B. Happy to look at fixing this, just wanted to check if anyone else has encountered, fixed and submitted. Had a quick look, but couldn't spot any submission. Thanks Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Collada (dae) Reader/Writer issue....
Hi J-S, Thanks for getting back to me. I'll look to test this today. I suppose a related question would be whether the Collada reader/writer should be amended to make use of this explicitly itself? The issue being that given this is a graphics driver issue, and our software may be run on hundreds of machines, it would be a nice comfort factor to know that this particular problem isn't likely to arise. Many thanks for thoughts and help, Neil. Jean-Sébastien Guay jean-sebastien.g...@cm-labs.com wrote: Hi Neil, I'm reordering your questions a bit to make the response more linear :-) Has anyone considered this already and come up with a really neat solution ? I recognise that coding around driver issues is not ideal, and may not be desireable. Yes, the osgUtil::Optimizer supports what you want. So you could use one of the osgUtil::Optimizer classes to the work for you (independently of whether your input model was COLLADA or any other format). Optionally you could just call osgconv with the appropriate OSG_OPTIMIZER environment variable on your model. Here are the relevant classes/environment options: One of the comments I read on the mail archive seemed to imply that in general to avoid these scaling problems, the model vertices should be scaled directly, and not via matrix operations. osgUtil::FlattenStaticTransformsVisitor FLATTEN_STATIC_TRANSFORMS I guess there are some complications here as if differing nodepaths to the same model produce differing scaling factors, the collada export is going to need to write geometry definitions for each of these nodepath scaled model instances. osgUtil::FlattenStaticTransformsDuplicatingSharedSubgraphsVisitor (wow what a mouthful :-) FLATTEN_STATIC_TRANSFORMS_DUPLICATING_SHARED_SUBGRAPHS I have never used those myself, but they should work. Let us know how you get on. The Optimizer also has a lot more things it can do to help you optimize your scene graph, so it's worth it to explore the options (osgconv --help-env) and the osgUtil/Optmizer header. Hope this helps, J-S -- __ Jean-Sebastien Guayjean-sebastien.g...@cm-labs.com 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 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Collada (dae) Reader/Writer issue....
Hi J-S, I've now had chance to have a look at your suggested optimisations, and I think they form the basis of a solution for me, so thanks for that. However there is one thing which troubles me a little, which is that the FLATTEN_STATIC_TRANSFORMS_DUPLICATING_SHARED_SUBGRAPHS optimisation, whilst flattening the static scalings, applies those scalings to the normals as well as the vertices. Whilst I can understand that this might be desirable for some strange effects, would you agree that the normals should at least be normalised after this process, else undesirable lighting anomalies can occur? Neil. Jean-Sébastien Guay jean-sebastien.g...@cm-labs.com wrote: Hi Neil, I'm reordering your questions a bit to make the response more linear :-) Has anyone considered this already and come up with a really neat solution ? I recognise that coding around driver issues is not ideal, and may not be desireable. Yes, the osgUtil::Optimizer supports what you want. So you could use one of the osgUtil::Optimizer classes to the work for you (independently of whether your input model was COLLADA or any other format). Optionally you could just call osgconv with the appropriate OSG_OPTIMIZER environment variable on your model. Here are the relevant classes/environment options: One of the comments I read on the mail archive seemed to imply that in general to avoid these scaling problems, the model vertices should be scaled directly, and not via matrix operations. osgUtil::FlattenStaticTransformsVisitor FLATTEN_STATIC_TRANSFORMS I guess there are some complications here as if differing nodepaths to the same model produce differing scaling factors, the collada export is going to need to write geometry definitions for each of these nodepath scaled model instances. osgUtil::FlattenStaticTransformsDuplicatingSharedSubgraphsVisitor (wow what a mouthful :-) FLATTEN_STATIC_TRANSFORMS_DUPLICATING_SHARED_SUBGRAPHS I have never used those myself, but they should work. Let us know how you get on. The Optimizer also has a lot more things it can do to help you optimize your scene graph, so it's worth it to explore the options (osgconv --help-env) and the osgUtil/Optmizer header. Hope this helps, J-S -- __ Jean-Sebastien Guayjean-sebastien.g...@cm-labs.com 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 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Texture object preparation...
Hi All, Is there a way I can get a Texture to perform its preparation at construction time, rather than when it is first required for rendering? Essentially I'm noticing a significant lag on the first frame that requires a Texture to be rendered. I've tracked it down to the Texture::apply() where it checks to see if it has a textureobject for the contextID. If it doesn't, then it goes off and does some work (I presume for sending the texture to the render pipeline). This is done once per context per Texture. What I would like to do is to take this hit when first constructing the Texture object, rather than when flying in space, as it causes a significant lag in frame rate (the simulation appears to stagger). Does anyone have any suggestions please ? Many thanks Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Image file PagedLOD...
Hi All, Has anyone come across a scenario where they wished to have PagedLOD that instead of handling Nodes, handle images ? Essentially I find myself in a position where I have a model that references a texture, but that texture is quite large, and even with reducing the texture size with little compromise in quality, the time to stream over the web is too large. So, what I'm considering is replacing the image load call, with a similar concept to that of PagedLOD's whereby I could give the PagedImageLOD object a list of ordered image file names that I want it to progressively load - in much the same way that apps like google map appears to work. The first image file could be a 2x2 pixel image, whilst the nth image file would be the full-on image. I would want the PagedImageLOD to return a valid Image node immediately - perhaps a hard coded small default image - whilst in the background the specified image filesnames are being downloaded, and replace the content of the created image node. The reason for this sort of functionality would be that I could then intercept all image load requests via my own callback handler, replace with a PagedImageLOD request, and still permit most plugins to work happily on the assumption that a model plugin that loads an image is probably only trying to reference that image in some stateset context. Sorry for the thought dump, but has anyone tried/needed to do a similar thing, and if so is there an example available ? Alternatively, if you can think of a better way that would be good as well. Most obvious alternative - that of have an LOD node on the model - would be fine if I only had one or two models. However, I have thousands of different models, so can't really ask our modellers to redo all these. Many thanks for any help/thoughts. Kind regards Neil. .to whose content was progressively replaced by the downloaded images. In this way I could ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Image file PagedLOD...
Thanks Jason, I'll take a look later today. Jason Beverage jasonbever...@gmail.com wrote: Hi Neil, The osgEarth project does essentially what you are looking for. Check it out at http://www.osgearth.org Thanks, Jason On Fri, Jun 12, 2009 at 5:08 AM, neil.hug...@tesco.net wrote: Hi All, Has anyone come across a scenario where they wished to have PagedLOD that instead of handling Nodes, handle images ? Essentially I find myself in a position where I have a model that references a texture, but that texture is quite large, and even with reducing the texture size with little compromise in quality, the time to stream over the web is too large. So, what I'm considering is replacing the image load call, with a similar concept to that of PagedLOD's whereby I could give the PagedImageLOD object a list of ordered image file names that I want it to progressively load - in much the same way that apps like google map appears to work. The first image file could be a 2x2 pixel image, whilst the nth image file would be the full-on image. I would want the PagedImageLOD to return a valid Image node immediately - perhaps a hard coded small default image - whilst in the background the specified image filesnames are being downloaded, and replace the content of the created image node. The reason for this sort of functionality would be that I could then intercept all image load requests via my own callback handler, replace with a PagedImageLOD request, and still permit most plugins to work happily on the assumption that a model plugin that loads an image is probably only trying to reference that image in some stateset context. Sorry for the thought dump, but has anyone tried/needed to do a similar thing, and if so is there an example available ? Alternatively, if you can think of a better way that would be good as well. Most obvious alternative - that of have an LOD node on the model - would be fine if I only had one or two models. However, I have thousands of different models, so can't really ask our modellers to redo all these. Many thanks for any help/thoughts. Kind regards Neil. .to whose content was progressively replaced by the downloaded images. In this way I could ___ 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] PluginStrData....
Hi Robert, I hope you're over your cold/flu now? I'm looking at the PluginStrData work , and wondered whether you think the setPluginStrData and getPluginStrData should be made case insensitive? I was thinking to force the 'set' parameters through the osgDB::convertToLowerCase function. It would make checking for specific option string settings a little cleaner, and more robust. Neil ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Regression in .zip loader?
