Sorry, I've just found this:
http://ironpython.codeplex.com/workitem/30228so I see that it is a
bug.
On Thu, Sep 20, 2012 at 3:48 PM, John Dickinson <jhcd.j...@gmail.com> wrote:
> I have at least two threads that end up calling struct.pack and
> struct.unpack, and regularly get two exceptions: "SystemError: Object
> reference not set to an instance of an object." and "SystemError: The
> LinkedList node does not belong to current LinkedList.". The following code
> reproduces the problem:
>
> import threading
> import struct
> import time
> import traceback
> import sys
>
> class TestThread(threading.Thread):
>
> def __init__(self):
> super(TestThread, self).__init__()
> self.setDaemon(True)
>
> def run(self):
> while True:
> try:
> struct.unpack('!H', '\x4f\x00')
> struct.pack('H', 20224)
> time.sleep(0.005)
> except Exception, e:
> print(str(e))
> traceback.print_exception(*sys.exc_info())
>
> if __name__ == '__main__':
>
> thread_1 = TestThread()
> thread_2 = TestThread()
>
> thread_1.start()
> thread_2.start()
>
> while True:
> print(".")
> time.sleep(1)
>
> (adding more TestThreads makes the error occur more regularly). Running
> the same code in CPython (2.7) I get no errors (presumably because of the
> GIL). It seems that struct isn't thread safe - is this a bug? I'm running
> IronPython "2.7.3 (2.7.0.40) on .NET 4.0.30319.269 (32-bit)" on Windows 7.
>
>
_______________________________________________
Ironpython-users mailing list
Ironpython-users@python.org
http://mail.python.org/mailman/listinfo/ironpython-users
