int vs s32

[pippy]

Is there really any difference between say double and f64? whats the point in the irrlicht version of these numbers?

I understand that on really old computers they had different sizes for numbers, but I don't think that Irrlicht would be able to run on those types of machines anyway

Is it just for potential, eg in a 386 a int has somewhere around 60k values, so when the user compiles on that machine he can specify the s32 to have four ints?

[vitek]

The idea is to make it so that you have control over the size [and range] of the values. If you want a 32-bit unsigned integer, you use u32. If you tried to use unsigned or unsigned long you might actually get a 64-bit quantity on some platforms.

Travis

[hybrid]

The f32 typdef could also be used for fixed point classes to come in on machines without a hardware FPU. And yeah, if you'd use a s32/u32 type with less than 32bits you'd be in great trouble.

[Halifax]

Umm, and also it is part of the API, and basically standard. Almost every other API either uses sXX uXX, or qword, dword etc.

I mean it is just a way to make a standard. Look at Insomniac Games code for Resistance. They also use a naming convention, and not C types.

[constchar*]

Another reason you'd use typedef's is for cross-platform compatibility
so say you port your application to platform X and say platform X doesn't
support a certain type so then you can just change the typedef to something
related without having to edit the entire engine.