Boost logo

Boost :

From: scleary_at_[hidden]
Date: 2000-12-20 11:22:54

> While looking at revamping STLport's iterator debugging code using
> iterator_adaptors as an example, it occurred to me that the wrapped
> type need not be an iterator at all. Given the right policies class
> and traits, you can make an iterator out of anything. It leaves me
> wondering whether we should loosen the constraints for usage, and if
> so, what exactly we're describing. Thoughts?

I totally agree. The Integer Range adaptor already does this. I see the
resulting semantic requirements for the underlying type as being quite
loose, mostly existing as prerequisites on the semantic descriptions of the


Boost list run by bdawes at, gregod at, cpdaniel at, john at