https://gcc.gnu.org/bugzilla/show_bug.cgi?id=117457
--- Comment #5 from Andrew Pinski ---
So the generated code on the gimple level looks fine, which means maybe _M_end
is incorrect and got over-written somehow.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=117457
--- Comment #4 from Andrew Pinski ---
So from my reading the sanitizier output, it seems like a check for `_M_current
== _M_end` is missing or is being optimized away incorrectly. I think the
latter.
0x5621f46983b4 is located 0 bytes after glob
