The SDG traits define a tag named Tag_has_bisector_constructions
that is either Tag_false or Tag_true.
If it is Tag_false, the bisector constructions are the hard-coded
ones from the algorithm (essentially they are the L2 ones).
If it is Tag_true, the bisector constructions are taken from the
traits. This is how the Linf traits are implemented.
This way, we avoid SFINAE.
Introduce pointers for functions that insert a point on a segment.
For L2, to keep the same functionality as before, the pointers
point to insert_exact_point_on_segment and insert_point_on_segment
functions.
Classes derived from Segment_Delaunay_graph_2 can change the
pointers to point to more specialized functions. For example, the
Linf implementation will point to different specialized function
defined in the Linf class.
Use an additional optional parameter (SDGLx) in the segment
Delaunay hierarchy template to be able to change the segment
Delaunay graph class that is used per level of the hierarchy.
Since this is the last optional parameter, there is no change
for users of the L2 segment Delaunau graph and hierarchy.
This parameter is necessary to support the Linf segment Delaunay
graph and hierarchy. Again, it is intended that the user of the
Linf classes does not have to be aware of this parameter.
Some additional classes are declared as friends of
Segment_Delaunay_graph_2.
If construction templates for bisectors (line, segment, ray)
exist in the traits, then use them, otherwise, use the L2 traits.
The check is implemented by a type Has_bisector_constructions_type
that might be included in the traits and using the
"Substitution failure is not an error" (SFINAE) principle.
The oriented_side predicate is used with a face corresponding to
the interior of a segment that is split by a vertex contained in
the segment. For an infinite such face the code has to become more
general in order to accomodate the Linf case. For example, in L2
the number of such infinite faces is 0, 2, or 4, whereas in Linf,
it can be any even non-negative integer.
Moreover, the existing code for infinite faces is L2 specific and
ignores the interior of the segment for its decision. In Linf the
interior of the segment is important. For this reason, there is a
new case of the predicate that is general enough to also
accomodate Linf:
// computes the oriented side of the Voronoi vertex of s1, s2, inf
// wrt the line that passes through the point p and its direction
// is the direction of the supporting line of s, rotated by 90
// degrees counterclockwise.
Oriented_side operator()(const Site_2& s1, const Site_2& s2,
const Site_2& s, const Site_2& p) const
The Oriented_side_C2 class is also adapted for L2 to use the new
form of the predicate, which internally calls the old form.
Use vertex conflict instead of orientation in insert_third for
the case of three points. This change is to make SDG more L2
agnostic.
Given are three point sites s1, s2, s3.
The orientation of (s1,s2,s3) is not collinear if and only if the
two vertex conflict quadruples (s1,s2,inf,s3) and (s2,s1,inf,s3)
are of opposite signs.
The orientation of (s1,s2,s3) is right turn if and only if the
vertex conflict quadruple (s2,s1,inf,s3) is NEGATIVE.
Successfully test in CGAL-4.3-Ic-102
This commit fixes a few issues in the doxygen documentation and in particular
the figure for the generation of latex doc.
Approved by the interim release manager
xplain why this merge is necessary,
this happens if you have an empty line or the text directly after /!*
to find them in the generated documentation, grep for "pre class=\"fragment\""
(a few of them are legitimate however)
This commit also updates the respective mock-headers when the
corresponding .tex documentation has changed. This has been the case for:
AABB_tree/doc_tex/AABB_tree_ref/AABBGeomTraits.tex
AABB_tree/doc_tex/AABB_tree_ref/AABBTraits.tex
AABB_tree/doc_tex/AABB_tree_ref/AABB_traits.tex
Mesh_3/doc_tex/Mesh_3_ref/parameters_features.tex
Mesh_3/doc_tex/Mesh_3_ref/parameters_no_features.tex
Number_types/doc_tex/NumberTypeSupport_ref/Interval_nt.tex
Number_types/doc_tex/NumberTypeSupport_ref/Lazy_exact_nt.tex
Spatial_searching/doc_tex/Spatial_searching_ref/Search_traits_2.tex
Spatial_searching/doc_tex/Spatial_searching_ref/Search_traits_3.tex
Triangulation_2/doc_tex/Triangulation_2_ref/Constrained_Delaunay_triangulation_2.tex
That is a followup-to my commit last year:
| ------------------------------------------------------------------------
| r63198 | lrineau | 2011-04-28 19:45:22 +0200 (Thu, 28 Apr 2011) | 5 lines
|
| Try to fix my last revision about cmake_policy, with CMake-2.6.x
|
| CMake gives an error if one tries to use cmake_policy(VERSION x.y.z) if
| x.y.z is greater than the current CMake version.
