https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118077
Andrew Pinski <pinskia at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |NEW
Last reconfirmed| |2024-12-17
Ever confirmed|0 |1
--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Andrew Pinski from comment #2)
> It is more closely related to PR 117665
yes it is.
Before EVRP:
;; basic block 3, loop depth 0
;; pred: 2
_15 = MEM[(struct b *)this_6(D)]._vptr.b;
_16 = *_15;
_17 = b::c (_1);
;; succ: 4
;; 5
Afterwards:
;; basic block 3, loop depth 0
;; pred: 2
_15 = MEM[(struct b *)this_6(D)]._vptr.b;
_16 = *_15;
;; succ: 4
;; 5
VRP should have cleaned up the abormal edge here.
Folding statement: _17 = b::c (_1);
Queued stmt for removal. Folds to: 0
Folding statement: _setjmp ();
Not folded
Folding statement: .ABNORMAL_DISPATCHER (0);
Not folded
Folding statement: return;
Not folded
Removing dead stmt:_17 = b::c (_1);
Removing dead stmt:_1 = &this_6(D)->D.2889;
Let me look into this.
