https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114245
Andrew Pinski <pinskia at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
See Also| |https://gcc.gnu.org/bugzill
| |a/show_bug.cgi?id=115050
Resolution|--- |WONTFIX
Status|UNCONFIRMED |RESOLVED
--- Comment #6 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
So I just debugged a bug report where the vtable storing in the deconstructor
found the bug in their code (PR 115050). It just happens with that store, it
causes a seg fault at runtime. Without the store the code seemly worked but
only on accident and I had to show that a deconstructor was called twice.
Having this store is definitely helpful for debugging code at -O0 and catching
undefined code early on.
So closing as won't fix. removing the vtable store is an optimization rather
than needed for correctness and having the vtable store there is very very
useful for debugabity of programs.
