For the header only version, the global variable default_random no more exist => we need to replace all occurences of default_random by a call to the global function get_default_random().
As usual, no modification for the non header only version.
There is const static variable, transformed into a static variable of a global function. This is not a problem since this is only used in debug mode to test the rounding mode at the end of the program. Testing it several times is ok.
All non-ASCII files should be encoded in UTF-8. So far, the ZSH script
`Scripts/developer_scripts/detect_wrong_encoding` detects wrong encoding
in source/header files of include/, src/, and doc/.
the previous version was hoping to have an = operator between the types
which is unlikely to happen in practice. We use the read from string
capabilities instead (which should be available for most NT)
Introduces the Surface modeling package that allows
to deform a triangulated surface mesh
Successfully tested in CGAL-4.5-Ic-90
Approved by the Release Manager
Conflicts:
Installation/changes.html
Reintroduce the possibility to use CGAL without GMP
(to use LEDA number types instead for example).
The cmake option to set to switch off GMP when building
the lib is CGAL_DISABLE_GMP.
Tested in CGAL-4.5-Ic-86
Approved by the release manager
ge is necessary,
When CORE_arithmetic_kernel and LEDA_arithmetic_kernel are used, then
one should have:
#include <CGAL/CORE_arithmetic_kernel.h>
#include <CGAL/LEDA_arithmetic_kernel.h>
Remove explicit interoperability between Interval_nt and int, so
Interval_nt becomes interoperable with every builtin type (in particular
long) via double.
Approved by the release manager
It shouldn't change the performance significantly (the time is spent in
malloc/free and the mpq_* calls), but at least I can follow the
(smaller) generated code.
Conflicts:
Kernel_d/doc_tex/Kernel_d/kernel_representation_d.tex
Kernel_d/doc_tex/Kernel_d_ref/Homogeneous_d.tex
Kernel_d/doc_tex/Kernel_d_ref/Kernel.tex
Kernel_d/doc_tex/Kernel_d_ref/main.tex
Resolved by dropping those files.
The code was incorrect, we should have used _FPU_GETCW(CW)&3 to keep
only the rounding mode bits. However, glibc comes with an optimized
inline version of fegetround and fesetround so there is no point using
anything else.
Guillaume Damiand
Marc Glisse
Laurent Rineau
Sébastien Loriot