[patch/rfc] Delete floatformat_unknown

[patch/rfc] Delete floatformat_unknown

[Andrew Cagney]

[-- Attachment #1: Type: text/plain, Size: 757 bytes --]

Hello,

The idea of doing this change has been kicking around for some time.

At present TARGET_LONG_DOUBLE_FORMAT defaults to floatformat_unknown 
when the target doesn't specify it.  This patch changes things so that 
it defaults to ieee_double and in doing so matches the default value of 
TARGET_LONG_DOUBLE_BIT.

Any target that didn't define it is going get this value defined. 
Playing around with the PPC (which doesn't define this).  I have a 
feeling this is a good move.

Before:

(top-gdb) print (long double)1
$2 = 0
(top-gdb) print (double)1
$3 = 1

after:

(gdb) print (long double)1
$1 = 1
(gdb) print (double)1
$2 = 1

thoughts.  Like IEEE_FLOAT this isn't ment to break things .....

Thoughts?  Probably commit it in about a week.

Andrew

[-- Attachment #2: diffs --]
[-- Type: text/plain, Size: 5309 bytes --]

2002-02-03  Andrew Cagney  <ac131313@redhat.com>

	* gdbarch.sh (TARGET_LONG_DOUBLE_FORMAT): Default to
	default_double_format.
	* gdbarch.h, gdbarch.c: Re-generate.
	* findvar.c (floatformat_unknown): Delete variable definition.
	* doublest.h (floatformat_unknown): Delete variable declaration.

Index: doublest.h
===================================================================
RCS file: /cvs/src/src/gdb/doublest.h,v
retrieving revision 1.8
diff -p -r1.8 doublest.h
*** doublest.h	2002/01/20 18:05:51	1.8
--- doublest.h	2002/02/04 06:08:37
***************
*** 35,42 ****
  
  #include "floatformat.h"	/* For struct floatformat */
  
- extern const struct floatformat floatformat_unknown;
- 
  /* Use `long double' if the host compiler supports it.  (Note that this is not
     necessarily any longer than `double'.  On SunOS/gcc, it's the same as
     double.)  This is necessary because GDB internally converts all floating
--- 35,40 ----
Index: findvar.c
===================================================================
RCS file: /cvs/src/src/gdb/findvar.c,v
retrieving revision 1.27
diff -p -r1.27 findvar.c
*** findvar.c	2002/01/05 04:30:17	1.27
--- findvar.c	2002/02/04 06:08:38
***************
*** 34,45 ****
  #include "symfile.h"		/* for overlay functions */
  #include "regcache.h"
  
- /* This is used to indicate that we don't know the format of the floating point
-    number.  Typically, this is useful for native ports, where the actual format
-    is irrelevant, since no conversions will be taking place.  */
- 
- const struct floatformat floatformat_unknown;
- 
  /* Basic byte-swapping routines.  GDB has needed these for a long time...
     All extract a target-format integer at ADDR which is LEN bytes long.  */
  
--- 34,39 ----
Index: gdbarch.c
===================================================================
RCS file: /cvs/src/src/gdb/gdbarch.c,v
retrieving revision 1.100
diff -p -r1.100 gdbarch.c
*** gdbarch.c	2002/01/20 18:05:51	1.100
--- gdbarch.c	2002/02/04 06:08:52
*************** verify_gdbarch (struct gdbarch *gdbarch)
*** 754,760 ****
    if (gdbarch->double_format == 0)
      gdbarch->double_format = default_double_format (gdbarch);
    if (gdbarch->long_double_format == 0)
!     gdbarch->long_double_format = &floatformat_unknown;
    /* Skip verify of convert_from_func_ptr_addr, invalid_p == 0 */
    /* Skip verify of addr_bits_remove, invalid_p == 0 */
    /* Skip verify of software_single_step, has predicate */
--- 754,760 ----
    if (gdbarch->double_format == 0)
      gdbarch->double_format = default_double_format (gdbarch);
    if (gdbarch->long_double_format == 0)
!     gdbarch->long_double_format = default_double_format (gdbarch);
    /* Skip verify of convert_from_func_ptr_addr, invalid_p == 0 */
    /* Skip verify of addr_bits_remove, invalid_p == 0 */
    /* Skip verify of software_single_step, has predicate */
Index: gdbarch.h
===================================================================
RCS file: /cvs/src/src/gdb/gdbarch.h,v
retrieving revision 1.75
diff -p -r1.75 gdbarch.h
*** gdbarch.h	2002/01/20 19:26:48	1.75
--- gdbarch.h	2002/02/04 06:09:09
*************** extern void set_gdbarch_double_format (s
*** 1982,1988 ****
  
