On 05/07/2012 07:16 PM, Jason Merrill wrote:
I don't think the diagnostic is meaningless; since a() is ill-formed, b() has type void.
Sure we do understand it ;)
But we ought to give a more helpful message when an expression used as a truth value has an unsuitable type.Thus something much more general than the lambda context of the PR. Ok, I can see what I can do.
Thanks! Paolo.