Hi Ulrich Robert, I've just done a cursory scan of the code again, and can't quite see what could be going wrong. However, I do note that the GetZipItem function can return an error state that the code isn't checking for. I'll try and have a look at it tomorrow, and set up a similar example for a 3ds node and a jpeg, and see what's happening there. However I would note in passing that I tested with a zip file that contained four 3ds models all zipped within the same file, and these worked file. Ulrich, could you send me an example zip file that causes this issue please? If you could exclude .x as I don't regularly deal with this format. 3ds/osg/ive/flt/jpg/tga would be fine if possible please. I could then investigate with a little more certainty that the example fails on your setup. Many thanks, Neil. Ulrich Hertlein u.hertl...@sandbox.de wrote: Hi Robert, On 25/5/09 10:24 AM, Robert Osfield wrote: On Sun, May 24, 2009 at 6:00 PM, Ulrich Hertleinu.hertl...@sandbox.de wrote: have there been any recent changes in the .zip loader? I just noticed that I can no longer load a bunch of zipped .x files. OSG_NOTIFY_LEVEL=DEBUG shows that all plugins are found (.zip, .x, .bmp) but I still get 'No data loaded' from osgviewer. If I unpack the .zip file and load the .x directly everything is fine. Arggh two steps forward, one back... As Neil mentioned I'ved merged his changes to the zip plugin to allow in memory unpacking, and it then use istreams to read the final modes, and it looks like the .x plugin doesn't support reading from streams yet. ... handle models zipped up and placed on http. The ideal solution would be to add streams support into .x plugin and any others that it's missing for. I've added stream support to the .x plugin but the problem is deeper than that. It seems like there's a serious bug in the zip loader. If there are multiple files in the archive only the first one is read properly. All the following ones are fed the same data. So for example if you have a .zip that contains a .jpg and an .ive the JPEG is read properly but then the JPEG data is fed to the .ive plugin which of course doesn't work. If the .ive is first then that is loaded and displayed but any following file is not read properly. Cheers, /ulrich ___ 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] Regression in .zip loader?
Hi Ulrich, If you've taken the recent trunk checkins, then the answer is yes, there have been some recent changes to the .zip loader. However, I had hoped that there would be no downsides to the amendments made. Essentially, where as previously the expansion of the .zip file was performed to the local filing system - ie your hard disk - now the unpacking process is all done in memory. There were some issues in getting the modification to compile on unix boxes, but I believe Robert solved that, and another maillist member made a modification to get it to compile on Mac's. So I thought it was all working. One issue that may now arise, and something that Robert and I are pondering, is where a model references a texture contained within the same zip file. At the moment, using the in-memory unpacking, this is not supported. Essentially we need to figure out how to use the .zip as an archive that can be interrogated for files during a model load operation. Hopefully Robert or I will have something on this in the near future. Until then...you could rename the original_readnode function in the .zip plugin to replace the new readnode (this would put you back to unpacking to the local file system). I understand that Robert is offline for a few days with a nasty cold/flu, but when he's back, we could look to make an amendment to provide an options switch to control which unpacking method is used. I may submit this anyway, and leave it to Robert to decide on acceptance. Hope the info helps, and sorry for any inconvenience that the amendment may have caused. Kind regards Neil. Ulrich Hertlein u.hertl...@sandbox.de wrote: Hi all, have there been any recent changes in the .zip loader? I just noticed that I can no longer load a bunch of zipped .x files. OSG_NOTIFY_LEVEL=DEBUG shows that all plugins are found (.zip, .x, .bmp) but I still get 'No data loaded' from osgviewer. If I unpack the .zip file and load the .x directly everything is fine. Cheers, /ulrich ___ 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] Taking it easy for a couple of days
Hi Robert, Hope you get well soon. Coincidently, this article appeared on the bbc today. So manflu does exist :-) http://news.bbc.co.uk/1/hi/health/8047321.stm Kind regards Neil. Robert Osfield robert.osfi...@gmail.com wrote: Hi All, I've gone down with a bad cold, so am struggling to focus, read and think clearly today. So... please don't expect much support from me today, some active threads will just have do without my contribution for a couple of days. Luckily human virus can't be transmitted by email/forums just yet. 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] Change cursor with object under mouse
Hi Romain, The way I did this was with a custom PickHandler. If you look at the PickHandler example - I think its the scribeFX one - you basically get the mouse events coming in. It gives you an opportunity to do your line intersections to determine the object under the mouse, at which point you can decide what cursor you wish to change to. Hope this helps. Neil. Romain Charbit romain.char...@gmail.com wrote: Hi, May I ask your guidance for a problem that I have. I want to change the cursor of the mouse in function of the object that is under of it. Example, when I'm passing the mouse over a dragger, I want to change the cursor to a Hand. I'm on the detection of the object for the moment. I'm using a handler which use a LineSegmentIntersector on every frame. It works but it's kind of slow. I've seen on the web about the GL_SELECT OpenGL renderMode which has been design for it. But I've also seen that it's not a good way to do that. Maybe someone has already done something like this or has an idea about it? Thank you! Cheers, Romain Charbit Romain Charbit -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=11520#11520 ___ 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] The mystery of the Disappearing Texture....
Hi All, I have an application that employs OSG in a window. I load a scene with textures and everything is fine. If I move the window to a different screen, the textures disappear ! I move the window back to the original screen, the textures reappear. Back-and-forth my window goes from screen to screen and the textures flash on and off accordingly. Clearly I've forgotten to do something, but I can't see what. Could someone shed some light on this please. Hair loss is increasing by the minute ;-( Thanks in advance. Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] The mystery of the Disappearing Texture....
Hi Robert, Thanks for the info. Guess I'll stick to one monitor for the moment ;-( Neil. Robert Osfield robert.osfi...@gmail.com wrote: Hi Neil, This will be a driver dependent issue, the driver/OS managing graphics contexts behind the scenes for you, but not doing it 100% correctly. Personally I much prefer window managers/drivers that don't attempt to be clever on your behalf, and just done allow such windows to be dragged unless they are shared on the same graphics card and with the same display properties as the tricks involved with duplicating contexts behind the scenes is both wasteful of resources and performance and driver complexity. Best thing you an do is try to update your drivers. Robert. On Tue, May 5, 2009 at 11:42 AM, neil.hug...@tesco.net wrote: Hi All, I have an application that employs OSG in a window. I load a scene with textures and everything is fine. If I move the window to a different screen, the textures disappear ! I move the window back to the original screen, the textures reappear. Back-and-forth my window goes from screen to screen and the textures flash on and off accordingly. Clearly I've forgotten to do something, but I can't see what. Could someone shed some light on this please. Hair loss is increasing by the minute ;-( Thanks in advance. Neil. ___ 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] custom search paths...
Hi Robert, Thanks for the reply, I'll look into this approach. Neil. Robert Osfield robert.osfi...@gmail.com wrote: Hi Neil, The ReaderWriter::Options structure has a data file path list built into it, this is checked before the main data file path list stored in the Registry is checked. So if you want a certainly directory checked just put it on your path, either via the ReaderWriter::Options or via the Registry using osgDB::s/getDataFilePathList. You'd typically use Options when you only want to locally set the search path. Robert. On Thu, Apr 30, 2009 at 5:29 PM, neil.hug...@tesco.net wrote: Hi All, I'd like to pick your collective brains for a moment in regards to search paths for files within OSG. I'm currently working on the 3DS plugin for OSG, looking at a little problem. The scenario is as follows. I have a 3DS file whose material table references a texture. When this model was created the texture resided in the same directory as the model, and hence no path information was stored - at least I guess that's why there is no path information for the texture referenced. Now when I want to load this model, the 3DS loader tries to locate the texture in the directory that the model resides in - using the findfileindirectory function - but as the texture is no longer in this directory, the function fails to find the file, and the 3DS reader decides not to put a texture on the geometry. Now, the thing is, I do know where the texture is relative to the model. Its in a parallel directory. In fact, if I ignore the result of the find file, and merely just try and load the image, my readfilecallback handler amends the path to the file, and the image loads. So, what I was wondering was whether there already exists a way by which I can direct the findfileindirectory function to take account of a custom search algorithm? I've had a look and I can't spot one, but wondered if others might know differently? On the assumption that no such methodology exists, my thoughts were that I could amend the existing function to call a custom defined function on the user supplied readfilecallback - if one has been supplied - in the event where the readfileindirectory function has failed. Any thoughts ? Thanks for any assistance/comments. Kind regards, Neil. ___ 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] custom search paths...
