[boost] Re: ublas type safety and bound check

From: Benedikt Weber (weber_at_[hidden])
Date: 2002-07-02 14:20:46

Joerg

Could you give me a hint why you have to compare floating point numbers
here.  I was just thinking of some simple checks like if the dimensions of
matrices are consistent for a matrix product or for assignment. For that you
would only need to compare indices. Or compile-time checks which would
prevent you to assign a full matrix to a triangular one as I did in the
example.

Benedikt

"Joerg Walter" <jhr.walter_at_[hidden]> wrote

> The corresponding checks exist in the packed matrix evaluation:
>
> #ifdef NUMERICS_BOUNDS_CHECK_EX
>             {
>                 // Need the const member dispatched.
>                 const M &cm = m;
>                 typename E::const_iterator1 it1e (e ().begin1 ());
>                 typename E::const_iterator1 it1e_end (e ().end1 ());
>                 while (it1e != it1e_end) {
>                     typename E::const_iterator2 it2e (it1e.begin ());
>                     typename E::const_iterator2 it2e_end (it1e.end ());
>                     while (it2e != it2e_end) {
>                         // FIXME: we need a better floating point
> comparison...
>                         check (*it2e == cm (it2e.index1 (), it2e.index2
()),
> bad_index ());
>                         ++ it2e;
>                     }
>                     ++ it1e;
>                 }
>             }
> #endif
>
> but are currently disabled due to the usual problems with floating point
> arithmetic when checking too strict. I'll discuss with Mathias, how to
> install a customizable comparator.