Yep, this sounds a bit strange - a single process pool? why would you need this??

What happened to me is that I wrote several scripts that use the (excellent) multiprocessing package for the very easy creation of a process pool.

from multiprocessing import Pool

The thing is that on some platforms, it does not run well. 

Not sure why is that, but I get an Issue 3770 (which should be closed http://bugs.python.org/issue3770) on a virtual machine, which basically means that I cannot use multiple processes using the process pool. Since I just want the same code to work on this vm, I wrote a simple class that mimics some of the functionality of the Pool:

class SingleApplyResult(object):

        def __init__(self,v):

                self.value=v

        def get(self):

                return self.value

class SingleProcessPool(object):

        def __init__(self):

                pass

        def apply_async(self, func, args=(), kwds={}, callback=None):

                value = func(*args, **kwds)

                return SingleApplyResult(value)

        def close(self):

                pass

        def join(self):

                pass

def get_pool( processes, force_singlethread=False ):

        try:

                if force_singlethread:

                        pool=SingleProcessPool()

                else:

                        pool=Pool(processes=processes)

        except:

                print "failed getting multi-process pool, getting single-process"

                pool=SingleThreadPool()

        return pool

That's it. When I want to create the pool, I just write:

pool = get_pool(processes=20)

and it gives me a pool which either supports many processes, or just one if the platform does not support it.