[v3] libstdc++/16612

[Paolo Carlini]

Paolo Carlini wrote:

> Nathan suggests avoiding sharing the empty rep: I think we should 
> explore that,
> is a  promising idea, indeed.

Today, while being out of town, I had the time to study this proposal in 
some detail:
seems absolutely feasible, actually. I'm now working on it and probably 
a draft will
be ready in 1-2 days. I expect a slight slow-down of the empty string 
constructor but
good performance in all the other cases.

Paolo.