IRC log of #zope3-dev for Sunday, 2006-06-25

*** batlogg has quit IRC00:04
*** mexiGOOOL is now known as philiGOOOL00:46
*** jinty has joined #zope3-dev01:38
*** jinty has quit IRC02:29
*** oferw has joined #zope3-dev02:47
*** oferw has quit IRC02:53
*** oferw_ has joined #zope3-dev02:53
*** oferw_ is now known as oferw02:53
*** flox has quit IRC03:14
*** yota has quit IRC03:49
*** philiGOOOL has quit IRC04:02
*** oferw has quit IRC04:29
*** hazmat has quit IRC05:21
*** lordMiUlEr has quit IRC05:48
*** tarek has quit IRC06:43
*** Aiste has quit IRC06:44
*** Aiste has joined #zope3-dev06:46
*** jbb has quit IRC06:47
*** SteveA has quit IRC06:47
*** tav has quit IRC06:47
*** dunny has quit IRC06:48
*** d2m has quit IRC06:48
*** genconc has quit IRC06:48
*** vinsci has quit IRC06:48
*** zagy_ has quit IRC06:52
*** TrevorP has quit IRC06:52
*** clasux has quit IRC06:52
*** benji has quit IRC06:52
*** nouri has quit IRC06:52
*** dokai has quit IRC06:52
*** Aiste has quit IRC06:52
*** rocky has quit IRC06:52
*** srichter has quit IRC06:52
*** tonico has quit IRC06:52
*** povbot has joined #zope3-dev07:12
*** sets mode: +ns 07:12 *** Notice -- TS for #zope3-dev changed from 1151208756 to 115120871907:19
*** sets mode: -s 07:19
*** tonico has joined #zope3-dev07:19
*** clasux has joined #zope3-dev07:19
*** dunny has joined #zope3-dev07:19
*** zagy has joined #zope3-dev07:19
*** Aiste has joined #zope3-dev07:19
*** nouri has joined #zope3-dev07:19
*** TrevorP has joined #zope3-dev07:20
*** tav has joined #zope3-dev07:30
*** jbb has joined #zope3-dev07:31
*** srichter has joined #zope3-dev08:01
*** ChanServ sets mode: +o srichter08:03 *** Notice -- TS for #zope3-dev changed from 1151208719 to 107682172408:03
*** sets mode: -o srichter08:03
*** jukart has joined #zope3-dev08:03
*** vinsci has joined #zope3-dev08:03
*** d2m has joined #zope3-dev08:03
*** SteveA has joined #zope3-dev08:03
*** genconc has joined #zope3-dev08:03
*** dokai has joined #zope3-dev08:03
*** sets mode: +b *!*Albesure*@*.comcast.net08:03
*** changes topic to "logs available at || paste code examples into"08:03
*** ChanServ sets mode: +o tav08:06
*** genconc has quit IRC08:47
*** jukart has quit IRC08:47
*** dokai has quit IRC08:47
*** vinsci has quit IRC08:47
*** SteveA has quit IRC08:47
*** d2m has quit IRC08:47
*** genconc has joined #zope3-dev08:47
*** dokai has joined #zope3-dev08:47
*** vinsci has joined #zope3-dev08:47
*** SteveA has joined #zope3-dev08:48
*** d2m has joined #zope3-dev08:48
*** jukart has joined #zope3-dev08:48
*** batlogg has joined #zope3-dev09:39
*** baijum has joined #zope3-dev09:40
*** RaFromBRC has joined #zope3-dev09:51
*** RaFromBRC has quit IRC09:52
*** oferw has joined #zope3-dev11:24
*** flox has joined #zope3-dev11:29
*** oferw has quit IRC11:30
*** philiGOOOL has joined #zope3-dev11:37
*** batlogg has quit IRC11:49
*** batlogg has joined #zope3-dev11:54
*** baijum has left #zope3-dev12:19
*** batlogg has quit IRC12:38
*** batlogg has joined #zope3-dev12:43
*** yota has joined #zope3-dev13:25
*** dunny has quit IRC13:32
d2manyone knows when zope.browsertestrecorder was removed from trunk ? svn up left me with an empty folder and a configfile in etc/package-includes13:48
philiGOOOLd2m, i removed it yesterday13:55
philiGOOOLthe checkin message says you should remove some stuff13:55
floxabout skinning... what is the best practice on Zope3.3+ ?13:57
d2mphiliGOOOL: ahh ;) thanks13:57
floxis it better to use layers (the same as 'Rotterdam' skin with 'rotterdam' layer)13:58
philiGOOOLflox, ppl like using viewlets13:58
floxwhen u declare a <browser:page> or a <browser:ressource>, you put attribute (the IBrowserSkinType)14:00
