IRC log of #zope3-dev for Wednesday, 2007-09-12

*** Nederhoed has left #zope3-dev00:08
*** mgedmin has joined #zope3-dev00:13
*** reco has joined #zope3-dev00:15
*** mgedmin has quit IRC00:16
*** tarek has joined #zope3-dev00:27
*** MJ has quit IRC00:43
WebMavenphiliKON: still planning to work on zopeproject this month?00:44
wiggymerging instance-less zopeproject into grokproject as well?00:44
philiKONmore like the other way around00:45
philiKONi have a grokproject branch that's based on zopeproject00:45
WebMavenOK. I am reminding you that I will have a window between the 18th and 22nd for testing and feedback.00:45
philiKONcool, thanks00:45
WebMavenafter that I'll be at the Foliage Sprint.00:46
*** ignas has quit IRC00:47
*** greenman has quit IRC00:49
*** greenman has joined #zope3-dev00:49
*** pelle_ has quit IRC01:07
WebMavenTheuni: I see you're doing some work on z3c.zalchemy01:12
*** nathany_ has quit IRC01:14
*** dunny has joined #zope3-dev01:15
*** ousado_ has joined #zope3-dev01:22
*** deo has quit IRC01:27
*** ousado has quit IRC01:39
*** Ariel_Calzada has joined #zope3-dev01:44
*** deo has joined #zope3-dev01:52
*** gstratton has quit IRC02:07
*** norro has quit IRC02:11
*** CSWookie has left #zope3-dev02:12
*** schwendinger has quit IRC02:14
*** benji has joined #zope3-dev02:20
*** b52laptop has quit IRC02:53
*** srichter has quit IRC02:56
*** benji has quit IRC02:59
*** greenman has quit IRC03:01
*** J1m has quit IRC03:25
*** edgordon has joined #zope3-dev03:33
*** rcrafton has quit IRC03:42
*** rcrafton has joined #zope3-dev03:46
*** Ariel_Calzada has quit IRC03:47
*** Ariel_Calzada has joined #zope3-dev03:48
*** jukart has joined #zope3-dev03:52
*** ousado_ has quit IRC04:16
*** ousado_ has joined #zope3-dev04:16
*** Theuni_ has joined #zope3-dev04:24
*** Theuni has quit IRC04:25
*** huajie has joined #zope3-dev04:25
*** Theuni has joined #zope3-dev04:38
*** Theuni_ has quit IRC04:38
*** netshade has joined #zope3-dev05:18
*** netshade has left #zope3-dev05:19
*** RaFromBRC has quit IRC05:20
*** Ariel_Calzada has quit IRC05:23
*** reco has quit IRC05:37
*** deo has quit IRC05:39
*** jukart has quit IRC05:46
*** niemeyer has quit IRC05:56
*** huajie has quit IRC05:58
*** fcorrea has left #zope3-dev06:01
*** stub has joined #zope3-dev06:14
*** ousado_ has quit IRC06:16
*** ousado_ has joined #zope3-dev06:16
*** alecm has quit IRC06:17
*** ousado_ has quit IRC07:09
*** ousado_ has joined #zope3-dev07:09
edgordonwhat is the package index url for the z3c packages? do i just use
edgordoni felt like there was another one, but I can't remember what it was07:15
*** philiKON_ has joined #zope3-dev07:20
edgordonalso, srichter had told me he worked on a package that made custom principal plugins easier. it was a z3c package, but he couldn't remember the name of it. anyone remember what this was?07:27
*** chacha_chaudhry has joined #zope3-dev07:28
*** netshade has joined #zope3-dev07:28
*** philiKON has quit IRC07:29
*** greenman has joined #zope3-dev07:51
*** netshade has quit IRC07:53
TheuniWebMaven: pong07:56
Theuniany questions regarding zalchemy?07:57
WebMavenSo I see you're doing some work on z3c.zalchemy07:57
TheuniHowever, I'm getting annoyed by some of the features of it that I don't need07:57
Theunithe whole container-enchilada for example07:58
WebMavenWell, I was thinking of doing some work on documentation and improving the demos at the Foliage Sprint07:58
Theunisounds good07:58
WebMavenby 'enchilada' you mean the demo code?07:58
WebMavenor something else?07:59
Theunifor me i'm currently not interested in the "high level" integration08:01
TheuniI merely want the transaction integration to be as good as possible08:01
Theunifor 0.4 that means integrating two-phase-commit and savepoints08:01
Theuni(And some ZODB-like object cache)08:01
TheuniThe whole keyreference, container, traversing, ... stuff is uninteresting for me08:02
