Fixing issue86 by normalizing unicode for IFromUnicode

2 files changed, 33 insertions, 2 deletions

diff --git a/src/zope/schema/_bootstrapfields.py b/src/zope/schema/_bootstrapfields.py
index e140090..0aec192 100644
--- a/src/zope/schema/_bootstrapfields.py
+++ b/src/zope/schema/_bootstrapfields.py
@@ -20,6 +20,7 @@ import fractions
 import numbers
 import sys
 import threading
+import unicodedata
 from math import isinf
 
 from zope.interface import Attribute
@@ -505,7 +506,8 @@ class Text(MinMaxLen, Field):
     """A field containing text used for human discourse."""
     _type = text_type
 
-    def __init__(self, *args, **kw):
+    def __init__(self, suppress_unicode_normalization=_NotGiven, *args, **kw):
+        self.suppress_unicode_normalization = suppress_unicode_normalization
         super(Text, self).__init__(*args, **kw)
 
     def fromUnicode(self, str):
@@ -528,8 +530,12 @@ class Text(MinMaxLen, Field):
         Traceback (most recent call last):
         ...
         zope.schema._bootstrapinterfaces.ConstraintNotSatisfied: (u'foo spam', '')
+        >>> [ unicodedata.name(c) for c in t.fromUnicode(unicodedata.normalize('NFD', 'ÄÖÜ'))]
+        ['LATIN CAPITAL LETTER A WITH DIAERESIS', 'LATIN CAPITAL LETTER O WITH DIAERESIS', 'LATIN CAPITAL LETTER U WITH DIAERESIS']
         """
         self.validate(str)
+        if self.suppress_unicode_normalization is _NotGiven:
+            str = unicodedata.normalize('NFC', str)
         return str
 
 
diff --git a/src/zope/schema/tests/test__bootstrapfields.py b/src/zope/schema/tests/test__bootstrapfields.py
index 8aee806..1293026 100644
--- a/src/zope/schema/tests/test__bootstrapfields.py
+++ b/src/zope/schema/tests/test__bootstrapfields.py
@@ -13,6 +13,7 @@
 ##############################################################################
 import doctest
 import unittest
+import unicodedata
 
 # pylint:disable=protected-access,inherit-non-class,blacklisted-name
 
@@ -958,11 +959,35 @@ class TextTests(EqualityTestsMixin,
         self.assertRaisesWrongType(txt.fromUnicode, txt._type, deadbeef)
 
     def test_fromUnicode_hit(self):
-
         deadbeef = u'DEADBEEF'
         txt = self._makeOne()
         self.assertEqual(txt.fromUnicode(deadbeef), deadbeef)
 
+    def test_normalization(self):
+        deadbeef = unicodedata.normalize('NFD', 'ÄÖÜ')
+        txt = self._makeOne()
+        self.assertEqual(
+            [unicodedata.name(c) for c in txt.fromUnicode(deadbeef)],
+            [
+                'LATIN CAPITAL LETTER A WITH DIAERESIS',
+                'LATIN CAPITAL LETTER O WITH DIAERESIS',
+                'LATIN CAPITAL LETTER U WITH DIAERESIS',
+            ]
+        )
+        txt = self._makeOne(suppress_unicode_normalization=True)
+        self.assertEqual(
+            [unicodedata.name(c) for c in txt.fromUnicode(deadbeef)],
+            [
+                'LATIN CAPITAL LETTER A',
+                'COMBINING DIAERESIS',
+                'LATIN CAPITAL LETTER O',
+                'COMBINING DIAERESIS',
+                'LATIN CAPITAL LETTER U',
+                'COMBINING DIAERESIS',
+            ]
+        )
+
+
 
 class TextLineTests(EqualityTestsMixin,
                     WrongTypeTestsMixin,