IRC log for #storm on 20080408

00:00.15*** join/#storm wallflower (n=wallflow@ip-205-246-113-216.pool.grokthis.net)
02:42.33*** join/#storm radix (n=radix@wordeology.com)
04:05.14*** join/#storm bigdo1 (n=scmikes@72-197-8-8-arpa.cust.cinci.current.net)
04:13.07*** join/#storm jml (n=jml@li2-200.members.linode.com)
05:22.28*** join/#storm jukart (n=jukart@85-124-221-45.static.xdsl-line.inode.at)
05:48.58*** join/#storm jukart (n=jukart@lsfw01.lovelysystems.com)
06:12.49*** join/#storm dobee (n=dobeee@lsfw01.lovelysystems.com)
07:07.14*** join/#storm mcella (n=mcella@ip-178-4.sn2.eutelia.it)
07:15.58*** join/#storm goschtl (n=goschtl@p5B0BD74F.dip.t-dialin.net)
09:16.57*** join/#storm jukart (n=jukart@lsfw01.lovelysystems.com)
09:50.46*** join/#storm oubiwann (n=oubiwann@209-181-152-131.omah.qwest.net)
11:15.50goschtldobee: you gave some suggestions for the zope stormcontainer! Can you remember?
11:16.12dobeejep goschtl
11:16.50goschtldobee: i´ve some questions/problems with the implementation of it!
11:16.59dobeek
11:17.20goschtldo you have a little bit time for questions?
11:17.43dobeejep
11:20.24dobeegoschtl: just write my name, so that i get notified here
11:22.44goschtldobee: This is your suggestion for __delitem__. http://paste.lisp.org/display/58800.  With self.values() i get a ResultSet. but this resultSet has no find attribute...
11:23.10goschtlif you need more info please ask
11:25.35dobeeups, jep thats true
11:25.56dobeebut for __delitem__ this is ok
11:27.01dobeecant't remeber, did i suggest to implemente a find method on the container?
11:27.01goschtlmoment i will paste your suggestions too
11:28.28goschtldobee: look here http://paste.lisp.org/display/58801
11:28.30dobeebtw: we found out that the whole dictionary pattern does not fit at all in pure rdb based applications
11:29.29dobeegoschtl in __delitem__ :
11:30.06dobeestore.find(self._class, pkvar=pk).remove()
11:30.28dobeefor convinience you could implement a find method
11:30.38dobeethat just adds self._class to *args
11:31.37goschtldobee: you don´t use a container for your rdb based applications?
11:31.44dobeeno
11:32.05dobeewe don't use container
11:32.46dobeejust for site objects that  are not in rdb
11:32.52dobeewe write traversers
11:33.12dobeefor example a users utility that may return profiles
11:33.20dobeeusers/<profileid>
11:33.27dobeethis traverser returns the profile
11:33.33goschtlyes i understand
11:33.59dobeegoschtl: the problem is, that if you use that dictionary pattern you get very inefficient
11:34.08goschtli found a container is very natural  easy to use for zope/grok
11:34.09dobeefor example with the rotterdam skin
11:34.58dobeeyou can implement a countainer without implementing the items, values, keys method
11:35.22dobeethese are senseless in rdb, because you just need some find method
11:35.44dobeethe traverser only uses __getitem__ or get
11:35.56dobeethes methods are efficicent with rdb
11:36.26goschtldobee: ok i understand
11:37.04goschtlbut i in the find method the results should then implement the ILocation interface???
11:38.23dobeegoschtl: the classes used for the object can implement the __name__ and __parent__ as properties
11:38.30dobeethis is what we do
11:38.37dobeefor the user utility example
11:39.00dobeethe __parent__ property of the profile objects does a utility lookup for IUsersUtility
11:39.52dobeeactually the values method is the find method
11:40.00dobeedef values(self, *args, **kwargs)
11:40.16dobeereturn store.find(self._class, *args, **kwargs)
11:41.09goschtlok and your User class implement the __name__ __parent__ as properties?
11:41.21dobeegoschtl: yes
11:41.47dobee__name__ is in most cases a string representation of the pk
11:45.19goschtldobee: I ´m not sure but i don´t like it when i have to implement the __name__ __parent__ in my "ContentClass". Is this not a problem for you?
11:47.06dobeegoschtl: the bigger problem is that if you want to return a resultset you cant because you have to instantiate objects
11:47.13dobeeto set __parent__ and __name__
11:48.12goschtlstimmt
11:48.13dobeegoschtl: the big benefit with resultsets is that you just fetch what you need
11:48.51dobeefor examle mycontainer.find(title="test")[:3]
11:49.24dobeeif you have to instantiate you have to fetch all items instead of just 3
11:49.59goschtlok now i understand.
11:52.26goschtldobee: i will refactor the container
11:55.40goschtldobee: Do you think it´s possible to make an adapter for the resultset which implements the __name__, __parent__ things?
11:58.01dobeegoschtl: i guess that is deep down inside
11:59.00dobeegoschtl: the load hook is what you need i guess
12:01.47goschtldobee: Thanks for your help. I try to refactor some things....
12:02.11dobeek
12:21.02*** join/#storm niemeyer (n=niemeyer@200-103-245-169.ctame705.dsl.brasiltelecom.net.br)
12:24.58*** join/#storm mcella_ (n=mcella@ip-178-4.sn2.eutelia.it)
14:05.31*** join/#storm bigdog (n=scmikes@72-197-8-8-arpa.cust.cinci.current.net)
14:31.06*** join/#storm festizio (n=greg@ool-4b7fed03.static.optonline.net)
15:19.08festizioHi, I'm having a problem adding objects to a ReferenceSet, does anyone have a minute?
