[PATCH 0/6] Make GMP data members private

[PATCH 0/6] Make GMP data members private

[Tom Tromey]

This series changes the GMP code in gmp-utils.h to add methods and
operators, and then to make the data members private.  Ages ago I told
Simon I'd do this, then I dropped it for years.

I've only added exactly the operators that are needed right now.  It's
simple to add more operators or more overloads as required.

Regression tested on x86-64 Fedora 36.

Tom

---
Tom Tromey (6):
      Clean up gmp-utils.h includes
      Add methods and operators to gdb_mpz
      Rename gdb_mpz::val and make contents private
      Add operators and methods to gdb_mpq
      Rename gdb_mpq::val and make contents private
      Hide the implementation of gdb_mpf

 gdb/ada-lex.l                       |   9 +-
 gdb/dwarf2/read.c                   |  46 +++----
 gdb/gmp-utils.c                     |  61 ++++-----
 gdb/gmp-utils.h                     | 248 ++++++++++++++++++++++++++++++------
 gdb/unittests/gmp-utils-selftests.c |  89 +++++++------
 gdb/valarith.c                      |  18 +--
 gdb/valops.c                        |  15 +--
 gdb/valprint.c                      |   2 +-
 gdb/value.c                         |   3 +-
 9 files changed, 325 insertions(+), 166 deletions(-)
---
base-commit: e98a23bfb3088fb129540d7fd26f3205e2de6f2a
change-id: 20230223-submit-gmp-hiding-8be8230612a0

Best regards,
-- 
Tom Tromey <tromey@adacore.com>

[PATCH 1/6] Clean up gmp-utils.h includes

[Tom Tromey]

gmp-utils.h includes "defs.h", but normally the rule in gdb is that
the .c files include this first.  This patch changes this code to
match the rest of gdb.
---
 gdb/gmp-utils.c                     | 1 +
 gdb/gmp-utils.h                     | 2 --
 gdb/unittests/gmp-utils-selftests.c | 1 +
 3 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/gdb/gmp-utils.c b/gdb/gmp-utils.c
index b9c9b495398..8150fa5dcdf 100644
--- a/gdb/gmp-utils.c
+++ b/gdb/gmp-utils.c
@@ -15,6 +15,7 @@
    You should have received a copy of the GNU General Public License
    along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
 
+#include "defs.h"
 #include "gmp-utils.h"
 
 /* See gmp-utils.h.  */
diff --git a/gdb/gmp-utils.h b/gdb/gmp-utils.h
index 0e57471b2ef..0522f3453aa 100644
--- a/gdb/gmp-utils.h
+++ b/gdb/gmp-utils.h
@@ -20,8 +20,6 @@
 #ifndef GMP_UTILS_H
 #define GMP_UTILS_H
 
-#include "defs.h"
-
 /* Include <stdio.h> and <stdarg.h> ahead of <gmp.h>, so as to get
    access to GMP's various formatting functions.  */
 #include <stdio.h>
diff --git a/gdb/unittests/gmp-utils-selftests.c b/gdb/unittests/gmp-utils-selftests.c
index cc525a13fc0..e2ccd2ca094 100644
--- a/gdb/unittests/gmp-utils-selftests.c
+++ b/gdb/unittests/gmp-utils-selftests.c
@@ -17,6 +17,7 @@
    You should have received a copy of the GNU General Public License
    along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
 
+#include "defs.h"
 #include "gmp-utils.h"
 #include "gdbsupport/selftest.h"
 

-- 
2.39.1

[PATCH 2/6] Add methods and operators to gdb_mpz

[Tom Tromey]

This adds various methods and operators to gdb_mpz, as a step toward
hiding the implementation.

This only adds the operators that were needed.  Many more could be
added as required.
---
 gdb/ada-lex.l                       |  9 ++--
 gdb/dwarf2/read.c                   | 46 +++++++++-----------
 gdb/gmp-utils.h                     | 87 +++++++++++++++++++++++++++++++++++++
 gdb/unittests/gmp-utils-selftests.c | 54 +++++++++++------------
 gdb/valops.c                        |  2 +-
 gdb/value.c                         |  3 +-
 6 files changed, 140 insertions(+), 61 deletions(-)

diff --git a/gdb/ada-lex.l b/gdb/ada-lex.l
index a5cfb84dcae..69fc14f7107 100644
--- a/gdb/ada-lex.l
+++ b/gdb/ada-lex.l
@@ -422,14 +422,14 @@ processInt (struct parser_state *par_state, const char *base0,
       int dig = fromhex (*num0);
       if (dig >= base)
 	error (_("Invalid digit `%c' in based literal"), *num0);
-      mpz_mul_ui (result.val, result.val, base);
-      mpz_add_ui (result.val, result.val, dig);
+      result *= base;
+      result += dig;
       ++num0;
     }
 
   while (exp > 0)
     {
-      mpz_mul_ui (result.val, result.val, base);
+      result *= base;
       exp -= 1;
     }
 
@@ -462,8 +462,7 @@ processInt (struct parser_state *par_state, const char *base0,
       return FLOAT;
     }
 
-  gdb_mpz maxval (ULONGEST_MAX);
-  if (mpz_cmp (result.val, maxval.val) > 0)
+  if (result > gdb_mpz (ULONGEST_MAX))
     error (_("Integer literal out of range"));
 
   int int_bits = gdbarch_int_bit (par_state->gdbarch ());
diff --git a/gdb/dwarf2/read.c b/gdb/dwarf2/read.c
index a5bc1d89a2d..cef79b6531b 100644
--- a/gdb/dwarf2/read.c
+++ b/gdb/dwarf2/read.c
@@ -14756,10 +14756,10 @@ get_mpz (struct dwarf2_cu *cu, gdb_mpz *value, struct attribute *attr)
 						   &len);
 	  if (ptr - blk->data + len <= blk->size)
 	    {
-	      mpz_import (value->val, len,
-			  bfd_big_endian (cu->per_objfile->objfile->obfd.get ())
-			  ? 1 : -1,
-			  1, 0, 0, ptr);
+	      value->read (gdb::make_array_view (ptr, len),
+			   bfd_big_endian (cu->per_objfile->objfile->obfd.get ())
+			   ? BFD_ENDIAN_BIG : BFD_ENDIAN_LITTLE,
+			   true);
 	      return;
 	    }
 	}
@@ -14770,10 +14770,10 @@ get_mpz (struct dwarf2_cu *cu, gdb_mpz *value, struct attribute *attr)
   else if (attr->form_is_block ())
     {
       dwarf_block *blk = attr->as_block ();
-      mpz_import (value->val, blk->size,
-		  bfd_big_endian (cu->per_objfile->objfile->obfd.get ())
-		  ? 1 : -1,
-		  1, 0, 0, blk->data);
+      value->read (gdb::make_array_view (blk->data, blk->size),
+		   bfd_big_endian (cu->per_objfile->objfile->obfd.get ())
+		   ? BFD_ENDIAN_BIG : BFD_ENDIAN_LITTLE,
+		   true);
     }
   else
     *value = gdb_mpz (attr->constant_value (1));
@@ -14825,19 +14825,19 @@ get_dwarf2_unsigned_rational_constant (struct die_info *die,
   gdb_mpz denom (1);
 
   get_dwarf2_rational_constant (die, cu, &num, &denom);
-  if (mpz_sgn (num.val) == -1 && mpz_sgn (denom.val) == -1)
+  if (num < 0 && denom < 0)
     {
-      mpz_neg (num.val, num.val);
-      mpz_neg (denom.val, denom.val);
+      num.negate ();
+      denom.negate ();
     }
-  else if (mpz_sgn (num.val) == -1)
+  else if (num < 0)
     {
       complaint (_("unexpected negative value for DW_AT_GNU_numerator"
 		   " in DIE at %s"),
 		 sect_offset_str (die->sect_off));
       return;
     }
-  else if (mpz_sgn (denom.val) == -1)
+  else if (denom < 0)
     {
       complaint (_("unexpected negative value for DW_AT_GNU_denominator"
 		   " in DIE at %s"),
@@ -14876,10 +14876,7 @@ ada_get_gnat_encoded_number (const char *encoding, int &k, gdb_mpz *result)
     return false;
 
   std::string copy (&encoding[start], k - start);
-  if (mpz_set_str (result->val, copy.c_str (), 10) == -1)
-    return false;
-
-  return true;
+  return result->set (copy.c_str (), 10);
 }
 
