Re: GCC optimizes integer overflow: bug or feature?

[Brooks Moses <brooks dot moses at codesourcery dot com>]

Andrew Pinski wrote:
> On Tue, 2006-12-19 at 06:54 +0100, Ralf Wildenhues wrote: 
[quoting Paul Eggert]
> > > Surely the GCC guys care about LIA-1.  After all, gcc has an
> > > -ftrapv option to enable reliable signal generation on signed
> > > overflow.  But I'd rather not go the -ftrapv route, since that
> > > will cause other problems. 
>
> Lets see, the C standard is very very specific about signed overflow as
> being undefined and LIA-1 says it is defined, well I guess I follow the
> C standard rather than LIA-1.

I have no position on the greater argument, but that seems like a very 
misleading reference to how standards work.  The choice is not one or 
the other; "undefined" does not mean that you are _required_ to make it 
break code if people use it.  And it is certainly a respected practice 
to define additional standards that go _on top of_ other standards, and 
add constraints that were not present in the original while retaining 
all of the original standard, and define things that would otherwise be 
undefined.

So the question here is _not_ one or the other.  It is assumed as given 
that GCC follows the C standard; the question is merely whether GCC 
should also follow the LIA-1 standard as well.

Now, if your argument is that following the LIA-1 standard will prevent 
optimizations that could otherwise be made if one followed only the C 
standard, that's a reasonable argument, but it should not be couched as 
if it implies that preventing the optimizations would not be following 
the C standard.

- Brooks