Hi Robert, I've had a look at this, and whilst it will address the specific scenario I described, I left out a little top spin to the scenario that has some bearing on the issue. Not only do I want to specify the directory where the search looks, but I also wish to amend the file extension. Essentially the issue is that when we originally started out modelling - circa 10 years ago now - we only used TGA files. Now we tend to prefer to load jpg files, unless there is an alpha channel in which case we'll load TGA's. For space reasons, and speed of load/download we therefore want to say that if a tga file is referenced, try loading, but if not found, try the jpg equivalent. I can handle this at the load image time within my supplied readfilecallback, if I circumvent the search results. However, I'd like the search routines to be able to handle this type of functionality as well. i.e. I request Pear.TGA, the search try's to locate it, but if it fails it tries some alternative extensions. Now I recognise that this is somewhat specific to myself, but rather than have customised 3DS readers, I was hoping to extend the search routines in a general way to permit an application to supply its own search algorithm, that could be used in the event that the file was not found with the current search implementations. I recognise that for the majority of people either the standard search path, or the options route you describe, is sufficient. Its just that I can't see a way to amend the extension/path according to some application defined rules using this approach. Robert, I think I'll code up a solution, and perhaps submit it for consideration if that's ok with you, and then we'll have a framework to bat about. Kind regards Neil. a jpgour models were producedmodelled, neil.hug...@tesco.net wrote: Hi Robert, Thanks for the reply, I'll look into this approach. Neil. Robert Osfield robert.osfi...@gmail.com wrote: Hi Neil, The ReaderWriter::Options structure has a data file path list built into it, this is checked before the main data file path list stored in the Registry is checked. So if you want a certainly directory checked just put it on your path, either via the ReaderWriter::Options or via the Registry using osgDB::s/getDataFilePathList. You'd typically use Options when you only want to locally set the search path. Robert. On Thu, Apr 30, 2009 at 5:29 PM, neil.hug...@tesco.net wrote: Hi All, I'd like to pick your collective brains for a moment in regards to search paths for files within OSG. I'm currently working on the 3DS plugin for OSG, looking at a little problem. The scenario is as follows. I have a 3DS file whose material table references a texture. When this model was created the texture resided in the same directory as the model, and hence no path information was stored - at least I guess that's why there is no path information for the texture referenced. Now when I want to load this model, the 3DS loader tries to locate the texture in the directory that the model resides in - using the findfileindirectory function - but as the texture is no longer in this directory, the function fails to find the file, and the 3DS reader decides not to put a texture on the geometry. Now, the thing is, I do know where the texture is relative to the model. Its in a parallel directory. In fact, if I ignore the result of the find file, and merely just try and load the image, my readfilecallback handler amends the path to the file, and the image loads. So, what I was wondering was whether there already exists a way by which I can direct the findfileindirectory function to take account of a custom search algorithm? I've had a look and I can't spot one, but wondered if others might know differently? On the assumption that no such methodology exists, my thoughts were that I could amend the existing function to call a custom defined function on the user supplied readfilecallback - if one has been supplied - in the event where the readfileindirectory function has failed. Any thoughts ? Thanks for any assistance/comments. Kind regards, Neil. ___ 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
Re: [osg-users] custom search paths...
Hi Robert, I suspect we may be saying the same thing, and perhaps I haven't explained myself clearly. All I'm suggesting is a slight amendment to the search functions to delegate to an application supplied readfilecallback the role of searching when the default search has failed to locate the file. OSG would be agnostic as far as the actual alternative search algorithm is concerned. It merely invokes it and returns the result. Neil. Robert Osfield robert.osfi...@gmail.com wrote: Hi Neil, This type of customization is exactly what ReadFileCallback is for. I don't feel it's appropriate for search routines to do anything other than search, having them remap file names is inappropriate. Robert. On Fri, May 1, 2009 at 11:32 AM, neil.hug...@tesco.net wrote: Hi Robert, I've had a look at this, and whilst it will address the specific scenario I described, I left out a little top spin to the scenario that has some bearing on the issue. Not only do I want to specify the directory where the search looks, but I also wish to amend the file extension. Essentially the issue is that when we originally started out modelling - circa 10 years ago now - we only used TGA files. Now we tend to prefer to load jpg files, unless there is an alpha channel in which case we'll load TGA's. For space reasons, and speed of load/download we therefore want to say that if a tga file is referenced, try loading, but if not found, try the jpg equivalent. I can handle this at the load image time within my supplied readfilecallback, if I circumvent the search results. However, I'd like the search routines to be able to handle this type of functionality as well. i.e. I request Pear.TGA, the search try's to locate it, but if it fails it tries some alternative extensions. Now I recognise that this is somewhat specific to myself, but rather than have customised 3DS readers, I was hoping to extend the search routines in a general way to permit an application to supply its own search algorithm, that could be used in the event that the file was not found with the current search implementations. I recognise that for the majority of people either the standard search path, or the options route you describe, is sufficient. Its just that I can't see a way to amend the extension/path according to some application defined rules using this approach. Robert, I think I'll code up a solution, and perhaps submit it for consideration if that's ok with you, and then we'll have a framework to bat about. Kind regards Neil. a jpgour models were producedmodelled, neil.hug...@tesco.net wrote: Hi Robert, Thanks for the reply, I'll look into this approach. Neil. Robert Osfield robert.osfi...@gmail.com wrote: Hi Neil, The ReaderWriter::Options structure has a data file path list built into it, this is checked before the main data file path list stored in the Registry is checked. So if you want a certainly directory checked just put it on your path, either via the ReaderWriter::Options or via the Registry using osgDB::s/getDataFilePathList. You'd typically use Options when you only want to locally set the search path. Robert. On Thu, Apr 30, 2009 at 5:29 PM, neil.hug...@tesco.net wrote: Hi All, I'd like to pick your collective brains for a moment in regards to search paths for files within OSG. I'm currently working on the 3DS plugin for OSG, looking at a little problem. The scenario is as follows. I have a 3DS file whose material table references a texture. When this model was created the texture resided in the same directory as the model, and hence no path information was stored - at least I guess that's why there is no path information for the texture referenced. Now when I want to load this model, the 3DS loader tries to locate the texture in the directory that the model resides in - using the findfileindirectory function - but as the texture is no longer in this directory, the function fails to find the file, and the 3DS reader decides not to put a texture on the geometry. Now, the thing is, I do know where the texture is relative to the model. Its in a parallel directory. In fact, if I ignore the result of the find file, and merely just try and load the image, my readfilecallback handler amends the path to the file, and the image loads. So, what I was wondering was whether there already exists a way by which I can direct the findfileindirectory function to take account of a custom search algorithm? I've had a look and I can't spot one, but wondered if others might know differently? On the assumption that no such methodology exists, my thoughts were that I could amend the existing function to call a custom defined
[osg-users] Cache system...
Hi All, I'm looking for some thoughts, and advice, on the cache process within OSG. Essentially a while ago (OSG1.2) I wrote an app that implemented its own node/texture cache in memory as I didn't want to be writing things to the file system as permissions were potentially a problem. I'm now looking at rewriting in OSG2.8, and trying to get a handle on whether the cacheing offered by OSG has changed at all. Essentially I want to cache objects in a memory cache, rather than to file. As far as I can see this isn't an option within OSG2.8, but I wanted to check if others new anything to the contrary ? Thanks for any help/comments. Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Cache system...
Hi All, I'm looking for some thoughts, and advice, on the cache process within OSG. Essentially a while ago (OSG1.2) I wrote an app that implemented its own node/texture cache in memory as I didn't want to be writing things to the file system as permissions were potentially a problem. I'm now looking at rewriting in OSG2.8, and trying to get a handle on whether the cacheing offered by OSG has changed at all. Essentially I want to cache objects in a memory cache, rather than to file. As far as I can see this isn't an option within OSG2.8, but I wanted to check if others new anything to the contrary ? Thanks for any help/comments. Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Cache system...
Hi Stephan, Thanks for coming back to me. I think I confused myself with the filecache concept. thanks for the hint. I've now tracked through and found out what is happening, so it looks like I don't need to implement my own cache, which is great. Thanks again. neil. Stephan Maximilian Huber ratzf...@digitalmind.de wrote: Hi neil, neil.hug...@tesco.net schrieb: Hi All, I'm looking for some thoughts, and advice, on the cache process within OSG. Essentially a while ago (OSG1.2) I wrote an app that implemented its own node/texture cache in memory as I didn't want to be writing things to the file system as permissions were potentially a problem. I'm now looking at rewriting in OSG2.8, and trying to get a handle on whether the cacheing offered by OSG has changed at all. Essentially I want to cache objects in a memory cache, rather than to file. As far as I can see this isn't an option within OSG2.8, but I wanted to check if others new anything to the contrary ? Thanks for any help/comments. Have you tried the internal cache of osg? I think it defaults to off, so you have to switch it on via: osgDB::ReaderWriter::Options* options = new osgDB::ReaderWriter::Options(); options-setObjectCacheHint(osgDB::ReaderWriter::CACHE_ALL); osgDB::Registry::instance()-setOptions(options); For more infos see the documentation of osgDB::ReaderWriter::Options HTH, Stephan ___ 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] ObjectType querry...
