Hi!

Meadhbh S. Hamrick (Infinity Linden) wrote:
well... it seems we have a couple of dependency recommendations:

    * ZCA
    * webob
    * grok
    * py.test

I have a vague understanding of why ZCA is advantageous... it gives us interfaces in python.

but why do we need grok again? or webob? or py.test?

grok makes registering those components easier. Please read part 4 of my tutorial on the registration options.

py.test is not really needed, I just mentioned it because it might have a nicer test interface than unit tests. But I never used it myself.

WebOb is a good generalization of Request/Response objects which prevents us from re-inventing the wheel. I wanted to use these in the interface of the network library. For an example use look in tests/base.py where I use it already in my test server.
(makes it easier than directly looking into the WSGI environment).

I'm still trying to resolve problems that have come up from using python eggs and buildout and now we want to add more dependencies?

so.. my take on this would be to ask:
* what objectives of the PyOGP project are met by using { grok | webob | py.test } ?
    * what features do we lose by not using them?

grok: We need to use XML for registering the components. Not a bad thing in itself but it seems easier to use Python syntax.

webob: we save time by not re-inventing the wheel but using an existing library which does the job well, is well tested and in production.

py.test see above. It was mainly a hint for you to have a look at it to decide if you like it or not :)

i'm trying to figure out what webob gives us that we don't get from httplib.

independance of httplib. You use one request object but you can exchange the part which uses it more easily, going from httplib to urllib2 to eventlet to a mockup server. Actually webob is really easy when it comes to calling WSGI applications.

what's the advantage of using py.test over pyunit?
why do we need an application server framework?

Oh, when we talk about grok, we actually mean grokcore.component which is not the server framework but just the part which helps registering the ZCA components.

Look here:
http://pypi.python.org/pypi/grokcore.component/1.4

So we need to call it the right way I think.


that being said...

i can see the advantages of making PyOGP work with grok, but that seems to me to be a sub-project. It just seems that if we _need_ an application server framework in order to make PyOGP work, then there's something seriously wrong with OGP.

Sorry, this was a misunderstanding, we really just want to use the small part of grok which was factored out and is useful in our case. Also look at my branch which I posted a while ago.

http://svn.secondlife.com/trac/linden/changeset/771



-cheers
-m/∞

P.S. - where did you list the reasons to use grok again?

http://www.mail-archive.com/pyogp@lists.secondlife.com/msg00017.html

this is also on the wiki somewhere.

-- christian



On Jul 18, 2008, at 5:28 AM, Christian Scholz wrote:

Hi!

I am just wondering how we get to a decision about grok?

Do we vote or do we first discuss a month about a governance model? ;-)

If we vote I vote +1 for all the reasons I have been giving recently.

Let's unblock this thing!

-- Christian



--
Christian Scholz                         video blog: http://comlounge.tv
COM.lounge                                   blog: http://mrtopf.de/blog
Luetticher Strasse 10                                    Skype: HerrTopf
52064 Aachen                              Homepage: http://comlounge.net
Tel: +49 241 400 730 0 E-Mail [EMAIL PROTECTED]
Fax: +49 241 979 00 850                               IRC: MrTopf, Tao_T

neue Show: TOPFtäglich (http://mrtopf.de/blog/category/topf-taglich/)

_______________________________________________
Click here to unsubscribe or manage your list subscription:
https://lists.secondlife.com/cgi-bin/mailman/listinfo/pyogp



------------------------------------------------------------------------

_______________________________________________
Click here to unsubscribe or manage your list subscription:
https://lists.secondlife.com/cgi-bin/mailman/listinfo/pyogp


--
Christian Scholz                         video blog: http://comlounge.tv
COM.lounge                                   blog: http://mrtopf.de/blog
Luetticher Strasse 10                                    Skype: HerrTopf
52064 Aachen                              Homepage: http://comlounge.net
Tel: +49 241 400 730 0                           E-Mail [EMAIL PROTECTED]
Fax: +49 241 979 00 850                               IRC: MrTopf, Tao_T

neue Show: TOPFtäglich (http://mrtopf.de/blog/category/topf-taglich/)

_______________________________________________
Click here to unsubscribe or manage your list subscription:
https://lists.secondlife.com/cgi-bin/mailman/listinfo/pyogp

Reply via email to