Re: [boost] Macros at namespace scope: trailing semicolon?

I'm amazed at how this particular thread has exploded. My $0.02: I've yet to see any arguments that I consider compelling to support the "always require camp". I expect direct textual substitution from a macro -- if it's expanding into a single statement, it ought to be expanding to a series of characters that includes the semicolon signalling the end of that statement. Other possibilities would only cause confusion. I'm firmly in the "never require a user to append a semicolon to a macro themself" camp. Dave