https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115138
--- Comment #4 from Iain Sandoe <iains at gcc dot gnu.org> --- so I am comparing the output of compiling gcc/d/dmd/opover.d with the stage1 and stage2 compilers. Using -fdump-tree-all. the .005t.original outputs are the same the .006t.gimple outputs already have differences in the numbering of some vars. the first dump file that mentions CSWTCH.15x is .032t.einline (which mentions a reference to the object). AFAICT the symbol CSWTCH.15x has the same value at all times; so it appears to have been created with a different id. I can upload the gimple if it's useful. ( one further experiment - and then I need to punt on this until there's some more focussed way to debug it - I am wondering if the issue has anything to do with the default c++ std being different for stage 1 and 2+ ) - to test this theory, will jam -std=c++11 into BOOT_CFLAGS.