Hi Robert, Please ignor last email, and this request. I've just realised what to do. Put it down to temporary moment of insanity on my part. Thanks for help. Neil. neil.hug...@tesco.net wrote: Hi Robert, Thanks for coming back to me so quickly, however I don't think I made myself clear enough, as I'm not sure that osgconv is what I'm looking for. I'm sitting within the zip plugin code, and I have the stream of bytes that represents one of the files within the zip file. I need to know whether I should be passing this stream into a ReadImage, readNode, readObject, or readArchive function - like the Curl plugin does. Now, I can get the appropriate reader as the registry provides functions for doing that. What I need to know, within code, is what type of object that stream represents, so that I can call the appropriate function within the plugin code at runtime. I was hoping that a plugin might publish the type of object it creates via a call, but I've looked at the plugin codes and all that is really available seems to be the extensions that a reader addresses, and not the type of object it returns. So, for example, I would like the JPG,TGA,BMP,PNG...(A.N.Other image format) plugins to return IMAGE as the object type they create creates, whilst the OSG, 3DS, FLT, IVE and (A.N.Other object format) would return node. Any thoughts ? Thanks in advance. Neil. Robert Osfield robert.osfi...@gmail.com wrote: Hi Neil, In OSG-2.8 you can do osgconv --formats to get a listed of all the capabilities of the plugins. Robert. On Wed, Apr 1, 2009 at 5:29 PM, neil.hug...@tesco.net wrote: Hi All, Can anyone tell me whether there is a way (call) in OSG to say what type of object a plugin will produce ? i.e. if I have a file A.jpg and I call osgDB::Registry::instance()-getReaderWriterForExtension passing in the file extension, I get the readerwriter. What I would like to know is what ObjectType that readerwrite would return. The reason for the question is I'm updating the ReaderWriterZip plugin to handle in-memory zip extraction, and I want to be able to give it a zip file that contains many files of varying format. Rippling through the zip file I can get the individual files, and their content, and unzip them in memory, but the issue I have is that I don't really want to have a big If statement in the plugin to look at the known file extensions and to say whether I should be calling ReadNode or ReadImage. I would like to be able to enquire the type of Object that the reader would produce, and then have a simple switch statement like the Curl plugin. One route I thought of would be to add to extend each plugin to have a function based that returns the type of object, and I guess this would be fairly straight forward, but before delving in, I just wanted to check if there is a method already available. Many thanks for any help. Neil. ___ 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] ObjectType querry...
Hi All, Can anyone tell me whether there is a way (call) in OSG to say what type of object a plugin will produce ? i.e. if I have a file A.jpg and I call osgDB::Registry::instance()-getReaderWriterForExtension passing in the file extension, I get the readerwriter. What I would like to know is what ObjectType that readerwrite would return. The reason for the question is I'm updating the ReaderWriterZip plugin to handle in-memory zip extraction, and I want to be able to give it a zip file that contains many files of varying format. Rippling through the zip file I can get the individual files, and their content, and unzip them in memory, but the issue I have is that I don't really want to have a big If statement in the plugin to look at the known file extensions and to say whether I should be calling ReadNode or ReadImage. I would like to be able to enquire the type of Object that the reader would produce, and then have a simple switch statement like the Curl plugin. One route I thought of would be to add to extend each plugin to have a function based that returns the type of object, and I guess this would be fairly straight forward, but before delving in, I just wanted to check if there is a method already available. Many thanks for any help. Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] SetAttribute crash...
Hi All, I have a standalone windows application (MDI) that embeds OSG in a window, loads an object, applies a texture, and displays the object for viewing. This works fine. No problems at all. I've also got a more complicated ActiveX control that does the same sort of thing, and this works fine as a release build running on an XP platform. However, when I move the ActiveX control to a Vista box, I get a crash in the Release build when trying to call setTextureAttributeAndModes on the stateset object of the drawable for my 3D object. In an attempt to identify what's happening, I've recompiled OSG (I should say OSG1.2, sorry) with a few protected: classifications removed so that I can look at the values a bit easier, and step into some code. Having done this, it looks like when the setAttribute function is eventually called from setTextureAttributeAndModes, as soon as it executes the line itr-second.first = attribute; the system dies a horrible death, taking down the web browser in which the activeX control is running. Looking at the call stack it seems to be occuring when trying to clean up the dereferencing of the previously assigned texture (this was assigned as part of the 3DS load operation). I can't believe that its the OSG code as such as the standalone version runs fine, however I was wondering whether anyone else has had a similar issue, and if so, could you give any tips as to what might be going wrong? Many thanks for any help. Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] SetAttribute crash...
Hi All, I have a standalone windows application (MDI) that embeds OSG in a window, loads an object, applies a texture, and displays the object for viewing. This works fine. No problems at all. I've also got a more complicated ActiveX control that does the same sort of thing, and this works fine as a release build running on an XP platform. However, when I move the ActiveX control to a Vista box, I get a crash in the Release build when trying to call setTextureAttributeAndModes on the stateset object of the drawable for my 3D object. In an attempt to identify what's happening, I've recompiled OSG (I should say OSG1.2, sorry) with a few protected: classifications removed so that I can look at the values a bit easier, and step into some code. Having done this, it looks like when the setAttribute function is eventually called from setTextureAttributeAndModes, as soon as it executes the line itr-second.first = attribute; the system dies a horrible death, taking down the web browser in which the activeX control is running. Looking at the call stack it seems to be occuring when trying to clean up the dereferencing of the previously assigned texture (this was assigned as part of the 3DS load operation). I can't believe that its the OSG code as such as the standalone version runs fine, however I was wondering whether anyone else has had a similar issue, and if so, could you give any tips as to what might be going wrong? Many thanks for any help. Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] SetAttribute crash...
Hi All, I have a standalone windows application (MDI) that embeds OSG in a window, loads an object, applies a texture, and displays the object for viewing. This works fine. No problems at all. I've also got a more complicated ActiveX control that does the same sort of thing, and this works fine as a release build running on an XP platform. However, when I move the ActiveX control to a Vista box, I get a crash in the Release build when trying to call setTextureAttributeAndModes on the stateset object of the drawable for my 3D object. In an attempt to identify what's happening, I've recompiled OSG (I should say OSG1.2, sorry) with a few protected: classifications removed so that I can look at the values a bit easier, and step into some code. Having done this, it looks like when the setAttribute function is eventually called from setTextureAttributeAndModes, as soon as it executes the line itr-second.first = attribute; the system dies a horrible death, taking down the web browser in which the activeX control is running. Looking at the call stack it seems to be occuring when trying to clean up the dereferencing of the previously assigned texture (this was assigned as part of the 3DS load operation). I can't believe that its the OSG code as such as the standalone version runs fine, however I was wondering whether anyone else has had a similar issue, and if so, could you give any tips as to what might be going wrong? Many thanks for any help. Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] SetAttribute crash...
Hi All, Email client going wrong again. Seems like three instances of message sent. Sorry. Neil. neil.hug...@tesco.net wrote: Hi All, I have a standalone windows application (MDI) that embeds OSG in a window, loads an object, applies a texture, and displays the object for viewing. This works fine. No problems at all. I've also got a more complicated ActiveX control that does the same sort of thing, and this works fine as a release build running on an XP platform. However, when I move the ActiveX control to a Vista box, I get a crash in the Release build when trying to call setTextureAttributeAndModes on the stateset object of the drawable for my 3D object. In an attempt to identify what's happening, I've recompiled OSG (I should say OSG1.2, sorry) with a few protected: classifications removed so that I can look at the values a bit easier, and step into some code. Having done this, it looks like when the setAttribute function is eventually called from setTextureAttributeAndModes, as soon as it executes the line itr-second.first = attribute; the system dies a horrible death, taking down the web browser in which the activeX control is running. Looking at the call stack it seems to be occuring when trying to clean up the dereferencing of the previously assigned texture (this was assigned as part of the 3DS load operation). I can't believe that its the OSG code as such as the standalone version runs fine, however I was wondering whether anyone else has had a similar issue, and if so, could you give any tips as to what might be going wrong? Many thanks for any help. Neil. ___ 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] SetAttribute crash...