 /* Scan two numbers from ENCODING at OFFSET, assuming the string is of
@@ -14959,16 +14956,16 @@ finish_fixed_point_type (struct type *type, const char *suffix,
   else if (attr->name == DW_AT_binary_scale)
     {
       LONGEST scale_exp = attr->constant_value (0);
-      gdb_mpz *num_or_denom = scale_exp > 0 ? &scale_num : &scale_denom;
+      gdb_mpz &num_or_denom = scale_exp > 0 ? scale_num : scale_denom;
 
-      mpz_mul_2exp (num_or_denom->val, num_or_denom->val, std::abs (scale_exp));
+      num_or_denom <<= std::abs (scale_exp);
     }
   else if (attr->name == DW_AT_decimal_scale)
     {
       LONGEST scale_exp = attr->constant_value (0);
-      gdb_mpz *num_or_denom = scale_exp > 0 ? &scale_num : &scale_denom;
+      gdb_mpz &num_or_denom = scale_exp > 0 ? scale_num : scale_denom;
 
-      mpz_ui_pow_ui (num_or_denom->val, 10, std::abs (scale_exp));
+      num_or_denom = gdb_mpz::pow (10, std::abs (scale_exp));
     }
   else if (attr->name == DW_AT_small)
     {
@@ -14992,10 +14989,7 @@ finish_fixed_point_type (struct type *type, const char *suffix,
 		 sect_offset_str (die->sect_off));
     }
 
-  gdb_mpq &scaling_factor = type->fixed_point_info ().scaling_factor;
-  mpz_set (mpq_numref (scaling_factor.val), scale_num.val);
-  mpz_set (mpq_denref (scaling_factor.val), scale_denom.val);
-  mpq_canonicalize (scaling_factor.val);
+  type->fixed_point_info ().scaling_factor = gdb_mpq (scale_num, scale_denom);
 }
 
 /* The gnat-encoding suffix for fixed point.  */
@@ -15076,7 +15070,7 @@ has_zero_over_zero_small_attribute (struct die_info *die,
 
   gdb_mpz num (1), denom (1);
   get_dwarf2_rational_constant (scale_die, cu, &num, &denom);
-  return mpz_sgn (num.val) == 0 && mpz_sgn (denom.val) == 0;
+  return num == 0 && denom == 0;
 }
 
 /* Initialise and return a floating point type of size BITS suitable for
diff --git a/gdb/gmp-utils.h b/gdb/gmp-utils.h
index 0522f3453aa..dcb998a1656 100644
--- a/gdb/gmp-utils.h
+++ b/gdb/gmp-utils.h
@@ -89,6 +89,21 @@ struct gdb_mpz
     return *this;
   }
 
+  /* Initialize this value from a string and a base.  Returns true if
+     the string was parsed successfully, false otherwise.  */
+  bool set (const char *str, int base)
+  {
+    return mpz_set_str (val, str, base) != -1;
+  }
+
+  /* Return a new value that is BASE**EXP.  */
+  static gdb_mpz pow (unsigned long base, unsigned long exp)
+  {
+    gdb_mpz result;
+    mpz_ui_pow_ui (result.val, base, exp);
+    return result;
+  }
+
   /* Convert VAL to an integer of the given type.
 
      The return type can signed or unsigned, with no size restriction.  */
@@ -115,6 +130,56 @@ struct gdb_mpz
   /* The destructor.  */
   ~gdb_mpz () { mpz_clear (val); }
 
+  /* Negate this value in place.  */
+  void negate ()
+  {
+    mpz_neg (val, val);
+  }
+
+  gdb_mpz &operator*= (long other)
+  {
+    mpz_mul_si (val, val, other);
+    return *this;
+  }
+
+  gdb_mpz &operator+= (unsigned long other)
+  {
+    mpz_add_ui (val, val, other);
+    return *this;
+  }
+
+  gdb_mpz &operator-= (unsigned long other)
+  {
+    mpz_sub_ui (val, val, other);
+    return *this;
+  }
+
+  gdb_mpz &operator<<= (unsigned long nbits)
+  {
+    mpz_mul_2exp (val, val, nbits);
+    return *this;
+  }
+
+  bool operator> (const gdb_mpz &other) const
+  {
+    return mpz_cmp (val, other.val) > 0;
+  }
+
+  bool operator< (int other) const
+  {
+    return mpz_cmp_si (val, other) < 0;
+  }
+
+  bool operator== (int other) const
+  {
+    return mpz_cmp_si (val, other) == 0;
+  }
+
+  bool operator== (const gdb_mpz &other) const
+  {
+    return mpz_cmp (val, other.val) == 0;
+  }
+
 private:
 
   /* Helper template for constructor and operator=.  */
@@ -166,6 +231,14 @@ struct gdb_mpq
     mpq_swap (val, from.val);
   }
 
+  gdb_mpq (const gdb_mpz &num, const gdb_mpz &denom)
+  {
+    mpq_init (val);
+    mpz_set (mpq_numref (val), num.val);
+    mpz_set (mpq_denref (val), denom.val);
+    mpq_canonicalize (val);
+  }
+
   /* Copy assignment operator.  */
   gdb_mpq &operator= (const gdb_mpq &from)
   {
@@ -179,12 +252,26 @@ struct gdb_mpq
     return *this;
   }
 
+  gdb_mpq &operator= (const gdb_mpz &from)
+  {
+    mpq_set_z (val, from.val);
+    return *this;
+  }
+
   /* Return a string representing VAL as "<numerator> / <denominator>".  */
   std::string str () const { return gmp_string_printf ("%Qd", val); }
 
   /* Return VAL rounded to the nearest integer.  */
   gdb_mpz get_rounded () const;
 
+  /* Return this value as an integer, rounded toward zero.  */
+  gdb_mpz as_integer () const
+  {
+    gdb_mpz result;
+    mpz_tdiv_q (result.val, mpq_numref (val), mpq_denref (val));
+    return result;
+  }
+
   /* Set VAL from the contents of the given byte array (BUF), which
      contains the unscaled value of a fixed point type object.
      The byte size of the data is the size of BUF.
diff --git a/gdb/unittests/gmp-utils-selftests.c b/gdb/unittests/gmp-utils-selftests.c
index e2ccd2ca094..8e028fec88c 100644
--- a/gdb/unittests/gmp-utils-selftests.c
+++ b/gdb/unittests/gmp-utils-selftests.c
@@ -43,8 +43,8 @@ gdb_mpz_as_integer ()
   /* Start with the smallest LONGEST  */
   l_expected = (LONGEST) 1 << (sizeof (LONGEST) * 8 - 1);
 
-  mpz_ui_pow_ui (v.val, 2, sizeof (LONGEST) * 8 - 1);
-  mpz_neg (v.val, v.val);
+  v = gdb_mpz::pow (2, sizeof (LONGEST) * 8 - 1);
+  v.negate ();
 
   SELF_CHECK (v.as_integer<LONGEST> () == l_expected);
 
@@ -53,13 +53,13 @@ gdb_mpz_as_integer ()
   for (int i = -256; i <= 256; i++)
     {
       l_expected = (LONGEST) i;
-      mpz_set_si (v.val, i);
+      v = i;
       SELF_CHECK (v.as_integer<LONGEST> () == l_expected);
 
       if (i >= 0)
 	{
 	  ul_expected = (ULONGEST) i;
-	  mpz_set_ui (v.val, i);
+	  v = ul_expected;
 	  SELF_CHECK (v.as_integer<ULONGEST> () == ul_expected);
 	}
     }
@@ -68,16 +68,16 @@ gdb_mpz_as_integer ()
   l_expected = LONGEST_MAX;
   ul_expected = (ULONGEST) l_expected;
 
