Re: [boost] Fw: [atomic] review results

Subject: Re: [boost] Fw: [atomic] review results
From: Peter Dimov (pdimov_at_[hidden])
Date: 2011-11-08 15:51:22

Andrey Semashev wrote:

> Surely there may be caveats, as well as in just about every design. But
> according to my practice one almost always compiles the application 
> against a
> fixed version of Boost. It is the other cases I would call contrived and 
> bad.

The version of Boost is not the issue. The code that operates on the atomics 
is. The normal way is for the module that creates the lock-free data 
structure to provide out of line functions which the other modules would 
call. Inlining these creates the possibility of the modules to get out of 
sync with each other. For plugin-type DLL uses, this is basically 
guaranteed.