https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90036
Jeffrey A. Law <law at redhat dot com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Target Milestone|10.0 |11.0
--- Comment #9 from Jeffrey A. Law <law at redhat dot com> ---
I still think extracting the predicate analysis bits out of tree-ssa-uninit.c
is the way to go. Essentially we'd build the predicate guarding the
problematical sprintf call and realize it can never be true.
There's a slight chance the new VRP code could help here as it does have a
backwards substitution/solver engine. I think we'd want to raise the query
what's the range of ownvptr_15 in bb4 and work backwards.
Aldy, want to take a looksie and see if your work could help here.
In either case, I don't see this being resolved in gcc-10.
