boost gil compile error because of type narrowing

[ryan.burn@…]

line 204 in boost/gil/channel_algorithm.hpp if (src > unsigned_integral_max_value<uintmax_t>::value - div2) gives me this error with clang error: non-type template argument

evaluates to -1, which cannot be narrowed to type 'unsigned long' [-Wc++11-narrowing]

struct unsigned_integral_max_value : public mpl::integral_c<UnsignedIntegralChannel,-1> {};

can you just replace with std::numeric_limits<unitmax_t>::max()?

[william@…]

This error persists in Boost 1.55 using clang of Xcode 5.1 on Mac OS 10.9 (Don't see how to select multiple values for the version field here!)

[William Gallafent <william@…>]

… incidentally, I went for a std:: numeric_limits <UnsignedIntegralChannel>:: max () instead … not sure which is more correct in this context.

[anonymous]

I am facing the same problem in the same file, with Boost 1.57, on line 54. Does someone knows the status of this ?

[frederic.bron@…]

patch

[frederic.bron@…]

This is easy to to fix. integral_c is defined by: template< typename T, T N > struct integral_c;

So the 2nd template parameter is of type T=UnsignedIntegralChannel.

Therefore, -1 can easily be replaced by either std::numeric_limits<UnsignedIntegralChannel>::max() or static_cast<UnsignedIntegralChannel>(-1)

See patch attached.

[Mateusz Loskot]

Duplicate of #7270 (already fixed)