Pickle in python2 then unpickle in python3 results in TypeError: an integer is required (got type str)

[Ronnie]

When pickling a datetime with tzinfo in python2 and later unpickle it in python3 it results in an TypeError: an integer is required (got type str)

== Versions ==
python2 -V
Python 2.7.12

python3 -V
Python 3.5.2

python2 -c "import pytz
print(pytz.__version__)"
2018.9

python3 -c "import pytz
print(pytz.__version__)"
2018.9

== Create the pickle and verify in python 2 ==
python2
>>> import pickle, pytz
>>> from datetime import datetime
>>> d = pickle.dumps(datetime.utcnow().replace(tzinfo=pytz.utc), protocol=pickle.HIGHEST_PROTOCOL)
>>> with open('data.pickle', 'w+') as f:
        f.write(d)

python2
>>> import pickle
>>> print(pickle.load(open('data.pickle', 'rb')))
2019-03-03 22:46:07.651744+00:00

python2
>>> import pickletools
>>> pickletools.dis(open('data.pickle', 'rb').read())
    0: \x80 PROTO 2
    2: c GLOBAL 'datetime datetime'
   21: q BINPUT 0
   23: U SHORT_BINSTRING '\x07\xe3\x03\x03\x16.\x07\t\xf1\xe0'
   35: q BINPUT 1
   37: c GLOBAL 'pytz _UTC'
   48: q BINPUT 2
   50: ) EMPTY_TUPLE
   51: R REDUCE
   52: q BINPUT 3
   54: \x86 TUPLE2
   55: q BINPUT 4
   57: R REDUCE
   58: q BINPUT 5
   60: . STOP
highest protocol among opcodes = 2

== Type to open the pickle in python 3 ==
python3
>>> import pickle
>>> print(pickle.load(open('data.pickle', 'rb'), encoding='latin-1'))
Traceback (most recent call last):
  File "<string>", line 2, in <module>
TypeError: an integer is required (got type str)

== There is also a small difference in the SHORT_BINSTRING: ã ==
python3
>>> import pickletools
>>> pickletools.dis(open('data.pickle', 'rb').read())
    0: \x80 PROTO 2
    2: c GLOBAL 'datetime datetime'
   21: q BINPUT 0
   23: U SHORT_BINSTRING '\x07ã\x03\x03\x16.\x07\tñà'
   35: q BINPUT 1
   37: c GLOBAL 'pytz _UTC'
   48: q BINPUT 2
   50: ) EMPTY_TUPLE
   51: R REDUCE
   52: q BINPUT 3
   54: \x86 TUPLE2
   55: q BINPUT 4
   57: R REDUCE
   58: q BINPUT 5
   60: . STOP
highest protocol among opcodes = 2

[Stuart Bishop (stub)]

I can reproduce the problem when pickling a datetime that uses a pytz timezone. However, I can successfully pickle and unpickle the timezone instance itself.

This looks like a Python bug, as it is reproducible without pytz:

Python2:

>>> from datetime import tzinfo, datetime
>>> import pickle
>>> d = pickle.dumps(datetime.utcnow().replace(tzinfo=tzinfo()), protocol=pickle.HIGHEST_PROTOCOL)
>>> with open('p4', 'wb') as f:
... f.write(d)

Python3:

>>> import pickle
>>> pickle.load(open('p4', 'rb'), encoding='latin-1')
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: an integer is required (got type str)

[Ronnie (ronnie.vd.c)]

Thanks, i see it's already fixed in Python 2.7