Re: [boost] [xint] Boost.XInt formal review

Subject: Re: [boost] [xint] Boost.XInt formal review
From: DE (satan66613_at_[hidden])
Date: 2011-03-07 11:13:10

on 07.03.2011 at 14:02
 Ivan Le Lann wrote :
> Without COW, the iterators used in another thread are not invalidated,
> since no data is shared.

> Anyway, this whole thing sums up to: let me choose the container that suits
> my needs.

either of us don't understand what we are talking about
suppose we have two cow implemented objects in different threads
sharing exactly the same data
lets assume that the ref counting is not atomic and to be fair that
only one thread modifies its object
then
thread#1                     thread#2
object a;                    object b;   //object share the same data
//"slim" iterators point to exactly the same data
iterator i = a.begin();      interator i = b.begin();
//a is not modified and      b.modify(); //modifying 'b' so its
//the iterator is still                  //iterator in this thread is
//valid in this thread                   //invalid now

in thread#2 'b' gets new storage so the declared iterator invalidates
but 'a' in thread#1 even after 'b' is modified still holds the same
data so there is no reason for iterator to be invalidated
i see no problems with iterators here
what do i miss?

-- 
Pavel
P.S.
if you notice a grammar mistake or weird phrasing in my message
please point it out