https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65683
--- Comment #18 from yuta tomino <demoonlit at panathenaia dot halfmoon.jp> --- Oh, I 'm sorry again. It's been a while for me too, and I seem to have forgotten a lot. I reread the RM and confirmed that the master of access type wasn't that kind of thing. Please allow me to retract my previous post. I also understand that the details of the type declared in “limited with” have not been analyzed at the time of the compiler decided to generate the master.
