Subject: CGAL users discussion list
List archive
- From: "Oleksiy Busaryev" <>
- To:
- Subject: [cgal-discuss] Monge fitting bug?
- Date: Fri, 29 Aug 2008 15:12:12 -0700
- Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=message-id:date:from:to:subject:mime-version:content-type; b=IlOcKVkFOfYwQXLMtGoBWJ3Z5o458uKY32/jQfSJYSgKTLPogiQAS5pNieNC+oWEO5 8qjUS5y3XC6Zk2k+cnzdDh84+MA0UP6rjh7kRM0dwBiMoEKXDhPjxaWAiTnMI4fN48LV K1UafG+p9K+YHS7z1UO7IjMxB+nGrUUh6bsXE=
Hello,
I found one small bug in monge_via_jet_fitting.h file.
In Monge_via_jet_fitting::operator() there is a lapack matrix and a vector declared:
LAMatrix M(nb_input_pts, nb_d_jet_coeff);
LAMatrix M(nb_input_pts, nb_d_jet_coeff);
LAVector Z(nb_input_pts);
LAMatrix and LAVector classes do not have destructors.
Not sure if it is a bug, maybe memory is supposed to be freed externally.
At least in the given example with operator() memory is not released, as a result,
for big input polyhedrons principal curvature estimation fails short of memory (as in my case).
Just adding
free(M.matrix());
free(Z.vector());
to the end of this function solved my problem.
--
Best regards,
Oleksiy Busaryev.
- [cgal-discuss] Monge fitting bug?, Oleksiy Busaryev, 08/30/2008
Archive powered by MHonArc 2.6.16.