d2manother question: i managed to run zope from wsgiref (simple_server), only oddity is, that "devmode on" starts unittests (maybe a sideffect of appsetup.config()) , "devmode off" works fine though14:00
philiGOOOLflox, yup14:00
philiGOOOLflox, or any other interface extendin gIBrowserRequest14:00
philiGOOOLflox, a layer doesn't have to be an IBRowserSkinType14:01
philiGOOOLd2m, weird14:01
floxphiliGOOOL: u mean it is better to use a class declaration like it is for the Rotterdam skin14:02
floxphiliGOOOL: class Rotterdam(rotterdam, IDefaultBrowserLayer):14:02
philiGOOOLclass declaration?14:02
philiGOOOLthere's no class created there14:02
philiGOOOLdon't let the 'class' statement fool you14:02
philiGOOOLthose are interfaces14:02
philiGOOOLand i don't understand you rquestion :)14:03
floxphiliGOOOL: factly, i saw examples where the pages and ressources are registered on a "IBrowserSkinType" directly14:05
philiGOOOLsure, you can do that14:05
floxphiliGOOOL: but when i see the Rotterdam skin, i see that the pages and ressources are registered on the layer 'class rotterdam(IBrowserRequest):'14:05
philiGOOOLit doesn't matter14:06
philiGOOOLyou decide what you want to do14:06
philiGOOOLif things are registered for rotterdam and Rotterdam extends rotterdam, things are also available for Rotterdam14:06
floxi understand that both examples work the same...14:07
floxi was wondering if one of this implementation was deprecated... or in the way to be deprecated14:08
floxi prefer the way implemented in "Rotterdam" skin... since all specific things are registered on an extra layer14:09
philiGOOOLno, the point is to give you flexibility14:10
philiGOOOLif you feel that you need the flexibility of having things on an extra layer, you can do so14:10
philiGOOOLif you dont' need that flexibility (and i would say that in most cases you don't), you can register things directly for the skin interface now14:10
philiGOOOLin fact, the way i look at it:14:10
philiGOOOLeverything extending IBrowserRequest is a layer14:11
philiGOOOLyou combine layers the way you want the overall layout to work14:11
philiGOOOLthen you select whichever layer is the one you want to "see", you turn it into a skin14:11
philiGOOOLin fact, you just register it with an additional interface14:11
floxthis is the 'class Rotterdam(rotterdam, IDefaultBrowserLayer):' line14:12
philiGOOOLthis just combines the layers14:12
philiGOOOLto another layer14:12
philiGOOOLrotterdam = layer14:12
philiGOOOLRotterdam = layer that combines rotterdam and IDefaultBrowserLayer14:12
philiGOOOLRotterdam is made a skin in ZCML14:12
floxthrough the  <intterface> directive14:13
philiGOOOL<interface interface='...Rotterdam' type='...IBrowserSkinType' name='Rotterdam' />14:13
floxok, i hv the point14:14
floxis there an easy way to have a custom login page? or an example?14:15
floxokok, i've found it on chapter 19...14:15
philiGOOOLchapter 19 is outdated, though14:15
philiGOOOLlook at
philiGOOOLyou want your own credentials plugin14:16
floxphiliGOOOL: i look for something similar to the 'Plone' login page14:16
philiGOOOLflox, tried the cookie auth plugin for PAU?14:17
floxphiliGOOOL: not yet...14:17
philiGOOOLtry it :)14:17
floxphiliGOOOL: i'll do, after reading source code14:18
philiGOOOLi'd try it first :)14:19
floxphiliGOOOL: i would like to hv different 'skins' for Anonymous access and Authenticated access...14:20
floxphiliGOOOL: is it easy to set up?14:21
philiGOOOLprobably best done in the credentials plugin14:21
philiGOOOLit's not very ahrd to do14:22
*** volvox has joined #zope3-dev14:26
*** stub has joined #zope3-dev14:34
floxdirective <browser:resourceDirectory> still exists?14:39
