DFP build failure

DFP build failure

[Daniel Jacobowitz]

void
decimal_from_floating (struct value *from, gdb_byte *to, int len)
{
  char *buffer;
  int ret;

  ret = asprintf (&buffer, "%.30Lg", value_as_double (from));
  ...

value_as_double returns a DOUBLEST, not necessarily a long double as
specified by %Lg.  And the system printf does not necessarily support
long double.  Is there some other way to do this?

(It failed to build on mingw32.)

-- 
Daniel Jacobowitz
CodeSourcery

Re: DFP build failure

[Thiago Jung Bauermann]

[-- Attachment #1: Type: text/plain, Size: 738 bytes --]

On Wed, 2008-01-09 at 11:49 -0500, Daniel Jacobowitz wrote:
> void
> decimal_from_floating (struct value *from, gdb_byte *to, int len)
> {
>   char *buffer;
>   int ret;
> 
>   ret = asprintf (&buffer, "%.30Lg", value_as_double (from));
>   ...
> 
> value_as_double returns a DOUBLEST, not necessarily a long double as
> specified by %Lg.  And the system printf does not necessarily support
> long double.  Is there some other way to do this?
> 
> (It failed to build on mingw32.)

Mmmm.... sorry about that. What do you think of the attached fix
(dangerous question to ask!)?

DOUBLEST_PRINT_FORMAT is currently not used, so not a big deal to change
it a bit.
-- 
[]'s
Thiago Jung Bauermann
Software Engineer
IBM Linux Technology Center

[-- Attachment #2: dfp-mingw-build-fix.diff --]
[-- Type: text/x-patch, Size: 1091 bytes --]

diff --git a/gdb/dfp.c b/gdb/dfp.c
index bf82114..4915543 100644
--- a/gdb/dfp.c
+++ b/gdb/dfp.c
@@ -237,7 +237,7 @@ decimal_from_floating (struct value *from, gdb_byte *to, int len)
   char *buffer;
   int ret;
 
-  ret = asprintf (&buffer, "%.30Lg", value_as_double (from));
+  ret = asprintf (&buffer, "%.30"DOUBLEST_PRINT_FORMAT, value_as_double (from));
   if (ret < 0)
     error (_("Error in memory allocation for conversion to decimal float."));
 
diff --git a/gdb/doublest.h b/gdb/doublest.h
index f3ab619..52e28df 100644
--- a/gdb/doublest.h
+++ b/gdb/doublest.h
@@ -49,11 +49,11 @@ struct floatformat;
 #if (defined HAVE_LONG_DOUBLE && defined PRINTF_HAS_LONG_DOUBLE \
      && defined SCANF_HAS_LONG_DOUBLE)
 typedef long double DOUBLEST;
-# define DOUBLEST_PRINT_FORMAT "%Lg"
+# define DOUBLEST_PRINT_FORMAT "Lg"
 # define DOUBLEST_SCAN_FORMAT "%Lg"
 #else
 typedef double DOUBLEST;
-# define DOUBLEST_PRINT_FORMAT "%g"
+# define DOUBLEST_PRINT_FORMAT "g"
 # define DOUBLEST_SCAN_FORMAT "%lg"
 /* If we can't scan or print long double, we don't want to use it
    anywhere.  */

Re: DFP build failure

[Daniel Jacobowitz]

On Wed, Jan 09, 2008 at 04:30:45PM -0200, Thiago Jung Bauermann wrote:
> Mmmm.... sorry about that. What do you think of the attached fix
> (dangerous question to ask!)?
> 
> DOUBLEST_PRINT_FORMAT is currently not used, so not a big deal to change
> it a bit.

OK, with changelog entry.  This matches the standard PRIx64 et al,
too.  Could you update DOUBLEST_SCAN_FORMAT to match?

> -  ret = asprintf (&buffer, "%.30Lg", value_as_double (from));
> +  ret = asprintf (&buffer, "%.30"DOUBLEST_PRINT_FORMAT, value_as_double (from));

Space between " and D.

