HI Geoff et al, Could you guys tests this new version of DelaunayTriangultor.cpp and if it works out well please say so. Geoff if everything looks solid could you post the file to osg-submissions.
Cheers, Robert. On 12/4/06, Geoff Michel <[EMAIL PROTECTED]> wrote:
Robert (and Amalric and jOan) this version of dealunay.cpp works with Amalric's case data and all the cases built into the original example. I still cant see why the previous version worked with VS6 though - it must have exercised the code that I have changed. The example does show however that the debug built into VS8 is quite comprehensive and correctly identified an attempt to read from an undefined element of an STL array (to prevent overrun viruses?). Anyone else who has seen this problem please attack it with your test cases! Geoff ----- Original Message ----- From: amalric alexandre To: osg users Sent: Friday, December 01, 2006 8:58 AM Subject: Re: [osg-users] Delaunay triangulator of concav geometry Hi Geoff, I'm agree with you on the fact that this first coordinate is far from others. But in my program I triangulate points clicked by mouse and it works under MS VS 2003 and not with MS VS 2005. The main goal of this example was just to demonstrate that under VS 2003 it works and not with MS 2005. Ideally it has to work under the both compilator. For me it's a little bug in the DelaunayTriangulator but I cannot find where. If you have any idea of the problem let me know. Thanks. 2006/11/30, Geoff Michel <[EMAIL PROTECTED]>: > > > Dear Alexandre > > I have tried your case, and it fails as you describe under vs8. However...... look at the first coordinate of your added constraint > > > > float x=-10000,y=600; > > bounds->push_back(osg::Vec3(x,y,getheight(x,y))); > > x=1610,y=900; > > all the other coordinates are near X=1600-2100. Your constraint is very (very, very, very....) long and thin, and moreover x=-10000 is way outside the terrain. I replaced the first coordinate with x=1000, and no problem the triangulation is fine. > > Is the -10000 correct (in terms of your problem)? I agree that ideally the triangulator should triangulate absolutely anything, but you are probably near the limits of floating point precision for such a long thin constraint. (other cases of this error are sought). > > Geoff > > > ----- Original Message ----- > From: amalric alexandre > To: osg users > Sent: Thursday, November 30, 2006 2:03 PM > Subject: Re: [osg-users] Delaunay triangulator of concav geometry > > > I found this morning that it's due to MS VS 2005 because I made a test modifying the example "osgDelaunay.cpp" (attached) compiling on MS VS 2003 there is no problem, but compiling with MS VS 2005 the program crashes in function lineBisects(delauneytriangulator.cpp) because of a "out of range" in accessing array "points". It doesn't depend on OSG version (same problem with OSG 1.2 and OSG 1.1). > Please see the attached files. > > Thanks. > > Alexandre AMALRIC > > > > 2006/11/30, jOan <[EMAIL PROTECTED]>: > > I have the same problem using DelaunayTriangulator with MS VS2005. > > > > > > > > 2006/11/30, amalric alexandre <[EMAIL PROTECTED]>: > > > > > > Hi, > > > > > > I want to triangulate with DelaunayTriangulator a geometry wich is concav. I give to the DelaunayTriangulator a set of point sorted and to the DelaunayConstraint the same set of point (edges of the geometry). But when i call the triangulate() function it crashes in lineBisects() function because of a value out of range. Is it a known bug of triangulator ? or is my method not appropriated ? If it is please explain how it is possible to triangulate a concave geometry ? > > > _______________________________________________ > > > osg-users mailing list > > > [email protected] > > > http://openscenegraph.net/mailman/listinfo/osg-users > > > http://www.openscenegraph.org/ > > > > > > > > > > > > > > -- > > jOan Abadie > > > > _______________________________________________ > > osg-users mailing list > > [email protected] > > http://openscenegraph.net/mailman/listinfo/osg-users > > http://www.openscenegraph.org/ > > > > > > > > ________________________________ > > _______________________________________________ > osg-users mailing list > [email protected] > http://openscenegraph.net/mailman/listinfo/osg-users > http://www.openscenegraph.org/ > > > > > > > _______________________________________________ > osg-users mailing list > [email protected] > http://openscenegraph.net/mailman/listinfo/osg-users > http://www.openscenegraph.org/ > > ________________________________ _______________________________________________ osg-users mailing list [email protected] http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ _______________________________________________ osg-users mailing list [email protected] http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
_______________________________________________ osg-users mailing list [email protected] http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
