Name of Fortran intrinsic types

Name of Fortran intrinsic types

[FX Coudert]

[-- Attachment #1: Type: text/plain, Size: 865 bytes --]

Hi all,

This is my first patch submission for gdb (I received confirmation of  
my copyright assignment a week ago). It's short and I hope it's  
straightforward: for intrinsic Fortran types (here, real and complex  
types), I think we should use the name the compiler produced, instead  
of trying to be clever.

The rationale is that this allows us to output Fortran names for the  
types, e.g. "real(kind=4)", instead gdb-specific, non-Fortran names  
(e.g. "int4"). The reason for doing that instead of simply replacing  
the format (currently "int%d") in gdb is that we have no way, in  
general, to know the kind of the variable in gdb, so it's better to  
simply trust the compiler!

Built and regtested on x86_64-linux (there are a fair number of  
FAILures in the testsuite, but none in the Fortran area, so maybe  
it's the normal state?).

Thanks,
FX



[-- Attachment #2: gdb-types.ChangeLog --]
[-- Type: application/octet-stream, Size: 214 bytes --]

2007-11-17  Francois-Xavier Coudert  <fxcoudert@gcc.gnu.org>

	* f-typeprint.c (print_equivalent_f77_float_type): Remove function.
	(f_type_print_base): Output the type names directly for real and
	complex types.


[-- Attachment #3: gdb-types.diff --]
[-- Type: application/octet-stream, Size: 1939 bytes --]

Index: f-typeprint.c
===================================================================
RCS file: /cvs/src/src/gdb/f-typeprint.c,v
retrieving revision 1.19
diff -p -u -r1.19 f-typeprint.c
--- f-typeprint.c	23 Aug 2007 18:08:31 -0000	1.19
+++ f-typeprint.c	17 Nov 2007 20:33:47 -0000
@@ -39,9 +39,6 @@
 static void f_type_print_args (struct type *, struct ui_file *);
 #endif
 
-static void print_equivalent_f77_float_type (int level, struct type *,
-					     struct ui_file *);
-
 static void f_type_print_varspec_suffix (struct type *, struct ui_file *,
 					 int, int, int);
 
@@ -254,17 +251,6 @@ f_type_print_varspec_suffix (struct type
     }
 }
 
-static void
-print_equivalent_f77_float_type (int level, struct type *type,
-				 struct ui_file *stream)
-{
-  /* Override type name "float" and make it the
-     appropriate real. XLC stupidly outputs -12 as a type
-     for real when it really should be outputting -18 */
-
-  fprintfi_filtered (level, stream, "real*%d", TYPE_LENGTH (type));
-}
-
 /* Print the name of the type (or the ultimate pointer target,
    function value or array element), or the description of a
    structure or union.
@@ -301,10 +287,7 @@ f_type_print_base (struct type *type, st
 
   if ((show <= 0) && (TYPE_NAME (type) != NULL))
     {
-      if (TYPE_CODE (type) == TYPE_CODE_FLT)
-	print_equivalent_f77_float_type (level, type, stream);
-      else
-	fputs_filtered (TYPE_NAME (type), stream);
+      fputs_filtered (TYPE_NAME (type), stream);
       return;
     }
 
@@ -365,14 +348,6 @@ f_type_print_base (struct type *type, st
 	goto default_case;
       break;
 
-    case TYPE_CODE_COMPLEX:
-      fprintfi_filtered (level, stream, "complex*%d", TYPE_LENGTH (type));
-      break;
-
-    case TYPE_CODE_FLT:
-      print_equivalent_f77_float_type (level, type, stream);
-      break;
-
     case TYPE_CODE_STRING:
       /* Strings may have dynamic upperbounds (lengths) like arrays. */
 

Re: Name of Fortran intrinsic types

[Daniel Jacobowitz]

On Sat, Nov 17, 2007 at 08:55:51PM +0000, FX Coudert wrote:
> Built and regtested on x86_64-linux (there are a fair number of FAILures 
> in the testsuite, but none in the Fortran area, so maybe it's the normal 
> state?).

OK.  Do you have commit access already (membership in the src group)?
If not, please ask overseers@ for it; CC me and say I approved it.
Then add yourself to gdb/MAINTAINERS in the obvious place.

Yes, a number of failures are normal.  I dislike this, but it's hard
to stay on top of all the different configurations.  Any time you want
to fix one of them or add appropriate XFAILs for conditions outside of
GDB's control, be my guest and I'll be happy to help.  Until then, I
recommend running the testsuite with and without your patches.

Also, if you could find a way to put your patches in plain text
instead of application/octet-stream... thanks.

-- 
Daniel Jacobowitz
CodeSourcery

Re: Name of Fortran intrinsic types

[FX Coudert]

> OK.  Do you have commit access already (membership in the src group)?
> If not, please ask overseers@ for it; CC me and say I approved it.
> Then add yourself to gdb/MAINTAINERS in the obvious place.

I won't be in a position to commit until mid-January. Feel free to  
commit it if you want, otherwise I'll do it then.

Thanks for reviewing the patch,
FX

Re: Name of Fortran intrinsic types

[Daniel Jacobowitz]

On Sun, Dec 16, 2007 at 09:14:41PM +0000, FX Coudert wrote:
>> OK.  Do you have commit access already (membership in the src group)?
>> If not, please ask overseers@ for it; CC me and say I approved it.
>> Then add yourself to gdb/MAINTAINERS in the obvious place.
>
> I won't be in a position to commit until mid-January. Feel free to commit 
> it if you want, otherwise I'll do it then.

I checked it in.

-- 
Daniel Jacobowitz
CodeSourcery