RE: regression in qmtest_libstdcpp

 
>Thanks for diagnosing this. Sounds like a TSD
>since the tests have been changed upstream.
>What do others think? If possible I'd like to
>get to closure on PR 80 soon

The tests changed to match an implementation change
(see my other message on this), so it's not necessarily
a TSD. 

I'd be somewhat surprised if we actually specified
this behavior, but it would be good if we could figure
that out.  I copied Jakub Jelinek on the other message
in the hopes he could enlighten us about the impacts.