[Bug tree-optimization/30252] [4.2 regression] miscompilation of sigc++-2.0 based code with -fstrict-aliasing

["bangerth at math dot tamu dot edu" <gcc-bugzilla at gcc dot gnu dot org>]

------- Comment #14 from bangerth at math dot tamu dot edu  2007-05-01 02:39 -------
Subject: Re:  [4.2 regression] miscompilation
 of sigc++-2.0 based code with -fstrict-aliasing


> >         typedef typed_slot_rep<T_functor> typed_slot;
> >         typed_slot *typed_rep = static_cast<typed_slot*>(rep);
> >         return (typed_rep->functor_)();
> >
> > This code could violate C++ aliasing rules.
>
> But:
>     template <class T_functor>
>     struct typed_slot_rep : public slot_rep
>
> so it looks like it might be valid.

But only, of course, if the pointer actually points to a typed_slot_rep.
Maybe someone who can reproduce the bug can make the following test:
- add a virtual destructor to slot_rep
- change the static_cast to a dynamic_cast
If the object pointed to is not a typed_slot_rep (in which the static_cast
in the original code would have been in error), then the code should
probably segfault or so...

W.

-------------------------------------------------------------------------
Wolfgang Bangerth                email:            bangerth@math.tamu.edu
                                 www: http://www.math.tamu.edu/~bangerth/


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30252