[Bug c++/101480] [11 Regression] Miscompiled code involving operator new

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101480

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to fail|11.1.0                      |

--- Comment #22 from Richard Biener <rguenth at gcc dot gnu.org> ---
(In reply to hubicka from comment #21)
> Hi,
> note that also tree-ssa-structalias has:
> /* If the call is to a replaceable operator delete and results            
>    from a delete expression as opposed to a direct call to                
>    such operator, then the effects for PTA (in particular                 
>    the escaping of the pointer) can be ignored.  */                       
> else if (fndecl                                                           
>          && DECL_IS_OPERATOR_DELETE_P (fndecl)                            
>          && gimple_call_from_new_or_delete (t))                           
>   ;                                                                       
> I wonder if this is safe...

I think the reasoning is that for a delete expression the storage passed
is released afterwards and thus the delete call cannot be an escape point
(as opposed to the 'new' call) for this very object.