RE: [boost] MPL fold algorithms

From: Aleksey Gurtovoy (agurtovoy_at_[hidden])
Date: 2002-11-15 08:05:01

Gennadiy Rozental wrote:
> Would it correct to state that following pseudo-code properly 
> reflect the essence of the appropriate algorithms:
> 
> iter_fold( sequence s, state st, binary_operator op ) {
>    iterator it = s.begin();
> 
>    while( it != s.end() )
>        st = op( st, it );

    return st;
> }

> fold( sequence s, state st, binary_operator op ) {
>    iterator it = s.begin();
> 
>    while( it != s.end() )
>        st = op( st, *it );

    return st;
> }

> iter_fold_backward( sequence s,
>                     state st,
>                     binary_operator backward_op,
>                     binary_operator forward_op = identity ) {
>    iterator it = s.begin();
> 
>    while( it != s.end() )
>        st = forward_op( st, it );
> 
>    while( --it != s.begin() )
>        st = forward_op( st, it );
> 
>     st = forward_op( st, it );

    return st;
> }

> 
> fold_backward( sequence s,
>                           state st,
>                           binary_operator backward_op,
>                           binary_operator forward_op = identity ) {
>    iterator it = s.begin();
> 
>    while( it != s.end() )
>        st = forward_op( st, *it );
> 
>    while( --it != s.begin() )
>        st = forward_op( st, *it );
> 
>     st = forward_op( st, *it );

    return st;
> }

Yep, that's basically what they do.

> P.S. MPL docs reference pages does not formatted for 
> printing. They does not fit the page width. See iter_fold or example.

Yep, it's a known problem. The new revision will come in PDF (properly
formatted) as well as HTML, so it won't be an issue then.

Aleksey