[Stackless] Problem with xmlrpc over stacklesssocket
omyar_hunt at yahoo.co.uk
Wed Sep 15 11:51:37 CEST 2010
No worries Richard, thanks for getting back so fast.
I've been poking around the python networking modules, and they're definitely doing some odd stuff, like spawning new threads or even procesess to handle a single http call on the server side.
Also, it looks like the various socket strategies for different network conditions make it a big job to try and rewrite the network layer. So I think I may fall back to a simpler strategy than xmlrpc, which probably wasn't the best solution for me anyway, it just seemed easy. Gonna see what I can get out of the other Stackless Networking examples.
Anyways, thanks for the feedback.
--- On Wed, 15/9/10, Richard Tew <richard.m.tew at gmail.com> wrote:
> From: Richard Tew <richard.m.tew at gmail.com>
> Subject: Re: [Stackless] Problem with xmlrpc over stacklesssocket
> To: "The Stackless Python Mailing List" <stackless at stackless.com>
> Date: Wednesday, 15 September, 2010, 4:24
> On Tue, Sep 14, 2010 at 5:53 PM, Ray
> Gallagher <omyar_hunt at yahoo.co.uk>
> > Howdy folks.
> > Was wondering if you could point me in the right
> direction. Not sure if stacklesssocket is your project
> exactly though.
> > I'm having trouble with the basic python xmlrpc client
> over stacklesssocket. Basically, whenever I try to connect,
> the tasklet hangs. Other network protocols (basic http and
> memcache connections) seem to work fine. Also, the
> SimpleXMLRPCServer works fine under stacklesssocket. Seems
> to only be the xmlrpc client.
> > Anyone know how to tweak stacklesssocket to work with
> xmlrpc? I've attached runnable example code.
> > (notes on the code: I added the timeout_func when
> stacklesssocket complained to me about not having it. Also,
> if the loop tasklet is taken out, the process exits
> naturally without finishing the xmlrpc call. So I'm assuming
> there's a channel waiting on a message that never comes
> > stacklesssocket is from here: http://stacklessexamples.googlecode.com/svn/trunk/examples/networking/stacklesssocket.py
> and the stackless version itself is 2.63
> You could try the older less correct version of this
> module, and see
> if that lets you get away with it.
> Your only other option is to try and debug this yourself.
> Unfortunately, this would potentially take more time than I
> can spare
> right now. It does indicate that the module should be
> run against
> more than just the standard library socket tests.
> Really, I guess, in
> an ideal world I would be running the entire Python test
> suite against
> a set of modules like it.
> I will try and get time to test the module against the unit
> tests for
> this, and similar higher level networking related modules
> in the next
> couple of weeks. But it could be anything.
> Maybe internally it makes
> use of select, maybe it uses threads, otherwise it should
> reasonable to expect it to just work.
> Stackless mailing list
> Stackless at stackless.com
More information about the Stackless