-  mpz_ui_pow_ui (v.val, 2, sizeof (LONGEST) * 8 - 1);
-  mpz_sub_ui (v.val, v.val, 1);
+  v = gdb_mpz::pow (2, sizeof (LONGEST) * 8 - 1);
+  v -= 1;
 
   SELF_CHECK (v.as_integer<LONGEST> () == l_expected);
   SELF_CHECK (v.as_integer<ULONGEST> () == ul_expected);
 
   /* Try with ULONGEST_MAX.  */
   ul_expected = ULONGEST_MAX;
-  mpz_ui_pow_ui (v.val, 2, sizeof (LONGEST) * 8);
-  mpz_sub_ui (v.val, v.val, 1);
+  v = gdb_mpz::pow (2, sizeof (LONGEST) * 8);
+  v -= 1;
 
   SELF_CHECK (v.as_integer<ULONGEST> () == ul_expected);
 }
@@ -116,9 +116,9 @@ gdb_mpz_as_integer_out_of_range ()
   gdb_mpz v;
 
   /* Try LONGEST_MIN minus 1.  */
-  mpz_ui_pow_ui (v.val, 2, sizeof (LONGEST) * 8 - 1);
-  mpz_neg (v.val, v.val);
-  mpz_sub_ui (v.val, v.val, 1);
+  v = gdb_mpz::pow (2, sizeof (LONGEST) * 8 - 1);
+  v.negate ();
+  v -= 1;
 
   check_as_integer_raises_out_of_range_error<ULONGEST> (v);
   check_as_integer_raises_out_of_range_error<LONGEST> (v);
@@ -131,14 +131,14 @@ gdb_mpz_as_integer_out_of_range ()
 
   /* Try LONGEST_MAX plus 1.  */
   v = LONGEST_MAX;
-  mpz_add_ui (v.val, v.val, 1);
+  v += 1;
 
   SELF_CHECK (v.as_integer<ULONGEST> () == (ULONGEST) LONGEST_MAX + 1);
   check_as_integer_raises_out_of_range_error<LONGEST> (v);
 
   /* Try ULONGEST_MAX plus 1.  */
   v = ULONGEST_MAX;
-  mpz_add_ui (v.val, v.val, 1);
+  v += 1;
 
   check_as_integer_raises_out_of_range_error<ULONGEST> (v);
   check_as_integer_raises_out_of_range_error<LONGEST> (v);
@@ -170,8 +170,8 @@ store_and_read_back (T val, size_t buf_len, enum bfd_endian byte_order,
   store_integer (buf, buf_len, byte_order, val);
 
   /* Pre-initialize ACTUAL to something that's not the expected value.  */
-  mpz_set (actual.val, expected.val);
-  mpz_sub_ui (actual.val, actual.val, 500);
+  actual = expected;
+  actual -= 500;
 
   actual.read ({buf, buf_len}, byte_order, !std::is_signed<T>::value);
 }
@@ -197,10 +197,10 @@ gdb_mpz_read_all_from_small ()
       gdb_mpz expected, actual;
 
       store_and_read_back (l, buf_len, BFD_ENDIAN_BIG, expected, actual);
-      SELF_CHECK (mpz_cmp (actual.val, expected.val) == 0);
+      SELF_CHECK (actual ==  expected);
 
       store_and_read_back (l, buf_len, BFD_ENDIAN_LITTLE, expected, actual);
-      SELF_CHECK (mpz_cmp (actual.val, expected.val) == 0);
+      SELF_CHECK (actual == expected);
     }
 
   /* Do the same as above, but with an unsigned type.  */
@@ -212,10 +212,10 @@ gdb_mpz_read_all_from_small ()
       gdb_mpz expected, actual;
 
       store_and_read_back (ul, buf_len, BFD_ENDIAN_BIG, expected, actual);
-      SELF_CHECK (mpz_cmp (actual.val, expected.val) == 0);
+      SELF_CHECK (actual == expected);
 
       store_and_read_back (ul, buf_len, BFD_ENDIAN_LITTLE, expected, actual);
-      SELF_CHECK (mpz_cmp (actual.val, expected.val) == 0);
+      SELF_CHECK (actual == expected);
     }
 }
 
@@ -232,11 +232,11 @@ gdb_mpz_read_min_max ()
 
   store_and_read_back (l_min, sizeof (LONGEST), BFD_ENDIAN_BIG,
 		       expected, actual);
-  SELF_CHECK (mpz_cmp (actual.val, expected.val) == 0);
+  SELF_CHECK (actual == expected);
 
   store_and_read_back (l_min, sizeof (LONGEST), BFD_ENDIAN_LITTLE,
 		       expected, actual);
-  SELF_CHECK (mpz_cmp (actual.val, expected.val) == 0);
+  SELF_CHECK (actual == expected);
 
   /* Same with LONGEST_MAX.  */
 
@@ -244,11 +244,11 @@ gdb_mpz_read_min_max ()
 
   store_and_read_back (l_max, sizeof (LONGEST), BFD_ENDIAN_BIG,
 		       expected, actual);
-  SELF_CHECK (mpz_cmp (actual.val, expected.val) == 0);
+  SELF_CHECK (actual == expected);
 
   store_and_read_back (l_max, sizeof (LONGEST), BFD_ENDIAN_LITTLE,
 		       expected, actual);
-  SELF_CHECK (mpz_cmp (actual.val, expected.val) == 0);
+  SELF_CHECK (actual == expected);
 
   /* Same with the smallest ULONGEST.  */
 
@@ -256,11 +256,11 @@ gdb_mpz_read_min_max ()
 
   store_and_read_back (ul_min, sizeof (ULONGEST), BFD_ENDIAN_BIG,
 		       expected, actual);
-  SELF_CHECK (mpz_cmp (actual.val, expected.val) == 0);
+  SELF_CHECK (actual == expected);
 
   store_and_read_back (ul_min, sizeof (ULONGEST), BFD_ENDIAN_LITTLE,
 		       expected, actual);
-  SELF_CHECK (mpz_cmp (actual.val, expected.val) == 0);
+  SELF_CHECK (actual == expected);
 
   /* Same with ULONGEST_MAX.  */
 
@@ -268,11 +268,11 @@ gdb_mpz_read_min_max ()
 
   store_and_read_back (ul_max, sizeof (ULONGEST), BFD_ENDIAN_BIG,
 		       expected, actual);
-  SELF_CHECK (mpz_cmp (actual.val, expected.val) == 0);
+  SELF_CHECK (actual == expected);
 
   store_and_read_back (ul_max, sizeof (ULONGEST), BFD_ENDIAN_LITTLE,
 		       expected, actual);
-  SELF_CHECK (mpz_cmp (actual.val, expected.val) == 0);
+  SELF_CHECK (actual == expected);
 }
 
 /* A helper function which creates a gdb_mpz object from the given
diff --git a/gdb/valops.c b/gdb/valops.c
index 3a1b14c3d44..e0936d45cb3 100644
--- a/gdb/valops.c
+++ b/gdb/valops.c
@@ -354,7 +354,7 @@ value_to_gdb_mpq (struct value *value)
       gdb_mpz vz;
       vz.read (value->contents (), type_byte_order (type),
 	       type->is_unsigned ());
-      mpq_set_z (result.val, vz.val);
+      result = vz;
 
       if (is_fixed_point_type (type))
 	mpq_mul (result.val, result.val,
diff --git a/gdb/value.c b/gdb/value.c
index d2c2a2d7859..672e0c7023a 100644
--- a/gdb/value.c
+++ b/gdb/value.c
@@ -2721,8 +2721,7 @@ unpack_long (struct type *type, const gdb_byte *valaddr)
 			     byte_order, nosign,
 			     type->fixed_point_scaling_factor ());
 
