ipyparallel broken for Python 3.8

[basnijholt]

In the Adaptive CI pipeline we notice that ipyparallel is broken for MacOS, Windows, and Ubuntu for Python 3.8.

Locally, on a MacOS machine, I can reproduce the failure mode using:

from ipyparallel import Client
import pexpect

child = pexpect.spawn("ipcluster start -n 1")
child.expect("Engines appear to have started successfully", timeout=35)
client = Client()
ex = client.executor()

def linear(x):
    return x

futs = ex.map(linear, range(10))
list(futs)

Which raises

---------------------------------------------------------------------------TypeError                                 Traceback (most recent call last)~/miniconda3/envs/py38/lib/python3.8/site-packages/ipyparallel/serialize/serialize.py in unpack_apply_message(bufs, g, copy)
    196     bufs = list(bufs) # allow us to pop
    197     assert len(bufs) >= 2, "not enough buffers!"
--> 198     f, bufs = deserialize_object(bufs, g)
    199     pinfo = buffer_to_bytes_py2(bufs.pop(0))
    200     info = pickle.loads(pinfo)
~/miniconda3/envs/py38/lib/python3.8/site-packages/ipyparallel/serialize/serialize.py in deserialize_object(buffers, g)
    143     bufs = list(buffers)
    144     pobj = buffer_to_bytes_py2(bufs.pop(0))
--> 145     canned = pickle.loads(pobj)
    146     if istype(canned, sequence_types) and len(canned) < MAX_ITEMS:
    147         for c in canned:
~/miniconda3/envs/py38/lib/python3.8/site-packages/ipyparallel/serialize/codeutil.py in code_ctor(*args)
     22 
     23 def code_ctor(*args):
---> 24     return types.CodeType(*args)
     25 
     26 def reduce_code(co):
TypeError: an integer is required (got type bytes)

This might be related to joblib/loky#240.

@minrk

Related #367 #396

edit: This issue has been fixed since 12 Aug 2019 #379. However, 6.2.4 was released on 13 May 2019.

[basnijholt]

Closed by e529d29, the release of 6.2.5 🎉