* eval.c (evaluate_subexp): Clean up handling of
[deliverable/binutils-gdb.git] / gdb / mdebugread.c
index 7b30ec30cb3f9160d737d12e3e2c1280372ed925..8d6c27dad96f02578ebf0be73d636c8858e9d22e 100644 (file)
@@ -274,8 +274,6 @@ static char stabs_symbol[] = STABS_SYMBOL;
    be using our own types thoughout this file, instead of sometimes using
    builtin_type_*.  */
 
-static struct type *mdebug_type_complex;
-static struct type *mdebug_type_double_complex;
 static struct type *mdebug_type_fixed_dec;
 static struct type *mdebug_type_float_dec;
 static struct type *mdebug_type_string;
@@ -1358,8 +1356,8 @@ parse_type (fd, ax, aux_index, bs, bigend, sym_name)
     0,                         /* btTypedef */
     0,                         /* btRange */
     0,                         /* btSet */
-    &mdebug_type_complex,      /* btComplex */
-    &mdebug_type_double_complex,       /* btDComplex */
+    &builtin_type_complex,     /* btComplex */
+    &builtin_type_double_complex,/* btDComplex */
     0,                         /* btIndirect */
     &mdebug_type_fixed_dec,    /* btFixedDec */
     &mdebug_type_float_dec,    /* btFloatDec */
@@ -4065,17 +4063,6 @@ _initialize_mdebugread ()
               0, "string",
               (struct objfile *) NULL);
 
-  mdebug_type_complex =
-    init_type (TYPE_CODE_ERROR,
-              TARGET_COMPLEX_BIT / TARGET_CHAR_BIT,
-              0, "complex",
-              (struct objfile *) NULL);
-  mdebug_type_double_complex =
-    init_type (TYPE_CODE_ERROR,
-              TARGET_DOUBLE_COMPLEX_BIT / TARGET_CHAR_BIT,
-              0, "double complex",
-              (struct objfile *) NULL);
-
   /* We use TYPE_CODE_INT to print these as integers.  Does this do any
      good?  Would we be better off with TYPE_CODE_ERROR?  Should
      TYPE_CODE_ERROR print things in hex if it knows the size?  */
This page took 0.027254 seconds and 4 git commands to generate.