bpo-39310: Add math.ulp(x)

Update documentation
Use math.ulp(x) in test_math
python · vstinner · vstinner · vstinner · Jan 12, 2020 · the-knights-who-say-ni
diff --git a/Modules/mathmodule.c b/Modules/mathmodule.c
+/*[clinic end generated code: output=f5207867a9384dd4 input=f1004db2d2436ab5]*/
+{
+    if (Py_IS_INFINITY(x) || Py_IS_NAN(x)) {
+        return x;
diff --git a/Lib/test/test_math.py b/Lib/test/test_math.py
+        self.assertEqual(math.ulp(0.0),
+                         sys.float_info.min * sys.float_info.epsilon)
+        self.assertEqual(math.ulp(1.0), sys.float_info.epsilon)
+        self.assertEqual(math.ulp(2.0 ** 52), 1.0)
diff --git a/Doc/library/math.rst b/Doc/library/math.rst
+.. function:: ulp(x)
+
+   Return the value of the least significant bit of a float *x*, such that the
+   first float bigger than *x* is ``x + ulp(x)``.  Then, given an expected
diff --git a/Doc/library/math.rst b/Doc/library/math.rst
+   Return the value of the least significant bit of a float *x*, such that the
+   first float bigger than *x* is ``x + ulp(x)``.  Then, given an expected
+   result *x* and a tolerance of ``n`` ulps, the result ``y`` should be such
+   that ``abs(y - x) <= n * ulp(x)``.
diff --git a/Modules/mathmodule.c b/Modules/mathmodule.c
+        return x - x2;
+    }
+    return x2 - x;
+}