Boost logo

Boost :

From: jsiek_at_[hidden]
Date: 2000-06-13 12:47:47

Ahh, your new version has the simplification I was asking for... your
implementation of reverse_iterators is now only a few lines, which is
what I wanted :)

What do you think about the friend instantiation workaround that I've
proposed? The one that specializes the wrapper class based on the
category of the iterator being wrapped. The code I've uploaded has
this workaround. There's only one thing wrong with this approach (that
I can think of). If the iterator uses a "custom" tag that derives from
one of the standard tags, then the correct adaptor may not be chosen.



David Abrahams writes:
> From: <jsiek_at_[hidden]>
> > I've uploaded the code as to the vault. All the
> > tests passed for g++ and KCC. I haven't tested the VC++
> > workaround yet.
> Still not sure what advantage we get from your version, but I hope someone
> will do a thorough analysis. Now associative_vector/wrap_iterator/ contains
> my latest versions. All tests pass with MSVC6.3, GCC 2.95.2 and Metrowerks
> CodeWarrior.
> -Dave
> ------------------------------------------------------------------------
> Take your development to new heights. Work with clients like Dell and
> pcOrder. Submit your resume to Visit us at
> ------------------------------------------------------------------------

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