On Sunday, 3 August 2014 at 21:36:44 UTC, Timon Gehr wrote:
Again, in other cases, the -release program will operate as expected.If I've grasped your point in the previous reply, yes, maybe. ...Ok. This was Johannes' point.
Taken, thanks for the explanation: I agree that you both that optimising on a wrongly written assert could turn in code breakage under the new '-release' semantic.
--- Paolo