[Bug c/104506] [12 Regression] ICE: tree check: expected class 'type', have 'exceptional' (error_mark) in useless_type_conversion_p, at gimple-expr.cc:87 on invalid symbol redeclaration

["pinskia at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104506

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
The NOP_EXPR comes from:

(gdb) up
#1  0x0000000000b99a72 in c_process_expr_stmt (loc=266752, expr=<parm_decl
0x7ffff7ff6080 x>) at
/home/apinski/src/upstream-gcc/gcc/gcc/c/c-typeck.cc:11356
11356         expr = build1 (NOP_EXPR, TREE_TYPE (expr), expr);
(gdb) l
11351
11352     /* If the expression is not of a type to which we cannot assign a
line
11353        number, wrap the thing in a no-op NOP_EXPR.  */
11354     if (DECL_P (expr) || CONSTANT_CLASS_P (expr))
11355       {
11356         expr = build1 (NOP_EXPR, TREE_TYPE (expr), expr);
11357         SET_EXPR_LOCATION (expr, loc);
11358       }


But if you had a cast to void, it would come from somewhere else.


The strip is:
      /* Strip away as many useless type conversions as possible
         at the toplevel.  */
      STRIP_USELESS_TYPE_CONVERSION (*expr_p);


Now do we want to add a check inside STRIP_USELESS_TYPE_CONVERSION for
error_mark_node, I am still deciding.