floxok, thks14:42
*** flox is now known as flox|away14:43
*** rocky has joined #zope3-dev14:44
*** volvox has quit IRC14:50
*** mexiKON has joined #zope3-dev14:54
*** mexiKON is now known as philiKON14:54
*** volvox has joined #zope3-dev15:10
*** philiGOOOL has quit IRC15:10
*** flox|away is now known as flox15:15
*** J1m has joined #zope3-dev15:38
*** jukart has left #zope3-dev16:10
*** batlogg has quit IRC16:34
*** hannosch has joined #zope3-dev16:47
*** volvox has quit IRC17:10
SteveAthanks srichter or philiKON :-)17:12
philiKONSteveA, you're welcome :)17:12
*** philiKON has left #zope3-dev17:13
*** philiKON has joined #zope3-dev17:13
*** volvox has joined #zope3-dev17:14
*** jinty has joined #zope3-dev17:28
floxfor the  <browser:resource> directive17:37
floxall resources are relative to current directory, by default..17:37
floxor to current object17:37
floxhow i can change the directive so that the resource are lookup at the same place for every object?17:38
floxex: if i visit, the icon is requested with URL
floxthen when i visit, the icon is requested with URL
floxi want to do that to improve caching.17:42
floxafter other tests, i see i a m wrong17:46
*** stub has quit IRC17:47
philiKONflox, resources are looked up at the nearest site17:47
floxfactly, it is weird...17:47
floxphiliKON: yes, i just figured out this thing17:47
floxphiliKON: except the 'pl.gif' and 'mi.gif' that are looked up at the root of the instance17:48
philiKONgotta, go, sorry17:50
*** epx has joined #zope3-dev17:50
floxat first i was lost, bec i set up 2 sites, and i did not  understand the logic. factly, i htink the behaviour is ok.17:51
floxto implement a skin selector, how i can build the list of available skins?18:02
*** flox is now known as flox|away18:09
*** hannosch has quit IRC18:10
*** oferw has joined #zope3-dev18:18
*** oferw has quit IRC18:59
*** alecm has joined #zope3-dev19:16
*** chen has joined #zope3-dev19:57
chenhow about ajax in zope3.3?19:58
*** zagy_ has joined #zope3-dev20:07
philiKONchen, what about it?20:07
chenis there any ajax capability builtin zope3.3?20:09
chenhi , philikon,20:09
chenor some official ajax package support zope3.3,please tell me ,thanks20:11
*** benji has joined #zope3-dev20:13
philiKONchen, no, no official ajax package20:13
chenok, thanks20:15
*** chen has quit IRC20:15
*** romanofski has joined #zope3-dev20:15
*** chenqiong has joined #zope3-dev20:16
*** zagy has quit IRC20:19
*** chenqiong has left #zope3-dev20:19
*** chenqiong has joined #zope3-dev20:20
*** batlogg has joined #zope3-dev20:27
*** sashav has joined #zope3-dev20:30
volvoxI've seen on the ancient proposal about "no owner role" that the core devs think it's important for principals to have only one role at a time. Is it still so?20:38
*** dobee has joined #zope3-dev20:41
philiKONvolvox, no20:43
philiKONvolvox, principals can have more than 1 role20:43
volvoxphiliKON: i know how it's implemented. i've also read an old philosophical stance by Jim about that, and didn't understand the reason  :-)20:48
volvox"Furthermore, having a role like this does not really fit into the Zope 3 security model, since we would like a user to have only one role at a time."20:48
philiKONvolvox, note that roles are just part of a particular (default) security policy20:49
philiKONthere is another security policy that doesn't use roles for zope3 (zc.sharing)20:49
volvoxok. i'm happy to know i'm doing nothing wrong, tnx20:50
