OK. Jason
On Tue, May 31, 2016 at 8:54 AM, Paolo Carlini <paolo.carl...@oracle.com> wrote: > Hi, > > the primary issue is already fixed, we don't ICE anymore, but the location > of the error message is (badly) incorrect, because > check_static_variable_definition doesn't use DECL_SOURCE_LOCATION. Just > doing that, consistently, plus a number of additional column checks in > existing testcases amounts to most of the patch. There is a subtlety though: > as shown by constexpr-static8.C we have been giving redundant diagnostics > about the out-of-class definitions of the erroneous in-class declarations. > This is particularly evident now because we would have to get right the > location of the latter too and DECL_SOURCE_LOCATION doesn't help much for > those. However, I think we simply want to suppress those messages, thus the > early return at the beginning of check_static_variable_definition. The > latter is called only by cp_finish_decl, thus conceivably the same logic > could be easily shuffled around.... > > Tested x86_64-linux. > > Thanks, > Paolo. > > ///////////////////////