TheuniZagy and I plan to sit down after our current project and extract some of the things we did there into zalchemy08:02
Theuniafk for a few minutes, i need to get some bread for my breakfast08:03
WebMavenmaybe z3c.zalchemy should be split.08:07
Theunii thought about that too08:10
Theunii wondered into which parts08:10
Theunione of those should stay z3c.zalchemy08:11
Theunithe question is which one :)08:11
*** cursor has joined #zope3-dev08:12
Theuniwe should probably talk to juergen about his plans08:13
WebMavenI know from older conversations that he was also not much interested in the 'higher level stuff'.08:15
TheuniHum. So who introduced that? =)08:15
WebMavenSo, I'd say the low level parts should stay in z3c.zalchemy.08:16
TheuniI have the feeling I should organise a zalchemy-sprint.08:16
WebMavenThat would be cool.08:16
WebMavenhe introduced it, but I think it was more of a demo.08:16
Theunii feel we need some discussion of the patterns we all are using08:17
WebMavenI kept testing and finding bugs.08:17
WebMavenAnd I added the other demos as well.08:17
WebMavenFor example, the crappy dublincore storage.08:17
TheuniWhen thinking about the people who stuff with sqlalchemy in zope I think of juergen, andreas jung, you and us gocept guys08:18
TheuniI wonder whether you'd be happy to join a sprint in europe...08:19
WebMavenI would be very happy. Would someone be able to sponsor me?08:19
WebMavenAlso, when were you thinking of doing this?08:20
TheuniDunno. Our immediate schedule is pretty packed. I'd like to do it this year though.08:21
TheuniMaybe november, early december.08:21
TheuniI guess when we find out who joins the sprint we can check for pooling some resources.08:21
WebMavenI am currently trying to sell my house and move to another city. Unless the sale happens soon, November may be too early.08:22
*** yvl has joined #zope3-dev08:22
WebMavenAt least, too early to be convenient.08:24
TheuniWhere are you going to? Weren't you in Vegas?08:24
WebMavenI am still in Vegas. I am tired of not having a local tech community and being the only Zope guy here, so I am moving.08:25
WebMavenBecause of my GF's asthma, my choices for another city were a bit limited, so we chose Albuquerque in New Mexico.08:25
WebMavenWe are going on a road trip next weekend to check the city out.08:26
WebMavenI'll be going to a BarCamp while I'm there.08:26
WebMavenThen back here, then the Foliage Sprint.08:27
WebMavenit's a busy month.08:27
WebMavenAnyway, I would *love* to attend a Sprint over there.08:33
TheuniAlright, I'll keep that in mind. I'll check whether gocept wants to host it (maybe friday to sunday or so)08:33
*** jukart has joined #zope3-dev08:33
TheuniIf we do that I can accomodate you in our guest room.08:33
WebMavenOther than an emergency layover in Paris when I was 13, I've never been to Europe.08:34
Theunijukart: perfect timing.08:34
jukartTheuni: hi08:34
WebMavenjukart: Hi08:34
*** cursor_ has joined #zope3-dev08:34
Theunijukart: WebMaven and I were just discussing organizing a zope3+sqlalchemy sprint to figure out some patterns how we are using sqlalchemy with zope 3 and to work on zalchemy.08:35
*** cursor has quit IRC08:35
Theuniwould you be interested in that?08:35
*** jodok has joined #zope3-dev08:36
jukartright now we don't use sqlalchemy, we are starting to use storm08:36
Theuniinteresting. does that mean you're not interested in zalchemy anymore?08:36
jukartwe have no use for it right now08:37
TheuniWhy the switch to storm?08:37
WebMavenjukart: an article comparing SA and Storm would be very interesting...08:37
jukartIcan't do that article, but dobee worked with storm the last weeks08:38
*** sorin has joined #zope3-dev08:39
jukartone thing that simplifies storm is that you don't need to have separate schemas08:39
WebMavenYeah, I know Canonical wrote Storm partially because of deficiencies in SA 0.208:39
jukartbut there is a lot missing08:39
