* ada-lang.c (assign_component): Use platform-specific integer type

	instead of builtin_type_int32 type.
	(ada_evaluate_subexp) [OP_ATR_SIZE]: Likewise.

	* ax-gdb.c (gen_expr) [UNOP_NEG]: Use platform-specific integer type
	instead of builtin_type_int8 type.
	* valarith.c (value_x_unop): Likewise.
	* python/python-value.c (valpy_absolute): Avoid reference to
	builtin_type_int8 type.

	* eval.c (evaluate_subexp_standard): Use platform-specific integer
	type instead of builtin_type_int8 as EVAL_SKIP return value type.
	* ada-lang.c (ada_evaluate_subexp): Likewise.
	* jv-lang.c (evaluate_subexp_java): Likewise.
	* m2-lang.c (evaluate_subexp_modula2): Likewise.
	* scm-lang.c (evaluate_exp): Likewise.

	* value.h (value_bitstring): Add INDEX_TYPE argument.
	* valops.c (value_bitstring): Add INDEX_TYPE argument, use it instead
	of builtin_type_int32 as base range type.
	* eval.c (evaluate_subexp_standard): Update value_bitstring call.

1 files changed, 2 insertions, 2 deletions

diff --git a/gdb/python/python-value.c b/gdb/python/python-value.c
index ab89842..ebc15dc 100644
--- a/gdb/python/python-value.c
+++ b/gdb/python/python-value.c
@@ -551,8 +551,8 @@ valpy_positive (PyObject *self)
 static PyObject *
 valpy_absolute (PyObject *self)
 {
-  if (value_less (((value_object *) self)->value,
-		  value_from_longest (builtin_type_int8, 0)))
+  struct value *value = ((value_object *) self)->value;
+  if (value_less (value, value_zero (value_type (value), not_lval)))
     return valpy_negative (self);
   else
     return valpy_positive (self);