[v3] c++0x <mutex> and <condition_variable>

[Paolo Carlini]

Kai Henningsen wrote:
> Instead of sprinkling #ifs, how about just supplying a single-thread
> implementation of what __GTHREADS offers? Most of those would
> presumably be dummy functions which'd vanish on inlining.
>   
I agree, I agree. However, some functions apparently don't use low-level 
gthread facilities directly and I'm wondering if a mix of the two 
strategies isn't the best solution (I mean, some code also in the 
trivial form corresponding to __GTHREADS undefined, protected by one 
single #ifdef, as you are saying, plus some code in common between the 
two cases)

Paolo.