Re: [rfc] [2/2] Remove macros TARGET_{FLOAT,DOUBLE,LONG_DOUBLE}_{BIT|FORMAT}

["Ulrich Weigand" <uweigand@de.ibm.com>]

Markus Deuling wrote:

> this patch removes following macros from gdbarch.sh: 
> 
> * TARGET_FLOAT_BIT
> * TARGET_FLOAT_FORMAT
> * TARGET_DOUBLE_BIT
> * TARGET_DOUBLE_FORMAT
> * TARGET_LONG_DOUBLE_BIT
> * TARGET_LONG_DOUBLE_FORMAT 


   /* NOTE: The new architecture variable is named ``current_gdbarch''
-     so that macros such as TARGET_DOUBLE_BIT, when expanded, refer to
+     so that macros such as gdbarch_double_bit, when expanded, refer to
      the current local architecture and not the previous global
      architecture.  This ensures that the new architectures initial
      values are not influenced by the previous architecture.  Once

Well, replacing the macro by a gdbarch function invalidates the whole point
the comment is making ;-)  You should either use another macro (that still
is a macro) as example, or -once there are no macros left- remove that
comment.


+  builtin_type_float = build_flt (gdbarch_float_bit (current_gdbarch), "float",
+				  gdbarch_float_format (current_gdbarch));
+  builtin_type_double = build_flt (gdbarch_double_bit (current_gdbarch),
+				   "double",
+				   gdbarch_double_format (current_gdbarch));
+  builtin_type_long_double = build_flt (gdbarch_long_double_bit
+					  (current_gdbarch),
+					"long double",
+					gdbarch_long_double_format
+					  (current_gdbarch));
It might be nicer to move the built_flt to its own line
   builtin_type_long_double
     = build_fld (gdbarch_long_double_bit (current_gdbarch),
              ... 

Otherwise this is OK, thanks!

Bye,
Ulrich

-- 
  Dr. Ulrich Weigand
  GNU Toolchain for Linux on System z and Cell BE
  Ulrich.Weigand@de.ibm.com