Implementing Exception Propagation (N2179)

[Sebastian Redl]

Peter Dimov wrote:
> FWIW, my general intent was for rethrow_exception to behave exactly as 
> throw; does in the "synchronous" case. In this case this would imply 
> acquiring the current handlers as mandated by the standard.
By "exactly as throw;", do you mean the argument-less rethrow statement?
In other words, do you believe that G++'s current behaviour for rethrow 
is in error?

Sebastian