Remove py2 support

.github/workflows/main.yml

@@ -11,7 +11,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python-version: ["3.8", "3.9", "3.10", "3.11", "3.12", "3.13"]
+        python-version: ["3.9", "3.10", "3.11", "3.12", "3.13"]
     steps:
       - uses: actions/checkout@v4
 
@@ -21,7 +21,7 @@ jobs:
           python-version: ${{ matrix.python-version }}
 
       - name: Run tests
-        run: uv run --with python-dateutil --with six test.py
+        run: uv run --with python-dateutil --with pytest pytest bson/tests test.py
 
   # https://github.com/marketplace/actions/alls-green#why used for branch protection checks
   check:

bson/codec.py

@@ -24,10 +24,6 @@
 from dateutil.tz import tzutc
 from binascii import b2a_hex
 
-from six import integer_types, iterkeys, text_type, PY3
-from six.moves import xrange
-
-
 utc = tzutc()
 
 class MissingClassDefinition(ValueError):
@@ -132,7 +128,7 @@ def encode_string(value):
 
 def encode_cstring(value):
     if not isinstance(value, bytes):
-        value = text_type(value).encode("utf-8")
+        value = str(value).encode("utf-8")
     if b"\x00" in value:
         raise ValueError("Element names may not include NUL bytes.")
         # A NUL byte is used to delimit our string, accepting one would cause
@@ -172,24 +168,26 @@ def encode_double_element(name, value):
 def encode_string_element(name, value):
     return b"\x02" + encode_cstring(name) + encode_string(value)
 
-
+# any need for this at all with py3??
 def _is_string(value):
-    if isinstance(value, text_type):
+    if isinstance(value, str):
         return True
-    elif isinstance(value, str) or isinstance(value, bytes):
-        try:
-            unicode(value, errors='strict')
-            return True
-        except:
-            pass
-    return False
+    # this never worked in py3
+    # elif isinstance(value, str) or isinstance(value, bytes):
+    #     try:
+    #         unicode(value, errors='strict')
+    #         return True
+    #     except:
+    #         pass
+    else:
+        return False
 
 
 def encode_value(name, value, buf, traversal_stack,
                  generator_func, on_unknown=None):
     if isinstance(value, bool):
         buf.write(encode_boolean_element(name, value))
-    elif isinstance(value, integer_types):
+    elif isinstance(value, int):
         if value < -0x80000000 or 0x7FFFFFFFFFFFFFFF >= value > 0x7fffffff:
             buf.write(encode_int64_element(name, value))
         elif value > 0x7FFFFFFFFFFFFFFF:
@@ -238,7 +236,7 @@ def encode_value(name, value, buf, traversal_stack,
 def encode_document(obj, traversal_stack, traversal_parent=None,
                     generator_func=None, on_unknown=None):
     buf = StringIO()
-    key_iter = iterkeys(obj)
+    key_iter = iter(obj)
     if generator_func is not None:
         key_iter = generator_func(obj, traversal_stack)
     for name in key_iter:
@@ -256,7 +254,7 @@ def encode_document(obj, traversal_stack, traversal_parent=None,
 def encode_array(array, traversal_stack, traversal_parent=None,
                  generator_func=None, on_unknown=None):
     buf = StringIO()
-    for i in xrange(0, len(array)):
+    for i in range(0, len(array)):
         value = array[i]
         traversal_stack.append(TraversalStep(traversal_parent or array, i))
         encode_value(str(i), value, buf, traversal_stack,
@@ -295,10 +293,7 @@ def decode_document(data, base, as_array=False):
 
         element_type = char_struct.unpack(data[base:base + 1])[0]
 
-        if PY3:
-            ll = data.index(0, base + 1) + 1
-        else:
-            ll = data.index("\x00", base + 1) + 1
+        ll = data.index(0, base + 1) + 1
         if decode_name:
             name = data[base + 1:ll - 1]
             try:

bson/network.py

@@ -1,7 +1,7 @@
 #!/usr/bin/env python
 from struct import unpack
 
-from six import BytesIO, b
+from io import BytesIO
 
 from . import dumps, loads
 
@@ -55,9 +55,10 @@ def recvbytes(self, bytes_needed, sock_buf = None):
         chunk = self.recv(min(bytes_needed - bytes_count, 32768))
         part_count = len(chunk)
 
