fixup! fixup! fixup! fixup! XMLExportImport: more support pickle protocol 3 🚧 ( repair python2 )

bt5/erp5_core_test/TestTemplateItem/portal_components/test.erp5.testXMLPickle.py

 ##############################################################################
-#
+# coding: utf-8
 # Copyright (c) 2008 Nexedi SA and Contributors. All Rights Reserved.
 #                    TAHARA Yusei <yusei@nexedi.com>
 #
@@ -26,16 +26,17 @@
 #
 ##############################################################################
 
+import base64
 import unittest
 import zodbpickle
 import zodbpickle.fastpickle as pickle
 import re
-import xml.sax
-from io import BytesIO, StringIO
-
-from Products.ERP5Type.XMLExportImport import ppml
+from io import BytesIO
+from six import StringIO
+from Products.ERP5Type.XMLExportImport import importXML, ppml
 import six
 
+
 class DummyClass:
   """
   A dummy data class
@@ -50,26 +51,37 @@ class XMLPickleTestCase(unittest.TestCase):
   def dump_to_xml(self, obj):
     pickled_string = pickle.dumps(obj, protocol=self._pickle_protocol)
     f = BytesIO(pickled_string)
-    return str(ppml.ToXMLUnpickler(f).load())
-
-  def load_from_xml(self, xml_string):
-    output = StringIO()
-    F=ppml.xmlPickler()
-    F.file = output
-    F.binary = 1 # XML pickle actually only supports the case of binary = 1
-
-    content_handler = xml.sax.handler.ContentHandler()
-    content_handler.startElement = F.unknown_starttag
-    content_handler.endElement = F.unknown_endtag
-    content_handler.characters = F.handle_data
-    xml.sax.parseString(xml_string, content_handler)
-
-    reconstructed_pickled_data = F._stack[0][0]
-    return pickle.loads(reconstructed_pickled_data)
+    xml = ppml.ToXMLUnpickler(f).load().__str__()
+    if six.PY2 and isinstance(xml, six.text_type):
+      xml = xml.encode('utf-8')
+    return xml
+
+  def load_from_xml(self, xml_string, persistent_load=None):
+    assertEqual = self.assertEqual
+    class DummyJar:
+      loaded = None
+      """follow interface expected by importXML"""
+      def importFile(self, file, clue):
+        assertEqual(clue, 'ignored')
+        assertEqual(file.read(4), b'ZEXP')
+        unpickler = pickle.Unpickler(file)
+        if persistent_load:
+          unpickler.persistent_load = persistent_load
+        self.loaded = unpickler.load()
+
+    jar = DummyJar()
+    xml_string = '<?xml version="1.0"?>\n<ZopeData>%s</ZopeData>' % xml_string
+    importXML(jar, StringIO(xml_string), clue='ignored')
+    return jar.loaded
 
   def dump_and_load(self, obj):
     return self.load_from_xml(self.dump_to_xml(obj))
 
+  def check_and_load(self, v):
+    reconstructed = self.dump_and_load(v)
+    self.assertEqual(reconstructed, v)
+    self.assertIs(type(reconstructed), type(v))
+
 
 class TestXMLPickle(XMLPickleTestCase):
 
@@ -101,76 +113,95 @@ class TestXMLPickle(XMLPickleTestCase):
     self.assertIs(self.dump_and_load(False), False)
 
   def test_int(self):
-    def check_int(v):
-      reconstructed = self.dump_and_load(v)
-      self.assertEqual(reconstructed, v)
-      self.assertIs(type(reconstructed), int)
-    check_int(-0)
-    check_int(1)
-    check_int(-1)
-    check_int(0xff)
-    check_int(0xff1)
-    check_int(0xffff)
-    check_int(0xffff1)
+    self.check_and_load(-0)
+    self.check_and_load(1)
+    self.check_and_load(-1)
+    self.check_and_load(0xff)
+    self.check_and_load(0xff1)
+    self.check_and_load(0xffff)
+    self.check_and_load(2**128)
+    # long4
+    # https://github.com/python/cpython/blob/4d4a6f1b/Lib/test/pickletester.py#L2049-L2050
+    self.check_and_load(12345678910111213141516178920 << (256*8))
+
+
+  if six.PY2:
+    def test_long(self):
+      self.check_and_load(long(-0))
+      self.check_and_load(long(1))
+      self.check_and_load(long(-1))
+      self.check_and_load(long(0xff))
+      self.check_and_load(long(0xff1))
+      self.check_and_load(long(0xffff))
+      self.check_and_load(long(2**128))
+      self.check_and_load(12345678910111213141516178920 << (256*8))
 
