Floating-Point Endianness
How do SDL's endianness functions handle floating-point numbers differently from integers?
Floating-point numbers require special handling for endianness because they use a more complex internal representation than integers.
SDL provides dedicated functions for handling floating-point endianness because the byte-swapping process needs to account for the IEEE 754 format.
Understanding Float Representation
A 32-bit float consists of:
- 1 bit for the sign
- 8 bits for the exponent
- 23 bits for the fraction (mantissa)
Here's a program that demonstrates how SDL handles float endianness:
#include <iomanip>
#include <iostream>
#include "SDL.h"
void PrintBytes(
const char* Label, float Value
) {
auto* Bytes = reinterpret_cast<Uint8*>(
&Value);
std::cout << Label << ": ";
for (int i = 0; i < sizeof(float); ++i) {
std::cout << std::hex << std::setw(2)
<< std::setfill('0')
<< static_cast<int>(Bytes[i]) << " ";
}
std::cout << "(" << std::dec << Value
<< ")\n";
}
int main() {
float Original{3.14f};
PrintBytes("Original ", Original);
// Swap endianness
float Swapped{SDL_SwapFloat(Original)};
PrintBytes("Swapped ", Swapped);
// Swap back
float Restored{SDL_SwapFloat(Swapped)};
PrintBytes("Restored ", Restored);
}Original : c3 f5 48 40 (3.14)
Swapped : 40 48 f5 c3 (3.14)
Restored : c3 f5 48 40 (3.14)Reading and Writing Floats to Files
When serializing floats, we can simply interpret them as bytes and use functions like SDL_ReadLE32() and SDL_WriteLE32() in the same way we would for any other 4-byte value:
#include <SDL.h>
#include <iostream>
// Write a float to an SDL_RWops stream in
// little-endian order
bool WriteFloat(SDL_RWops* rw, float value) {
Uint32 data = *reinterpret_cast<Uint32*>(&
value);
return SDL_WriteLE32(rw, data) > 0;
}
// Read a float from an SDL_RWops stream in
// little-endian order
bool ReadFloat(SDL_RWops* rw, float& value) {
Uint32 data = SDL_ReadLE32(rw);
if (data == 0 && SDL_RWtell(rw) == 0) {
// Check for read error
return false;
}
value = *reinterpret_cast<float*>(&data);
return true;
}
int main(int argc, char** argv) {
SDL_RWops* rw = SDL_RWFromFile(
"example.bin", "wb+");
if (!rw) {
std::cerr << "Failed to open file: "
<< SDL_GetError() << std::endl;
return 1;
}
float valueToWrite = 3.14159f;
if (!WriteFloat(rw, valueToWrite)) {
std::cerr << "Failed to write float."
<< std::endl;
}
SDL_RWseek(rw, 0, RW_SEEK_SET);
// Reset file pointer
float valueRead;
if (!ReadFloat(rw, valueRead)) {
std::cerr << "Failed to read float."
<< std::endl;
} else {
std::cout << "Read value: " << valueRead
<< std::endl;
}
SDL_RWclose(rw);
return 0;
}Read value: 3.14159Key points to remember:
- Use
SDL_SwapFloat()for in-memory byte swapping - For file I/O, it is safe to byte-swap floats to achieve the desired endianness
Byte Order and Endianness
Learn how to handle byte order in using SDL's endianness functions