Pickling of functions in a file loaded in a Sage Worksheet fails about cocalc HOT 5 CLOSED

Problem occurs with functions and classes defined at top level in worksheet as well as loaded from file:

function:

import pickle
def f():
    pass
s = pickle.dumps(f, protocol = 2)
==>
PicklingError: Can't pickle <function f at 0x7f1ebb091500>: it's not found as __main__.f

class:

class Klas():
    def kf():
        pass
sk = pickle.dumps(Klas, protocol = 2)
==>
PicklingError: Can't pickle <class __builtin__.Klas at 0x7f1eba3e1a10>: it's not found as __builtin__.Klas

from cocalc.

WON'T FIX.

Decided not to deploy a fix for this problem, because there is negligible value to saving/pickling a top-level function or class definition at the cost of complexity added. See last few paragraphs here: What can be pickled and unpickled?

A workaround function to save a function or class at top level is here, in case it is needed in some edge case:
76c386b

from cocalc.

Despite the "wont fix", I'm reopening this. It's unexpected "buggy" behavior, and it's 100% caused by the sagews layer.

Dear future reader: you can use a jupyter sage kernel talking "more directly" with sage. then this works.

from cocalc.

@haraldschilly:

Comment from @williamstein on 2016-07-22 was that the proposed fix was not useful in any real-life situation and should therefore be reverted. So I think we have working code, the decision was whether to deploy it.

from cocalc.

Obviously we aren't fixing this and nobody else complained in 8 years. So closing again.

from cocalc.