*** sorin is now known as sorindregan08:39
WebMavenYou mean Storm uses Z3 schemas?08:40
jukartWebMaven: no08:40
WebMavenHow do you get that advantage, then?08:42
*** jukart has quit IRC08:43
*** jukart has joined #zope3-dev08:43
WebMavenJukart: How do you get that advantage, then?08:44
WebMavenTheuni: there is one other person to add to your list: Kapil.08:44
TheuniThe part that struck me about storm when I saw it presented at EP was: a) they wouldn't name the deficiencies they found and b) your application contains SQL-related code all over the place08:47
*** jodok has quit IRC08:47
*** jodok has joined #zope3-dev08:47
WebMavenTheuni: Oh, I can shed some light on that.08:47
WebMavena lot of their frustrations with SA (and to a greater extent, SQLObject) was due to code quality.08:48
WebMaventhe other part was composing objects from data that came from different databases.08:49
WebMavenI had a number of long talks with them about all this last year.08:50
WebMavenbut that's what it boiled down to.08:53
WebMavenOK, so where were we...08:56
WebMavenTheuni: shall we touch base again in a few days?08:57
*** zagy has joined #zope3-dev09:01
*** zagy_ has joined #zope3-dev09:02
*** zagy has quit IRC09:02
Theunisorry, i was moving around09:03
TheuniAh, well. The different database stuff is in SQLAlchemy by now.09:04
TheuniRight, touching base in a few days is good, maybe middle of next week? I'll be busy until then.09:04
WebMavenYes. As I said, they started Storm when SA was at 0.209:04
Theuni(which must have been quite a while ago already)09:04
Theunianyway, gotta do other stuff :)09:05
WebMavenTheuni: OK, don't forget to add one other person to your list of SA users: Kapil.09:05
WebMavenBye, talk to you next week.09:05
*** ousado_ has quit IRC09:10
*** sorindregan has quit IRC09:10
*** ousado_ has joined #zope3-dev09:10
*** jodok has quit IRC09:12
*** jodok has joined #zope3-dev09:12
*** rcrafton has quit IRC09:15
*** hdima has joined #zope3-dev09:25
*** dunny has quit IRC09:34
*** greenman has quit IRC09:36
*** schwendinger has joined #zope3-dev09:37
*** markusleist has quit IRC09:48
*** dobee has joined #zope3-dev09:48
*** tarek has quit IRC09:49
*** timte has joined #zope3-dev09:53
*** pelle_ has joined #zope3-dev10:03
*** norro has joined #zope3-dev10:07
*** pelle_ has quit IRC10:11
*** baijum has joined #zope3-dev10:16
*** gstratton has joined #zope3-dev10:20
*** pyqwer has joined #zope3-dev10:24
*** norro has quit IRC10:26
*** tarek has joined #zope3-dev10:36
*** zagy_ is now known as zagy10:37
*** sorin has joined #zope3-dev10:38
*** sorin is now known as sorindregan10:39
*** naro has joined #zope3-dev10:42
*** pelle_ has joined #zope3-dev10:48
*** pelle_ has joined #zope3-dev10:48
*** schwendinger has quit IRC10:56
*** huajie has joined #zope3-dev10:56
*** jukart_ has joined #zope3-dev10:58
*** jukart has quit IRC10:59
*** jodok has quit IRC11:07
*** jodok has joined #zope3-dev11:08
*** schwendinger has joined #zope3-dev11:17
*** greenman has joined #zope3-dev11:20
*** jodok has quit IRC11:32
*** romanofski has joined #zope3-dev11:38
*** MJ has joined #zope3-dev11:49
*** MJ has joined #zope3-dev11:53
*** sorindregan has quit IRC11:59
*** dunny has joined #zope3-dev11:59
*** sorin has joined #zope3-dev12:00
*** sorin is now known as sorindregan12:00
*** philiKON_ has quit IRC12:00
*** philiKON has joined #zope3-dev12:01
*** mgedmin has joined #zope3-dev12:15
*** huajie has quit IRC12:24
*** b52laptop has joined #zope3-dev12:33
*** huajie has joined #zope3-dev12:41
*** rocky has quit IRC12:58
*** pelle_ has quit IRC13:01
*** pelle_ has joined #zope3-dev13:02
*** greenman has quit IRC13:03
*** greenman has joined #zope3-dev13:03
*** pelle_ has quit IRC13:04
*** pelle_ has joined #zope3-dev13:04
*** mkerrin has joined #zope3-dev13:05
*** stub has quit IRC13:09
*** cursor_ has quit IRC13:21
*** cursor has joined #zope3-dev13:22
*** benji has joined #zope3-dev13:23
*** dunny has quit IRC13:24
*** pelle_ has quit IRC13:40
*** MJ is now known as MJ|lunch13:44
*** rocky has joined #zope3-dev13:45
*** dobee has quit IRC13:50
*** dobee has joined #zope3-dev13:50
*** pelle_ has joined #zope3-dev14:04
*** benji has quit IRC14:05
*** ignas has joined #zope3-dev14:14
*** MJ|lunch is now known as MJ14:55
*** bigkevmcd has quit IRC14:57
*** netshade has joined #zope3-dev15:00
*** rcrafton has joined #zope3-dev15:00
*** schwendinger has quit IRC15:03
*** schwendinger has joined #zope3-dev15:04
*** ThomasKarlRichte has joined #zope3-dev15:04
*** bigkevmcd has joined #zope3-dev15:05
* ThomasKarlRichte is away: Zur Zeit nicht da.15:06
*** ThomasKarlRichte is now known as gamtaway15:06
* gamtaway is back.15:06
*** gamtaway is now known as ThomasKarlRichte15:06
*** netshade has quit IRC15:06
*** greenman has quit IRC15:17
*** niemeyer has joined #zope3-dev15:23
*** benji has joined #zope3-dev15:26
*** norro_work has joined #zope3-dev15:27
*** huajie has quit IRC15:30
*** alga has joined #zope3-dev15:33
*** chacha_chaudhry has quit IRC15:39
*** J1m has joined #zope3-dev15:54
*** cursor has quit IRC16:05
*** netshade has joined #zope3-dev16:08
*** netshade has left #zope3-dev16:08
*** mgedmin has quit IRC16:20
*** alga has quit IRC16:20
*** baijum has quit IRC16:24
*** gstratton_ has joined #zope3-dev16:35
*** lucielejard has joined #zope3-dev16:46
*** goschtl has joined #zope3-dev16:46
*** gstratton has quit IRC16:52
*** yvl has quit IRC16:53
*** zagy has quit IRC16:56
*** zagy has joined #zope3-dev16:56
*** srichter has joined #zope3-dev17:01
*** ChanServ sets mode: +o srichter17:02
*** lucielejard has quit IRC17:11
*** netshade has joined #zope3-dev17:19
*** netshade has left #zope3-dev17:20
*** lucielejard has joined #zope3-dev17:21
*** b52laptop has quit IRC17:21
*** markusleist has joined #zope3-dev17:22
*** mgedmin has joined #zope3-dev17:24
*** sorindregan has quit IRC17:25
*** spython has joined #zope3-dev17:28
spythonhi, what's a good way to have an attribute in a content component that is going to reference a container? weakref/intid?17:30
spythonhmm, i might just do  >>> myitem.containers.append(container)17:33
*** b52laptop has joined #zope3-dev17:33
*** goschtl has quit IRC17:33
*** hdima has quit IRC17:37
*** gstratton_ has left #zope3-dev17:41
*** hazmat has quit IRC17:43
mgedminyou can do that, as long as you're aware of the consequences17:50
mgedmin(i.e. if you remove container but keep myitem somehow, it will keep container alive but without a valid __parent__)17:50
mgedminthis is probably not interesting for your use case (if you remove a container, you'll remove all the items within)17:51
spythonbut myitem will not be located in the container that it references17:51
spythonI'm thinking of using intids isntead.. feels a bit safer17:52
ignasspython: so your container will not have any references to the item?17:52
ignasinvalid intids stored in a list are not that much better than references to parentless objects17:53
ignasi'd go for references, and a container removed subscriber that cleans up, or intids if you like indirection and a container removed subscriber that cleans up17:54
*** pelle_ has quit IRC17:54
ignasor you can add code that "handles" outdated intids i guess17:54
spythonignas: myitem will be located in /items ..  and in /groups i have a bunch of containers that an item can have references to17:54
ignaswhy containers will not have any references to the item?17:55
ignashow will items be "located" in containers then?17:55
ignasand if containers will point to items (have items in themselves) in some way, then a subscriber that iterates through all items in a container17:56
spythonwell, they do in a way.. a container contains a local catalog of items that "belongs" to a group17:56
ignasand removes the container from container list of the item17:56
spythoncontainer = group17:56
