Module: ppl/ppl Branch: floating_point Commit: e382dfd1d7fa78031d49b05b7bbc0fac51ca3aa9 URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=e382dfd1d7fa7...
Author: Fabio Bossi bossi@cs.unipr.it Date: Mon Sep 21 15:41:16 2009 +0200
Added some more documentation.
---
src/Polyhedron.defs.hh | 60 ++++++++++++++++++++++++++++++++++++++++++++++++ 1 files changed, 60 insertions(+), 0 deletions(-)
diff --git a/src/Polyhedron.defs.hh b/src/Polyhedron.defs.hh index 3c3a9d5..7ecc2cf 100644 --- a/src/Polyhedron.defs.hh +++ b/src/Polyhedron.defs.hh @@ -2628,6 +2628,25 @@ protected: //@} // Exception Throwers #endif // defined(PPL_DOXYGEN_INCLUDE_IMPLEMENTATION_DETAILS)
+ //! Helper function that overapproximates an interval linear form. + /*! + \param lf + The linear form on intervals with floating point boundaries to approximate. + + \param lf_dimension + Must be the space dimension of \p lf. + + \param store + The interval abstract store in which the approximation is performed. + + \param result + Used to store the result. + + This function makes \p result become a linear form that is a correct + approximation of \p lf in \p store. The resulting linear form has the + property that all of its variable coefficients have a non-significative + upper bound and can thus be considered as singletons. + */ template <typename FP_Format, typename Interval_Info> static void overapproximate_linear_form( const Linear_Form<Interval <FP_Format, Interval_Info> >& lf, @@ -2635,12 +2654,53 @@ protected: const std::map< dimension_type, Interval<FP_Format, Interval_Info> >& store, Linear_Form<Interval <FP_Format, Interval_Info> >& result);
+ /*! \brief + Helper function that makes \p result become a Linear_Expression obtained + by normalizing the denominators in \p lf. + + \param lf + The linear form on intervals with floating point boundarie to normalize. + It should be the result of an application of static method + <CODE>overapproximate_linear_form</CODE>. + + \param l_dimension + Must be the space dimension of \p lf. + + \param result + Used to store the result. + + This function ignores the upper bound of intervals in \p lf, + so that in fact the result can be seen as \p lf multiplied by a proper + normalization constant. + */ template <typename FP_Format, typename Interval_Info> static void convert_to_integer_expression( const Linear_Form<Interval <FP_Format, Interval_Info> >& lf, const dimension_type lf_dimension, Linear_Expression& result);
+ //! Normalization helper function. + /*! + \param lf + The linear form on intervals with floating point boundaries to normalize. + It should be the result of an application of static method + <CODE>overapproximate_linear_form</CODE>. + + \param lf_dimension + Must be the space dimension of \p lf. + + \param res + Stores the normalized linear form, except its inhomogeneous term. + + \param res_low_coeff + Stores the lower boundary of the inhomogeneous term of the result. + + \param res_hi_coeff + Stores the higher boundary of the inhomogeneous term of the result. + + Results are obtained by normalizing denominators in \p lf, ignoring + the upper bounds of variable coefficients in \p lf. + */ template <typename FP_Format, typename Interval_Info> static void convert_to_integer_expressions( const Linear_Form<Interval <FP_Format, Interval_Info> >& lf,