On Thu, Apr 21, 2011 at 10:49:05PM -0400, Jason Merrill wrote: > On 04/21/2011 08:50 PM, Nathan Froyd wrote: >> On Fri, Apr 08, 2011 at 01:50:24PM -0400, Jason Merrill wrote: >>> As jsm and richi said, using input_location like this is a regression. >>> Can we use DECL_SOURCE_LOCATION (label_decl) instead? >> >> I went off and tried that; some callers provide a NULL label_decl. > > Hunh. How does that work? They fill in CASE_LABEL later? Can that be > changed?
Yeah, tree-eh.c:lower_try_finally_switch. I don't know how easy it is to fix; it certainly looks non-trivial. -Nathan