-	gdb_mpz vz;
-	mpz_tdiv_q (vz.val, mpq_numref (vq.val), mpq_denref (vq.val));
+	gdb_mpz vz = vq.as_integer ();
 	return vz.as_integer<LONGEST> ();
       }
 

-- 
2.39.1

[PATCH 3/6] Rename gdb_mpz::val and make contents private

[Tom Tromey]

This changes gdb_mpz to hide its data, and renames the data member
from 'val' to 'm_val', following gdb convention.
---
 gdb/gmp-utils.c | 46 ++++++++++++++++++-------------------
 gdb/gmp-utils.h | 70 +++++++++++++++++++++++++++++++--------------------------
 2 files changed, 61 insertions(+), 55 deletions(-)

diff --git a/gdb/gmp-utils.c b/gdb/gmp-utils.c
index 8150fa5dcdf..aaa075e236e 100644
--- a/gdb/gmp-utils.c
+++ b/gdb/gmp-utils.c
@@ -46,7 +46,7 @@ void
 gdb_mpz::read (gdb::array_view<const gdb_byte> buf, enum bfd_endian byte_order,
 	       bool unsigned_p)
 {
-  mpz_import (val, 1 /* count */, -1 /* order */, buf.size () /* size */,
+  mpz_import (m_val, 1 /* count */, -1 /* order */, buf.size () /* size */,
 	      byte_order == BFD_ENDIAN_BIG ? 1 : -1 /* endian */,
 	      0 /* nails */, buf.data () /* op */);
 
@@ -57,9 +57,9 @@ gdb_mpz::read (gdb::array_view<const gdb_byte> buf, enum bfd_endian byte_order,
 	 was in fact negative, we need to adjust VAL accordingly.  */
       gdb_mpz max;
 
-      mpz_ui_pow_ui (max.val, 2, buf.size () * HOST_CHAR_BIT - 1);
-      if (mpz_cmp (val, max.val) >= 0)
-	mpz_submul_ui (val, max.val, 2);
+      mpz_ui_pow_ui (max.m_val, 2, buf.size () * HOST_CHAR_BIT - 1);
+      if (mpz_cmp (m_val, max.m_val) >= 0)
+	mpz_submul_ui (m_val, max.m_val, 2);
     }
 }
 
@@ -81,7 +81,7 @@ gdb_mpz::safe_export (gdb::array_view<gdb_byte> buf,
 {
   gdb_assert (buf.size () > 0);
 
-  if (mpz_sgn (val) == 0)
+  if (mpz_sgn (m_val) == 0)
     {
       /* Our value is zero, so no need to call mpz_export to do the work,
 	 especially since mpz_export's documentation explicitly says
@@ -100,19 +100,19 @@ gdb_mpz::safe_export (gdb::array_view<gdb_byte> buf,
     {
       lo = 0;
 
-      mpz_ui_pow_ui (hi.val, 2, max_usable_bits);
-      mpz_sub_ui (hi.val, hi.val, 1);
+      mpz_ui_pow_ui (hi.m_val, 2, max_usable_bits);
+      mpz_sub_ui (hi.m_val, hi.m_val, 1);
     }
   else
     {
-      mpz_ui_pow_ui (lo.val, 2, max_usable_bits - 1);
-      mpz_neg (lo.val, lo.val);
+      mpz_ui_pow_ui (lo.m_val, 2, max_usable_bits - 1);
+      mpz_neg (lo.m_val, lo.m_val);
 
-      mpz_ui_pow_ui (hi.val, 2, max_usable_bits - 1);
-      mpz_sub_ui (hi.val, hi.val, 1);
+      mpz_ui_pow_ui (hi.m_val, 2, max_usable_bits - 1);
+      mpz_sub_ui (hi.m_val, hi.m_val, 1);
     }
 
-  if (mpz_cmp (val, lo.val) < 0 || mpz_cmp (val, hi.val) > 0)
+  if (mpz_cmp (m_val, lo.m_val) < 0 || mpz_cmp (m_val, hi.m_val) > 0)
     error (_("Cannot export value %s as %zu-bits %s integer"
 	     " (must be between %s and %s)"),
 	   this->str ().c_str (),
@@ -121,17 +121,17 @@ gdb_mpz::safe_export (gdb::array_view<gdb_byte> buf,
 	   lo.str ().c_str (),
 	   hi.str ().c_str ());
 
-  gdb_mpz exported_val (val);
+  gdb_mpz exported_val (m_val);
 
-  if (mpz_cmp_ui (exported_val.val, 0) < 0)
+  if (mpz_cmp_ui (exported_val.m_val, 0) < 0)
     {
       /* mpz_export does not handle signed values, so create a positive
 	 value whose bit representation as an unsigned of the same length
 	 would be the same as our negative value.  */
       gdb_mpz neg_offset;
 
-      mpz_ui_pow_ui (neg_offset.val, 2, buf.size () * HOST_CHAR_BIT);
-      mpz_add (exported_val.val, exported_val.val, neg_offset.val);
+      mpz_ui_pow_ui (neg_offset.m_val, 2, buf.size () * HOST_CHAR_BIT);
+      mpz_add (exported_val.m_val, exported_val.m_val, neg_offset.m_val);
     }
 
   /* Do the export into a buffer allocated by GMP itself; that way,
@@ -147,7 +147,7 @@ gdb_mpz::safe_export (gdb::array_view<gdb_byte> buf,
   size_t word_countp;
   gdb::unique_xmalloc_ptr<void> exported
     (mpz_export (NULL, &word_countp, -1 /* order */, buf.size () /* size */,
-		 endian, 0 /* nails */, exported_val.val));
+		 endian, 0 /* nails */, exported_val.m_val));
 
   gdb_assert (word_countp == 1);
 
@@ -170,19 +170,19 @@ gdb_mpq::get_rounded () const
      towards zero.  */
 
   gdb_mpz quotient, remainder;
-  mpz_fdiv_qr (quotient.val, remainder.val,
+  mpz_fdiv_qr (quotient.m_val, remainder.m_val,
 	       mpq_numref (abs_val.val), mpq_denref (abs_val.val));
 
   /* Multiply the remainder by 2, and see if it is greater or equal
      to abs_val's denominator.  If yes, round to the next integer.  */
 
-  mpz_mul_ui (remainder.val, remainder.val, 2);
-  if (mpz_cmp (remainder.val, mpq_denref (abs_val.val)) >= 0)
-    mpz_add_ui (quotient.val, quotient.val, 1);
+  mpz_mul_ui (remainder.m_val, remainder.m_val, 2);
+  if (mpz_cmp (remainder.m_val, mpq_denref (abs_val.val)) >= 0)
+    mpz_add_ui (quotient.m_val, quotient.m_val, 1);
 
   /* Re-apply the sign if needed.  */
   if (mpq_sgn (val) < 0)
-    mpz_neg (quotient.val, quotient.val);
+    mpz_neg (quotient.m_val, quotient.m_val);
 
   return quotient;
 }
@@ -197,7 +197,7 @@ gdb_mpq::read_fixed_point (gdb::array_view<const gdb_byte> buf,
   gdb_mpz vz;
   vz.read (buf, byte_order, unsigned_p);
 
-  mpq_set_z (val, vz.val);
+  mpq_set_z (val, vz.m_val);
   mpq_mul (val, val, scaling_factor.val);
 }
 
diff --git a/gdb/gmp-utils.h b/gdb/gmp-utils.h
index dcb998a1656..601d9a6a1ec 100644
--- a/gdb/gmp-utils.h
+++ b/gdb/gmp-utils.h
@@ -31,25 +31,26 @@
 
 std::string gmp_string_printf (const char *fmt, ...);
 
