1 files changed, 47 insertions, 3 deletions
diff --git a/src/grt/grt-fcvt.ads b/src/grt/grt-fcvt.ads
index 82ed94363..e1e272850 100644
--- a/src/grt/grt-fcvt.ads
+++ b/src/grt/grt-fcvt.ads
@@ -23,16 +23,60 @@
 --  however invalidate any other reasons why the executable file might be
 --  covered by the GNU Public License.
 
+--  IMPORTANT: this unit can also be used by the front-end, so it must NOT
+--  depend on anything else from grt (except the grt parent package), including
+--  grt.types.
+
 with Interfaces; use Interfaces;
 
 package Grt.Fcvt is
-   --  Convert V to RADIX number stored (in ASCII) in STR/LEN [using at most
+   --  Convert V to 10-based number stored (in ASCII) in STR/LEN [using at most
    --  NDIGITS digits.]
    --  LEN is the number of characters needed (so it may be greater than
    --   STR'Length).
    --  Requires STR'First = 1.
    procedure To_String (Str : out String;
                         Len : out Natural;
-                        V : IEEE_Float_64;
-                        Radix : Positive := 10);
+                        V : IEEE_Float_64);
+
+   --  Input format is [+-]int[.int][e[+-]int]
+   --  where int is digit { _ digit }
+   --    and [+-] means optional + or -.
+   --  The input string must be correctly formatted.
+   function From_String (Str : String) return IEEE_Float_64;
+
+   --  Ad-hoc implementation of bignums, with the minimal features to support
+   --  radix conversion.
+   type Bignum is private;
+
+   --  Compute L * R + CARRY_IN.
+   procedure Bignum_Mul_Int
+     (L : in out Bignum; R : Positive; Carry_In : Natural := 0);
+
+   --  Multiply two bignums.
+   function Bignum_Mul (L, R : Bignum) return Bignum;
+
+   --  Compute L**N
+   function Bignum_Pow (L : Natural; N : Natural) return Bignum;
+
+   --  Create a bignum from a natural.
+   procedure Bignum_Int (Res : out Bignum; N : Natural);
+
+   --  Convert N to RES.  OK is set to True if the number fits in RES.
+   procedure Bignum_To_Int
+     (N : Bignum; Res : out Unsigned_64; OK : out Boolean);
+
+   --  Return (-1)**Neg * F * BASE**EXP to a float.
+   function To_Float_64
+     (Neg : Boolean; F : Bignum; Base : Positive; Exp : Integer)
+     return IEEE_Float_64;
+private
+   type Unsigned_32_Array is array (Natural range <>) of Unsigned_32;
+
+   type Bignum is record
+      --  Number of digits.  Must be 0 for number 0.
+      N : Natural;
+      --  Digits.  The leading digit V(N + 1) must not be 0.
+      V : Unsigned_32_Array (1 .. 37);
+   end record;
 end Grt.Fcvt;