Re: [avr-gcc-list] Question about gcc preprocessing and port/pinassignments

[Dave Hansen]

From: David Kelly <address@hidden>
> On Dec 9, 2004, at 7:43 PM, Dave Hylands wrote:
>
> > In the example I showed, a = ( b = 3 ) + 4; we know that a will be
> > assigned 7 and b will be assigned 3. We don't know the order that these
> > assignments will take place in.

It's worse than that.  Given

  b = 3;
  a = b + 4;

we don't know in which order the assignments will take place.  Unless 'a' 
and 'b' are volatile.

> A classic example is that the following won't evaluate the same for all C 
> compilers. We know nothing of the order each a++ will be executed. "b" will 
> usually be either 9 or 11, depending on your compiler and its options altho 
> 8 is also possible.
>
> a = 2;
> b = a++ + 3 * a++;

Technically, other answers are possible, like 42, or your hard drive being 
formatted, launching nuclear missiles, or the proverbial demons flying from 
your nose.

Practically, because the compiler's optimizer is allowed to assume you 
aren't breaking the rules (even though it seems obvious it should be able to 
see that, it isn't required to detect it), you can get some very strange 
answers numerically.  I wouldn't be at all surprised by 12 or 16, for 
example.  And I certainly wouldn't be surprised if 'a' was 3 after the 
statement executed.  I suppose I would be surprised by 42, however...

Regards,
  -=Dave