+struct gdb_mpq;
+struct gdb_mpf;
+
 /* A class to make it easier to use GMP's mpz_t values within GDB.  */
 
 struct gdb_mpz
 {
-  mpz_t val;
-
   /* Constructors.  */
-  gdb_mpz () { mpz_init (val); }
+  gdb_mpz () { mpz_init (m_val); }
 
   explicit gdb_mpz (const mpz_t &from_val)
   {
-    mpz_init (val);
-    mpz_set (val, from_val);
+    mpz_init (m_val);
+    mpz_set (m_val, from_val);
   }
 
   gdb_mpz (const gdb_mpz &from)
   {
-    mpz_init (val);
-    mpz_set (val, from.val);
+    mpz_init (m_val);
+    mpz_set (m_val, from.m_val);
   }
 
   /* Initialize using the given integral value.
@@ -59,26 +60,26 @@ struct gdb_mpz
   template<typename T, typename = gdb::Requires<std::is_integral<T>>>
   explicit gdb_mpz (T src)
   {
-    mpz_init (val);
+    mpz_init (m_val);
     set (src);
   }
 
   explicit gdb_mpz (gdb_mpz &&from)
   {
-    mpz_init (val);
-    mpz_swap (val, from.val);
+    mpz_init (m_val);
+    mpz_swap (m_val, from.m_val);
   }
 
   
   gdb_mpz &operator= (const gdb_mpz &from)
   {
-    mpz_set (val, from.val);
+    mpz_set (m_val, from.m_val);
     return *this;
   }
 
   gdb_mpz &operator= (gdb_mpz &&other)
   {
-    mpz_swap (val, other.val);
+    mpz_swap (m_val, other.m_val);
     return *this;
   }
 
@@ -93,14 +94,14 @@ struct gdb_mpz
      the string was parsed successfully, false otherwise.  */
   bool set (const char *str, int base)
   {
-    return mpz_set_str (val, str, base) != -1;
+    return mpz_set_str (m_val, str, base) != -1;
   }
 
   /* Return a new value that is BASE**EXP.  */
   static gdb_mpz pow (unsigned long base, unsigned long exp)
   {
     gdb_mpz result;
-    mpz_ui_pow_ui (result.val, base, exp);
+    mpz_ui_pow_ui (result.m_val, base, exp);
     return result;
   }
 
@@ -125,59 +126,59 @@ struct gdb_mpz
 	      bool unsigned_p) const;
 
   /* Return a string containing VAL.  */
-  std::string str () const { return gmp_string_printf ("%Zd", val); }
+  std::string str () const { return gmp_string_printf ("%Zd", m_val); }
 
   /* The destructor.  */
-  ~gdb_mpz () { mpz_clear (val); }
+  ~gdb_mpz () { mpz_clear (m_val); }
 
   /* Negate this value in place.  */
   void negate ()
   {
-    mpz_neg (val, val);
+    mpz_neg (m_val, m_val);
   }
 
   gdb_mpz &operator*= (long other)
   {
-    mpz_mul_si (val, val, other);
+    mpz_mul_si (m_val, m_val, other);
     return *this;
   }
 
   gdb_mpz &operator+= (unsigned long other)
   {
-    mpz_add_ui (val, val, other);
+    mpz_add_ui (m_val, m_val, other);
     return *this;
   }
 
   gdb_mpz &operator-= (unsigned long other)
   {
-    mpz_sub_ui (val, val, other);
+    mpz_sub_ui (m_val, m_val, other);
     return *this;
   }
 
   gdb_mpz &operator<<= (unsigned long nbits)
   {
-    mpz_mul_2exp (val, val, nbits);
+    mpz_mul_2exp (m_val, m_val, nbits);
     return *this;
   }
 
   bool operator> (const gdb_mpz &other) const
   {
-    return mpz_cmp (val, other.val) > 0;
+    return mpz_cmp (m_val, other.m_val) > 0;
   }
 
   bool operator< (int other) const
   {
-    return mpz_cmp_si (val, other) < 0;
+    return mpz_cmp_si (m_val, other) < 0;
   }
 
   bool operator== (int other) const
   {
-    return mpz_cmp_si (val, other) == 0;
+    return mpz_cmp_si (m_val, other) == 0;
   }
 
   bool operator== (const gdb_mpz &other) const
   {
-    return mpz_cmp (val, other.val) == 0;
+    return mpz_cmp (m_val, other.m_val) == 0;
   }
 
 private:
@@ -202,6 +203,11 @@ struct gdb_mpz
     being exported.  */
   void safe_export (gdb::array_view<gdb_byte> buf,
 		    int endian, bool unsigned_p) const;
+
+  friend struct gdb_mpq;
+  friend struct gdb_mpf;
+
+  mpz_t m_val;
 };
 
 /* A class to make it easier to use GMP's mpq_t values within GDB.  */
@@ -234,8 +240,8 @@ struct gdb_mpq
   gdb_mpq (const gdb_mpz &num, const gdb_mpz &denom)
   {
     mpq_init (val);
-    mpz_set (mpq_numref (val), num.val);
-    mpz_set (mpq_denref (val), denom.val);
+    mpz_set (mpq_numref (val), num.m_val);
+    mpz_set (mpq_denref (val), denom.m_val);
     mpq_canonicalize (val);
   }
 
@@ -254,7 +260,7 @@ struct gdb_mpq
 
   gdb_mpq &operator= (const gdb_mpz &from)
   {
-    mpq_set_z (val, from.val);
+    mpq_set_z (val, from.m_val);
     return *this;
   }
 
@@ -268,7 +274,7 @@ struct gdb_mpq
   gdb_mpz as_integer () const
   {
     gdb_mpz result;
-    mpz_tdiv_q (result.val, mpq_numref (val), mpq_denref (val));
+    mpz_tdiv_q (result.m_val, mpq_numref (val), mpq_denref (val));
     return result;
   }
 
@@ -340,7 +346,7 @@ template<typename T>
 void
 gdb_mpz::set (T src)
 {
-  mpz_import (val, 1 /* count */, -1 /* order */,
+  mpz_import (m_val, 1 /* count */, -1 /* order */,
 	      sizeof (T) /* size */, 0 /* endian (0 = native) */,
 	      0 /* nails */, &src /* op */);
   if (std::is_signed<T>::value && src < 0)
@@ -350,8 +356,8 @@ gdb_mpz::set (T src)
 	 the correct negative value.  */
       gdb_mpz neg_offset;
 
-      mpz_ui_pow_ui (neg_offset.val, 2, sizeof (T) * HOST_CHAR_BIT);
-      mpz_sub (val, val, neg_offset.val);
+      mpz_ui_pow_ui (neg_offset.m_val, 2, sizeof (T) * HOST_CHAR_BIT);
+      mpz_sub (m_val, m_val, neg_offset.m_val);
     }
 }
 

-- 
2.39.1

[PATCH 4/6] Add operators and methods to gdb_mpq

[Tom Tromey]

This adds some operators and methods to gdb_mpq, in preparation for
making its implementation private.

