Cross Platform float

[Foaly]

In config.hpp SFML defines a bunch of fixed-point data types in a cross-platform way (sf::Int16 etc.). I was wondering why are there no floating point data types (i.e. float or double)? I read somewhere that they are not consistent across all platforms either.

[Laurent]

The C++ standard doesn't define a specific encoding for floating point numbers, but on all the platforms that SFML supports you can expect floats to use the IEEE-754 representation with fixed sizes.

[Foaly]

Meaing on windows, Mac and Linux float and double will always be the same size?

[Laurent]

Yes.

[Foaly]

sweet thank you very much!

[MorleyDev]

I feel the need to note that long double will vary from compiler to compiler. On Visual Studio it's no different from a double (8 bytes in size), whilst on gcc it's typically 12 bytes (float is 4 and double is 8 ).

[Mario]

Also keep in mind that there might be different precision/rounding if you're using floats for anything that should stay in sync between platforms.