Hi robert thanks for the response. I sense the bullet heading my way ;-( will ponder a little longer. Neil. Robert Osfield robert.osfi...@gmail.com wrote: Hi Neil, The best thing you could look at is to check that the object you are trying to set is a valid object, and hasn't been deleted. Beyond this I don't think there is much we could do to help. And as gentle, perhaps it's time to updated to OSG-2.8, as it time passes it become increasingly difficult to support OSG-1.2 users as the knowledge of this old rev is a bit of distant memory for most of us. If you want to use these old rev's then you'll need to increasingly take responsibility for support of it, or... just bite the bullet and upgrade, it needn't be a difficult experience. Robert. On Mon, Mar 23, 2009 at 9:59 AM, neil.hug...@tesco.net wrote: Hi All, I have a standalone windows application (MDI) that embeds OSG in a window, loads an object, applies a texture, and displays the object for viewing. This works fine. No problems at all. I've also got a more complicated ActiveX control that does the same sort of thing, and this works fine as a release build running on an XP platform. However, when I move the ActiveX control to a Vista box, I get a crash in the Release build when trying to call setTextureAttributeAndModes on the stateset object of the drawable for my 3D object. In an attempt to identify what's happening, I've recompiled OSG (I should say OSG1.2, sorry) with a few protected: classifications removed so that I can look at the values a bit easier, and step into some code. Having done this, it looks like when the setAttribute function is eventually called from setTextureAttributeAndModes, as soon as it executes the line itr-second.first = attribute; the system dies a horrible death, taking down the web browser in which the activeX control is running. Looking at the call stack it seems to be occuring when trying to clean up the dereferencing of the previously assigned texture (this was assigned as part of the 3DS load operation). I can't believe that its the OSG code as such as the standalone version runs fine, however I was wondering whether anyone else has had a similar issue, and if so, could you give any tips as to what might be going wrong? Many thanks for any help. Neil. ___ 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] PNG not displaying correctly..
Hi Paul, That worked fine. I shall investigate further. Thanks for the help Neil. Paul Martz pma...@skew-matrix.com wrote: Try osgviewer --image filename and see if you get different results. Paul Martz Skew Matrix Software LLC 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 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] PNG not displaying correctly..
Hi Stephan, I'm not sure. I shall investigate further. Paul's suggestion has demonstrated that the osgviewer displays the png correctly, so I guess that, coupled with you suggestion, gives me a good starting point for investigating what's wrong at my end. Many thanks. Neil Stephan Maximilian Huber ratzf...@digitalmind.de wrote: neil.hug...@tesco.net schrieb: Hi All, Please find attached a png file that has a purple circle, black boarder, and an alpha channel. I'm using OSG1.2, and trying to read it in via the png pluggin, and apply it to an object. The PNG loads, but the alpha channel is being ignored, and I get solid white instead of the underlying object color/texture. Please could someone tell me is there something wrong with the png file, the reader, or am I doing something that's not possible. Many thanks for any help. there was no attachment. Have you enabled GL_BLEND for your geometry? node-getOrCreateStateSet()-setMode(GL_BLEND, osg::StateAttribute::ON); cheers, Stephan ___ 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] PNG not displaying correctly..
Hi All, Please find attached a png file that has a purple circle, black boarder, and an alpha channel. I'm using OSG1.2, and trying to read it in via the png pluggin, and apply it to an object. The PNG loads, but the alpha channel is being ignored, and I get solid white instead of the underlying object color/texture. Please could someone tell me is there something wrong with the png file, the reader, or am I doing something that's not possible. Many thanks for any help. Kind regards Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] memory leak false positives on Windows
Hi All,
As a simple user of OSG in a Microsoft VS environment, I would like to say that
this one issue is the most frustrating concern I have about continuing to use
OSG. I recognise that for those who don't use Microsoft VS the problem doesn't
occur, but, for a large proportion of OSG users, I suspect, this is an on going
problem that they, like I, encounter every single day.
I note that recently the OSG pages were updated to welcome over 2000
subscribers/users to the mailing list. I think that's wonderful, but I can't
help thinking that as that number grows, more and more will be developers on a
Micrsoft platform, and whose first attempts at using OSG are going to be
plagued with false positive memory leaks. I can't see this as a good advert for
the product - which I happen to think is seriously good.
It appears that whenever this issue is raised, readers are directed to the
archives to read up on the discussions, only to come away with the conclusion
that we as programmers are being told that to fix the issue we really need to
invest in a piece of software that doesn't fall fowl of this problem. However,
I for one am deafened by the silence when it comes to someone giving a concrete
recommendation as to which IDE software we should therefore purchase. As a
programmer I simply want to write programmes that use Microsoft libraries,
which make use of OSG, and which do not report false positives. I want a nice
front end to the IDE as I am inherrantly lazy, and like point-and-click rather
than command line RSI.
Now, if I understand Cory's comments correctly, he has tracked some of the
issues down to these singletons. Could someone please explain why his
question regarding a restructing of these problem structures/objects doesn't
deserve serious consideration ?
I have wasted days - if not weeks (no exageration here at all) - in trying to
track through programmes to identify what may be a memory leak, only to find
that its another false positive. If a little restructuring could address the
failing of the MS tools, whilst not imposing a significant overhead on other
platforms, why is this not a good thing to consider ?
Sorry to go on about this, but I'm currently faced with a nasty little problem
as to inconsistent bahaviour between compilations of my activeX viewer that
uses OSG. I'm facing the prospect of trying to track through what is happening
with this nagging doubt that I'm not releasing memory when I should be, but I
am swamped by the false positives.
Could someone please help us with these false positives ?
Thanks for listenning. If I have offended, it was not intended.
Neil.
Jean-Sébastien Guay jean-sebastien.g...@cm-labs.com wrote:
Hi Cory,
Can anybody recommend a Windows-based memory tracking tool?
I have not found any tool on Windows that gave me useful results.
Perhaps that was through lack of Google-skills or whatever, but I ended
up changing osg::Referenced to keep track of where objects were
allocated/destroyed. For a start, you can enable
#define DEBUG_OBJECT_REF_UNREF
in src/osg/Referenced.cpp and then recompile OSG and relink with your
app. This will tell you, on the command line, each time an
osg::Referenced object (or subclass) is allocated and destroyed, as well
as a count of the number of objects that are currently allocated. You
will most likely see that the count goes down to 0 at the end of your
program, meaning there are no memory leaks.
Those static objects you saw are not leaks, of course, and in my testing
I got anywhere between 40 and 300 of them in various places. You can
test by, for example, changing the osgviewer application so that the
whole main() function is enclosed in a block ( { } ) and then put a
breakpoint before and after that block. You will probably get a few
objects allocated before anything OSG is called in the main(), and about
47 OSG objects still live after the block. But all those are deleted
once you exit the app.
If you find you do have memory leaks (the memory usage of your app grows
at run time even though you shouldn't be allocating anything), then
check the osg-submissions archives for a modified osg::Referenced
header/implementation which I posted about 2 weeks ago. You can use
this to create an AllocationObserver which will be called at each ref(),
unref() and unref_nodelete() so that you can know where the call was
made. These leaks generally come from some object keeping a ref to an
object when it shouldn't, like a list/vector/map not being cleared or
whatever, so you can print out the call stack at those points, and
inspect all the results to see where the object's last ref is coming
from, for example.
But in general, as Robert said, VS's memory leak check is crap. There
are apparently ways to fix it somewhat, but even then I wouldn't trust it.
Hope this helps,
J-S
--
[osg-users] How does DEEP_COPY_ALL treat textures?
Hi All, Can anyone assist with an issue/misunderstanding I'm having when using DEEP_COPY_ALL please ? The issue is that if I have a cube that I apply a texture to and add to the scenegraph. This displays very quickly and no problems. If then do a clone operation on the cube - say 400 times - adding each cloned node to the scenegraph with a changed parent transform so that I effectively have an arrangement of tiles in a grid like pattern, the realisation of the first frame is very slow. Once first frame is realised, subsequent frame rate is fine. Its almost as if, even though the DEEP_COPY_ALL operation has completed, the nodes are referencing something that the first frame needs to resolve. The only thing I can think of -perhaps erroneously I accept - is that the textures are having to be loaded. Is this the case, and if so, what can I do to speed up realisation of the first frame? Thanks for any help. Kind regards Neil ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] How does DEEP_COPY_ALL treat textures?
Hi All, Can anyone assist with an issue/misunderstanding I'm having when using DEEP_COPY_ALL please ? The issue is that if I have a cube that I apply a texture to and add to the scenegraph. This displays very quickly and no problems. If then do a clone operation on the cube - say 400 times - adding each cloned node to the scenegraph with a changed parent transform so that I effectively have an arrangement of tiles in a grid like pattern, the realisation of the first frame is very slow. Once first frame is realised, subsequent frame rate is fine. Its almost as if, even though the DEEP_COPY_ALL operation has completed, the nodes are referencing something that the first frame needs to resolve. The only thing I can think of -perhaps erroneously I accept - is that the textures are having to be loaded. Is this the case, and if so, what can I do to speed up realisation of the first frame? Thanks for any help. Kind regards Neil ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] How does DEEP_COPY_ALL treat textures?
