Re: [boost] [fusion] More quanfusion

From: Andy Little (andy_at_[hidden])
Date: 2006-09-14 06:40:26

"Andy Little" <andy_at_[hidden]> wrote in message
news:eeb265$m3u$1_at_sea.gmane.org...
>
> "David Abrahams" wrote
>> Joel de Guzman  writes:
>>
>>> advance_iterator looks general enough.
>>
>> What *is* an advance_iterator?
>
> FWIW:
>
> fusion::advance_iterator<
>    Sequence,                      // the sequence to advance over
>    Stride,                             // number of positions in underlying
>                                          //sequence to move on 'next'
>    StridedPosition,            //  underlying sequence::position *
>                                          //    Stride + Offset
>    Offset                             // offset in underlying sequence units
>>
>
> The main use is to  return a column in a linear sequence representing a
> matrix.

And here FWIW is an example use.

The example sets up a 2D rotation matrix  comprised partly of static_values and
partly of runtime values( here with a runtime settable 90 degree rotation ) and
a homogeneous coordinate again comprised partly of static and partly of runtime
values. Because some members are static_values, calcs on them will be
completetly optimised away by the compiler , hence, even though the matrix and
coordinate directly represent length quantities using quan::length's , the only
runtime calcs are on double values and even there the calcs will be reduced
compared to the equivalent double only matrix. (The matrix will of course
survive serialisation, so that the static_values will survive)

http://tinyurl.com/e7729

regards
Andy Little