On 03/01/14 07:55 PM, Rodney Rushing wrote: > I don't think there is anything invalid about reinterpret_cast of > point/int - that's the whole purpose.
No that was just ugly and in that case invalid since sizeof(void*) != sizeof(int) But generally speaking I have seen MSVC rejecting valid C/C++ code and accepting invalid C/C++ code. gcc is more strict about that. The current solution seems to work. > My point was that the actual > LeoCAD compiler settings between the IDE, command line, and platforms > is not equivalent so I get different warnings depending on where I'm > building it - there's probably some qmake massaging that would be > necessary to make that more consistent. Even then. Trust me. Not my first rodeo in cross platform land. > Even then some of the warnings are mutually exclusive. For example, > unreferenced formal parameters - if I implement an interface and > don't use an argument VS will warn about unreferenced args. If I put > a dummy reference in to make VS happy then gcc warns that the > reference does nothing. That's because you should just comment or remove the variable name from the argument list. Perfectly valid C++ even for Microsoft. Hub _______________________________________________ Leocad mailing list [email protected] https://list.gerf.org/listinfo/leocad