ignasis very easy to write17:56
* ignas can't see the benefit of having catalogs and int ids over simple btrees and a containers persistent list for each item17:58
*** ousado_ has quit IRC17:58
ignasyou need subscribers for cleanup anyway17:58
*** ousado_ has joined #zope3-dev17:58
*** xbeanx has quit IRC17:58
ignasI use schooltool.relationships for these cases though, so not like i have ever had to solve such problems ...17:59
spythonbasically, in my case, an item can 'belong' to several groups.. ie an item that belongs to "zope" & "python" should be viewable in both /groups/zope and /groups/python18:01
ignasso you just add the same object to both containers18:01
ignasand add both containers to it's container list18:01
philiKONexcept you don't, because containers expect a single __parent__ hierarchy18:01
philiKONbut you can certainly use several BTrees18:01
* spython keeps forgetting the reference stuff in python18:02
ignasphiliKON: having intids will not really solve url generation problems with multiple parents anwyay18:02
spythonignas: ok, I'll do that :)18:02
spythonoh, ok.. so i can't do this because of the __parent__ thing?18:03
ignasphiliKON: would LocationProxyying objects when they are being retrieved from the container by the traverser work?18:04
philiKONi suppose so18:04
ignasspython: in all the cases you will still have this problem of objects needing to have 1 __parent__ for url generation18:04
ignaswhich you will have to solve in some way18:05
ignasif you want to have 1 object accessible through multiple urls AND have these urls resemble the way objects were accessed, it's a bit complicated18:06
spythonno, it's ok if they are accessed via one url18:06
ignasso you have 1 single container for all items18:07
spythonyes.. /items18:07
ignasand have "tags" for them18:07
ignasand if one of the tags is removed - it gets removed from all the items tagged that way18:08
ignasand if object is removed, tags do not reference that object anymore18:08
ignasand you will probably want set operations for tags18:10
ignaslike "and" and "or"18:10
ignasin that case i would go for18:10
ignasa btree for items18:10
ignasa btree for tag objects18:10
ignasand a list of intids in every item18:10
ignaswith object removed subscribers for items and for tags that take care of invalid int ids18:11
ignashaving a list of item intids in every tag18:11
ignasshould give you the benefit of being able to find objects that have tags "A" and "B" for example very efficiently18:11
ignasno need for catalogs though i'd guess18:12
ignasyou can just have persistent lists of intids18:12
ignaslike - some_item.tag_ids and some_tag.item_ids18:13
ignaswith tags and items properties that would list actual tag and item objects18:13
spythonit'd still be nice to filter out items by some attribute.. so I think i'll have a catalog18:13
spythonoh wait, reading your stuff.. i'm a bit slow18:14
spythonhmm, bit confused but i think i got it now.. :)18:15
ignashmm, it is possible to implement tags and object through catalogs only without any direct references18:16
ignasbut that is a bit complicated18:16
ignasi mean - without even having intids of tags or items in lists18:16
*** norro_work has quit IRC18:18
spythonso, kind of like this...18:19
ignasotoh - mgedmin just suggested me that if you don't have any additional information on your tags except for the title string18:19
ignasyou can implement tags without a tag container18:19
ignasjust through a catalog and items container18:19
ignasoh, you have descriptions18:20
ignasso yes18:20
ignasjust not list18:20
ignaspersistent list18:20
ignaslist.append will get you in trouble ;)18:20
spythoni read something about inefficiency with long lists.. does that apply with just a simple int-list?18:21
spythonpersistent lists that is18:21
ignasif it will be slow, you will refactor it18:22
spythonyea.. :)18:23
