Hi, it would be quite nice, if the different implementations could somehow be united (I've written the pypy one, but never used it, so there are indeed lots of errors lurking). pypy's stackless is actually implemented using pypy's coroutine implementation, which, in case CPython is used, is emulated by greenlets.
pypy's stackless http://codespeak.net/svn/pypy/trunk/pypy/lib/stackless.py has a corresponding test file http://codespeak.net/svn/pypy/trunk/pypy/lib/app_test/test_stackless.py which is supposed to be compatible with regular stackless as well. Maybe we could come up with a set of tests that cover all of stackless interface and therefore can be used to test our greenlet based implementations against. Cheers, Stephan Richard Tew wrote: > ---------- Forwarded message ---------- > From: Richard Tew <[email protected]> > Date: Mon, Feb 16, 2009 at 11:21 AM > Subject: Re: [Stackless] Stackless on top of greenlets > To: Henk Punt <[email protected]> > > > On Mon, Feb 16, 2009 at 10:33 AM, Henk Punt <[email protected]> wrote: >> Can you give some pointers to those implementations?, >> >> The only one I could find is the one included in PyPy and that one is >> not-working, e.g. it lacks certains methods in the API and also it >> does not behave >> exactly as stackless does, especially w.r.t. the order of task-scheduling. >> Also it contains quite a lot of bugs, for instance raising an >> exception in a task that is >> blocked on a channel does not restore the correct balance and similar >> issues... >> > > None of the implementations I have seen have been perfect, including my own. > > http://www.stackless.com/pipermail/stackless/2005-May/002416.html > > Cheers, > Richard. > > _______________________________________________ > Stackless mailing list > [email protected] > http://www.stackless.com/mailman/listinfo/stackless _______________________________________________ Stackless mailing list [email protected] http://www.stackless.com/mailman/listinfo/stackless
