Fix: add ``Text.unicode_normalization = 'NFC'`` as defaultadamg-fix-text

3 files changed, 26 insertions, 4 deletions

diff --git a/CHANGES.rst b/CHANGES.rst
index a117f0f..15539e2 100644
--- a/CHANGES.rst
+++ b/CHANGES.rst
@@ -2,10 +2,12 @@
  Changes
 =========
 
-5.1 (unreleased)
-================
+5.0.1 (unreleased)
+==================
 
-- Nothing changed yet.
+- Fix: add ``Text.unicode_normalization = 'NFC'`` as default, because some are
+  persisting schema fields. Setting that attribute only in ``__init__``
+  breaks loading old objects.
 
 
 5.0 (2020-03-06)
diff --git a/src/zope/schema/_bootstrapfields.py b/src/zope/schema/_bootstrapfields.py
index 7152b82..9a64a2d 100644
--- a/src/zope/schema/_bootstrapfields.py
+++ b/src/zope/schema/_bootstrapfields.py
@@ -505,9 +505,11 @@ class MinMaxLen(object):
 class Text(MinMaxLen, Field):
     """A field containing text used for human discourse."""
     _type = text_type
+    unicode_normalization = 'NFC'
 
     def __init__(self,  *args, **kw):
-        self.unicode_normalization = kw.pop('unicode_normalization', 'NFC')
+        self.unicode_normalization = kw.pop(
+            'unicode_normalization', self.unicode_normalization)
         super(Text, self).__init__(*args, **kw)
 
     def fromUnicode(self, value):
diff --git a/src/zope/schema/tests/test__bootstrapfields.py b/src/zope/schema/tests/test__bootstrapfields.py
index 34bdfc8..027944c 100644
--- a/src/zope/schema/tests/test__bootstrapfields.py
+++ b/src/zope/schema/tests/test__bootstrapfields.py
@@ -967,6 +967,7 @@ class TextTests(EqualityTestsMixin,
     def test_normalization(self):
         deadbeef = unicodedata.normalize('NFD', b'\xc3\x84\xc3\x96\xc3\x9c'.decode('utf-8'))
         txt = self._makeOne()
+        self.assertEqual(txt.unicode_normalization, 'NFC')
         self.assertEqual(
             [unicodedata.name(c) for c in txt.fromUnicode(deadbeef)],
             [
@@ -976,6 +977,7 @@ class TextTests(EqualityTestsMixin,
             ]
         )
         txt = self._makeOne(unicode_normalization=None)
+        self.assertEqual(txt.unicode_normalization, None)
         self.assertEqual(
             [unicodedata.name(c) for c in txt.fromUnicode(deadbeef)],
             [
@@ -988,6 +990,22 @@ class TextTests(EqualityTestsMixin,
             ]
         )
 
+    def test_normalization_after_pickle(self):
+        # test an edge case where `Text` is persisted
+        # see https://github.com/zopefoundation/zope.schema/issues/90
+        import pickle
+        orig_makeOne = self._makeOne
+        def makeOne(**kwargs):
+            result = orig_makeOne(**kwargs)
+            if not kwargs:
+                # We should have no state to preserve
+                result.__dict__.clear()
+
+            result = pickle.loads(pickle.dumps(result))
+            return result
+
+        self._makeOne = makeOne
+        self.test_normalization()
 
 
 class TextLineTests(EqualityTestsMixin,