(*((int8 *)&x+1))

[georpo]

Can someone please explain this code?

It is in 24128.c driver

[Ttelmah]

It's potentially a slightly inefficient way of returning the high byte of the
variable 'x'.
Inefficient, since you are taking the address of 'x', incrementing this
and then taking the byte addressed by this. However the compiler
optimiser is quite smart and in fact will just automatically solve all this
at compile time and just return the next byte, so it works OK.

[georpo]

I understood what is does but I can not even imagine what the syntax says.
I guess the make8() function would be better?
Or a simple >>=8 & 0xFF would be even better?

Thanks!
_________________
George.

[Ttelmah]

make8 yes.
The 'simple' >>=8 &0xFF is potentially nearly as inefficient, but again
the optimiser is good and tidies this to match the make8.
It's inefficient, since it implies accessing the whole 16bit variable and
rotating this. On a chip which is '16bit', this would be OK, but on the
8bit PIC, this potentially means doing several operations....

Key is that all these are made 'efficient', when dealing with a variable at
a fixed address, by the optimiser 'knowing' that they just mean to access
the next byte, and using the direct single byte access.....