Good morning, > > sorry but I don't understand that about compatible and incompatble types. > > > > My second source is the same as the first one, except that I have removed > > one level of indirection. > > Which was in both cases declared as 'const'. > I'm aware of that. Read what I wrote again, slowly. Can you explain me what is ment with compatible and incompatible? As related to why the one is a compatible and the other is not? For me, there is still no difference in both examples .... > > You say that to be correct according to (that piece of) standard, > > the must also be 2 warnings in the case of single direction? > Well, none of this is absolutely required, as far as I can see: these > are warnings, not errors. > Note that I am *not* claiming that we are being pefectly consistent in > what we do -- I'm explaining what we do. This is not necessarliy > totally consistent, but AFAICS it is a good compromise between too few > and too many warnings. Yeah, OK. Not all warnings are telling me that do something wrong. Some only say: "he, look at what you doing, is it correct?". I was not trying to offend GCC, only have a discussion about why it is done or not. If we come to the conclusion that this warning could be superflous, that won't mean it is a bug .... Jacob.
