http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57880
--- Comment #8 from Paolo Carlini <paolo.carlini at oracle dot com> --- If we have a straightforward explanation for why the very weird code isn't causing problems, I think the fix almost qualifies as obvious, if properly tested. Do we? Does it only amount to a small memory leak or a small buffer overrun, something similar? In any case, yes, gcc-patches@ CC Jason Merrill (there aren't *that* many C++ maintainers on gcc-patches named Jason ;)
