> Anybody know how to simplify (or more specifically
 > optimise) the code in the raw_to_mystruct function?  I
 > was thinking along the lines of

 >   x->thang = (raw[3] << 8) | (raw[2] << 8) | (raw[1]
<< 8) | raw[0];

Nope, won't work. You're ORing all 4 bytes into the one byte.

x->thang = ((unsigned long)raw[3] << 24) |
           ((unsigned long)raw[2] << 16) |
           ((unsigned long)raw[1] << 8)  |
           (unsigned long)raw[0];

The cast prevents overflow since the default conversion before the bitshift
will be to an int, which will overflow on < 32 bit systems.

Whether or not the code ends up being faster depends on your compiler's
optimisation capabilities.

--- MaltEd/2 1.0.b6