spythonthere's probably coming some more attributes in IItem so I probably need a catalog18:25
*** nathany has joined #zope3-dev18:25
spythonis it worth making each IGroup a site to add a local catalog.. avoiding one enourmous in the main site?18:26
spythonsure you index some objects several times, but we got space :)18:27
*** pelle_ has joined #zope3-dev18:27
ignasspython: emm, i'd not go into that at the moment ;) as i don't think you *need* search by tag title or tag description18:28
ignasbut i guess you can18:28
*** schwendinger has quit IRC18:29
spythonbut while you're browsing /groups/python and only want to see items in there where attribute x = 'foo'.. you'd certainly need a catalog for that?18:30
spythonor maybe i misunderstood something you said earlier :)18:30
spythonignas: true, i don't need search by tag (IGroup).. only need to do searches on items18:35
ignasso you only need a catalog for items then18:35
spythonyes, but how is performance when it gets big?18:36
spython'good enough' i guess :)18:37
ignasdepends on how big, and on the UI that will be displaying data18:41
spythonusers will probably do a lot of filtering by certain attributes in item, hmm18:43
ignasso you will be doing a lot of integer list intersections18:43
ignaswhich are fast enough18:44
ignasand then you can perform a hack that retrieves all the titles from the catalog using the remaining list of intids18:44
ignasto show first 10-50 by alphabetic order18:44
ignaswhich is fast enough as you are not loading any objects18:45
spythonyeah :)18:46
spythonand i rewrote IntIds._generateId so that new objects get increasing ids (oldest = lowest id), should be easy to order by oldest/newest18:47
*** pyqwer has quit IRC18:48
ignasand how do you convert from intid to object?18:48
ignasoh, it's stored in the utility18:49
spythonuhm , queryObject=? My IntId class inherits from IntIds.. only rewrote one func :)18:49
ignasand the intid <-> title system is at least fast enough for 10k of objects18:51
spythonnow I'm really hungry.. thx for the help ignas :)18:51
*** timte has quit IRC18:58
*** norro has joined #zope3-dev18:59
*** edgordon has quit IRC19:00
*** pelle_ has quit IRC19:01
*** tarek has quit IRC19:05
*** dobee has quit IRC19:07
*** MJ has quit IRC19:08
*** jhauser has joined #zope3-dev19:10
*** J1m has quit IRC19:21
*** J1m has joined #zope3-dev19:21
*** jhauser has quit IRC19:21
*** J1m has quit IRC19:21
*** jukart_ has quit IRC19:25
*** J1m has joined #zope3-dev19:27
*** faassen has joined #zope3-dev19:27
*** mgedmin has quit IRC19:37
*** RaFromBRC has joined #zope3-dev19:58
*** reco has joined #zope3-dev20:00
*** reco has quit IRC20:01
*** reco has joined #zope3-dev20:01
*** thruflo has joined #zope3-dev20:01
*** fcorrea has joined #zope3-dev20:12
*** dobee has joined #zope3-dev20:15
*** faassen has quit IRC20:22
*** jsadjohnson has quit IRC20:25
*** jsadjohnson has joined #zope3-dev20:25
*** Ariel_Calzada has joined #zope3-dev20:30
*** J1m has quit IRC20:33
*** xbeanx has joined #zope3-dev20:36
*** J1m has joined #zope3-dev20:39
*** supton has joined #zope3-dev20:47
*** alecm has joined #zope3-dev20:53
*** b52laptop has quit IRC20:54
*** mkerrin has quit IRC20:56
*** naro has left #zope3-dev21:00
*** brandon_rhodes has joined #zope3-dev21:00
brandon_rhodesI get a buildout error when attempting to follow the "formdemo" instructions in!-t4108543.html21:01
brandon_rhodesDoes anyone know whether those instructions (an svn checkout of formdemo followed by bootstrap followed by builtout) should still work today?21:01
thrufloi get the same error21:01
*** thruflo has quit IRC21:02
brandon_rhodesThank you, thruflo.  I was hoping it wasn't just a quirk of my machine!21:02
*** ignas has quit IRC21:03
spythonhmm, is it really necessary to use PersistentList if you store a list of intids?21:03
philiKONbrandon_rhodes: you haven't said yet what error it is21:03
spython...on a content object21:04