-- 
Daniel Jacobowitz
CodeSourcery

Re: DFP build failure

[Thiago Jung Bauermann]

[-- Attachment #1: Type: text/plain, Size: 741 bytes --]


On Wed, 2008-01-09 at 13:35 -0500, Daniel Jacobowitz wrote:
> On Wed, Jan 09, 2008 at 04:30:45PM -0200, Thiago Jung Bauermann wrote:
> > Mmmm.... sorry about that. What do you think of the attached fix
> > (dangerous question to ask!)?
> > 
> > DOUBLEST_PRINT_FORMAT is currently not used, so not a big deal to change
> > it a bit.
> 
> OK, with changelog entry.  This matches the standard PRIx64 et al,
> too.  Could you update DOUBLEST_SCAN_FORMAT to match?
> 
> > -  ret = asprintf (&buffer, "%.30Lg", value_as_double (from));
> > +  ret = asprintf (&buffer, "%.30"DOUBLEST_PRINT_FORMAT, value_as_double (from));
> 
> Space between " and D.


What about this?
-- 
[]'s
Thiago Jung Bauermann
Software Engineer
IBM Linux Technology Center

[-- Attachment #2: fix-mingw-build-with-scanf.diff --]
[-- Type: text/x-patch, Size: 4101 bytes --]

2008-01-09  Thiago Jung Bauermann  <bauerman@br.ibm.com>

	* doublest.h (DOUBLEST_PRINT_FORMAT): Remove '%' from definition.
	(DOUBLEST_SCAN_FORMAT): Likewise.
	* dfp.c (decimal_from_floating): Use DOUBLEST_PRINT_FORMAT.
	* ada-lex.l (processReal): Prepend "%" to use of DOUBLEST_SCAN_FORMAT.
	* c-exp.y (parse_number): Likewise.
	* jv-exp.y (parse_number): Likewise.
	* objc-exp.y (parse_number): Likewise.
	* p-exp.y (parse_number): Likewise.

diff --git a/gdb/ada-lex.l b/gdb/ada-lex.l
index 6be64ab..cd6de8c 100644
--- a/gdb/ada-lex.l
+++ b/gdb/ada-lex.l
@@ -373,7 +373,7 @@ processInt (const char *base0, const char *num0, const char *exp0)
 static int
 processReal (const char *num0)
 {
-  sscanf (num0, DOUBLEST_SCAN_FORMAT, &yylval.typed_val_float.dval);
+  sscanf (num0, "%" DOUBLEST_SCAN_FORMAT, &yylval.typed_val_float.dval);
 
   yylval.typed_val_float.type = type_float ();
   if (sizeof(DOUBLEST) >= gdbarch_double_bit (current_gdbarch)
diff --git a/gdb/c-exp.y b/gdb/c-exp.y
index 22bb4c6..b520721 100644
--- a/gdb/c-exp.y
+++ b/gdb/c-exp.y
@@ -1125,7 +1125,7 @@ parse_number (p, len, parsed_float, putithere)
 	  return (DECFLOAT);
 	}
 
-      num = sscanf (p, DOUBLEST_SCAN_FORMAT "%s",
+      num = sscanf (p, "%" DOUBLEST_SCAN_FORMAT "%s",
 		    &putithere->typed_val_float.dval, s);
       p[len] = saved_char;	/* restore the input stream */
 
diff --git a/gdb/dfp.c b/gdb/dfp.c
index bf82114..e184ecd 100644
--- a/gdb/dfp.c
+++ b/gdb/dfp.c
@@ -237,7 +237,8 @@ decimal_from_floating (struct value *from, gdb_byte *to, int len)
   char *buffer;
   int ret;
 
-  ret = asprintf (&buffer, "%.30Lg", value_as_double (from));
+  ret = asprintf (&buffer, "%.30" DOUBLEST_PRINT_FORMAT,
+		  value_as_double (from));
   if (ret < 0)
     error (_("Error in memory allocation for conversion to decimal float."));
 
diff --git a/gdb/doublest.h b/gdb/doublest.h
index f3ab619..247eb5d 100644
--- a/gdb/doublest.h
+++ b/gdb/doublest.h
@@ -49,12 +49,12 @@ struct floatformat;
 #if (defined HAVE_LONG_DOUBLE && defined PRINTF_HAS_LONG_DOUBLE \
      && defined SCANF_HAS_LONG_DOUBLE)
 typedef long double DOUBLEST;
-# define DOUBLEST_PRINT_FORMAT "%Lg"
-# define DOUBLEST_SCAN_FORMAT "%Lg"
+# define DOUBLEST_PRINT_FORMAT "Lg"
+# define DOUBLEST_SCAN_FORMAT "Lg"
 #else
 typedef double DOUBLEST;
-# define DOUBLEST_PRINT_FORMAT "%g"
-# define DOUBLEST_SCAN_FORMAT "%lg"
+# define DOUBLEST_PRINT_FORMAT "g"
+# define DOUBLEST_SCAN_FORMAT "lg"
 /* If we can't scan or print long double, we don't want to use it
    anywhere.  */
 # undef HAVE_LONG_DOUBLE
diff --git a/gdb/jv-exp.y b/gdb/jv-exp.y
index 1db33a6..14a7bf8 100644
--- a/gdb/jv-exp.y
+++ b/gdb/jv-exp.y
@@ -713,7 +713,7 @@ parse_number (p, len, parsed_float, putithere)
       char saved_char = p[len];
 
