* Beef-up tests for str.count().

rhettinger · rhettinger · commit 57e7447c4472 · 2005-02-20T09:54:53.000Z
* Speed-up str.count() by using memchr() to fly between first char matches.
diff --git a/Lib/test/string_tests.py b/Lib/test/string_tests.py
@@ -114,6 +114,33 @@ def test_count(self):
         self.checkraises(TypeError, 'hello', 'count')
         self.checkraises(TypeError, 'hello', 'count', 42)
 
+        # For a variety of combinations,
+        #    verify that str.count() matches an equivalent function
+        #    replacing all occurrences and then differencing the string lengths
+        charset = ['', 'a', 'b']
+        digits = 7
+        base = len(charset)
+        teststrings = set()
+        for i in xrange(base ** digits):
+            entry = []
+            for j in xrange(digits):
+                i, m = divmod(i, base)
+                entry.append(charset[m])
+            teststrings.add(''.join(entry))
+        teststrings = list(teststrings)
+        for i in teststrings:
+            i = self.fixtype(i)
+            n = len(i)
+            for j in teststrings:
+                r1 = i.count(j)
+                if j:
+                    r2, rem = divmod(n - len(i.replace(j, '')), len(j))
+                else:
+                    r2, rem = len(i)+1, 0
+                if rem or r1 != r2:
+                    self.assertEqual(rem, 0)
+                    self.assertEqual(r1, r2)
+
     def test_find(self):
         self.checkequal(0, 'abcdefghiabc', 'find', 'abc')
         self.checkequal(9, 'abcdefghiabc', 'find', 'abc', 1)
@@ -135,6 +162,7 @@ def test_find(self):
                 i, m = divmod(i, base)
                 entry.append(charset[m])
             teststrings.add(''.join(entry))
+        teststrings = list(teststrings)
         for i in teststrings:
             i = self.fixtype(i)
             for j in teststrings:
diff --git a/Objects/stringobject.c b/Objects/stringobject.c
@@ -2145,7 +2145,7 @@ interpreted as in slice notation.");
 static PyObject *
 string_count(PyStringObject *self, PyObject *args)
 {
-	const char *s = PyString_AS_STRING(self), *sub;
+	const char *s = PyString_AS_STRING(self), *sub, *t;
 	int len = PyString_GET_SIZE(self), n;
 	int i = 0, last = INT_MAX;
 	int m, r;
@@ -2186,11 +2186,16 @@ string_count(PyStringObject *self, PyObject *args)
 		} else {
 			i++;
 		}
+		if (i >= m)
+			break;
+		t = memchr(s+i, sub[0], m-i);
+		if (t == NULL)
+			break;
+		i = t - s;
 	}
 	return PyInt_FromLong((long) r);
 }
 
-
 PyDoc_STRVAR(swapcase__doc__,
 "S.swapcase() -> string\n\
 \n\
