--- In boost@[...].. , Kevlin Henney <kevlin@[...]..>  wrote:
>  In message <01022215243600.04969@[...]..>, Vladimir Prus
>  <ghost@[...]..> writes
>  >Kevlin Henney wrote:
>  >
>  >>>At the moment lexical_cast<string>(whatever) will fail if
'whatever'
>  >decides
>  >>>to insert spaces in its representation. I think that it is not
very
>  >>>convenient.
>  >
>  >>You're right that it's not convenient, but I would hesitate to
make
>  >>std::string a special case.
>  >
>  >I really think that making std::string a special case is OK,
because:
>  >1. It will always be a special case, either inside lexical_cast or
in user
>  >code.
> 
>  I was referring to std::string as a special case of
std::basic_string,
>  but that's something that can be accommodated without a problem, and
>  also to other string classes. However, in the latter case if the
'API'
>  is opened up, that offers a hook for extension.
> 
>  >3. Actually, string's operator>> has no choice but to stop when it
encounters
>  >space, for it has no knowledge what was put in the stream.
> 
>  Well, it does have a choice, but that's a historical design decision
we
>  can't easily retake :->
> 
>  >> A more general solution would be to replace the expression
>  >
>  >> ... || !(interpreter >> result) || ...
>  >
>  >> by
>  >
>  >> ... || !interpret(interpreter, result) || ...
>  >
>  >Yes, this is shurely better. What are chances that something like
this will
>  >be added?
> 
>  High :-) When I get a moment I will try to roll this change in with
the
>  proposed floating point mods, plus tests and changes to docs.
> 
>  Kevlin

Have you gotten anywhere on this Kevlin? I got bit by passing in an
empty string yesterday and it really made me wish for a string to
string specialization (overload?).