https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118865
Andrew Pinski <pinskia at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Component|c++ |tree-optimization
Summary|[15 regression] |[12/13/14/15 regression]
|-Wfree-nonheap-object false |-Wfree-nonheap-object false
|positive |positive
Target Milestone|--- |12.5
--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
So there is a false positive but there is also true positive here.
The true positive is due to processJetContainer assumes `jbeg != bend`. Adding:
if (jbeg == jend) __builtin_unreachable();
To the begining of the function gets rid of one reason for the warning but
there is still another.
Note the difference between GCC 14 and 15 is:
# _34 = PHI <&listJets.D.39335._M_impl._M_node.D.18822(4), &MEM[(struct
_List_node_header *)&listJets].D.18822(2)>
vs &MEM[(struct _List_node_header *)&listJets].D.18258 directly.
And getting rid of the true positive makes GCC 14 warn on the false positive.
