On 15/02/13 10:27, Michiel Konstapel wrote:
> And for fun, try the quiz at http://blog.regehr.org/archives/721.
> Then cry as you realize how much broken code there must be out
> there...
>
There are a number of inaccuracies in that quiz. Question 3 asks for
the evaluation of "(unsigned short) 1
I traced this down to the value range propagation phase in gcc, which is a
tree optimization that occurs long before anything that's MSP430-specific.
This phase replaces the value of i8 with a constant 127 after the first
output.
So I'm gonna have to agree with Michiel that the code relies on unde
mspgcc-users@lists.sourceforge.net
> Subject: Re: [Mspgcc-users] msp430-gcc 4.6.3 optimizer killing my code
>
> Signed overflow is actually undefined behaviour in C/C++, and the
> compiler is free to do whatever it wants in that case. For gcc, there's
> a flag to make signe
Signed overflow is actually undefined behaviour in C/C++, and the compiler is
free to do whatever it wants in that case. For gcc, there's a flag to make
signed integer arithmetic wrap around, though:
-fwrapv
This option instructs the compiler to assume that signed arithmetic
ov
