Le 28-avr.-10 à 22:06, Konstantinos Poulios a écrit :
After some further experimentation I tend to conclude that there is a
logical error in the way the avl tree is used in the case of
"allNonDuplicatedPoints". To my understanding, the comparison function
"comparePosition" which is used by this tree is not an appropriate
comparison function. The following 2D example demonstrates what I
mean:
yes, true, we should use bounding boxes and multidimensional trees...
yet, we're working now on a complete re-writing of Gmsh's modeler, that
will be a true modeler, so be patient, that kind of thing should not
happen
anymore ;-)
JF
Consider the points A(0,0), B(0.5,-1), C(1,-2) and eps = 0.6, in this
case using the function "comparePosition" to compare A, B and C in
pairs would yield to:
A > B (because A.X ~= B.X and A.Y > A.Y)
B > C (because B.X ~= C.X and B.Y > C.Y)
A < C (because A.X < C.X)
which is a logical error.
I will try to find some fix for the "comparePosition" function and try
it in praxis.
If I am indeed right with my considerations, it would be strange that
I am the first to experience crashes due to this problem.
Best Regards
Kostas
_______________________________________________
gmsh mailing list
[email protected]
http://www.geuz.org/mailman/listinfo/gmsh
------------------------------------------------------------------
Prof. Jean-Francois Remacle
Universite catholique de Louvain (UCL)
Ecole Polytechnique de Louvain (EPL) - Louvain School of Engineering
Institute of Mechanics, Materials and Civil Engineering (iMMC)
Center for Systems Engineering and Applied Mechanics (CESAME)
Tel : +32-10-472352 -- Mobile : +32-473-909930
_______________________________________________
gmsh mailing list
[email protected]
http://www.geuz.org/mailman/listinfo/gmsh
