https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106922
--- Comment #27 from Jan Žižka <jan.zizka at nokia dot com> --- (In reply to rguent...@suse.de from comment #26) > > They are clearly necessary to fix this bug. What I'm unsure yet about > is the risk of generally enhancing VN for this diagnostic regression. > The enhancement itself is quite small and self-contained which is > why I'm still considering it ;) Understood. And I didn't report this, maybe I should have, with these fixes we had another uninitialized value warning ... it was actually bug so I did fix it ... but the interesting was that same/similar bug was couple lines below and this one the compiler didn't complain about. I'm not sure if all that could be related to your considerations but if it would help I could create a reproducer for this particular case. I guess it is false negative :-) of some sorts.