philiKONso i'm wondering how thruflo could be so sure that he got "the same error"21:04
philiKONspython: yes21:04
*** dobee has quit IRC21:05
philiKONspython: though if that list gets significantly large (say, ~100 items), you really want to use a TreeSet21:05
philiKONspython: just read my book, will ya :)21:05
philiKONbrandon_rhodes: looks like somebody either checked in a wrong develop= line by mistake21:06
philiKONbrandon_rhodes: or somebody forgot to add an svn external21:06
spythonit's easier when you answer :D21:06
suptonhmm... I'm wondering if there is a simple way to put a constraint on a schema List field (where value_type=Object()) that all objects in sequence implement a particular interface?21:06
philiKONspython: chapter 6, dude21:06
spython;) yeah, i'll look more into 621:06
philiKONsupton: you already got it...21:07
*** gumpa has joined #zope3-dev21:11
suptonphiliKON: is this as simple as List(value_type=Object, constraint=IFoo.providedBy) ??21:11
philiKONthe value_type things always need to be *instances* of schema fields21:12
suptonah, okay.21:12
*** markusleis1 has joined #zope3-dev21:14
*** markusleis1 has quit IRC21:17
*** markusleist has quit IRC21:27
suptonanyone know what status, current ideas are on Grok getting some kind of N->N relationships (e.g. some kind of higher level piece on top of zc.relationship indexes)?21:27
suptonsomething to lure people used to Archetypes ReferenceFields or Django ForeignKeys...21:27
*** dunny has joined #zope3-dev21:28
supton...which are useful / required for most non-trivial applications.21:28
philiKONi thought zc.relation{ship} was pretty high level21:29
suptonmy mental gap with zc.relationship is bending it to what I want to do, which is define relationship "fields" on a grok.Model subclass with names without explicitly needing to create indexes for each relationship21:33
suptonI just cannot break Archetypes-induced habits, I guess, as I'm trying to learn zope321:34
philiKONthe question is whether those relationships are intrinsic properties of the model21:35
suptonI'd like some colleagues who have been working with Django to take a look at Grok, but much of the applications they work on now have Django ForeignKeys across a bunch of items...21:35
philiKONin most cases they aren't21:35
suptonI agree.21:35
suptonbut the simple app developer views them this way21:35
philiKONeven in sqlalchemy or whatever django uses, foreign key references are not part of the actual schema declaration i believe21:35
philiKONi agree that we should have a simple (iow, grokkish) way to deal with them21:36
philiKONbut i also think the machinery is in place21:36
philiKONit may yet need more typing21:36
*** santiag1 has joined #zope3-dev21:36
santiag1hi there21:37
santiag1hello ?? is any body out there ??21:38
suptonin Django, they are definitely intrinsic on schema/model declaration:  example paste of part of a model class for high school sports application I have worked on with a colleage in Django:
*** deo has joined #zope3-dev21:39
derek_richardsonif I want to replace use of old-style ftests (default layer) with new-style ftests (specified layers) but I want *all* the packages that used to be loaded by the default layer to still be loaded, what's the easiest way to do this? Or am I taking the wrong approach to migrating from old-style ftests? The main thing that annoys me is that the ability to run the test depends on what directory you run them from, not the lack of isolation (21:40
*** santiag1 has left #zope3-dev21:42
*** santiag1 has joined #zope3-dev21:43
*** whit is now known as whit|lunch21:46
santiag1can someone help with the examples in the zope3Book?21:48
*** santiag1 has left #zope3-dev21:49
philiKONderek_richardson: you write an ftesting.zcml that basically defines the setup for that layer21:51
derek_richardsonphiliKON: how will it differ from the one generated by buildout, then? Should I just copy the old one so that it's explicit? If so, how do I solve the 'results are relative to where I run the tests from' problem?21:54