Hi All, Can anyone assist with an issue/misunderstanding I'm having when using DEEP_COPY_ALL please ? The issue is that if I have a cube that I apply a texture to and add to the scenegraph. This displays very quickly and no problems. If then do a clone operation on the cube - say 400 times - adding each cloned node to the scenegraph with a changed parent transform so that I effectively have an arrangement of tiles in a grid like pattern, the realisation of the first frame is very slow. Once first frame is realised, subsequent frame rate is fine. Its almost as if, even though the DEEP_COPY_ALL operation has completed, the nodes are referencing something that the first frame needs to resolve. The only thing I can think of -perhaps erroneously I accept - is that the textures are having to be loaded. Is this the case, and if so, what can I do to speed up realisation of the first frame? Thanks for any help. Kind regards Neil ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] [RE] How does Deep_Copy work - take 2.
Hi Robert et al, First please accept my apologee for the previous email getting sent 3 times. I don't quite know what happened there, but it certainly wasn't intended. Gremlins in my system somewhere I would guess. Sorry. In regards to the clone operation, is there a more optimum alternative that I should take ? Many thanks for the help. Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] How to deduce the implied size of texture from coordinate mappings....
Hi All, I've had a look through the mail archives, and I don't think this has been asked before, so I was hoping someone might be able give some guidence to me on a question I'm bashing my head with at the moment. Its not strictly an osg question - though I am using osg to try and solve it. Question == I have a general triangle in 3D space, known coordinates, with known S,T texture coordinates. Knowing only this information, is it possible to calculate the size of the texture? Background The background to this problem is that I have an object with texture coordinates, that has applied a texture. I now wish to apply a new texture to this object where I know the size that this texture is mean't to represent in the 3D space. Thus I believe I need to scale the texture coordinates on the object such that the size of the new texture is correctly represented relative to the object. [Brief asside - I suspect that there might be a way to do this other than altering all the texture coordinates, perhaps a scale measure on a texture object but I haven't delved deep enough into osg::Texture2D to know.] In order to scale the texture coordinates I need to know the size that the original texture represented relative to the object/universe it was in. Each geometry could have its own texture, therefore I'm at the geometry level, and I have the triangle with the largest variation in the S,T domain. I can transform this triangle to a cordinate system that is co-planar with the triangle, where one edge forms the first of a pair of orthogonal basis vectors in the triangle plane (e1,e2), and hence I can represent my triangle in this space. The S,T domain is coplanar for my requirements, and with a shift of origin, (0,0) in the S,T domain maps to (0,0) in the e1,e2 domain. So I get to a point where I have the two sets of orthogonal axis in the same triangle plane, where a rotation of the e1,e2 axis would align them to the U,V axis. However, my next step is not clear. Can any one assist please. I may be barking up the wrong tree here, and if so, please accept my apologee. Thanks for any help Neil Hughes. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] RE How to deduce texture size....
Hi Art and Gordon, Thanks for coming back to me. I guess I need to try and get some more info about the the texture that was originally applied. Great support, as ever, from OSG users. Kind regards Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Re How to deduce texture size ....
Hi Art and Gordon, A further thought on this was that I omitted to mention that I do know that the 3D coordinates of the triangles on my object are all in mm. Given that I therefore know the real world size of the triangle, and for each vertex I know the S,T co-ordinate in the texture domain, would this help at all in the math to get a size for the texture applied ? Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Re How to deduce texture size ....
Hi Art and Gordon, A further thought on this was that I omitted to mention that I do know that the 3D coordinates of the triangles on my object are all in mm. Given that I therefore know the real world size of the triangle, and for each vertex I know the S,T co-ordinate in the texture domain, would this help at all in the math to get a size for the texture applied ? Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Re Hoe to deduce texture size...
Hi Gordon, Many thanks for that. I'll need to scratch head and see if more info on texture is available. Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Seeking help in understanding HEADLIGHT light...
Hi Robert, thanks for coming back to me. One thing I left out of my mail was that I'm still using OSG 1.2 - though looking to migrate very soon. Is your response still valid for OSG 1.2? Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Seeking help in understanding HEADLIGHT light...
Hi All, I'm trying to adjust the light that is created as part of the HEADLIGHT option, as it appears to be washing out an another light that I have in my scene. Essentially I'd like to take its effect down so that it only comes into play when I'm close to an object. I've tried adjusting the light on the scenedecorator, and I'm monitoring the apply function on the Light, and I see my values coming through, but I'm also getting the original values coming through. Its as if I have two lights in the room, but I haven't added any, other than the headlight option. In an attempt to track what was happening I've discovered that when the camera is set up, it also has a light created. I don't know if this is important or not. Can anyone help please ? Many thanks Neil ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Coordinate system mapping from WTK to OSG...
Hi Jason, thanks for the comments. I think I've got to the bottom of it now. There is some rather irritating issues associated with which rotation is performed first. Finally got an ordering (trial-and-error) that did what it was supposed to do. Coupled with the rotation about X I feel significantly older now ;-( Once again, thanks for the comments. Regards Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Coordinate system mapping from WTK to OSG...
Hi All, At the risk of appearing very dense, could I ask for a little help with a problem that I have been banging my head against a wall with for the the past week? The problem is that I have a position and orientation of an observer described by a position vector, and a vector of euler angles that represent the orientation of the observer. To complicate matters, the coordinate system in operation when this information was created was that of WorldToolKit (WTK) - in fact we were using WTK when creating this information. In this system positive Y is down, and positive Z is in towards the screen, with positive x traversing left to right. Now, I would like to represent the position and direction of view (orientation) in OSG's coordinate space. Now, as I understand it, a rotation of -90 degrees will take a point in OSG space, and represent it in WTK space - and the inverse doing the reverse. Cool. The issue I have is what should happen to the orientation vector. The euler angles were obtained from WTK's quaternion to euler function. I've looked at the order of multiplications that occur in WTK when creating the resultant rotation, yet I still can't seem to get the correct result. So, if anyone out there has experience of mapping from WTK to OSG, specifically in regards to what should happen to rotation angles, and quarternians, I would be most grateful - and perhaps my wall could then be repainted beige instead of red ;-( Many thanks for any help Neil. I am trying to bring this into OSG. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] What's the mpr.dll ?
Hi, You've just encountered the wonderful world of Windows. If you look up the mpr.dll problem on the internet, you eventually get to find out that the Depends.exe of windows has problems with late binding dependencies. This is really an issue with Depends, not your programme. The best advice I can give - having spent many an hour trying to resolve this myself - is that you can safely ignore this warning, and any to do with shlapi(?) as well. At run time, the relavent stubs run, and everything is fine. If you application is failing on another machine, it is highly unlikely that these warnings can contribute to tracking down what the problem is. If the app is an osg one, did you copy across all the osg dlls? I usualy forget openthreads myself ;-( Hope this shortens your search. Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Highly detailed closed scene navigations...
Hi Jason Robert, Sorry for not getting back to you sooner. Many thanks for the info. I guess I'm going to have to just suck-and-see. The occluder route looks to have the greatest gains for me. Once again, many thanks to yourselves and the forum. Kind regards Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Highly detailed closed scene navigations...
Hi All, Not sure if I've got the correct title for this question, but hopefully the forum might be able to help. Suppose that I wish to model a complete office in OSG. The office is on two floors, the downstairs split into many rooms, the upstairs open plan. Desks, computers, windows, doors and everything that you would find in an office needs to be modelled, down to the pencil on the desk and the clutter in the corner. Resource wise we are talking potentially a few thousand models, each with two LOD's, all with highly detailed, high resolution textures, and many many facets. In fact the design is so big that I can't load it all into memory. What are my options ? My client won't accept any delay in transiting from one area of a design to the next. So multiple design files for the individual areas seems a non starter, unless someone can tell me how I could do this without a delay moving from one area to the next. Some form of preemptive loading springs to mind, but is there an existing mechanism within OSG for handling that sort of thing ? My thoughts led to the database pager, but I haven't used this at all, and would hate to start looking at it only to find a few weeks later that it really isn't suitable for my needs. Can anyone suggest an approach given the constraints please ? Many thanks Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] RE FLT export submission imminent
Hi Paul, Bob Robert, Once again many thanks for the work you guys are doing. Robert, thanks for the advice. I'd thought of this route myself, and have consequently been trying to get the latest SVN, but for some reason it keeps hanging. Probably my company network causing grief ;-( Bob Paul, I take your point, and I would love to be in a position to take you up on a support contract, however, if I can prove the conversion route, I'll invest some time and get a colleague to migrate our work from 1.2 to the latest stable release with patches. Looking to include the shadow work that's been done as well, so this is likely to be the most productive route for ourselves. Once again, thanks for your help. Kind regards Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Building Collada plugin in OSG 1.2....
Hi Steve, Thanks for the info. I'll give Dom1.3 a look just to get up to speed with Collada. Cheers Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] RE FLT export submission imminent (Paul Martz)
Hi Paul, Will your work integrate with OSG1.2 ? Irrespective of that, I will give it a go soonest and give you some feedback. Many thanks for doing this work, as you know I've been champing-at-the-bit for it. Thanks also to your funding source :-) Kind regards Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] RE OSG to Multi-Gen Creator...