   def test_float(self):
-    def check_float(v):
-      reconstructed = self.dump_and_load(v)
-      self.assertEqual(reconstructed, v)
-      self.assertIs(type(reconstructed), float)
-    check_float(-0.0)
-    check_float(1.0)
-    check_float(-1.0)
-    check_float(.1 + .2)
+    self.check_and_load(-0.0)
+    self.check_and_load(1.0)
+    self.check_and_load(-1.0)
+    self.check_and_load(.33)
 
   def test_None(self):
     self.assertIs(
       self.dump_and_load(None), None)
 
   def test_bytes(self):
-    self.assertEqual(self.dump_and_load(b"bytes"), b"bytes")
-    self.assertEqual(self.dump_and_load(b"long bytes" * 100), b"long bytes" * 100)
-    self.assertEqual(
-      self.dump_and_load(zodbpickle.binary(b"bytes")),
-      zodbpickle.binary(b"bytes"))
-    self.assertIs(type(self.dump_and_load(zodbpickle.binary(b"bytes"))), zodbpickle.binary)
+    self.check_and_load(b"bytes")
+    self.check_and_load(b"long bytes" * 100)
+    self.check_and_load(zodbpickle.binary(b"bytes"))
+    self.check_and_load(zodbpickle.binary(b""))
 
-  def test_unicode(self):
+  def test_unicode(self):  # BBB PY2
     self.assertIs(type(self.dump_and_load(u"OK")), six.text_type)
-    self.assertEqual(self.dump_and_load(u"short"), u"short")
-    self.assertEqual(self.dump_and_load(u"unicode 👍"), u"unicode 👍")
-    self.assertEqual(self.dump_and_load(u"long" * 100), u"long" * 100)
-    self.assertEqual(self.dump_and_load(u"long…" * 100), u"long…" * 100)
-    self.assertEqual(self.dump_and_load(u">"), u">")
-    self.assertEqual(self.dump_and_load(u"a\nb"), u"a\nb")
+    self.check_and_load(u"short")
+    self.check_and_load(u"unicode 👍")
+    self.check_and_load(u"long" * 100)
+    self.check_and_load(u"long…" * 100)
+    self.check_and_load(u">")
+    self.check_and_load(u"a\nb")
+    self.check_and_load(u" with spaces ")
+    self.check_and_load(u"\twith\ttabs\t")
+    self.check_and_load(u"")
+
+  def test_str(self):
+    self.assertIs(type(self.dump_and_load("OK")), str)
+    self.check_and_load("short")
+    self.check_and_load("unicode 👍")
+    self.check_and_load("long" * 100)
+    self.check_and_load("long…" * 100)
+    self.check_and_load(">")
+    self.check_and_load("a\nb")
+    self.check_and_load(" with spaces ")
+    self.check_and_load("\twith\ttabs\t")
+    self.check_and_load("")
 
   def test_dict(self):
-    self.assertEqual(
-      self.dump_and_load({'a': 1, 'b': 2}), {'a': 1, 'b': 2})
+    self.check_and_load({'a': 1, 'b': 2})
+    self.check_and_load({'hé': 'ho'})
+    self.check_and_load(dict.fromkeys(range(3000)))
 
   def test_tuple(self):
-    self.assertEqual(
-      self.dump_and_load((1, )), (1, ))
-    self.assertEqual(
-      self.dump_and_load((1, 'two')), (1, 'two'))
-    self.assertEqual(
-      self.dump_and_load((1, 'two', 3.0)), (1, 'two', 3.0))
-    self.assertEqual(
-      self.dump_and_load(tuple([1] * 1000)), tuple([1] * 1000))
-    self.assertEqual(
-      self.dump_and_load(()), ())
+    self.check_and_load((1, ))
+    self.check_and_load((1, 'two'))
+    self.check_and_load((1, 'two', 3.0))
+    self.check_and_load(tuple([1] * 1000))
+    self.check_and_load(())
+    self.check_and_load(('hé',))
+    self.check_and_load(('hé', 'hé'))
+    self.check_and_load(('hé', 'hé', 'hé'))
+    self.check_and_load(('hé', 'hé', 'hé', 'hé'))
 