  /* Default (value) for non- multi-arch platforms. */
  #if (!GDB_MULTI_ARCH) && !defined (TARGET_LONG_DOUBLE_FORMAT)
! #define TARGET_LONG_DOUBLE_FORMAT (&floatformat_unknown)
  #endif
  
  extern const struct floatformat * gdbarch_long_double_format (struct gdbarch *gdbarch);
--- 1982,1988 ----
  
  /* Default (value) for non- multi-arch platforms. */
  #if (!GDB_MULTI_ARCH) && !defined (TARGET_LONG_DOUBLE_FORMAT)
! #define TARGET_LONG_DOUBLE_FORMAT (default_double_format (current_gdbarch))
  #endif
  
  extern const struct floatformat * gdbarch_long_double_format (struct gdbarch *gdbarch);
Index: gdbarch.sh
===================================================================
RCS file: /cvs/src/src/gdb/gdbarch.sh,v
retrieving revision 1.105
diff -p -r1.105 gdbarch.sh
*** gdbarch.sh	2002/01/20 19:26:48	1.105
--- gdbarch.sh	2002/02/04 06:09:11
*************** v:2:PARM_BOUNDARY:int:parm_boundary
*** 540,546 ****
  #
  v:2:TARGET_FLOAT_FORMAT:const struct floatformat *:float_format::::::default_float_format (gdbarch)
  v:2:TARGET_DOUBLE_FORMAT:const struct floatformat *:double_format::::::default_double_format (gdbarch)
! v:2:TARGET_LONG_DOUBLE_FORMAT:const struct floatformat *:long_double_format::::::&floatformat_unknown
  f:2:CONVERT_FROM_FUNC_PTR_ADDR:CORE_ADDR:convert_from_func_ptr_addr:CORE_ADDR addr:addr:::core_addr_identity::0
  # On some machines there are bits in addresses which are not really
  # part of the address, but are used by the kernel, the hardware, etc.
--- 540,546 ----
  #
  v:2:TARGET_FLOAT_FORMAT:const struct floatformat *:float_format::::::default_float_format (gdbarch)
  v:2:TARGET_DOUBLE_FORMAT:const struct floatformat *:double_format::::::default_double_format (gdbarch)
! v:2:TARGET_LONG_DOUBLE_FORMAT:const struct floatformat *:long_double_format::::::default_double_format (gdbarch)
  f:2:CONVERT_FROM_FUNC_PTR_ADDR:CORE_ADDR:convert_from_func_ptr_addr:CORE_ADDR addr:addr:::core_addr_identity::0
  # On some machines there are bits in addresses which are not really
  # part of the address, but are used by the kernel, the hardware, etc.

Re: [patch/rfc] Delete floatformat_unknown

[Andrew Cagney]

> Thoughts?  Probably commit it in about a week.


Here we go ...

> 2002-02-03  Andrew Cagney  <ac131313@redhat.com>
> 
> * gdbarch.sh (TARGET_LONG_DOUBLE_FORMAT): Default to
> 	default_double_format.
> 	* gdbarch.h, gdbarch.c: Re-generate.
> 	* findvar.c (floatformat_unknown): Delete variable definition.
> 	* doublest.h (floatformat_unknown): Delete variable declaration.
> 
> 
I'm currently checking this in.


Andrew

Re: [patch/rfc] Delete floatformat_unknown

[Richard Earnshaw]

> > Thoughts?  Probably commit it in about a week.
> 
> 
> Here we go ...
> 
> > 2002-02-03  Andrew Cagney  <ac131313@redhat.com>
> > 
> > * gdbarch.sh (TARGET_LONG_DOUBLE_FORMAT): Default to
> > 	default_double_format.
> > 	* gdbarch.h, gdbarch.c: Re-generate.
> > 	* findvar.c (floatformat_unknown): Delete variable definition.
> > 	* doublest.h (floatformat_unknown): Delete variable declaration.
> > 
> > 
> I'm currently checking this in.
> 
> 
> Andrew
> 
> 
> 

Given that we now assume that long double == double, why is 
TARGET_LONG_DOUBLE_BIT = 128 by default?

R.

Re: [patch/rfc] Delete floatformat_unknown

[Andrew Cagney]

> I'm currently checking this in.
>> 
>> 
>> Andrew
>> 
>> 
>> 
> 
> 
> Given that we now assume that long double == double, why is 
> TARGET_LONG_DOUBLE_BIT = 128 by default?

Oops!  Fixing it now.

Andrew