15:20.33festizioclass Product(Storm):
15:20.33festizio__storm_table__ = "inventory_product"
15:20.33festizioproduct_id = Int(primary=True)
15:20.33festizioproduct = Unicode()
15:20.33festiziogroup = Reference(product_id, "Product_Group.group_id")
15:20.33festiziocomponents = ReferenceSet(product_id, "Component.component_id")
15:22.05radixfestizio: What's the problem?
15:25.08festizioI have a Product() that is made of Components()
15:26.16festizioI can add Components to Product.components
15:27.06festiziofor example
15:27.35festiziomodel500 = Product(u'500')
15:28.46festiziopreppedchassis500 = (store.find(Component, Component.component.like(u'500 chassis'), Component.prepped==1).one()
15:29.23festiziocfcard = (store.find(Component, Component.component.like(u'compact flash card')).one()
15:29.38festiziomodel500.components.add(preppedchassis500)
15:29.43radixgreat
15:29.47festiziomodel500.components.add(cfcard)
15:30.02radixso what's the problem?
15:30.08festiziobut if i do model500.components.count()
15:30.16festizioit returns 1
15:31.03radixfestizio: What does list(model500.components) return?
15:32.23festizio[<product_classes.Component object at 0xb77b29ec>]
15:37.30radixfestizio: hmm
15:37.37radixfestizio: I wonder if one of those objects you're adding is None
15:37.41radixfestizio: can you print them before you add them?
15:38.04festizioI also have a method in Products() called num_buildable that does self.components.count()
15:38.09festizioand it returns None
15:38.40radixfestizio: sorry, what? count() returns None?
15:38.52festiziodef num_buildable(self):
15:39.28festiziooh wait, hold on a minute
15:40.32festizionevermind, count() returns one, I had a typo
15:41.24radixok :)
15:41.33radixfestizio: ok, so can you please check to see if one of the objects you're adding to the reference set is None
15:41.38radixfestizio: i.e., model500 and cfcard
15:41.38festiziook
15:46.09festiziowhen I print model500 i get <product_classes.Product object at 0xb77e254c>
15:46.28festizioprint cfcard
15:46.34festizio<product_classes.Component object at 0xb77e292c>
15:46.47festizioprint prepped500chassis
15:46.51festizio<product_classes.Component object at 0xb77e26ec>
15:48.57radixhmm. strange.
15:49.21radixfestizio: I guess I can't figure out what the problem is. If you can post a fully self-contained example program that reproduces the problem, I'll be able to help more.
15:49.46festiziook
15:50.03festizioyou want it to build the tables also?
15:50.26radixfestizio: that would be ideal, because it's possible the schema would affect things.
15:50.30festiziook
15:51.09radixniemeyer: http://rafb.net/p/SYbQRv82.html
15:52.46niemeyerradix: Ouch
15:53.05radixniemeyer: yeah, I thought it might have been festizio's problem, but he's not adding None after all
16:16.05festizioi have a mysqldump and 2 python files, do you want them all in one pastebin, or seperated
16:20.39festizioradix: http://rafb.net/p/ZfkbjS15.html
16:21.38festizioradix: http://rafb.net/p/lB8b7w58.html
16:22.02radixfestizio: hum. I wonder if it's because you're using MyISAM instead of InnoDB
16:23.05radixalthough I don't immediately see why that would be a problem, in general storm relies on transactional behavior
16:24.49festizioradix: http://rafb.net/p/4ylFSa95.html
16:25.13radixfestizio: that last paste seems to have some syntax errors
16:25.27festiziook, one sec
16:26.32festiziohttp://rafb.net/p/WkQGo871.html
16:26.40festizioi just pasted it again
16:27.24radixthanks
16:27.37radixI don't have a mysql instance running so I'll have to set one up
16:27.44festiziook, thanks
16:38.37festizioradix: I'm going to grab some lunch, I'll be back in an hour or less
16:44.49*** join/#storm marianom (n=marianom@ubuntu/member/marianom)
17:33.20festizioradix: I'm back, any thoughts?
18:02.58*** join/#storm dobee (n=dobeee@85-124-200-100.static.xdsl-line.inode.at)
18:25.37festizioradix: you still there?
18:26.04radixfestizio: Oh, hi
18:26.16radixfestizio: I haven't gotten around to checking out your code, I went to lunch and am now working
18:26.58festizioradix: ok
18:47.53*** part/#storm marianom (n=marianom@ubuntu/member/marianom)
18:59.35*** join/#storm marianom (n=marianom@ubuntu/member/marianom)
20:52.14*** join/#storm mcella (n=mcella@ip-176-82.sn2.eutelia.it)
20:59.52festizioradix: i'll be back later
21:43.16*** join/#storm harnak (n=harnak@bilbo.netconcepts.com)
22:35.52*** join/#storm harnak_ (n=harnak@bilbo.netconcepts.com)
22:36.39*** join/#storm harnak (n=harnak@bilbo.netconcepts.com)
23:59.42oubiwannradix: ping on storm.twisted...
23:59.55radixoubiwann: argh
23:59.56radixhi

Generated by irclog2html.pl Modified by Tim Riker to work with infobot.