From: "Jeremy Siek"

Here's another idea for a "hello world" example. A rewrite of Matt Austern's line_iterator example from his Generic Programming and the STL book. The line_iterator call getline() and then return a std::string from operator*.

On Fri, 26 Jul 2002, David Abrahams wrote: dave> dave> That's neat, but we already have the permutation adaptor, and I dave> don't think it's really a simple enough example. Beman's example dave> just makes a simple iterator, not an adaptor, and I think that's an dave> important feature for a "hello, world"-type introduction. dave> dave> -Dave

Don't you think people will find an iterator which traverses the elements of a real sequence to be more-obvious? That's why I'm suggesting the singly-linked list iterator. ----------------------------------------------------------- David Abrahams * Boost Consulting dave@boost-consulting.com * http://www.boost-consulting.com

OK, here's my (untested) singly-linked-list iterator: typedef std::string slist_data; struct slist_node { slist_node* next; slist_data data; }; struct slist_policies : default_iterator_policies { template <class IteratorAdaptor> void increment(IteratorAdaptor& x) { x.base() = x.base()->next; } template <class IteratorAdaptor> typename IteratorAdaptor::reference dereference(IteratorAdaptor& x) { return x.base()->data; } // initialize, equal covered by default policies }; typedef iterator_adaptorstd::forward_iterator_tag > slist_iterator; typedef iterator_adaptorstd::forward_iterator_tag > slist_const_iterator; -Dave ----------------------------------------------------------- David Abrahams * Boost Consulting dave@boost-consulting.com * http://www.boost-consulting.com ----- Original Message ----- From: "Jeremy Siek" To: Cc: Sent: Friday, July 26, 2002 10:58 AM Subject: Re: [Boost-Users] Re: [boost] Iterator Adaptor Poll

On Fri, 26 Jul 2002, David Abrahams wrote: dave> dave> Don't you think people will find an iterator which traverses the dave> elements of a real sequence to be more-obvious? That's why I'm dave> suggesting the singly-linked list iterator.

The singly-linked list example was the second thing that popped into my mind... I like that a lot too. Perhaps we should do both :)

Cheers, Jeremy

---------------------------------------------------------------------- Jeremy Siek http://php.indiana.edu/~jsiek/ Ph.D. Student, Indiana Univ. B'ton email: jsiek@osl.iu.edu C++ Booster (http://www.boost.org) office phone: (812) 855-3608 ----------------------------------------------------------------------

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

At 11:10 AM 7/26/2002, David Abrahams wrote:

OK, here's my (untested) singly-linked-list iterator:

Yes! Nice! Having this would have saved me a lot of bumbling around. How about supplying slist_node with an operator++ (and eliminating slist_policies::increment), so that you are illustrating the two different possibilities for supplying a required operation. That was something that confused me initially. Another thing that still confuses me is why you need to specify Value, but can let Reference, Pointer, and Distance default. A little cometary in the docs which go with the example to explain the rationale for your choice of template parameter values would be much appreciated. --Beman

typedef std::string slist_data; struct slist_node { slist_node* next; slist_data data; };

struct slist_policies : default_iterator_policies { template <class IteratorAdaptor> void increment(IteratorAdaptor& x) { x.base() = x.base()->next; }

template <class IteratorAdaptor> typename IteratorAdaptor::reference dereference(IteratorAdaptor& x) { return x.base()->data; }

// initialize, equal covered by default policies };

typedef iterator_adaptorstd::forward_iterator_tag > slist_iterator;

typedef iterator_adaptorstd::forward_iterator_tag > slist_const_iterator;