[cpp-threads] static locals

[Peter Dimov]

Boehm, Hans wrote:

> 2) (More important to me personally.)  Distinguishing the static
> and dynamic case here seems very ugly to me.

The standard already distinguishes between static and dynamic 
initialization. See 6.7/4 and 3.6.2. We aren't inventing anything new, we 
are just clarifying the semantics of "the first time control passes through 
its declaration" by saying what happens in the presence of control races, 
and specifying the visibility semantics of "is considered initialized" 
(citations are from 6.7/4).