[TIP] pytest, multiprocessing and pyzmq's ioloop

Pere Martir pere.martir4 at gmail.com
Tue Oct 25 13:45:23 PDT 2011


On Tue, Oct 25, 2011 at 2:34 PM, Stefan Scherfke
<stefan.scherfke at offis.de> wrote:
> If no exceptions occurs, everything runs fine. However, if the
> `multiprocessing.Process` raises an exception and thus never sends a reply,
> py.test waits forever without noticing the exception. The only thing I can
> do about this is to run py.test with "-s" so that I can see exception's
> traceback in the captured stderr and can then abort pytest using Ctrl+C.

I have a similar problem but no better solution yet. It's already
known that py.test simply doesn't catch the exception in the other
threads:

  https://bitbucket.org/hpk42/pytest/issue/56/feature-report-exceptions-in-threads-as

You may be interested in the workaround posted on August of 2009 here:

  http://bugs.python.org/issue1230540

I've not tried it yet. What's your opinion ? Does it workaround also
work for threading2, greenlet, etc ?

I would probably give Twisted Trial a shoot:

  http://twistedmatrix.com/trac/wiki/TwistedTrial

According to this recent post, it's said that your application doesn't
has to be a Twisted one to use Trial:

  http://lists.idyll.org/pipermail/testing-in-python/2011-October/004483.html

I have not tried it yet neither.



More information about the testing-in-python mailing list