Hangs when trying to call `os.close`

[jakirkham]

Not sure what the issue is here.

Traceback (most recent call last):
  File "/zopt/conda/envs/nanshenv/lib/python2.7/site-packages/billiard/context.py", line 138, in Queue
    return Queue(maxsize, ctx=self.get_context())
  File "/zopt/conda/envs/nanshenv/lib/python2.7/site-packages/billiard/queues.py", line 45, in __init__
    self._rlock = ctx.Lock()
  File "/zopt/conda/envs/nanshenv/lib/python2.7/site-packages/billiard/context.py", line 103, in Lock
    return Lock(ctx=self.get_context())
  File "/zopt/conda/envs/nanshenv/lib/python2.7/site-packages/billiard/synchronize.py", line 182, in __init__
    SemLock.__init__(self, SEMAPHORE, 1, 1, ctx=ctx)
  File "/zopt/conda/envs/nanshenv/lib/python2.7/site-packages/billiard/synchronize.py", line 101, in __init__
    register(self._semlock.name)
  File "/zopt/conda/envs/nanshenv/lib/python2.7/site-packages/billiard/semaphore_tracker.py", line 71, in register
    self._send('REGISTER', name)
  File "/zopt/conda/envs/nanshenv/lib/python2.7/site-packages/billiard/semaphore_tracker.py", line 78, in _send
    self.ensure_running()
  File "/zopt/conda/envs/nanshenv/lib/python2.7/site-packages/billiard/semaphore_tracker.py", line 62, in ensure_running
    os.close(w)
OSError: [Errno 9] Bad file descriptor

[ask]

Are you on Windows or are you using FORCE_EXECV/forking_enable?

[jakirkham]

This is on Mac, actually. I should add that I am using billiard.get_context("spawn") to create a spawn context that I use for launching a single billiard.Process.

[ask]

Ok, pretty sure I have only tested fork so far, so that probably requires some work

[jakirkham]

Alright, let me know if I can help, I am very interested in getting this to work.

[thedrow]

@jakirkham Feel free to contribute a PR.

[jakirkham]

Unfortunately I don't know the cause. So I don't know how to fix it either.

[thedrow]

We need a test case at the very least to debug this. I'm not sure you'll be able to provide one since so much time has passed. Sorry about that,

If you can, please do.