RE: [boost] Re: Enhanced call_once()

From: Anthony Williams (anthwil_at_[hidden])
Date: 2002-08-07 07:37:06

> From: Peter Dimov [mailto:pdimov_at_[hidden]]
> Sent: 07 August 2002 12:01
> From: "Anthony Williams" <anthwil_at_[hidden]>
> > 
> > However, I'm not sure we want POSIX semantics --- it poses 
> an unnecessary
> > burden on the users to have to specify a static initializer 
> or call a
> > dynamic init function, it is much preferable to just have 
> Mutex() do the
> > Right Thing in all cases.
> 
> Is this possible? Without language/memory architecture 
> support, I mean.

See my other post for Win32Mutex and POSIXMutex skeletons that just have a
default constructor.

My point is that you can have thread-safe lazy dynamic initialization, which
works for static objects without causing init-order problems, if you use the
appropriate primitives for each platform. On POSIX, pthread_once provides
such a mechanism; on Win32, Named Mutexes provide such a mechanism. I expect
that other platforms have other primitives, but I would be surprised if
there wasn't _any_ mechanism for this on a particular platform.

However, you are right that it requires some form of platform support, just
not necessarily at the language level.

Anthony