-        if type(chunk) == str:
-            chunk = b(chunk)
-
+        if type(chunk) == str: # this probably never occurs ...
+            # chunk = b(chunk)
+            # six.b() encoded to latin-1
+            chunk = chunk.encode("latin-1")
         if part_count < 1:
             return None
 

bson/objectid.py

@@ -27,15 +27,14 @@
 import threading
 import time
 
-from bson.py3compat import PY3, bytes_from_hex, string_type, text_type
+# from bson.py3compat import bytes_from_hex #string_type
 from bson.tz_util import utc
 
 
 # fnv_1a_24 adaptation taken from MongoDB Python Driver at https://github.com/mongodb/mongo-python-driver/commit/61850357a0e0eeec1a30e1adc0bbf7ebee807358
-if PY3:
-    _ord = lambda x: x
-else:
-    _ord = ord
+# is this needed at all for Py3?
+_ord = lambda x: x
+
 # http://isthe.com/chongo/tech/comp/fnv/index.html#FNV-1a
 def _fnv_1a_24(data, _ord=_ord):
     """FNV-1a 24 bit hash"""
@@ -212,18 +211,17 @@ def __validate(self, oid):
         """
         if isinstance(oid, ObjectId):
             self.__id = oid.binary
-        # bytes or unicode in python 2, str in python 3
-        elif isinstance(oid, string_type):
+        elif isinstance(oid, str):
             if len(oid) == 24:
                 try:
-                    self.__id = bytes_from_hex(oid)
+                    self.__id = bytes.fromhex(oid)
                 except (TypeError, ValueError):
                     _raise_invalid_id(oid)
             else:
                 _raise_invalid_id(oid)
         else:
             raise TypeError("id must be an instance of (bytes, %s, ObjectId), "
-                            "not %s" % (text_type.__name__, type(oid)))
+                            "not %s" % ('str', type(oid)))
 
     @property
     def binary(self):
@@ -261,15 +259,13 @@ def __setstate__(self, value):
         # ObjectIds pickled in python 2.x used `str` for __id.
         # In python 3.x this has to be converted to `bytes`
         # by encoding latin-1.
-        if PY3 and isinstance(oid, text_type):
+        if isinstance(oid, str):
             self.__id = oid.encode('latin-1')
         else:
             self.__id = oid
 
     def __str__(self):
-        if PY3:
-            return binascii.hexlify(self.__id).decode()
-        return binascii.hexlify(self.__id)
+        return binascii.hexlify(self.__id).decode()
 
     def __repr__(self):
         return "ObjectId('%s')" % (str(self),)

bson/tests/test_array.py

@@ -2,7 +2,6 @@
 from unittest import TestCase
 
 from bson import dumps, loads
-from six import PY3
 
 
 class TestArray(TestCase):
@@ -74,5 +73,5 @@ def test_long_array(self):
 
