V3 patch causes EH failures in ILP32 mode

[Gabriel Dos Reis]

You Wrote Nathan Myers
> On Wed, Jun 23, 2004 at 11:32:35AM -0700, Mark Mitchell wrote:
>> The other difference is that the compiler may use a bitwise copy in many
>> situations.   ...
>>
>> I would advise against making these kinds of functions explicit, except
>> when necessary.
>
> I agree.  It would be better to remove that check from -Weff, or
> (failing that) add -Weff-anal.  There's nothing quite like having to
> make your code more fragile just to satisfy a compiler warning to
> turn users away from use of compiler warnings.

And that is a real nuisance.  I would like to see -Weffec++ go
out of the picture.

   I have set of abstract classes, in pure classic OO
tradition -- pure virtual functions, no data member; no
deletion through base classes.  Yet the compiler insists
on sending non-sensical remarks about classes having virtual
functions but no virtual destructors.  Yes.  But what would
such a destructor would be good for?  Yes, I know I can say
-Wno-non-virtual-dtor.  But then, the set of flags is getting
too complicated, just to get classic codes :-(