On Wed, Apr 4, 2018 at 8:48 PM, Paolo Carlini <paolo.carl...@oracle.com> wrote: > I'm really happy to report that these 5 ugly lines are causing an actual > bug. Seriously, not considering the formatting, the problem is that we > really want to keep 'type' in sync, because we are using it below before > returning. Note that we don't regress location-wise because either > explicitly or implicitly we pass input_location anyway. Finishing testing on > x86_64-linux.
Hmm, I wonder why we need to make the type error_mark_node at all, given that the problem is with the declarator. But this patch is OK. Jason