*** mkerrin has joined #zope3-dev21:05
*** chenqiong has quit IRC21:15
*** chenqiong has joined #zope3-dev21:20
*** dobee has quit IRC21:40
*** flox|away is now known as flox21:52
*** enterloper has joined #zope3-dev22:00
enterloperHi all, if there is a kind soul with knowledge of Zope3 authentication machinery (esp. sessioning) around here, i am a developer working on an extension of zope who could really use a few good pointers! mostly in regards to the publisher module under z3. in particular, when in the publish function, IClientIdManager seems not to be initialized without a request.traverse into the ZODB. why would this be? (have a few peripheral22:01
philiKONenterloper, your msg is too long, got truncated22:01
enterlopersorry, let me try again..22:02
philiKONbreak it up into specific questions22:02
enterloperin regards to the publisher module under z3. in particular, when in the publish function, IClientIdManager seems not to be initialized without a request.traverse into the ZODB. why would this be? (have a few peripheral guesses)22:02
enterloperalways gotta have an intro sentence :p22:02
enterloperpardon if it's not a square question. I'm playing in the publisher function in zope/publisher/publish.py22:03
philiKONhmm, why?22:03
enterloperan experimental extension of zope22:03
philiKONuh huh22:04
enterloperthat does things that would make you upset, but it's fun and exciting22:04
philiKONclient id managers are typically persistent22:04
philiKONat least the one for HTTP clients is22:04
enterloperand so it's a proven thing in zope2 and i am trying to move what i have to zope 322:04
philiKONhence, you'll need to find sites first where a client id manager is stored22:04
philiKONin order to be able to look it up22:04
enterloperright. thought so - can i make a few statements, then see what ya think? here goes22:05
enterloperwhat i'm doing is testing request.get('PATH_INFO') and IF it's a url i want to munge - i divert it into another system22:06
enterlopernow when in zope2 i did this, the credentials (user/session/etc) for the current user didn't exist until (i found out trial/error) AFTER i did a request.traverse22:06
philiKONenterloper, why not implement this as a wsgi middleware?22:06
benjienterloper, you might be better off with wsgi22:06
philiKONbenji, :)22:06
philiKONhey benji22:07
benjihey philiKON22:07
enterloperi will look, hope to get this out first22:07
enterloperto you guys22:07
philiKONwe'll probably just ocntinue to tell you that you're hooking into the wrong places22:07
enterloperIm cool with that. here is what i did in zope2 -22:07
enterloperdummyObj=request.traverse('', validated_hook=validated_hook)22:08
enterloperafter i did this - i had good auth info from the security machinery in zope2 - i suppose this is because it was acquring from a toplevel user folder object22:09
enterloperin zope3 it looks almost the same22:09
enterloperbut now i have to traverse to a an actual object somewhere before it will pick this info up22:10
enterloperso having looked over all the code..22:10
enterloperat least alot of it :)22:10
enterloperi was wondering how one would 'fake' it - how could i get authenication info setup for a request (against perhaps a preset manager in the instance)22:11
enterloper(whatever object does that biz)22:11
philiKONwhat's your (high-level) use case?22:11
enterloperso that i can have my cake and eat it too - diversion to my system - authentication from zope22:11
philiKONaha. you want authentication from zope but stuff come from another system?22:12
enterloperyeah -- and previously with zope2 i did this - all in the same process which is very cool and important to the project22:12
enterloperthat 'other system' is really just a python module i load into the namespace22:13
enterloperso imagine a url '/1/2/3' and the system is set up to recognize 1 as my area - then the response to 1/2/3 comes from me and returns.22:14
