On May 8, 2011, at 1:54 PM, Dave Abrahams wrote:

On Sat, May 7, 2011 at 5:50 PM, Neil Groves wrote:

...

Additionally I wondered what you thought about moving the Boost.Range internal any_iterator implementation into Boost.Iterator.

A capital idea!

+1 I've my own simplistic any_iterator, with a pending todo of adding SBO. I recently ran across Neil's any_iterator in a detail namespace of a boost version I'm not yet using, and was hoping it might become a public interface.

On Mon, May 9, 2011 at 1:24 PM, Jeff Flinn wrote:

Kim Barrett wrote:

...

On May 8, 2011, at 1:54 PM, Dave Abrahams wrote:

...

On Sat, May 7, 2011 at 5:50 PM, Neil Groves wrote:

...

Additionally I wondered what you thought about moving the Boost.Range internal any_iterator implementation into Boost.Iterator.

A capital idea!

+1

I've my own simplistic any_iterator, with a pending todo of adding SBO. I recently ran across Neil's any_iterator in a detail namespace of a boost version I'm not yet using, and was hoping it might become a public interface.

+1

I didn't know about it either.

I have discussed the promotion of any_iterator from the Boost.Range internals into the Boost.Iterator library with Dave Abrahams. We both think that this would be useful. I have started implementing test cases to cover the additional use-cases that are not currently tested as part of Boost.Range. I shall then write appropriate documentation before any_iterator can then be reviewed by Dave before the final promotion to the public interface. I wanted to let everyone know the intent, so to enable decisions with-respect to using the existing non-public implementation.

Jeff

Regards, Neil Groves