     def test_encoded_order(self):
         serialized = dumps(self.doc)
-        expected = repr(serialized)[1:] if PY3 else repr(serialized)
+        expected = repr(serialized)[1:]
         self.assertEqual(expected, '\'\\xea\\x08\\x00\\x00\\x04lyrics\\x00\\xdd\\x08\\x00\\x00\\x020\\x00\\x14\\x00\\x00\\x00Viva La Vida lyrics\\x00\\x021\\x00\\x01\\x00\\x00\\x00\\x00\\x022\\x00!\\x00\\x00\\x00        I used to rule the world\\x00\\x023\\x00-\\x00\\x00\\x00        Seas would rise when I gave the word\\x00\\x024\\x00)\\x00\\x00\\x00        Now in the morning I sleep alone\\x00\\x025\\x00(\\x00\\x00\\x00        Sweep the streets I used to own\\x00\\x026\\x00\\x01\\x00\\x00\\x00\\x00\\x027\\x00 \\x00\\x00\\x00        I used to roll the dice\\x00\\x028\\x00)\\x00\\x00\\x00        Feel the fear in my enemy\\\'s eyes\\x00\\x029\\x00\\\'\\x00\\x00\\x00        Listen as the crowd would sing\\x00\\x0210\\x008\\x00\\x00\\x00        "Now the old king is dead! Long live the king!"\\x00\\x0211\\x00\\x01\\x00\\x00\\x00\\x00\\x0212\\x00"\\x00\\x00\\x00        One minute I held the key\\x00\\x0213\\x00)\\x00\\x00\\x00        Next the walls were closed on me\\x00\\x0214\\x00/\\x00\\x00\\x00        And I discovered that my castles stand\\x00\\x0215\\x001\\x00\\x00\\x00        Upon pillars of salt and pillars of sand\\x00\\x0216\\x00\\x01\\x00\\x00\\x00\\x00\\x0217\\x00)\\x00\\x00\\x00        I hear Jerusalem bells a ringing\\x00\\x0218\\x00)\\x00\\x00\\x00        Roman Cavalry choirs are singing\\x00\\x0219\\x00*\\x00\\x00\\x00        Be my mirror, my sword and shield\\x00\\x0220\\x00+\\x00\\x00\\x00        My missionaries in a foreign field\\x00\\x0221\\x00\\x01\\x00\\x00\\x00\\x00\\x0222\\x00(\\x00\\x00\\x00        For some reason I can\\\'t explain\\x00\\x0223\\x00$\\x00\\x00\\x00        Once you go there was never\\x00\\x0224\\x00\\x1d\\x00\\x00\\x00        Never an honest word\\x00\\x0225\\x00,\\x00\\x00\\x00        And that was when I ruled the world\\x00\\x0226\\x00\\x01\\x00\\x00\\x00\\x00\\x0227\\x00(\\x00\\x00\\x00        It was the wicked and wild wind\\x00\\x0228\\x00)\\x00\\x00\\x00        Blew down the doors to let me in\\x00\\x0229\\x001\\x00\\x00\\x00        Shattered windows and the sound of drums\\x00\\x0230\\x000\\x00\\x00\\x00        People couldn\\\'t believe what I\\\'d become\\x00\\x0231\\x00\\x01\\x00\\x00\\x00\\x00\\x0232\\x00\\x1d\\x00\\x00\\x00        Revolutionaries wait\\x00\\x0233\\x00&\\x00\\x00\\x00        For my head on a silver plate\\x00\\x0234\\x00)\\x00\\x00\\x00        Just a puppet on a lonely string\\x00\\x0235\\x00+\\x00\\x00\\x00        Oh who would ever want to be king?\\x00\\x0236\\x00\\x01\\x00\\x00\\x00\\x00\\x0237\\x00)\\x00\\x00\\x00        I hear Jerusalem bells a ringing\\x00\\x0238\\x00)\\x00\\x00\\x00        Roman Cavalry choirs are singing\\x00\\x0239\\x00*\\x00\\x00\\x00        Be my mirror, my sword and shield\\x00\\x0240\\x00+\\x00\\x00\\x00        My missionaries in a foreign field\\x00\\x0241\\x00\\x01\\x00\\x00\\x00\\x00\\x0242\\x00(\\x00\\x00\\x00        For some reason I can\\\'t explain\\x00\\x0243\\x00.\\x00\\x00\\x00        I know Saint Peter won\\\'t call my name\\x00\\x0244\\x00\\x1d\\x00\\x00\\x00        Never an honest word\\x00\\x0245\\x00,\\x00\\x00\\x00        But that was when I ruled the world\\x00\\x0246\\x00\\x01\\x00\\x00\\x00\\x00\\x0247\\x00)\\x00\\x00\\x00        I hear Jerusalem bells a ringing\\x00\\x0248\\x00)\\x00\\x00\\x00        Roman Cavalry choirs are singing\\x00\\x0249\\x00*\\x00\\x00\\x00        Be my mirror, my sword and shield\\x00\\x0250\\x00+\\x00\\x00\\x00        My missionaries in a foreign field\\x00\\x0251\\x00\\x01\\x00\\x00\\x00\\x00\\x0252\\x00(\\x00\\x00\\x00        For some reason I can\\\'t explain\\x00\\x0253\\x00.\\x00\\x00\\x00        I know Saint Peter won\\\'t call my name\\x00\\x0254\\x00\\x1d\\x00\\x00\\x00        Never an honest word\\x00\\x0255\\x00,\\x00\\x00\\x00        But that was when I ruled the world\\x00\\x00\\x00\'')

