On Tue, 2018-05-01 at 07:18 -0400, Nathan Sidwell wrote: > On 04/30/2018 08:29 PM, David Malcolm wrote: > > Following on from the thread on the "gcc" list here: > > > > https://gcc.gnu.org/ml/gcc/2018-04/msg00172.html > > > > here's an updated version of Jonathan's patch, which: > > + { > > + tree valtype = TREE_TYPE (DECL_RESULT (fndecl)); > > + if (TREE_CODE (valtype) == REFERENCE_TYPE > > + && same_type_ignoring_top_level_qualifiers_p > > + (TREE_TYPE (valtype), TREE_TYPE > > (current_class_ref))) > > While this is probably close enough, you could > *) use convert_to_base to include cases where the return type's a > base > of the current object. > *) convert_to_base would also allow dropping the REFERENCE_TYPE > check > here, so icky code returning by-value could also be caught. > > Up to you. But otherwise ok.
Sorry about the belated response; this fell off my radar for some reason. I looked at updating it to support the cases you suggest, but I wasn't happy with issuing the fix-it hint for them, so I've gone with the patch as-is. Committed to trunk as r263298 (after rebasing and re-testing) Thanks Dave
