Michael Ekstrand wrote: > Michele Simionato wrote: > > Michael Ekstrand wrote: > >> After reading AMK's survey of what's new in Python 2.5, I am suitably > >> impressed. As usual, I can't wait to start using the cool new > >> features... extended generators? (mind is currently swimming with the > >> question of "can I implement Scheme's call-with-current-continuation > >> using extended generators".) > > > > No. > > Oh well. I'll try not to think too hard about it then. > > The day Python (without using Stackless) has true continuations will be > a happy day.
It is interesting that the support for full continuations was removed in recent versions of Stackless (I think there was support in version 1, not in versions 2 and 3, but I am not a Stackless user so please correct me if I am wrong). Coroutines give you more control on your program flow, but not as much as full continuations. With full continuations you can store the current state of your program (with some restrictions, the program should not have side effects or interact with an external environment) and at a later time go back to to that state. In a sense you can go back in time (but only in points that your program has already travelled) whereas with coroutine you can go just in one direction in the time. A nice thing you can do with full continuations is a modal Web server (seehttp://www.double.co.nz/scheme/modal-web-server.html). But this is definitely OT for this thread, so let me stop here ;) Michele Simionato -- http://mail.python.org/mailman/listinfo/python-list