Re: Fw: [patch] Make std::tr1::shared_ptr thread-safe.

[Robert Zeh <rzeh at efs-us dot com>]

Paolo Carlini wrote:

> Alexander and Peter,
>
> I'm trying to understand the basic implications of your interesting
> analyses. Can you possibly point out some specific broken codepaths?
> Much better, prepare, pragmatically, one or more C++ snippet using the
> library that fails in a consistent way on a specific arch (!= IA32, of
> course). Sooner or later, after all the deep analyses we have to start
> *doing* something and that would be a nice starting point, in my
> opinion: right now we are simply *totally* lacking feedback from the
> users about the brokeness.
>
> Thanks in advance,
> Paolo.
>  
As a user who reads the mailing list and has experience with 
multi-threaded applications,
my plea is for the first version of libstdc++'s tr1::shared_ptr to use a 
mutex.  I can use a
mutex version with confidence in my multi-threaded applications.  The 
discussions
I've seen on this list have eroded my confidence in a libstdc++ lock 
free implementation.

A mutex version should be as fast as the currently released 
boost::shared_ptr,
which sees widespread use.  The mutex version won't be as fast as a lock 
free version,
and it will probably be slower then the upcoming lock free 
boost::shared_ptr.
But it stands a greater chance of working correctly.

Just my $.02.

Robert Zeh
http://home.earthlink.net/~rzeh