I'm seeing strange errors when my C++ code has min() or max() calls. I'm using Visual C++ compilers.
Check if your code is including the windows.h header file and either your code or other third-party headers have their own min()/max() definitions. If yes, then prepend your windows.h inclusion with a definition of NOMINMAX like this:
#define NOMINMAX #include <windows.h>
Ugh... scope it, dude:
I haven't used it in years but from memory boost assigns min and max too, possibly?
Another possibility could be from side effects. Most min/max macros will include the parameters multiple times and may not do what you expect. Errors and warnings could also be generated.
max(a,i++) expands as ((a) > (i++) ? (a) : (i++))The () in the expansion are to avoid problems if you call it with formulae. Try expanding max(a,b+c)
afterwards i is either plus 1 or plus 2
Honestly, when it comes to min/max, I find it best to just define my own:
#define min(a,b) ((a) < (b) ? (a) : (b)) #define max(a,b) ((a) >= (b) ? (a) : (b))
Since Windows defines this as a function-style macro, the following workaround is available:
int i = std::min<int>(3,5);
This works because the macro
min() is expanded only when
min is followed by
(, and not when it's followed by