4 years later the test is still skipped. That's a good example of "temporary fix" :-)

David: Should we just close the issue, or is there anyone around interested to work on a fix? Sadly, the issue doesn't provide the error message and I'm not sure how to reproduce the bug.

Maybe the bug has been fixed in the meanwhile? Example of SocketServer fix:

commit ba8474b77dd86d8dde40eaa7a4a6715a476d6242
Author: Martin Panter <vadmium+py@gmail.com>
Date:   Thu Feb 18 10:43:55 2016 +0000

    Issue #26309: Shut down SocketServer request if verify_request() is false
    
    Based on patch by Aviv Palivoda.

--

On Fedora 29, I tried to reproduce the bug by running many tests in parallel:

(*) Run test_linetoolong in a loop, 4 processes:

./python -m test -u all -F -m test.test_imaplib.ThreadedNetworkedTests.test_linetoolong test_imaplib 

(*) Run test_imaplib in a loop, 4 processes:

./python -m test -u all -v -F test_imaplib

(*) Stress the system (run the Python test suite in a loop), 2 processes:

./python -m test -r -u all,-gui -j0 -F

(*) Stress the CPU using my script (to ensure that the system load is at least 10)

system_load.py 10

7 minutes later, test_linetoolong has been run 11629 times x 4 processes and it's still pass.