Hi Robert, Sorry for the noise. In part I couldn't remember what I'd asked before, but also the emphasis of my request was slightly different in that, as I don't know creator at all and don't even have the application, I don't know what other formats it can take in. I do, however, have a client who needs to get models into that package. I was merely asking whether there was an alternative route to getting an osg scene into creator - even via a third party if necessary - so that I could at least get something for my client whilst the OSG pluggin was being developed further by Paul. I shall assume, from your response, that there isn't really an alternative that I can take, and I shall await the release of the new pluggin from Paul. Once again, sorry for the noise. Kind regards Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] RE OSG to Multi-Gen Creator...
Hi Robert and Gordon, Many thanks for the responses. I think I failed to make my problem clear. Sorry. I have an OSG scene. My client needs to load this into Creator to do some additional work on it. Once he has finished his work, he exports to FLT from Creator, and runs osgconv to convert it to an IVE file. I'm looking for a means of doing the transfer from OSG to Creator. The rest is already sorted. Whilst the FLT export pluggin was being developed I hoped to achieve the export from OSG to Creator via another format. My colleague looked at .obj but the results from the OSG export weren't good (apparently - though I haven't seen them myself). I don't think we tried the COLLADA, but given your response, Gordon, I suspect I'm not going to be successful there. Looked at Polytrans as we already have a version here, but didn't spot any support for COLLADA. Thanks for your help on this. Regards Neil ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] RE. Philips WOW TV displays...
Hi Robert and Kim Many thanks for the info. This should make life a little easier I hope. In regards to the depth resolution, Kim's observations tie up with what we've observed. As we fly closer into the room - a kitchen in our case - the back half of the room suddenly looses depth information, and hence the depth effect is lost. A colleague of mine was wondering whether the depth information is only being used on first frame, and after that is never reset ? I'm not sure myself. Looking for guidance really. Any thoughts ? Once again many thanks for the info so far. Will let you know if we discover anything. Is anyone talking to philips about these issues of loss of depth? Kind regards Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Philips WOW TV displays...
Hi All, Not sure I've asked this before. If I have, sorry. I've got access to a couple of Philips WOW TV's for a few days, and a colleague has been playing with them for the past week. One of the supporting applications is written in OSG, and generates a split screen image. The left half is the standard 3D view, whilst the right half is the depth map. Now I know that I can set this up myself, however we suspect that there is more to it than this, as we've taken a screen grab of this screen image and sent it to the display, and it still doesn't show it as a combined 3D with depth feel. Does anyone know anything about this display, and perhaps could shed some light on what we need to do to get a normal OSG split screen to work with it? We suspect that there are a few calls that we are missing. Any pointers would be really useful. By the way, using the supplied OSG viewing with our room it looks really quite cool. I was a bit of a sceptic at first, but it seems that Philips may have something really neat here. If you get a chance, check it out. Regards Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] RE Plugin loading problems....
Hi Robert Gordon, Thanks for your interest in this. I'm as certain as I can be that the dll is failing to load because its usage is within an ActiveX control. Under Visual studio you can switch on extended information for seeing what happens when the LoadLibrary is performed. The output records the fact that when trying to load the osgdb_ive.dll file, the loader failed to find the osgSim.dll. This dll is in the same location as the ive dll, and was built at the same time. The output also shows the path that was searched. Within the path was the . directory. I take this to mean ie the current execution path directory. As the LoadLibrary found the osgdb_ive.dll, I can only assume that its knowledge of where it is running from is correct, as the ActiveX control is also located, and registered, within the directory that the osg dll's are. So the only conclusion I can come to is that when dynamically loading a dll that has dependencies on other, as yet not loaded dlls, within an ActiveX control, there appears to be an issue. When I get a chance I'll set up a little test case, and assuming this proves to be the case, I'll submit it to the forum - only for the Windows users thought ;-( If you want more info I can send the extended output from Visual Studio. Its quite explicit in where it fails. Kind regards Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] RE Plugin loading problems....
Hi All, I think I can now shed some more light on this problem, and perhaps I have found a potential bug. In the last update to this issue I had tracked the problem down to something to do with osgSim not wanting to load as a dependency of osgDB_ive.dll - and osgDB_flt.dll. On further investigation I think this is ultimately the issue. If in my ActiveX control I force an explicit, fully qualified dynamic load to osgSim, and then try saving as an ive file, everything is fine and all works. However, if I don't force the explicit load, the save fails as the plugin dlls for the ive format fails to load because it fails to load the osgSim dll. In order to get around this issue I forced an explicit linkage to osgSim in my activeX control - merely by creating and deleteing an osgSim::LightPoint object. The ActiveX control subsequently needed to link to osgSim, and hence when the control starts, osgSim is loaded - not as part of the plugin process , but as a consequence of the control loading up. When I now save as an ive file everything is fine. Clearly there is an issue, with osgSim loading as a consequence of a plugin, rather than an explicit linkage. Any thoughts ? Neil ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] RE Plugin loading problems....
Hi Robert, thanks for coming back to me. I think you're correct. I think this is an issue within MS and ActiveX controls. It does appear that that the LoadLibrary has issues resolving dependencies of libraries that are being loaded, when issued from within an ActiveX control. Obviously this isn't one for OSG as you can only use the call that is available - LoadLibrary. Having now got to a point where I can fix the issue with a relatively small alteration to the ActiveX control, I'm not too concerned. Hopefully if anyone else encounters this problem, this thread might be of use. Kind regards Neil Perhaps ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Plugin loading problems....
Hi All, I asked this question a couple of weeks ago, and Paul and others were kind enough to offer some advice, but I'm still having problems so I thought I would reword it a little to see if it sparks any more ideas with the forum. The issue I have is that I have an ActiveX control that wraps the OSG functionality - well the bits I want. This control loads 3DS models, applies textures (targas and jpegs), and all this occurs over the internet. Clearly for this to be working the osgdb_net, osgdb_jpg, osgdb_3ds and osgdb_tga plugins must all be loading correctly, and indeed they are. I now come to the issue which is that I would like to be able to save a .ive file. I can save (thanks to Roberts help) a .osg file with all the information I need for OSGViewer to display the scene, however I thought it would be good to look at saving the equivalent .ive format. The issue that I have is that when I tell OSG to save the .ive file it eventually gets to the routine DynamicLibrary::HANDLE DynamicLibrary::getLibraryHandle( const std::string libraryName) and fails to load the osgdb_ive.dll file. The curious thing is that the file is exactly where all the other dll's are, and the path that the LoadLibrary function uses is the path to the dll. In the previous post I made I was trying to load the osgdb_flt.dll and this also failed. I now have the same problem with at least two plugin dlls and I don't see what the difference is. Can anyone shed some light on this please ? I think it might be to do with dll dependencies that may not be loaded but I can't see why. Many thanks Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] OSG file format and textures...
Hi Robert, thanks for getting back to me. I think there is probably a whole load of confussion at my on this point. I had mistakenly assumed that the information I was looking at in the OSG file was the embedded textures. Doh! Evidently I should try reading the node names in the osg file. Sorry. You are correct, I think it really is just the texture references that I should have told it to export as well. Thanks for the help. Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] OSG file format and textures...
Hi All, Sorry if this is an old chestnut, but I got a question about the .osg file format. I've got a scene created, and I save the universe out as a .osg file. I've passed no options through in the the call, so the file that's created appears to have the textures that the scene uses actually embedded within the osg file. For my purposes this is great. However, when I double click on the osg file, and view it with the osgviewer, no textures appear. I should add at this point that I'm still using OSG 1.2 - would love to move on but no time at the moment. So, has anyone come across this problem before ? The textures that have been used are targas and jpegs if that's any help. It appears that using the options parameter I could force the save to merely record a reference to the texture, and this would solve my immediate problem, however I thought that I may have stumbled across a minor bug, and wanted to check if anyone knew about it. Thanks for any comments Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Help with OpenFlight....
Hi All, Encountered a slight problem that I don't quite understand. I have a scene (or universe if you like) and I want to save it as an openflight (.FLT) format. Tracking through the call to write the universe node, I get to the point where it needs to load in the osgdb_OpenFlightd.dll library. It makes the call, but returns a null handle. Help? I've tried saving a .osg format and that works fine. Any thoughts ? Thanks Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Help with OpenFlight....
Hi All, Thanks for the comments about openflight. I guess I have to wait for Paul's work to become available. Would be happy to beta test if that is any help, although I confess to not knowing anything about the details of the flt format. Having said all this, I guess the point of my mail is still open. Why would the LoadLibrary fail for the openflight osg dll given that its all compiled on my machine, and dll is in the correct place? A bit concerning really. I can accept that once loaded I'm not going to get far with the save command, but it should at least load the library. Any thoughts would be most welcome. Kind regards Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Re How to override plugins...
