[boost] Re: Re: [shared_ptr] Interlocked* - possible bug?

From: Peter Dimov (pdimov_at_[hidden])
Date: 2004-09-27 14:36:16

John Torjo wrote:
> Ben Hutchings wrote:
>
>> expired() may read the count as still being 1 and so return false,
>> but I believe the wnd_shared_ptr constructor will catch the fact
>> that the pointer really has expired.  Unfortunately I can't yet
>> see the code to confirm this!
>
> By looking at the code, I would think not. But I may be wrong.

As the comment says, the if( expired() ) test is only an optimization. 
Remove it if it makes you feel better. ;-)

The main idea is that a weak_ptr can never transition by itself from an 
expired() state back to a !expired() state unless you explicitly assign to 
it.

Because you are in weak_ptr::lock at the moment - a read access - you know 
that no other threads assign to *this, because this would be a violation of 
the "basic thread safety" requirements.

Therefore, if expired() returns true, it will continue returning true for 
the foreseeable future, and you can skip the rest of lock().