Removing UTF-16 aware Unicode comparison code. This kind of compare

function (together with other locale aware ones) should into a new collation
support module. See python-dev for a discussion of this removal.

Note: This patch should also be applied to the 1.6 branch.

Lib/test/output/test_unicode

@@ -1,6 +1,5 @@
 test_unicode
 Testing Unicode comparisons... done.
-Testing UTF-16 code point order comparisons... done.
 Testing Unicode contains method... done.
 Testing Unicode formatting strings... done.
 Testing builtin codecs... done.

Lib/test/test_unicode.py

@@ -168,56 +168,59 @@ assert 'abc' < u'abcd'
 assert u'abc' < u'abcd'
 print 'done.'
 
-print 'Testing UTF-16 code point order comparisons...',
+if 0:
-#No surrogates, no fixup required.
+    # Move these tests to a Unicode collation module test...
-assert u'\u0061' < u'\u20ac'
-# Non surrogate below surrogate value, no fixup required
-assert u'\u0061' < u'\ud800\udc02'
 
-# Non surrogate above surrogate value, fixup required
+    print 'Testing UTF-16 code point order comparisons...',
-def test_lecmp(s, s2):
+    #No surrogates, no fixup required.
-  assert s <  s2 , "comparison failed on %s < %s" % (s, s2)
+    assert u'\u0061' < u'\u20ac'
-  
+    # Non surrogate below surrogate value, no fixup required
-def test_fixup(s):
+    assert u'\u0061' < u'\ud800\udc02'
-  s2 = u'\ud800\udc01'
-  test_lecmp(s, s2)
-  s2 = u'\ud900\udc01'
-  test_lecmp(s, s2)
-  s2 = u'\uda00\udc01'
-  test_lecmp(s, s2)
-  s2 = u'\udb00\udc01'
-  test_lecmp(s, s2)
-  s2 = u'\ud800\udd01'
-  test_lecmp(s, s2)
-  s2 = u'\ud900\udd01'
-  test_lecmp(s, s2)
-  s2 = u'\uda00\udd01'
-  test_lecmp(s, s2)
-  s2 = u'\udb00\udd01'
-  test_lecmp(s, s2)
-  s2 = u'\ud800\ude01'
-  test_lecmp(s, s2)
-  s2 = u'\ud900\ude01'
-  test_lecmp(s, s2)
-  s2 = u'\uda00\ude01'
-  test_lecmp(s, s2)
-  s2 = u'\udb00\ude01'
-  test_lecmp(s, s2)
-  s2 = u'\ud800\udfff'
-  test_lecmp(s, s2)
-  s2 = u'\ud900\udfff'
-  test_lecmp(s, s2)
-  s2 = u'\uda00\udfff'
-  test_lecmp(s, s2)
-  s2 = u'\udb00\udfff'
-  test_lecmp(s, s2)
 
-test_fixup(u'\ue000')
+    # Non surrogate above surrogate value, fixup required
-test_fixup(u'\uff61')
+    def test_lecmp(s, s2):
+      assert s <  s2 , "comparison failed on %s < %s" % (s, s2)
 
-# Surrogates on both sides, no fixup required
+    def test_fixup(s):
-assert u'\ud800\udc02' < u'\ud84d\udc56'
+      s2 = u'\ud800\udc01'
-print 'done.'
+      test_lecmp(s, s2)
+      s2 = u'\ud900\udc01'
+      test_lecmp(s, s2)
+      s2 = u'\uda00\udc01'
+      test_lecmp(s, s2)
+      s2 = u'\udb00\udc01'
+      test_lecmp(s, s2)
+      s2 = u'\ud800\udd01'
+      test_lecmp(s, s2)
+      s2 = u'\ud900\udd01'
+      test_lecmp(s, s2)
+      s2 = u'\uda00\udd01'
+      test_lecmp(s, s2)
+      s2 = u'\udb00\udd01'
+      test_lecmp(s, s2)
+      s2 = u'\ud800\ude01'
+      test_lecmp(s, s2)
+      s2 = u'\ud900\ude01'
+      test_lecmp(s, s2)
+      s2 = u'\uda00\ude01'
+      test_lecmp(s, s2)
+      s2 = u'\udb00\ude01'
+      test_lecmp(s, s2)
+      s2 = u'\ud800\udfff'
+      test_lecmp(s, s2)
+      s2 = u'\ud900\udfff'
+      test_lecmp(s, s2)
+      s2 = u'\uda00\udfff'
+      test_lecmp(s, s2)
+      s2 = u'\udb00\udfff'
+      test_lecmp(s, s2)
+
+    test_fixup(u'\ue000')
+    test_fixup(u'\uff61')
+
+    # Surrogates on both sides, no fixup required
+    assert u'\ud800\udc02' < u'\ud84d\udc56'
+    print 'done.'
 
 test('ljust', u'abc',  u'abc       ', 10)
 test('rjust', u'abc',  u'       abc', 10)

Objects/unicodeobject.c

@@ -3169,6 +3169,12 @@ unicode_center(PyUnicodeObject *self, PyObject *args)
     return (PyObject*) pad(self, left, marg - left, ' ');
 }
 
+#if 0
+
+/* This code should go into some future Unicode collation support
+   module. The basic comparison should compare ordinals on a naive
+   basis (this is what Java does and thus JPython too).
+
 /* speedy UTF-16 code point order comparison */
 /* gleaned from: */
 /* http://www-4.ibm.com/software/developer/library/utf16.html?dwzone=unicode */
@@ -3213,6 +3219,33 @@ unicode_compare(PyUnicodeObject *str1, PyUnicodeObject *str2)
     return (len1 < len2) ? -1 : (len1 != len2);
 }
 
+#else
+
+static int
+unicode_compare(PyUnicodeObject *str1, PyUnicodeObject *str2)
+{
+    register int len1, len2;
+
+    Py_UNICODE *s1 = str1->str;
+    Py_UNICODE *s2 = str2->str;
+
+    len1 = str1->length;
+    len2 = str2->length;
+    
+    while (len1 > 0 && len2 > 0) {
+	register long diff;
+
+        diff = (long)*s1++ - (long)*s2++;
+        if (diff)
+            return (diff < 0) ? -1 : (diff != 0);
+        len1--; len2--;
+    }
+
+    return (len1 < len2) ? -1 : (len1 != len2);
+}
+
+#endif
+
 int PyUnicode_Compare(PyObject *left,
 		      PyObject *right)
 {