bson/tests/test_object.py

@@ -3,8 +3,8 @@
 
 from bson import BSONCoding, dumps, loads, import_class
 
-
-class TestData(BSONCoding):
+# named with underscore so as not to confuse pytest
+class _TestData(BSONCoding):
     def __init__(self, *args):
         self.args = list(args)
         self.nested = None
@@ -17,7 +17,7 @@ def bson_init(self, raw_values):
         self.nested = raw_values["nested"]
 
     def __eq__(self, other):
-        if not isinstance(other, TestData):
+        if not isinstance(other, _TestData):
             return NotImplemented
         if self.args != other.args:
             return False
@@ -31,12 +31,12 @@ def __ne__(self, other):
 
 class TestObjectCoding(TestCase):
     def test_codec(self):
-        import_class(TestData)
-        data = TestData(u"Lorem ipsum dolor sit amet",
+        import_class(_TestData)
+        data = _TestData(u"Lorem ipsum dolor sit amet",
                         "consectetur adipisicing elit",
                         42)
 
-        data2 = TestData(u"She's got both hands in her pockets",
+        data2 = _TestData(u"She's got both hands in her pockets",
                          "and she won't look at you won't look at you eh",
                          66,
                          23.54,

bson/tests/test_objectid.py

@@ -25,7 +25,6 @@
 
 from bson.objectid import ObjectId, _fnv_1a_24
 from bson.objectid import InvalidId
-from bson.py3compat import PY3, _unicode
 from bson.tz_util import (FixedOffset,
                           utc)
 
@@ -467,7 +466,7 @@ def test_fnv_1a_24(self):
 
     def test_unicode(self):
         a = ObjectId()
-        self.assertEqual(a, ObjectId(_unicode(a)))
+        self.assertEqual(a, ObjectId(a))
         self.assertEqual(ObjectId("123456789012123456789012"),
                          ObjectId(u"123456789012123456789012"))
         self.assertRaises(InvalidId, ObjectId, u"hello")
@@ -511,18 +510,19 @@ def test_pid(self):
         self.assertTrue(oid_generated_on_client(ObjectId()))
 
     def test_generation_time(self):
-        d1 = datetime.datetime.utcnow()
+        d1 = datetime.datetime.now(utc)
         d2 = ObjectId().generation_time
 
         self.assertEqual(utc, d2.tzinfo)
-        d2 = d2.replace(tzinfo=None)
         self.assertTrue(d2 - d1 < datetime.timedelta(seconds=2))
 
     def test_from_datetime(self):
         if 'PyPy 1.8.0' in sys.version:
             # See https://bugs.pypy.org/issue1092
             raise SkipTest("datetime.timedelta is broken in pypy 1.8.0")
-        d = datetime.datetime.utcnow()
+        # note: could use tz aware datetimes here?
+        d = datetime.datetime.now(utc).replace(tzinfo=None)
+
         d = d - datetime.timedelta(microseconds=d.microsecond)
         oid = ObjectId.from_datetime(d)
         self.assertEqual(d, oid.generation_time.replace(tzinfo=None))
@@ -559,13 +559,9 @@ def test_pickle_backwards_compatability(self):
             b"object\np2\nNtp3\nRp4\n"
             b"S'M\\x9afV\\x13v\\xc0\\x0b\\x88\\x00\\x00\\x00'\np5\nb.")
 
-        if PY3:
-            # Have to load using 'latin-1' since these were pickled in python2.x.
-            oid_1_9 = pickle.loads(pickled_with_1_9, encoding='latin-1')
-            oid_1_10 = pickle.loads(pickled_with_1_10, encoding='latin-1')
-        else:
-            oid_1_9 = pickle.loads(pickled_with_1_9)
-            oid_1_10 = pickle.loads(pickled_with_1_10)
+        # Have to load using 'latin-1' since these were pickled in python2.x.
+        oid_1_9 = pickle.loads(pickled_with_1_9, encoding='latin-1')
+        oid_1_10 = pickle.loads(pickled_with_1_10, encoding='latin-1')
 
         self.assertEqual(oid_1_9, ObjectId("4d9a66561376c00b88000000"))
         self.assertEqual(oid_1_9, oid_1_10)