This only adds the operators currently needed by gdb.  More could be
added as necessary.
---
 gdb/gmp-utils.h                     | 72 +++++++++++++++++++++++++++++++++++++
 gdb/unittests/gmp-utils-selftests.c | 34 ++++++++----------
 gdb/valarith.c                      | 18 +++++-----
 gdb/valops.c                        | 13 +++----
 4 files changed, 100 insertions(+), 37 deletions(-)

diff --git a/gdb/gmp-utils.h b/gdb/gmp-utils.h
index 601d9a6a1ec..dfb61302f6b 100644
--- a/gdb/gmp-utils.h
+++ b/gdb/gmp-utils.h
@@ -245,6 +245,13 @@ struct gdb_mpq
     mpq_canonicalize (val);
   }
 
+  gdb_mpq (long num, long denom)
+  {
+    mpq_init (val);
+    mpq_set_si (val, num, denom);
+    mpq_canonicalize (val);
+  }
+
   /* Copy assignment operator.  */
   gdb_mpq &operator= (const gdb_mpq &from)
   {
@@ -264,6 +271,67 @@ struct gdb_mpq
     return *this;
   }
 
+  gdb_mpq &operator= (double d)
+  {
+    mpq_set_d (val, d);
+    return *this;
+  }
+
+  /* Return the sign of this value.  This returns -1 for a negative
+     value, 0 if the value is 0, and 1 for a positive value.  */
+  int sgn () const
+  { return mpq_sgn (val); }
+
+  gdb_mpq operator+ (const gdb_mpq &other) const
+  {
+    gdb_mpq result;
+    mpq_add (result.val, val, other.val);
+    return result;
+  }
+
+  gdb_mpq operator- (const gdb_mpq &other) const
+  {
+    gdb_mpq result;
+    mpq_sub (result.val, val, other.val);
+    return result;
+  }
+
+  gdb_mpq operator* (const gdb_mpq &other) const
+  {
+    gdb_mpq result;
+    mpq_mul (result.val, val, other.val);
+    return result;
+  }
+
+  gdb_mpq operator/ (const gdb_mpq &other) const
+  {
+    gdb_mpq result;
+    mpq_div (result.val, val, other.val);
+    return result;
+  }
+
+  gdb_mpq &operator*= (const gdb_mpq &other)
+  {
+    mpq_mul (val, val, other.val);
+    return *this;
+  }
+
+  gdb_mpq &operator/= (const gdb_mpq &other)
+  {
+    mpq_div (val, val, other.val);
+    return *this;
+  }
+
+  bool operator== (const gdb_mpq &other) const
+  {
+    return mpq_cmp (val, other.val) == 0;
+  }
+
+  bool operator< (const gdb_mpq &other) const
+  {
+    return mpq_cmp (val, other.val) < 0;
+  }
+
   /* Return a string representing VAL as "<numerator> / <denominator>".  */
   std::string str () const { return gmp_string_printf ("%Qd", val); }
 
@@ -278,6 +346,10 @@ struct gdb_mpq
     return result;
   }
 
+  /* Return this value converted to a host double.  */
+  double as_double () const
+  { return mpq_get_d (val); }
+
   /* Set VAL from the contents of the given byte array (BUF), which
      contains the unscaled value of a fixed point type object.
      The byte size of the data is the size of BUF.
diff --git a/gdb/unittests/gmp-utils-selftests.c b/gdb/unittests/gmp-utils-selftests.c
index 8e028fec88c..384ca2df212 100644
--- a/gdb/unittests/gmp-utils-selftests.c
+++ b/gdb/unittests/gmp-utils-selftests.c
@@ -399,8 +399,8 @@ read_fp_test (int unscaled, const gdb_mpq &scaling_factor,
 
   actual.read_fixed_point ({buf, len}, byte_order, 0, scaling_factor);
 
-  mpq_set_si (expected.val, unscaled, 1);
-  mpq_mul (expected.val, expected.val, scaling_factor.val);
+  expected = gdb_mpq (unscaled, 1);
+  expected *= scaling_factor;
 }
 
 /* Perform various tests of the gdb_mpq::read_fixed_point method.  */
@@ -409,38 +409,37 @@ static void
 gdb_mpq_read_fixed_point ()
 {
   gdb_mpq expected, actual;
-  gdb_mpq scaling_factor;
 
   /* Pick an arbitrary scaling_factor; this operation is trivial enough
      thanks to GMP that the value we use isn't really important.  */
-  mpq_set_ui (scaling_factor.val, 3, 5);
+  gdb_mpq scaling_factor (3, 5);
 
   /* Try a few values, both negative and positive... */
 
   read_fp_test (-256, scaling_factor, BFD_ENDIAN_BIG, expected, actual);
-  SELF_CHECK (mpq_cmp (actual.val, expected.val) == 0);
+  SELF_CHECK (actual == expected);
   read_fp_test (-256, scaling_factor, BFD_ENDIAN_LITTLE, expected, actual);
-  SELF_CHECK (mpq_cmp (actual.val, expected.val) == 0);
+  SELF_CHECK (actual == expected);
 
   read_fp_test (-1, scaling_factor, BFD_ENDIAN_BIG, expected, actual);
-  SELF_CHECK (mpq_cmp (actual.val, expected.val) == 0);
+  SELF_CHECK (actual == expected);
   read_fp_test (-1, scaling_factor, BFD_ENDIAN_LITTLE, expected, actual);
-  SELF_CHECK (mpq_cmp (actual.val, expected.val) == 0);
+  SELF_CHECK (actual == expected);
 
   read_fp_test (0, scaling_factor, BFD_ENDIAN_BIG, expected, actual);
-  SELF_CHECK (mpq_cmp (actual.val, expected.val) == 0);
+  SELF_CHECK (actual == expected);
   read_fp_test (0, scaling_factor, BFD_ENDIAN_LITTLE, expected, actual);
-  SELF_CHECK (mpq_cmp (actual.val, expected.val) == 0);
+  SELF_CHECK (actual == expected);
 
   read_fp_test (1, scaling_factor, BFD_ENDIAN_BIG, expected, actual);
-  SELF_CHECK (mpq_cmp (actual.val, expected.val) == 0);
+  SELF_CHECK (actual == expected);
   read_fp_test (1, scaling_factor, BFD_ENDIAN_LITTLE, expected, actual);
-  SELF_CHECK (mpq_cmp (actual.val, expected.val) == 0);
+  SELF_CHECK (actual == expected);
 
   read_fp_test (1025, scaling_factor, BFD_ENDIAN_BIG, expected, actual);
-  SELF_CHECK (mpq_cmp (actual.val, expected.val) == 0);
+  SELF_CHECK (actual == expected);
   read_fp_test (1025, scaling_factor, BFD_ENDIAN_LITTLE, expected, actual);
-  SELF_CHECK (mpq_cmp (actual.val, expected.val) == 0);
+  SELF_CHECK (actual == expected);
 }
 
 /* A helper function which builds a gdb_mpq object from the given
@@ -463,9 +462,7 @@ write_fp_test (int numerator, unsigned int denominator,
   gdb_byte buf[len];
   memset (buf, 0, len);
 
-  gdb_mpq v;
-  mpq_set_si (v.val, numerator, denominator);
-  mpq_canonicalize (v.val);
+  gdb_mpq v (numerator, denominator);
   v.write_fixed_point ({buf, len}, byte_order, 0, scaling_factor);
 
   return extract_unsigned_integer (buf, len, byte_order);
@@ -479,8 +476,7 @@ gdb_mpq_write_fixed_point ()
   /* Pick an arbitrary factor; this operations is sufficiently trivial
      with the use of GMP that the value of this factor is not really
      all that important.  */
-  gdb_mpq scaling_factor;
-  mpq_set_ui (scaling_factor.val, 1, 3);
+  gdb_mpq scaling_factor (1, 3);
 
   gdb_mpq vq;
 
diff --git a/gdb/valarith.c b/gdb/valarith.c
index a6a5f5102a2..f8ce508e825 100644
--- a/gdb/valarith.c
+++ b/gdb/valarith.c
@@ -883,43 +883,43 @@ fixed_point_binop (struct value *arg1, struct value *arg2, enum exp_opcode op)
   switch (op)
     {
     case BINOP_ADD:
-      mpq_add (res.val, v1.val, v2.val);
+      res = v1 + v2;
       val = fixed_point_to_value (res);
       break;
 
     case BINOP_SUB:
-      mpq_sub (res.val, v1.val, v2.val);
+      res = v1 - v2;
       val = fixed_point_to_value (res);
       break;
 
     case BINOP_MIN:
-      val = fixed_point_to_value (mpq_cmp (v1.val, v2.val) < 0 ? v1 : v2);
+      val = fixed_point_to_value (std::min (v1, v2));
       break;
 