Hi Robert, Thanks for that. I thought it would be easier to derive from existing zip reader, but evidently that wasn't the way to do it as I then ended up creating two zip reader instances as the proxy construction and the class definition for the base zip reader are in the cpp file. I've fixed my problem by simply deriving from ReaderWriter instead. This brings me to a more general question about readers. Is there a reason why some employ the .h .cpp conventions, whilst others roll them all into one - either .h or .cpp ? Thanks again for the info. Neil ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] How to override default plugins....
Hi all, I wish to supply my own plugin for the zip format, which unzips to memory. I've created my plugin derived from the existing zip plugin, overridden the readnode functions for file and stream, and at the bottom of my class I've called the RegisterReaderWriterProxy template for my class. It all appears to work, but I now have two zip file format plugins. The original, and mine. My one gets called second due to the way the RegisterReaderWriterProxy template appends to the existing list of known plugins, the plugin it is given. Ideally I would like to remove the existing plugin from the known list, thus there would only be my zip file plugin, and life is happy. However I am unclear how to do this, or even whether it is possible. Can anyone suggest what to do please ? I can see that there is a call removeReaderWriter in the registry, but can't see how I could call it with the original zip plugin reader, as this would surely have problems at shut down when the proxy class was destroyed. Please help Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] How to exclude nodes from a pick list
Hi Andreas, Many, many thanks for that. My problem is now sorted, and I can look forward to the cheesey wall paper decorating my walls. Kind regards Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] How to exclude nodes from a pick list
Hi All, I have a quick question. I'm playing with picking objects in a scene using the pickhandler example. The scene is essentially a big box whose faces face inwards. Within the box I have a smaller box whose faces face outwards. If I am outside both boxes, looking into the outer box, the pickhandler returns both the outer box, face that I have hit, and the inner box faces that I have hit. As the outer box face is actually facing away from me, I want to exclude it. Is there an easy way to do this? My thinking was that I essentially need to look at the normal of the face that I hit, and the direction of view vector for the camera, take the dot product, and if positive exclude it. Does this sound right? Is there an easier way/structure that I should use? Many thanks for any help Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] How to exclude nodes from a pick list
Hi All, I have a quick question. I'm playing with picking objects in a scene using the pickhandler example. The scene is essentially a big box whose faces face inwards. Within the box I have a smaller box whose faces face outwards. If I am outside both boxes, looking into the outer box, the pickhandler returns both the outer box, face that I have hit, and the inner box faces that I have hit. As the outer box face is actually facing away from me, I want to exclude it. Is there an easy way to do this? My thinking was that I essentially need to look at the normal of the face that I hit, and the direction of view vector for the camera, take the dot product, and if positive exclude it. Does this sound right? Is there an easier way/structure that I should use? Many thanks for any help Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] How to exclude nodes from a pick list
Hi Robert, sorry, I didn't really explain myself very clearly. The issue isn't so much that I don't want the face from the outer box to be hit, I merely want to be able to identify that it is facing away from the camera, and hence I can ignore it when try to identify that the inner box has been hit. The outer box essentially represents the walls of a room. From outside the room, looking in, I can see the interrior walls which is correct. If I am outside my room looking in, when I click on the box that sits on the floor of my room, I get a pick list that includes the wall that I am looking through. The difficulty is that I wish to discard this, and get the next item in the pick list which is the box that sits on the floor. I can't simply skip the first intersection, as if I were inside the room, then the box on the floor would be the first intersection and I would miss it. Given that I've got a hit on the wall that I'm looking through - which faces away from me and hence isn't visible as the culling gets rid of it for rendering purposes -what I need to do is essentially marry the culling that occurs for rendering with a pick handler so that the pick handler only considers faces that a rendering cull would want to render. My thinking was that, rather than go to the headache of trying to do the marry it would be simpler to take my hit list, and perform the calculation that basically says whether the face is facing towards the camera, or away from it. The hit records gives me a world normal of the face that is hit. I thought that the simplest thing to do would be to get the vector of the direction of look for the camera, multiply the two vectors, and discard those with a positive component in the direction of the camera view. My difficulty is in getting the vector that represents the direction of view of the camera. I don't know how to do that, and I guess thats what I really need the help with. Could you tell me how I get this vector, or perhaps suggest an alternative approach ? Surely the node mask approach would prevent me selecting the interior wall that is facing towards me, which isn't right as I may wish to select this wall and put a really cheesey wall paper texture on it :-) I would be grateful for any help on this. Kind regards Neil ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Function decoration....
Hi All, I'm having a slight problem building a release version of OSG1.2. I know I should move to 2.2 - and a colleague of mine has - however for various reasons I'm unable to at this moment. I'm running VS2005, and the problem I have is that I can build a debug version, link it against my app, and everything is fine. When I build the Release version of the OSG1.2 and link this against the release version of my app I get 5 unknown function calls - as follows. 1axtest1ctrl.obj : error LNK2001: unresolved external symbol public: virtual bool __thiscall osgProducer::Viewer::realize(enum Producer::CameraGroup::ThreadingModel) ([EMAIL PROTECTED]@osgProducer@@[EMAIL PROTECTED]@Producer@@@Z) 1axtest1ctrl.obj : error LNK2001: unresolved external symbol public: virtual bool __thiscall osg::Group::removeChild(unsigned int,unsigned int) ([EMAIL PROTECTED]@osg@@[EMAIL PROTECTED]) 1axtest1ctrl.obj : error LNK2001: unresolved external symbol public: virtual bool __thiscall osg::Group::removeChild(class osg::Node *) ([EMAIL PROTECTED]@osg@@[EMAIL PROTECTED]@@Z) 1axtest1ctrl.obj : error LNK2001: unresolved external symbol public: virtual bool __thiscall osg::Geode::removeDrawable(unsigned int,unsigned int) ([EMAIL PROTECTED]@osg@@[EMAIL PROTECTED]) 1axtest1ctrl.obj : error LNK2019: unresolved external symbol __declspec(dllimport) public: void __thiscall osg::CameraNode::setRenderOrder(enum osg::CameraNode::RenderOrder) ([EMAIL PROTECTED]@osg@@[EMAIL PROTECTED]@@Z) referenced in function public: unsigned long __thiscall CAxTest1Ctrl::CreateHUDNode(void) ([EMAIL PROTECTED]@@QAEKXZ) 1CorrectInitialUnitPartMatrixVisitor.obj : error LNK2019: unresolved external symbol __declspec(dllimport) class osg::Matrixd __cdecl osg::computeLocalToWorld(class std::vectorclass osg::Node *,class std::allocatorclass osg::Node * const ) ([EMAIL PROTECTED]@@[EMAIL PROTECTED]@[EMAIL PROTECTED]@osg@@[EMAIL PROTECTED]@osg@@@std@@@std@@@Z) referenced in function public: virtual void __thiscall CCorrectInitialUnitPartMatrixVisitor::apply(class osg::Transform ) ([EMAIL PROTECTED]@@[EMAIL PROTECTED]@@@Z) Sorry about that. Anyway, on closer inspection it appears that my app is trying to link to osg::Group::removeChild(unsigned int,unsigned int) ([EMAIL PROTECTED]@osg@@[EMAIL PROTECTED]) however when I examine the OSG dll produced when building a release version it contains the decorated function osg::Group::removeChild(unsigned int,unsigned int) ([EMAIL PROTECTED]@osg@@[EMAIL PROTECTED]) (in case you missed it, the bit that's different is the decoration). I've trawled through the MS documentation, played with a few compiler/linker settings associated with inline functions, all to no avail. Has anyone got any ideas ? Many thanks for any help. Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] [RE] Function decoration...
Hi All, I've sorted out my problem. No doubt most of you guessed, but I was picking up the wrong header file. Doh! Sorry to have bothered you. Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Osg blocking on application exit...
Hi Robert, Sorry about that last email. I wasn't trying to Dis the OSG multithreading, my comment was in regards to the recent exchange about toggling textures on/off and changing threading modes of operation dynamically at run time. There seemed to be forum comments regarding some potential instability there, and I was merely highlighting this to one of my colleagues. Sorry if it came across as something else. To be totally frank about it, the email wasn't event mean't to go to the list, but I missed the recipient list when I replied to Chris. Doh! Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] osgViewerMFC example fails...
Hi All, Has anyone tried the osgViewerMFC example in OSG2.0? I tried to build it two days ago, and it failed. The issue is in regards to an assignment of a CString to a std::string. Trawling internet solved problem, but has the example been corrected for the 2.2 release ? Neil. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
