Issue python#5828 (Invalid behavior of unicode.lower): Fixed bogus logic in

doerwalter · doerwalter · commit 5d98ec76bb54 · 2009-04-25T14:03:16.000Z
makeunicodedata.py and regenerated the Unicode database (This fixes
u'\u1d79'.lower() == '\x00').
diff --git a/Lib/test/test_unicodedata.py b/Lib/test/test_unicodedata.py
@@ -20,7 +20,7 @@
 class UnicodeMethodsTest(unittest.TestCase):
 
     # update this, if the database changes
-    expectedchecksum = 'aef99984a58c8e1e5363a3175f2ff9608599a93e'
+    expectedchecksum = 'b7db9b5f1d804976fa921d2009cbef6f025620c1'
 
     def test_method_checksum(self):
         h = hashlib.sha1()
@@ -257,6 +257,19 @@ def test_ucd_510(self):
         # the upper-case mapping: as delta, or as absolute value
         self.assert_(u"a".upper()==u'A')
         self.assert_(u"\u1d79".upper()==u'\ua77d')
+        self.assert_(u".".upper()==u".")
+
+    def test_bug_5828(self):
+        self.assertEqual(u"\u1d79".lower(), u"\u1d79")
+        # Only U+0000 should have U+0000 as its upper/lower/titlecase variant
+        self.assertEqual(
+            [
+                c for c in range(sys.maxunicode+1)
+                if u"\x00" in unichr(c).lower()+unichr(c).upper()+unichr(c).title()
+            ],
+            [0]
+        )
+
 
 def test_main():
     test.test_support.run_unittest(
diff --git a/Misc/NEWS b/Misc/NEWS
@@ -773,6 +773,10 @@ Library
 - Issue #2703: SimpleXMLRPCDispatcher.__init__: Provide default values for
   new arguments introduced in 2.5.
 
+- Issue #5828 (Invalid behavior of unicode.lower): Fixed bogus logic in
+  makeunicodedata.py and regenerated the Unicode database (This fixes
+  u'\u1d79'.lower() == '\x00').
+
 Tools/Demos
 -----------
 
diff --git a/Objects/unicodetype_db.h b/Objects/unicodetype_db.h
@@ -118,7 +118,7 @@ const _PyUnicode_TypeRecord _PyUnicode_TypeRecords[] = {
     {0, 0, 0, 0, 7, 4},
     {0, 0, 0, 0, 8, 4},
     {0, 0, 0, 0, 9, 4},
-    {42877, 0, 42877, 0, 0, 265},
+    {42877, 7545, 42877, 0, 0, 265},
     {3814, 0, 3814, 0, 0, 9},
     {65477, 0, 65477, 0, 0, 9},
     {0, 57921, 0, 0, 0, 129},
@@ -159,7 +159,7 @@ const _PyUnicode_TypeRecord _PyUnicode_TypeRecords[] = {
     {0, 54787, 0, 0, 0, 129},
     {0, 54753, 0, 0, 0, 129},
     {58272, 0, 58272, 0, 0, 9},
-    {0, 7545, 0, 0, 0, 385},
+    {42877, 7545, 42877, 0, 0, 385},
     {0, 40, 0, 0, 0, 129},
     {65496, 0, 65496, 0, 0, 9},
 };
diff --git a/Tools/unicode/makeunicodedata.py b/Tools/unicode/makeunicodedata.py
@@ -371,33 +371,32 @@ def makeunicodetype(unicode, trace):
                 flags |= UPPER_MASK
             # use delta predictor for upper/lower/title if it fits
             if record[12]:
-                upper = int(record[12], 16) - char
-                if  -32768 <= upper <= 32767 and delta:
-                    upper = upper & 0xffff
-                else:
-                    upper += char
-                    delta = False
+                upper = int(record[12], 16)
             else:
-                upper = 0
+                upper = char
             if record[13]:
-                lower = int(record[13], 16) - char
-                if -32768 <= lower <= 32767 and delta:
-                    lower = lower & 0xffff
-                else:
-                    lower += char
-                    delta = False
+                lower = int(record[13], 16)
             else:
-                lower = 0
+                lower = char
             if record[14]:
-                title = int(record[14], 16) - char
-                if -32768 <= lower <= 32767 and delta:
-                    title = title & 0xffff
-                else:
-                    title += char
-                    delta = False
+                title = int(record[14], 16)
+            else:
+                # UCD.html says that a missing title char means that
+                # it defaults to the uppercase character, not to the
+                # character itself. Apparently, in the current UCD (5.x)
+                # this feature is never used
+                title = upper
+            upper_d = upper - char
+            lower_d = lower - char
+            title_d = title - char
+            if -32768 <= upper_d <= 32767 and \
+               -32768 <= lower_d <= 32767 and \
+               -32768 <= title_d <= 32767:
+                # use deltas
+                upper = upper_d & 0xffff
+                lower = lower_d & 0xffff
+                title = title_d & 0xffff
             else:
-                title = 0
-            if not delta:
                 flags |= NODELTA_MASK
             # decimal digit, integer digit
             decimal = 0