     case BINOP_MAX:
-      val = fixed_point_to_value (mpq_cmp (v1.val, v2.val) > 0 ? v1 : v2);
+      val = fixed_point_to_value (std::max (v1, v2));
       break;
 
     case BINOP_MUL:
-      mpq_mul (res.val, v1.val, v2.val);
+      res = v1 * v2;
       val = fixed_point_to_value (res);
       break;
 
     case BINOP_DIV:
-      if (mpq_sgn (v2.val) == 0)
+      if (v2.sgn () == 0)
 	error (_("Division by zero"));
-      mpq_div (res.val, v1.val, v2.val);
+      res = v1 / v2;
       val = fixed_point_to_value (res);
       break;
 
     case BINOP_EQUAL:
       val = value_from_ulongest (language_bool_type (language, gdbarch),
-				 mpq_cmp (v1.val, v2.val) == 0 ? 1 : 0);
+				 v1 == v2 ? 1 : 0);
       break;
 
     case BINOP_LESS:
       val = value_from_ulongest (language_bool_type (language, gdbarch),
-				 mpq_cmp (v1.val, v2.val) < 0 ? 1 : 0);
+				 v1 < v2 ? 1 : 0);
       break;
 
     default:
diff --git a/gdb/valops.c b/gdb/valops.c
index e0936d45cb3..2cef24fc4c5 100644
--- a/gdb/valops.c
+++ b/gdb/valops.c
@@ -341,11 +341,7 @@ value_to_gdb_mpq (struct value *value)
 
   gdb_mpq result;
   if (is_floating_type (type))
-    {
-      double d = target_float_to_host_double (value->contents ().data (),
-					      type);
-      mpq_set_d (result.val, d);
-    }
+    result = target_float_to_host_double (value->contents ().data (), type);
   else
     {
       gdb_assert (is_integral_type (type)
@@ -357,8 +353,7 @@ value_to_gdb_mpq (struct value *value)
       result = vz;
 
       if (is_fixed_point_type (type))
-	mpq_mul (result.val, result.val,
-		 type->fixed_point_scaling_factor ().val);
+	result *= type->fixed_point_scaling_factor ();
     }
 
   return result;
@@ -386,7 +381,7 @@ value_cast_to_fixed_point (struct type *to_type, struct value *from_val)
   /* Divide that value by the scaling factor to obtain the unscaled
      value, first in rational form, and then in integer form.  */
 
-  mpq_div (vq.val, vq.val, to_type->fixed_point_scaling_factor ().val);
+  vq /= to_type->fixed_point_scaling_factor ();
   gdb_mpz unscaled = vq.get_rounded ();
 
   /* Finally, create the result value, and pack the unscaled value
@@ -559,7 +554,7 @@ value_cast (struct type *type, struct value *arg2)
 
 	  struct value *v = value::allocate (to_type);
 	  target_float_from_host_double (v->contents_raw ().data (),
-					 to_type, mpq_get_d (fp_val.val));
+					 to_type, fp_val.as_double ());
 	  return v;
 	}
 

-- 
2.39.1

[PATCH 5/6] Rename gdb_mpq::val and make contents private

[Tom Tromey]

This changes gdb_mpq to hide its data, and renames the data member
from 'val' to 'm_val', following gdb convention.
---
 gdb/gmp-utils.c | 18 +++++++--------
 gdb/gmp-utils.h | 72 ++++++++++++++++++++++++++++++---------------------------
 2 files changed, 47 insertions(+), 43 deletions(-)

diff --git a/gdb/gmp-utils.c b/gdb/gmp-utils.c
index aaa075e236e..d134bc32a1d 100644
--- a/gdb/gmp-utils.c
+++ b/gdb/gmp-utils.c
@@ -162,8 +162,8 @@ gdb_mpq::get_rounded () const
   /* Work with a positive number so as to make the "floor" rounding
      always round towards zero.  */
 
-  gdb_mpq abs_val (val);
-  mpq_abs (abs_val.val, abs_val.val);
+  gdb_mpq abs_val (m_val);
+  mpq_abs (abs_val.m_val, abs_val.m_val);
 
   /* Convert our rational number into a quotient and remainder,
      with "floor" rounding, which in our case means rounding
@@ -171,17 +171,17 @@ gdb_mpq::get_rounded () const
 
   gdb_mpz quotient, remainder;
   mpz_fdiv_qr (quotient.m_val, remainder.m_val,
-	       mpq_numref (abs_val.val), mpq_denref (abs_val.val));
+	       mpq_numref (abs_val.m_val), mpq_denref (abs_val.m_val));
 
   /* Multiply the remainder by 2, and see if it is greater or equal
      to abs_val's denominator.  If yes, round to the next integer.  */
 
   mpz_mul_ui (remainder.m_val, remainder.m_val, 2);
-  if (mpz_cmp (remainder.m_val, mpq_denref (abs_val.val)) >= 0)
+  if (mpz_cmp (remainder.m_val, mpq_denref (abs_val.m_val)) >= 0)
     mpz_add_ui (quotient.m_val, quotient.m_val, 1);
 
   /* Re-apply the sign if needed.  */
-  if (mpq_sgn (val) < 0)
+  if (mpq_sgn (m_val) < 0)
     mpz_neg (quotient.m_val, quotient.m_val);
 
   return quotient;
@@ -197,8 +197,8 @@ gdb_mpq::read_fixed_point (gdb::array_view<const gdb_byte> buf,
   gdb_mpz vz;
   vz.read (buf, byte_order, unsigned_p);
 
-  mpq_set_z (val, vz.m_val);
-  mpq_mul (val, val, scaling_factor.val);
+  mpq_set_z (m_val, vz.m_val);
+  mpq_mul (m_val, m_val, scaling_factor.m_val);
 }
 
 /* See gmp-utils.h.  */
@@ -208,9 +208,9 @@ gdb_mpq::write_fixed_point (gdb::array_view<gdb_byte> buf,
 			    enum bfd_endian byte_order, bool unsigned_p,
 			    const gdb_mpq &scaling_factor) const
 {
-  gdb_mpq unscaled (val);
+  gdb_mpq unscaled (m_val);
 
-  mpq_div (unscaled.val, unscaled.val, scaling_factor.val);
+  mpq_div (unscaled.m_val, unscaled.m_val, scaling_factor.m_val);
 
   gdb_mpz unscaled_z = unscaled.get_rounded ();
   unscaled_z.write (buf, byte_order, unsigned_p);
diff --git a/gdb/gmp-utils.h b/gdb/gmp-utils.h
index dfb61302f6b..7889756a0ea 100644
--- a/gdb/gmp-utils.h
+++ b/gdb/gmp-utils.h
@@ -214,126 +214,124 @@ struct gdb_mpz
 
 struct gdb_mpq
 {
-  mpq_t val;
-
   /* Constructors.  */
-  gdb_mpq () { mpq_init (val); }
+  gdb_mpq () { mpq_init (m_val); }
 
   explicit gdb_mpq (const mpq_t &from_val)
   {
-    mpq_init (val);
-    mpq_set (val, from_val);
+    mpq_init (m_val);
+    mpq_set (m_val, from_val);
   }
 
   gdb_mpq (const gdb_mpq &from)
   {
-    mpq_init (val);
-    mpq_set (val, from.val);
+    mpq_init (m_val);
+    mpq_set (m_val, from.m_val);
   }
 
   explicit gdb_mpq (gdb_mpq &&from)
   {
-    mpq_init (val);
-    mpq_swap (val, from.val);
+    mpq_init (m_val);
+    mpq_swap (m_val, from.m_val);
   }
 
   gdb_mpq (const gdb_mpz &num, const gdb_mpz &denom)
   {
-    mpq_init (val);
-    mpz_set (mpq_numref (val), num.m_val);
-    mpz_set (mpq_denref (val), denom.m_val);
-    mpq_canonicalize (val);
+    mpq_init (m_val);
+    mpz_set (mpq_numref (m_val), num.m_val);
+    mpz_set (mpq_denref (m_val), denom.m_val);
+    mpq_canonicalize (m_val);
   }
 
   gdb_mpq (long num, long denom)
   {
-    mpq_init (val);
-    mpq_set_si (val, num, denom);
-    mpq_canonicalize (val);
+    mpq_init (m_val);
+    mpq_set_si (m_val, num, denom);
+    mpq_canonicalize (m_val);
   }
 
