[patch] Fix GC issue triggered by arithmetic overflow checking

Eric Botcazou ebotcazou@adacore.com
Thu Oct 13 10:59:00 GMT 2016


> I guess it depends on whether temp$8 and temp$low overlap or not.

I don't think so, the code is identical so that would be a more serious bug:

struct cpp_num
{
  cpp_num_part high;
  cpp_num_part low;
  bool unsignedp;
  bool overflow;
};

I think it's just dwarf2out_var_location being sensitive to the ordering.

-- 
Eric Botcazou



More information about the Gcc-patches mailing list