philiKONbuildout generates an ftesting.zcml?21:54
philiKONprobably just for BBB21:54
derek_richardsonit can....21:54
philiKONthe "results are relative to where I run the tests from" problem is due to the fact that the standard ftesting layer (the old one) simply picks up ftesting.zcml from '.'21:55
derek_richardson - search for 'Legacy Functional Testing Support'21:55
philiKONbtw, *please* don't say buildout generates this or that21:55
philiKONit's never buildout21:55
philiKONit's a particular recipe21:55
derek_richardsonphiliKON: true.21:55
philiKONand i hate zc.zope3recipes :)21:55
derek_richardsonheh. you like zopeproject, right?21:56
philiKONright :)21:56
philiKONwhen you have your own layer that's based on an ftesting.zcml file, there's no PWD problem21:56
derek_richardsonI'm missing something here. Is creating layers covered in your book? I only see the two pages in chapter 12 that use the old default layer, afaict.21:58
philiKONno, it isn't21:58
derek_richardsonah. bummer.21:58
philiKONbut if you go to pretty much any eggified package that has ftests you'll find it21:58
philiKONusually in or testing.py21:58
* derek_richardson goes to look. thanks!21:59
philiKONit's pretty much the usual dance21:59
philiKONlayer = ZCMLayer('layer', '/path/to/ftesting.zcml') or so21:59
philiKON(of course, you make the path to ftesting.zcml relative to __file__, etc.)21:59
*** hazmat has joined #zope3-dev22:05
*** ChanServ sets mode: +o hazmat22:05
brandon_rhodesphiliKON: thanks for suggesting earlier that an SVN external was missing in formdemo; I did a manual "svn co" and buildout got farther.  Now it's dying because it can't compile something requiring some XML C libraries, but that's another issue.22:09
*** cbcunc has quit IRC22:23
*** b52laptop has joined #zope3-dev22:27
*** rcrafton has quit IRC22:34
*** rcrafton has joined #zope3-dev22:34
*** whit|lunch is now known as whit22:43
*** lucielejard has quit IRC22:57
*** greenman has joined #zope3-dev23:03
*** reco has quit IRC23:09
fcorreabrandon_rhodes: Just checked a new z3c.fomdemo checkout here. The only thing I needed to do was checking out z3c.form in the root of the buildout, everything is running out. Weird it's not working for you23:16
brandon_rhodesfcorrea: should the z3c.form be provided by an svn:external?23:17
brandon_rhodesfcorrea: My buildout has just succeeded, but returns four times the message "Download error: unknown url type: svn -- Some packages may not be found!"23:18
fcorreabrandon_rhodes: I think it is supposed to23:19
brandon_rhodesfcorrea: all of my other errors were because I needed to (on this Debian box, at least) "aptitude install libxml2-dev libxslt1-dev" and then, once those two packages were installed, "cd /usr/include; ln -s libxml2/libxml ."23:21
*** alecm has quit IRC23:21
*** alecm has joined #zope3-dev23:21
fcorreaMy steps: 1) svn co svn:// z3c.form, 2) cd formdemo; svn co svn:// z3c.form; 3) python 4) ./bin/buildout23:21
fcorreammm...I am not sure if I have those here. OS X here23:21
brandon_rhodesI didn't know to do step 2) when I started this morning.23:22
fcorrealemme check23:22
fcorreaYep, that's wrong. Will report23:22
brandon_rhodesThe compile errors that lead me to track down those XML libraries happened while compiling something called "src/lxml/etree.c"23:23
*** Ariel_Calzada has quit IRC23:28
fcorreabrandon_rhodes: just posted in zope3-users23:29
brandon_rhodesThanks, fcorrea23:33
*** reco has joined #zope3-dev23:33
*** ousado has joined #zope3-dev23:34
*** brandon_rhodes has quit IRC23:37
philiKONJ1m: could you please give me, philikon, pypi access to thx23:44
*** edgordon has joined #zope3-dev23:49
*** ousado_ has quit IRC23:51
*** benji has quit IRC23:52
*** ousado_ has joined #zope3-dev23:59

Generated by 2.15.1 by Marius Gedminas - find it at!