[Bug c++/58781] Unicode strings broken in a strange way

[3dw4rd at verizon dot net]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58781

Ed Smith-Rowland <3dw4rd at verizon dot net> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #31061|0                           |1
        is obsolete|                            |

--- Comment #2 from Ed Smith-Rowland <3dw4rd at verizon dot net> ---
Created attachment 31064
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=31064&action=edit
If decltype of user-defined lieteral of a function call could res


Question: Does the decltype problem happen with user-defined numeric literals?
Answer: No.

Question: If I call the operator""_s explicitly in decltype is there a problem?
Answer: No.

Remaining questions...

Question #1: Is this a regression?  There was work for decltype(auto).

Question #2: in cp_parser_decltype_expr what path should be taken for a normal
function call?