       p[len] = 0;	/* null-terminate the token */
-      num = sscanf (p, DOUBLEST_SCAN_FORMAT "%c",
+      num = sscanf (p, "%" DOUBLEST_SCAN_FORMAT "%c",
 		    &putithere->typed_val_float.dval, &c);
       p[len] = saved_char;	/* restore the input stream */
       if (num != 1) 		/* check scanf found ONLY a float ... */
diff --git a/gdb/objc-exp.y b/gdb/objc-exp.y
index d421441..a4a42d9 100644
--- a/gdb/objc-exp.y
+++ b/gdb/objc-exp.y
@@ -1023,7 +1023,7 @@ parse_number (p, len, parsed_float, putithere)
 
       /* It's a float since it contains a point or an exponent.  */
 
-      sscanf (p, DOUBLEST_SCAN_FORMAT "%c",
+      sscanf (p, "%" DOUBLEST_SCAN_FORMAT "%c",
 	      &putithere->typed_val_float.dval, &c);
 
       /* See if it has `f' or `l' suffix (float or long double).  */
diff --git a/gdb/p-exp.y b/gdb/p-exp.y
index 857a69f..9cab7cd 100644
--- a/gdb/p-exp.y
+++ b/gdb/p-exp.y
@@ -797,7 +797,7 @@ parse_number (p, len, parsed_float, putithere)
       char saved_char = p[len];
 
       p[len] = 0;	/* null-terminate the token */
-      num = sscanf (p, DOUBLEST_SCAN_FORMAT "%c",
+      num = sscanf (p, "%" DOUBLEST_SCAN_FORMAT "%c",
 		    &putithere->typed_val_float.dval, &c);
       p[len] = saved_char;	/* restore the input stream */
       if (num != 1) 		/* check scanf found ONLY a float ... */

Re: DFP build failure

[Daniel Jacobowitz]

On Wed, Jan 09, 2008 at 05:11:25PM -0200, Thiago Jung Bauermann wrote:
> 2008-01-09  Thiago Jung Bauermann  <bauerman@br.ibm.com>
> 
> 	* doublest.h (DOUBLEST_PRINT_FORMAT): Remove '%' from definition.
> 	(DOUBLEST_SCAN_FORMAT): Likewise.
> 	* dfp.c (decimal_from_floating): Use DOUBLEST_PRINT_FORMAT.
> 	* ada-lex.l (processReal): Prepend "%" to use of DOUBLEST_SCAN_FORMAT.
> 	* c-exp.y (parse_number): Likewise.
> 	* jv-exp.y (parse_number): Likewise.
> 	* objc-exp.y (parse_number): Likewise.
> 	* p-exp.y (parse_number): Likewise.

OK, and thanks.

-- 
Daniel Jacobowitz
CodeSourcery

Re: DFP build failure

[Thiago Jung Bauermann]

On Wed, 2008-01-09 at 14:14 -0500, Daniel Jacobowitz wrote:
> On Wed, Jan 09, 2008 at 05:11:25PM -0200, Thiago Jung Bauermann wrote:
> > 2008-01-09  Thiago Jung Bauermann  <bauerman@br.ibm.com>
> > 
> > 	* doublest.h (DOUBLEST_PRINT_FORMAT): Remove '%' from definition.
> > 	(DOUBLEST_SCAN_FORMAT): Likewise.
> > 	* dfp.c (decimal_from_floating): Use DOUBLEST_PRINT_FORMAT.
> > 	* ada-lex.l (processReal): Prepend "%" to use of DOUBLEST_SCAN_FORMAT.
> > 	* c-exp.y (parse_number): Likewise.
> > 	* jv-exp.y (parse_number): Likewise.
> > 	* objc-exp.y (parse_number): Likewise.
> > 	* p-exp.y (parse_number): Likewise.
> 
> OK, and thanks.

No problem. Committed.
-- 
[]'s
Thiago Jung Bauermann
Software Engineer
IBM Linux Technology Center