On 02/06/08 01:21, dan marsden wrote:

Larry Evans <cppljevans@suddenlink.net> wrote:

<snip exampling showing make list returns a cons cell list, rather than fusion::list>

...

IOW, instead of:

boost::fusion::vector<...>

there should be:

boost::fusion::list<...>

as the 1st line of the above result type for as_list<source_type>::type.

Does this lead to a problem in your code? The sequence of cons cells is a list (fusion::list can be seen as just a convenient way of building a long cons cell sequence). The cons sequence will behave exactly as a list. Maybe these functions should be called as_random_access_sequence, as_bidirectional_sequence etc? Possibly that would be less misleading?

Thanks dan, you're right, it doesn't make a difference, as I just found out. It appears the original problem was using mpl::filter_view then mpl::transform on that result, and this lead to a wrong result. I'd jumped to the conclusion that the problem was the fusion cons list, but my recent test results indicate that the problem is in mpl::filter_view.