Approved by the release manager
Conflicts:
Documentation/doc/Documentation/dependencies
Documentation/resources/how_to_cite_cgal.bib.in
Installation/changes.html
The BGL is not const correct by design choice. Passing a const Graph as
template parameter to the primitive type is error prone.
This will make the life easier to the user.
That is a followup to the following commit:
| commit 3e93d0bfc6
| Author: Laurent Rineau <laurent.rineau@cgal.org>
| Date: Tue Dec 31 15:15:24 2013 +0100
|
| Fix test of AABB_tree (use of an epsilon)
|
| The test 'aabb_naive_vs_tree_distance_triangle_test' failed
| something. It was due to a test using an epsilon, with a too-low epsilon
| value (1e-8). I managed to reproduce the bug, and the quantity compared
| to the epsilon value was about 1.8e-8. I increase the epsilon value to
| 1e-7, and add a few lines that show the compared values, when the test
| fails.
1/ There was one more epsilon value with 1e-8 in the tests.
2/ There was a test that check the equality of two points, in a case,
but the test cannot be fully correct with a non-exact kernel.
That function set_shared_data is never tested in CGAL, happened not to
be callable using its documented API. I do not understand why there is
an additional template parameter 'PrimtiveType'. That commit removes it.
(cherry picked from commit 99790a0d39864c3d2e72d18876f41603ab94a5a5)
The test 'aabb_naive_vs_tree_distance_triangle_test' failed
something. It was due to a test using an epsilon, with a too-low epsilon
value (1e-8). I managed to reproduce the bug, and the quantity compared
to the epsilon value was about 1.8e-8. I increase the epsilon value to
1e-7, and add a few lines that show the compared values, when the test
fails.
Trivial bug-fix.
This fixes the graph primitives that were only model of
AABBPrimitiveWithSharedData, and also remove the artificial compatibility
with deprecated primitives that are not needed since we want to make users
use the new method.
Successfully tested in CGAL-4.3-Ic-107
One example on the platform i686_Linux-2.6_g++-4.1.2_CentOS-5.1-O3 indicates
an error, but after checking it is a floating point computation threshold
issue. The example expects an error to be below 1e-10 and it is 1.1e-10.
Approved by the interim release manager