   /* Copy assignment operator.  */
   gdb_mpq &operator= (const gdb_mpq &from)
   {
-    mpq_set (val, from.val);
+    mpq_set (m_val, from.m_val);
     return *this;
   }
 
   gdb_mpq &operator= (gdb_mpq &&from)
   {
-    mpq_swap (val, from.val);
+    mpq_swap (m_val, from.m_val);
     return *this;
   }
 
   gdb_mpq &operator= (const gdb_mpz &from)
   {
-    mpq_set_z (val, from.m_val);
+    mpq_set_z (m_val, from.m_val);
     return *this;
   }
 
   gdb_mpq &operator= (double d)
   {
-    mpq_set_d (val, d);
+    mpq_set_d (m_val, d);
     return *this;
   }
 
   /* Return the sign of this value.  This returns -1 for a negative
      value, 0 if the value is 0, and 1 for a positive value.  */
   int sgn () const
-  { return mpq_sgn (val); }
+  { return mpq_sgn (m_val); }
 
   gdb_mpq operator+ (const gdb_mpq &other) const
   {
     gdb_mpq result;
-    mpq_add (result.val, val, other.val);
+    mpq_add (result.m_val, m_val, other.m_val);
     return result;
   }
 
   gdb_mpq operator- (const gdb_mpq &other) const
   {
     gdb_mpq result;
-    mpq_sub (result.val, val, other.val);
+    mpq_sub (result.m_val, m_val, other.m_val);
     return result;
   }
 
   gdb_mpq operator* (const gdb_mpq &other) const
   {
     gdb_mpq result;
-    mpq_mul (result.val, val, other.val);
+    mpq_mul (result.m_val, m_val, other.m_val);
     return result;
   }
 
   gdb_mpq operator/ (const gdb_mpq &other) const
   {
     gdb_mpq result;
-    mpq_div (result.val, val, other.val);
+    mpq_div (result.m_val, m_val, other.m_val);
     return result;
   }
 
   gdb_mpq &operator*= (const gdb_mpq &other)
   {
-    mpq_mul (val, val, other.val);
+    mpq_mul (m_val, m_val, other.m_val);
     return *this;
   }
 
   gdb_mpq &operator/= (const gdb_mpq &other)
   {
-    mpq_div (val, val, other.val);
+    mpq_div (m_val, m_val, other.m_val);
     return *this;
   }
 
   bool operator== (const gdb_mpq &other) const
   {
-    return mpq_cmp (val, other.val) == 0;
+    return mpq_cmp (m_val, other.m_val) == 0;
   }
 
   bool operator< (const gdb_mpq &other) const
   {
-    return mpq_cmp (val, other.val) < 0;
+    return mpq_cmp (m_val, other.m_val) < 0;
   }
 
   /* Return a string representing VAL as "<numerator> / <denominator>".  */
-  std::string str () const { return gmp_string_printf ("%Qd", val); }
+  std::string str () const { return gmp_string_printf ("%Qd", m_val); }
 
   /* Return VAL rounded to the nearest integer.  */
   gdb_mpz get_rounded () const;
@@ -342,13 +340,13 @@ struct gdb_mpq
   gdb_mpz as_integer () const
   {
     gdb_mpz result;
-    mpz_tdiv_q (result.m_val, mpq_numref (val), mpq_denref (val));
+    mpz_tdiv_q (result.m_val, mpq_numref (m_val), mpq_denref (m_val));
     return result;
   }
 
   /* Return this value converted to a host double.  */
   double as_double () const
-  { return mpq_get_d (val); }
+  { return mpq_get_d (m_val); }
 
   /* Set VAL from the contents of the given byte array (BUF), which
      contains the unscaled value of a fixed point type object.
@@ -374,7 +372,13 @@ struct gdb_mpq
 			  const gdb_mpq &scaling_factor) const;
 
   /* The destructor.  */
-  ~gdb_mpq () { mpq_clear (val); }
+  ~gdb_mpq () { mpq_clear (m_val); }
+
+private:
+
+  friend struct gdb_mpf;
+
+  mpq_t m_val;
 };
 
 /* A class to make it easier to use GMP's mpf_t values within GDB.
@@ -405,7 +409,7 @@ struct gdb_mpf
     gdb_mpq tmp_q;
 
     tmp_q.read_fixed_point (buf, byte_order, unsigned_p, scaling_factor);
-    mpf_set_q (val, tmp_q.val);
+    mpf_set_q (val, tmp_q.m_val);
   }
 
   /* The destructor.  */

-- 
2.39.1

[PATCH 6/6] Hide the implementation of gdb_mpf

[Tom Tromey]

This renames the data member of gdb_mpf and makes it private.  It also
adds a single new method to aid in this change.  Unlike the earlier
changes here, I did this one all together because gdb_mpf has very few
uses.
---
 gdb/gmp-utils.h | 17 ++++++++++++-----
 gdb/valprint.c  |  2 +-
 2 files changed, 13 insertions(+), 6 deletions(-)

diff --git a/gdb/gmp-utils.h b/gdb/gmp-utils.h
index 7889756a0ea..274c28c0ce8 100644
--- a/gdb/gmp-utils.h
+++ b/gdb/gmp-utils.h
@@ -388,10 +388,8 @@ struct gdb_mpq
 
 struct gdb_mpf
 {
-  mpf_t val;
-
   /* Constructors.  */
-  gdb_mpf () { mpf_init (val); }
+  gdb_mpf () { mpf_init (m_val); }
 
   DISABLE_COPY_AND_ASSIGN (gdb_mpf);
 
@@ -409,11 +407,20 @@ struct gdb_mpf
     gdb_mpq tmp_q;
 
     tmp_q.read_fixed_point (buf, byte_order, unsigned_p, scaling_factor);
-    mpf_set_q (val, tmp_q.m_val);
+    mpf_set_q (m_val, tmp_q.m_val);
   }
 
+  /* Convert this value to a string.  FMT is the format to use, and
+     should have a single '%' substitution.  */
+  std::string str (const char *fmt) const
+  { return gmp_string_printf (fmt, m_val); }
+
   /* The destructor.  */
-  ~gdb_mpf () { mpf_clear (val); }
+  ~gdb_mpf () { mpf_clear (m_val); }
+
+private:
+
+  mpf_t m_val;
 };
 
 /* See declaration above.  */
diff --git a/gdb/valprint.c b/gdb/valprint.c
index 357db3815b0..b05d90dd26b 100644
--- a/gdb/valprint.c
+++ b/gdb/valprint.c
@@ -856,7 +856,7 @@ generic_val_print_fixed_point (struct value *val, struct ui_file *stream,
 			  type->fixed_point_scaling_factor ());
 
       const char *fmt = type->length () < 4 ? "%.11Fg" : "%.17Fg";
-      std::string str = gmp_string_printf (fmt, f.val);
+      std::string str = f.str (fmt);
       gdb_printf (stream, "%s", str.c_str ());
     }
 }

-- 
2.39.1

Re: [PATCH 0/6] Make GMP data members private

[Tom Tromey]

>>>>> "Tom" == Tom Tromey via Gdb-patches <gdb-patches@sourceware.org> writes:

Tom> This series changes the GMP code in gmp-utils.h to add methods and
Tom> operators, and then to make the data members private.  Ages ago I told
Tom> Simon I'd do this, then I dropped it for years.

Tom> I've only added exactly the operators that are needed right now.  It's
Tom> simple to add more operators or more overloads as required.

Tom> Regression tested on x86-64 Fedora 36.

I rebased this and I am going to check it in now.

Tom