Question

I am writing a c application for decoding binary files and I need to be sure about the sizes of my chunks. Reading the documentation I understood that only the minimum size is stated while the maximum depends on the compiler or/and architecture…

so could I do something like : ***PSEUDOCODE

unsigned char byte;
if((byte = ~0) > 0xff){
    typedef (unsigned char & 0xff) byte; /* I know.. ;P */
}else{
    typedef unsigned char byte;
}

should I just apply the bit mask every time I use unsigned char to be sure or is there another way to hard code a size to a type that I don’t know off?

ps: The reason this is important for me is because I going to be doing allot of shifting..
Thank you 😉

Answer 1

Use the types intN_t and uintN_t from stdint.h added in C99 (common values for N are 8, 16, 32, 64). They’re guaranteed to have fixed size.