Re: "new" statement and -fno-exceptions

[Paolo Carlini <pcarlini at suse dot de>]

Paul Koning wrote:

> > > > > > "Steven" == Steven L Zook <SLZook@Qualstar.com> writes:
> > > > > >            
> Steven> If you compile with -fno-exceptions, and you do not compile
> Steven> with -fcheck-new, don't constructors on new class objects
> Steven> blow up with NULL pointers when the new for that class (or a
> Steven> more global new) fails?  
>
> No, that's what surprised me too.  And, from the sound of it, Paolo
> reacted the same way.
>  
Yes, sorry for a little bit of confusion, but...

> I suppose one could argue that the design is ok as it stands -- if so,
> then the documentation clearly needs help.
>  
... in hindisght, this is a nice occasion to improve the documentation 
and also,
perhaps, give more (and more consistent ;) options to the users.

Thanks,
Paolo.