https://gcc.gnu.org/bugzilla/show_bug.cgi?id=56578
--- Comment #6 from Richard Biener <rguenth at gcc dot gnu.org> ---
We now do output 'malloc', with and without -fno-builtin-malloc:
> gcc-nm-15 libxxx.a
a.o:
00000000 T malloc
without -fno-builtin-malloc we FAIL, with -fno-builtin-malloc we segfault
in puts("OK") ....
So I think this is somehow fixed when properly using -fno-builtin-malloc
(which we could imply when seeing definitions of builtins ...). But sth
is still odd.