enterloperelse 2/3/4 comes from zope as normal22:14
philiKONi think hacking is not a good option22:15
enterloperbut when i 'divert' in the module, something about calling request.traverse is what finally fills out the session and client info (cookies and all) and i wanted to find out how to fake it22:15
philiKONtry a custom traversal adapter22:15
*** sashav has quit IRC22:15
*** sashav has joined #zope3-dev22:15
enterloperin my mind it is good because i want to filter all outgoing content - and it all leaves out of there. it's a fulcrum point.22:16
enterlopera custom traversal adapter is new news to me.22:16
enterloperthat is filter all outgoing html, actually.22:16
philiKONyou either want a wsgi middleware, a custom traversal adapter, or a combination22:17
philiKONthe hacks you're trying to do sound just too wild22:17
enterloperthe funny thing is that this actually works man22:17
enterloperi've been using the exact same thing in zope2 for quite a while22:17
enterloperworks well.22:17
philiKONgood for you22:18
philiKONi wouldn't advise the same thing in z3 though22:18
philiKONbecause there are just simpler measure22:18
enterloperhmm gotcha.22:18
enterlopernotes to take are wsgi and custom traversals huh? the idea of course is that i did this in 4 lines the old way, but if there is 'simpler' in z3 i'm all about it.22:19
philiKONwell, it might be more lines22:19
philiKONbut it'd be less brittle i guess22:19
philiKONplus, i'd be surprised if it was just 4 line sin z3 :)22:19
philiKONafter all, if the hack worked in zope 2, who is to say that it still works in z322:20
enterloperi do notice that it probably worked in z2 cause it just acquired a auth source from the top level - but not in z3 cause it needs a site folder -22:20
enterloperi already have it working half way in z3 - all i'm missing is the step where auth info in created and wished to know where how to force caling a auth source in a site folder to trick it22:21
philiKONauth is done during traversal22:21
enterlopernow - if i have a site22:21
philiKONsearch for maybeplacefullyauthenticate or so22:21
enterloperhow would i traverse to it with out having an actual request to that area - that's what's trickin me22:22
enterloper(i had looked there previously, good long night with that code)22:22
philiKONyou need to get hold of the root object22:22
philiKONroot database object22:22
philiKONwhich isn't the root folder, though22:22
philiKONthe root db object contains the root folder22:22
philiKONthe root folder usually is a site22:22
enterloperi saw that in the maybeauthplacefully -22:23
enterloperi can dupe that, then take that object and traverse to it perhaps?22:24
philiKONsorry, need to get back to work...22:24
enterloperfine lead anyway22:24
enterlopertotally perverse!22:36
enterloper    publication.beforeTraversal(request)22:36
enterloper    object = publication.getApplication(request)22:36
enterloper    object = request.traverse(object)22:36
enterloper    publication.afterTraversal(request, object)22:36
enterloper    interface.divert(request)22:36
enterloperworks thanks philiKON22:37
* philiKON shivers22:37
enterloperyessir i know 20 pushups right now22:37
*** chenqiong has quit IRC22:37
*** romanofski has quit IRC22:53
*** enterloper has quit IRC22:59
*** zagy_ has quit IRC23:05
*** dunny has joined #zope3-dev23:08
*** flox has quit IRC23:18
*** flox has joined #zope3-dev23:19
*** alecm has quit IRC23:21
*** RaFromBRC has joined #zope3-dev23:43
floxabout the Rotterdam skin, again23:47
floxthere's 2 templates, '' and  ''23:49
floxbut 2nd one is not  used23:49
floxis it because it is buggy, or other reason?23:50

Generated by 2.15.1 by Marius Gedminas - find it at!