https://gcc.gnu.org/bugzilla/show_bug.cgi?id=17314
--- Comment #25 from Anthony Sharp <anthonysharp15 at gmail dot com> --- Hopefully (pending approval) the original bug is now fixed, even if it did take a long time! I agree that compilers should match the standard where reasonable to do so (like in this case), but I can't comment much specifically on the issue of whether deriving from a virtual private inheritance should be ill-formed (it seems the general consensus is that it should be, although one can find people debating the opposite on various forums e.g. https://stackoverflow.com/questions/2371176/c-private-virtual-inheritance-problem). Although I could probably go and try find out, I think it'd be best to leave that to those who know a bit more about the topic than me; this is only my first ever patch submission after all :)
