"David Abrahams" wrote in message news:u1xhk3ohg.fsf@boost-consulting.com...

Matt S Trentini writes:

...

Heya Boosters,

I ran into a problem using boost::bind yesterday and, although I found a workaround, I've got a few questions. For the record I'm using boost 1.30.

I stumbled onto the issue that boost::bind doesn't define the argument_type typedef (I was trying to use std::not1). After trawling

Don't the boost::functional implementations of these function object adaptors play nicer with boost::bind? I think I've used these just for this reason in the past. Jeff F

David Abrahams wrote:

Because it doesn't know what you will pass to it ;-)

Consider:

struct polymorphic_function_object { template <class T> T operator()(T x) const; };

What should argument_type be in the result of

bind(polymorphic_function_object(), _1)

Ah. It's all become clear. :) Makes perfect sense.

Have you looked at the Boost lambda library? It incorporates all the operators.

Unfortunately I'm stuck working with VC7.0 which, as I understand, won't cut the mustard for lambda... Thanks, Matt

Peter Dimov wrote:

...

o Has the operator! addition been incorporated into later versions of boost?

Not yet, but you can always use the snippet in the link you posted to add it yourself.

Yep, I've done just that. Thanks for providing such a neat little snippet! :)

...

- If not, why not?

It will probably be in the release after 1.32. While operator! is trivial to implement, the relational operators that I also want to add are a little more involved because they interact with another experimental feature, equality comparisons of bind() function objects (required by function<>::contains).

Sounds interesting! :) Cheers, Matt