   def test_list(self):
-    self.assertEqual(
-      self.dump_and_load([1]), [1])
-    self.assertEqual(
-      self.dump_and_load([]), [])
-    self.assertEqual(
-      self.dump_and_load([1] * 1000), [1] * 1000)
+    self.check_and_load([1])
+    self.check_and_load([])
+    self.check_and_load([1] * 1000)
+    self.check_and_load(['hé'])
 
   def test_set(self):
-    self.assertEqual(
-      self.dump_and_load(set('abc')), set('abc'))
+    self.check_and_load(set('abc'))
+    self.check_and_load(set('hé'))
+    self.check_and_load(set([]))
 
   def test_reference(self):
     ref = []
@@ -178,6 +209,14 @@ class TestXMLPickle(XMLPickleTestCase):
     self.assertEqual(reconstructed, [ref, ref, ref])
     self.assertIs(reconstructed[0], reconstructed[1])
 
+  def test_reference_long(self):
+    # same as reference (which is using BINPUT/BINGET but with large data
+    # to use LONG_BINPUT/LONG_BINGET)
+    ref = [list() for _ in range(256)]
+    reconstructed = self.dump_and_load([ref, ref, ref])
+    self.assertEqual(reconstructed, [ref, ref, ref])
+    self.assertIs(reconstructed[0], reconstructed[1])
+
 
 class TestXMLPickleStringEncoding(XMLPickleTestCase):
   def test_string_base64(self):
@@ -214,19 +253,39 @@ class TestXMLPickleStringEncoding(XMLPickleTestCase):
 class TestXMLPickleStringHeuristics(XMLPickleTestCase):
   """Heuristics to map python2 str to unicode or bytes in business templates.
   """
-  def test_oid_base64(self):
-    # if it looks like an oid, it's bytes
-    self.assertEqual(
-      self.load_from_xml("""
-        <pickle><string encoding="base64">AAAAAAAAAAE=</string></pickle>
-      """),
-      b"\x00\x00\x00\x00\x00\x00\x00\x01")
 
   def test_bytes_base64(self):
-    # if it does not decode as utf-8 it's bytes
+    # if it does not decode as utf-8, it's bytes
     self.assertEqual(
       self.load_from_xml("""
         <pickle><string encoding="base64">/wA=</string></pickle>
       """),
       b"\xFF\x00")
 
+  def test_long_bytes_base64(self):
+    # if it does not decode as utf-8, it's bytes
+    long_bytes = b"\xFF\x00" * 256
+    self.assertEqual(
+      self.load_from_xml("""
+        <pickle><string encoding="base64">%s</string></pickle>
+      """ % base64.b64encode(long_bytes).decode()),
+      long_bytes)
+
+  def test_string_persistent_id_base64(self):
+    # persistent ids are loaded as bytes
+    persistent_ids = []
+    def persistent_load(oid):
+      persistent_ids.append(oid)
+    self.assertEqual(
+      self.load_from_xml("""
+      <pickle>
+        <persistent>
+          <string encoding="base64">AAAAAAAAAAE=</string>
+        </persistent>
+      </pickle>
+      """,
+      persistent_load=persistent_load),
+      None)
+    self.assertEqual(
+      persistent_ids,
+      [b'\x00\x00\x00\x00\x00\x00\x00\x01'])

product/ERP5Type/XMLExportImport/__init__.py

@@ -363,12 +363,6 @@ def exportXML(jar, oid, file=None):
         p = getReorderedPickle(oid)
         write(XMLrecord(oid, len(p), p, id_mapping))
     write('</ZopeData>\n')
-    if 0:
-      try:
-        print(file.getvalue())
-      except AttributeError:
-        pass
-      import pdb; pdb.set_trace()
     return file
 
 class zopedata:
@@ -421,7 +415,6 @@ def importXML(jar, file, clue=''):
         F.end_handlers['record'] = save_record
         F.end_handlers['ZopeData'] = save_zopedata
         F.start_handlers['ZopeData'] = start_zopedata
-        F.binary=1
         F.file=outfile
         # <patch>
         # Our BTs XML files don't declare encoding but have accented chars in them