https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68281
--- Comment #1 from Josh Stone <jistone at redhat dot com> --- I suspect this may actually be acceptable, NOTABUG, because I can't think of any uninitialized data that would make the effective behavior different than if it had been checked in the proper order. Both conditions still have to be met, and there's no bad pointer deref or anything. I guess the optimizer just took a guess which side is most likely to be the deciding factor. But I'm not sure if valgrind can possibly recognize that this uninit condition is actually ok. (I can just give these values a default init as a workaround, of course.)
