Assignment float-to-rational

mpir.net/mpir.net-tests/HugeRationalTests/Assignment.cs

@@ -86,6 +86,17 @@ namespace MPIR.Tests.HugeRationalTests
                 Assert.AreEqual(d * 5, b.Denominator);
             }
         }
+
+        [TestMethod]
+        public void RationalAssignFloat()
+        {
+            using(var a = new HugeFloat("222509832503450298342455029.125"))
+            using(var b = new HugeRational("1/3"))
+            {
+                b.SetTo(a);
+                Assert.AreEqual("1780078660027602386739640233/8", b.ToString());
+            }
+        }
         //more tests coming here
     }
 }

mpir.net/mpir.net/HugeFloat.cpp

@@ -427,7 +427,7 @@ namespace MPIR
 
     #pragma endregion
 
-    #pragma region methods in other classes with rational parameters
+    #pragma region methods in other classes with float parameters
 
     void HugeInt::SetTo(MPEXPR_NAME^ value)
     {
@@ -435,5 +435,11 @@ namespace MPIR
         mpz_set_f(_value, CTXT(0));
     }
 
+    void HugeRational::SetTo(FloatExpression^ value)
+    {
+        IN_CONTEXT(value);
+        mpq_set_f(_value, CTXT(0));
+    }
+
     #pragma endregion
 };

mpir.net/mpir.net/HugeRational.h

@@ -1162,6 +1162,13 @@ namespace MPIR
                 mpz_set(&_value->_mp_den, CTXTI(1));
             }
 
+            /// <summary>
+            /// Sets the value of the rational object.  There is no rounding, this conversion is exact.
+            /// <para>Do not change the value of an object while it is contained in a hash table, because that changes its hash code.
+            /// </para></summary>
+            /// <param name="value">new value for the object</param>
+            void SetTo(FloatExpression^ value);
+
             /// <summary>
             /// Swaps the values of two rationals.
             /// <para>This operation is a pointer swap and doesn't affect allocated memory.