[Cubicweb] meeting with Paul Everitt "a Pyramid guy" on July 31st 2014 at Logilab

Paul Everitt paul at agendaless.com
Mon Jul 28 09:35:14 CEST 2014

I still have a couple of days to prepare before coming this week. Comments below...

On Jul 17, 2014, at 10:43 PM, Christophe de Vienne <christophe at unlish.com> wrote:

> Hi everyone,
> I'm very happy that this meeting can take place so soon after the beginning of our experimentation with pyramid and cubicweb. Thanks Logilab for supporting this initiative!
> To help deciding on what we should spend this day, below are the things we need for the prototype to become an official part of the next cubicweb.
> Finish what has been started :
> Bypass totally CubicWebPublisher
> Carefully check that all the errors are properly handled
> Use error views of cubicweb in pyramid error handlers
> Provide sane defaults for the application configuration so the default behavior is very close to what cubicweb. Here I am mainly thinking of the authentication policy(ies).
> Test the prototype with existing cubes: we want full compatibility (except for authentication)
> Transform the prototype into a cubicweb patch
> Probably add pyramid 1.5+ to wheezy-backports, unless we decide to base our work on v1.4.x.
> Eventually, drop the twisted and wsgi handlers and their dependencies.
> Other things that needs care :
> Decide what else can be done without modifying cubicweb too much, like reimplementing the static content controller
> Decide how we expose the pyramid API to the cubes.
> See if we can (and I think we can) use cubicweb predicates and resource locator to have pyramid route the request directly to the right cubicweb view. We may even be able to make the cubicweb maintemplate into a pyramid renderer.
When I heard the description of the way views in CubicWeb are looked up, it made me think of Pyramid's traversal, which is designed for graphs. It's possible that traversal is precisely the right approach for CubicWeb and, if used, would bring in a number of benefits:


I have an introduction-to-traversal tutorial that I could give when I'm there, if you are interested.
> Get an idea of what can be re-done the "pyramid way", at what cost, and choose the next subjects. I think most of the general ideas were already listed in my previous emails.
> +1 for a bit of training for developers who may contribute. The more we are to code the better (we want more sprints!!)

I have far more material than time available. :) Much of my existing material is for beginners, but I plan to make a new tutorial for add-on developers. Let me know how much time you would like to spend on a tutorial, and who is the audience, and I can recommend some topics.


> -1 for the javascript subject outside of a few quick talks : one day is short and we should focus on the pyramid subject if we want to get results shortly.
> What I would love, at the end of the day, is to have a 2-3 month roadmap to a pyramid-based version of cubicweb in the wild.
> Regards,
> Christophe
> Le 16/07/2014 16:41, Nicolas Chauvat a écrit :
>> Hi List,
>> Logilab will welcome Paul Everitt [1] on July 31st 2014 in its offices
>> in Paris. The idea is to spend the day hacking on an implementation of
>> CubicWeb based on Pyramid.
>> If you are interested, you are welcome to join us, just reply to this
>> invitation to say you want to come (no fees involved, Logilab is
>> taking care of Paul's invoice).
>> The exact plan for the day is left for you to discuss here with
>> Paul. He offered to mix training and consulting about Pyramid, with
>> maybe a bit of how he uses javascript for the frontend.
>> I am happy that we managed to set this up[2] and that Christophe and
>> Aurélien made progress yesterday on the pyramid+cubicweb prototype[3].
>> Join us for that hackathon and let's see how far we can get with this!
>> Notes:
>> [1] https://agendaless.com/people
>> [2] Paul lives in the US and will use one day of his vacation in
>> France to come work with us. With everyone coming and going, it has
>> been difficult to find a day when most people involved could be in the
>> same room.
>> [3] http://hg.logilab.org/review/pyramid_cubicweb/
> _______________________________________________
> Cubicweb mailing list
> Cubicweb at lists.cubicweb.org
> http://lists.cubicweb.org/mailman/listinfo/cubicweb

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cubicweb.org/pipermail/cubicweb/attachments/20140728/964351e1/attachment-0186.html>

More information about the Cubicweb mailing list