Re: [boost] ublas interoperability

From: Kresimir Fresl (fresl_at_[hidden])
Date: 2002-09-27 07:11:11

Michael Stevens wrote:

> Hmm.. Interesting. I've given this a bit of thought before and concluded 
> that the .begin() is better. When used where a pointer is required, it 
> will ONLY compile if the iterator is a pointer. This is a good thing; it 
> almost certainly is not a good idea to use the linear memory after the 
> first entry unless this is true.
> As you say all this only works if storage type has continuous storage. 
> To keep things simple I would simply define these two properties 
> (iterator==pointer, linear storage) as being equivilent!

I don't agree. ``iterator==pointer'' does not follow from ``storage is 
linear''.

Issue 69 of ``C++ Standard Library Issues List'' is
``Must the elements of a vector be in contiguous memory?''
(http://anubis.dkuug.dk/JTC1/SC22/WG21/docs/lwg-defects.html#69).

Proposed resolution, which has status TC (``The full WG21 committee
has voted to accept the Defect Report's Proposed Resolution as a
Technical Corrigenda''), is:

``Add the following text to the end of 23.2.4, paragraph 1:

The elements of a vector are stored contiguously, meaning that if v
is a vector<T, Allocator> where T is some type other than bool, then
it obeys the identity &v[n] == &v[0] + n for all 0 <= n < v.size().

Rationale: The LWG feels that as a practical matter the answer is
clearly "yes". ''

But there is also a note:

``(Please note that this is entirely separate from the question of
whether a vector iterator is required to be a pointer; the answer to
that question is clearly "no," as it would rule out debugging
implementations)''

fres