|
| ------------------------------------------------------------------------
The following check:
if("${CMAKE_MAJOR_VERSION}.${CMAKE_MINOR_VERSION}" VERSION_GREATER 2.6)
is useless just after a call to:
cmake_minimum_required(VERSION 2.6.2)
The script used to fix that was:
#!/usr/bin/env perl
$replacement=<<'END';
if("${CMAKE_MAJOR_VERSION}.${CMAKE_MINOR_VERSION}.${CMAKE_PATCH_VERSION}" VERSION_GREATER 2.8.3)
cmake_policy(VERSION 2.8.4)
else()
cmake_policy(VERSION 2.6)
endif()
END
while(<>) {
if(/if\("\${CMAKE_MAJOR_VERSION}.\${CMAKE_MINOR_VERSION}" VERSION_GREATER 2.6\)/) {
while(<>) {
if(/^endif\(\)/) {
print "$replacement";
while(<>) {
print;
}
exit 0
}
}
}
print;
}
Wextra results in a dreadful wall of yellow in the test-suite and is
scary when using CGAL. Reduce some of the noise by silencing the
obvious cases with CGAL_USE or comments.
features/doxy-pmoeller and add changes for ccRequires on top of it.
ATTN: This changeset should be unmerged before the whole branch goes
back into next.
still some headers contain FORTH as copyright headers;
this was done (incorrectly on the trunk (r65713) and reverted by Laurent; it is now properly done in ^/branches/next
We declare
cmake_minimum_required(VERSION 2.6.2)
but we also use
cmake_policy(VERSION 2.8.4)
to declare that our CMake scripts are OK with all the defaults of CMake policies
as of CMake-2.8.4. That shuts down the warnings of CMake-2.8.4.
That way, we no longer need any declaration of specific policies.
Those two lines must be present and maintained in all our CMakeLists.txt
files (the one for the libraries, and also the one for examples and demos,
and maybe tests).
The documentation of CGAL says that the minimal version of CMake must be
CMake-2.6.2.
- change cmake_minimum_required to VERSION 2.6.2
- no longer any need for the policy CMP0003 (was for 2.4.x)
- no longer any need for CMAKE_ALLOW_LOOSE_LOOP_CONSTRUCTS (was for 2.4)
In particular remove UNTESTED_XXXXXXXXXXX unused variable that possibly hide true warnings.
In those cases, the string printed while executed now starts with "NOTE: ".
*CGAL internal code no longer rely on depecrated features
by namespace CGAL { and } //namespace CGAL. in all .h and .cpp files
in a directory.
Apply it to all packages in the trunk
Remove macro definition from the config.h file.
when possible;
benchmark_nox.cpp uses now the filtered traits with static filters by default;
this is activated by the CGAL_SDG_TRAITS_WITH_STATIC_FILTERS macro;
Vertex_conflict_C2.h: added operator() that takes four points;
Traits_wrapper_2.h: the Side_of_oriented_circle_2 class is now implemented
using the Vertex_conflict_2 class from the base traits;
added some types and some assertions in edge lists (both regular and in-place);
fixed bugs in post-increment operators of edge list circulators (edge_list.h)
non-intersecting segments; the code is activated by defining the
CGAL_SDG_USE_SIMPLIFIED_ARRANGEMENT_TYPE_PREDICATE macro, which is currently done in the benchmark program
added support for in-place storage (at the faces) of the faces that define the conflict region; also the sign of the incircle test is cached there; this uses the Face_data that now exists in TDS2;
to activate the in-place storage of the faces of conflict regions define the macros: CGAL_TDS2_DATA and CGAL_SDG_NO_FACE_MAP
the file Triangulation_face_base_with_in_place_edge_list_2.h implements an in-place edge list for the boundary of the conflict region
we are in the special case where the Voronoi vertex is degenerate (which happens
when the Voronoi vertex is defined by two segments and a point, where the point
is the common endpoint of the two segments)
removed file Incircle_sqrt_field_C2.h: the code in there is superseeded by the
code in Voronoi_vertex_sqrt_field_new_C2.h;
the code used now by default for computing the incircle test is that in the file
Voronoi_vertex_sqrt_field_new_C2.h; the old code can be activated by defining
the macro CGAL_SDG_USE_OLD_INCIRCLE; also by defining the macro
CGAL_SDG_CHECK_INCIRCLE_CONSISTENCY the incircle test is evaluated by both codes
and if they are different an error is shown in std::cerr
Andreas Fabri
Sébastien Loriot
Philipp Möller
Sebastien Loriot
Laurent Rineau
Panagiotis Cheilaris
Ross Hemsley
Philipp Möller
Philipp Möller
Laurent Rineau
Menelaos Karavelas