[boost] Re: Move semantics (was: Boost.Threads Locking Dilemma)

From: Alexander Terekhov (terekhov_at_[hidden])
Date: 2002-05-18 15:19:25

Peter Dimov wrote:
[...]
> An entity that is that close to std::auto_ptr<M> (only the deallocator is
> different) should be simpler to build using the framework (which presumably
> can duplicate std::auto_ptr out of the box) - or am I missing something?

Nope, IMO. BTW:

http://groups.google.com/groups?selm=c29b5e33.0201290818.2744d2e6%40posting.google.com

"....
 For *mutable* 
 "lazy-singletons" you need a lock anyway, 
 so I think that pthread_once does not provide 
 you much at all in addition to a statically 
 initialized mutex, which could be used for 
 both construction and further usage sync. 
 (static get() method could just "return" a 
 thread-private auto-"smart_ref/ptr" which 
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 would just automatically unlock on its 
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 destruction or perhaps even a thread-shareable 
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 class instance that would provide extra lock/
 unlock calls in addition to the singleton's 
 interface, for example).
 ...."

regards,
alexander.