Re: [boost] stride_iterator_policy

From: Jeremy Siek (jsiek_at_[hidden])
Date: 2001-08-27 09:46:59

On Mon, 27 Aug 2001, David Abrahams wrote:

david.> This is a good idea, but it is common to need to select
david.> the stride at runtime.

Ahh, I missed the compile-time aspect while skimming the code. Yes, I
agree that the stride is often times determined at run time, and should
therefore be a data-member, not a template parameter.

david.> A couple of my colleagues made such a policy and
david.> associated generators; perhaps you'd like to take a look?
david.> If so, I'll see if I can convince them to post it.
david.>
david.> An intersting fact about striding over matrices: to avoid
david.> undefined behavior you usually need to allocate an extra
david.> row (or column) so that there's a valid end iterator
david.> position.

Ugh.

Ciao,
Jeremy

----------------------------------------------------------------------
 Jeremy Siek                          http://php.indiana.edu/~jsiek/
 Ph.D. Student, Indiana Univ. B'ton   email: jsiek_at_[hidden]
 C++ Booster (http://www.boost.org)   office phone: [phone]
----------------------------------------------------------------------