OK.
On Wed, Aug 2, 2017 at 5:15 PM, Paolo Carlini <paolo.carl...@oracle.com> wrote: > Hi, > > by and large this issue seems already fixed in mainline, ie, we don't ICE > anymore during error recovery, but I noticed that we try to prettyprint > constructor and destructor as expressions and we end up with ugliness like: > > error: taking address of constructor ‘((C<int>*)this)->*A::__ct ’ > > thus the below, tested x86_64-linux. Ok? > > Thanks, > Paolo. > > ////////////////////
