IRC log of #zope3-dev for Friday, 2005-12-16

*** klaus_ has joined #zope3-dev00:02
*** elro has quit IRC00:13
*** sashav_ has joined #zope3-dev00:22
*** sashav has quit IRC00:22
*** sashav_ is now known as sashav00:22
*** klaus_ has quit IRC00:41
srichterprojekt01: barely :-)00:41
*** GaryPoster has quit IRC00:46
projekt01srichter, do you know that the bugtracker depends on the old pluggableauth service?00:52
*** andres__ has joined #zope3-dev00:55
*** andres__ is now known as andres00:55
srichterprojekt01: could be00:57
srichterno clue00:57
srichterI have not worked on it for a long time00:57
projekt01bugtracker uses the getPrincipals method in the user vocabulary which doesn't exist in PAU00:58
projekt01srichter, does SchoolTool have a widget or vocabulary for lookup principals?01:00
srichterwell, we can always get to the ST application and from their to our persons01:01
srichterI think we have a utility that uses our persons to provide a principal source01:01
srichterre bugtracker: Oh, I see; right it should use the source instead01:03
srichterbtw, I like ST's approach in this respect01:04
srichterwe can alsways get to the application root and go from there01:04
srichterit's very easy01:04
projekt01And what's about local PAU?01:05
projekt01Ah, you mean the site (application root)01:05
srichterwe don't care about the local PAU that much, since we know that all people provided by the local PAU are also listed in the person container01:07
projekt01person container? You mean in ST?01:07
projekt01Ok, it really turns me crazy to use this queriables.01:08
*** andres_home has quit IRC01:09
projekt01Now I go to implement different principal widgets and vocabularies which will render a better UI for selecting principals.01:09
srichteryeah, I don't quite understand them yet either01:09
srichterI think for small amounts of users we will eventually want to support listing them all01:10
srichterand have a JS-based filter or something01:10
*** ruda_porto has quit IRC01:12
projekt01I think of a tree like UI where all IAuthentication utilities are listed and principals as child nodes. This would make it expandable.01:12
*** J1m has quit IRC01:29
*** dman13 has quit IRC01:45
*** philiKON has joined #zope3-dev01:51
*** sashav has quit IRC01:57
*** srichter has quit IRC01:58
*** jinty has quit IRC02:06
*** zbir has joined #zope3-dev02:24
*** newpers has joined #zope3-dev02:28
newpersphiliKON:  thank you so much for you post --
*** niemeyer has quit IRC02:30
*** ruda_porto has joined #zope3-dev02:36
*** philiKON has quit IRC02:41
*** GaryPoster has joined #zope3-dev02:44
*** srichter has joined #zope3-dev02:47
*** dman13 has joined #zope3-dev02:53
*** ChanServ sets mode: +o srichter03:10
*** GaryPoster has quit IRC03:24
*** GaryPoster has joined #zope3-dev03:24
*** andres_ has quit IRC03:27
*** newpers has quit IRC03:37
*** projekt01 has quit IRC03:51
*** ruda_porto has quit IRC03:56
*** jfmoxley has joined #zope3-dev05:09
jfmoxleychanges in my css should be reflected immediately, and not require a zope restart, right?05:10
*** GaryPoster has quit IRC06:06
*** adamSummers has joined #zope3-dev06:35
*** dobee has joined #zope3-dev08:04
*** dobee has quit IRC08:15
*** dobee has joined #zope3-dev08:39
*** hazmat has joined #zope3-dev08:57
*** rom|zzZ is now known as romanofski08:59
*** romanofski has quit IRC09:21
*** Theuni has joined #zope3-dev09:24
*** natea_ has quit IRC09:41
*** tarek has joined #zope3-dev09:43
*** romanofski has joined #zope3-dev09:56
*** sashav has joined #zope3-dev10:14
*** stub has joined #zope3-dev10:15
*** xenru has quit IRC10:21
*** xenru has joined #zope3-dev10:22
*** yota has quit IRC10:26
*** MJ has quit IRC10:39
*** povbot` has joined #zope3-dev10:47
*** povbot has quit IRC10:47
*** ignas has joined #zope3-dev10:50
*** tahara has joined #zope3-dev10:51
*** MJ has joined #zope3-dev11:08
*** agroszer has joined #zope3-dev11:12
taharaHi.I want to join bugday tommorow, but I don't know how to find important issue in collector...11:13
taharaDoes anyone where is todo-list or something?11:19
*** mgedmin has joined #zope3-dev11:23
*** hazmat has quit IRC11:34
*** bradb has joined #zope3-dev11:50
*** projekt01 has joined #zope3-dev11:53
Theunitahara:  important issues in the collector are marked "critical"11:55
Theuniif there are no critical issues, you can just move on to "urgent" or "normal" bugs.11:56
Theuni(for me, the bugday is already today ;))11:56
taharaI see, thanks!11:57
taharaI'll try.11:57
taharaAh, the bugday is today for me too.12:00
*** bradb has left #zope3-dev12:04
SteveAis there a bug day happening?12:06
SteveATheuni: is the bug day today?12:08
Theunilooks like it12:10
Theunii just noticed it this morning that stefan send an announcement yesterday12:10
Theunii'll try to spend some time later fixing some stuff12:10
* Theuni works fulltime on the common criteria now12:11
*** philiKON has joined #zope3-dev12:12
SteveAthat's dedication12:12
projekt01Theuni, did you see my answer about Residual Information Protection?12:20
Theunidid you see mine? :)12:25
projekt01Yes, it's a answer to yours.12:25
Theunii mean: did you see my answer to your answer? ;)12:25
projekt01Theuni, Ok, I see, good usecases. I think it's a good idea to use a identifier for principals and make sure that they get information from previous settings because of the naming.12:31
projekt01We really have o add something like that.12:31
Theunii'll make more in-depth proposal for that in january12:32
projekt01Your question about sublocation: Take a look at dispatchToSublocations()12:33
projekt01Theuni, sounds good to me.12:34
* Theuni looks12:36
projekt01Theuni, btw, are you interested to join a sprint early next year in switzerland or germany?12:36
Theunithe answer is: maybe12:37
Theunii'm working on the certification during the whole of january12:37
Theuniif it's in there, i might have time12:38
Theuniotherwise it's likely that i'm in some customer project12:38
Theuniand i'll be at pycon in february ...12:38
agroszerprojekt01: when, where?12:38
projekt01agroszer, we don't know the date right now, we only decide to hold sprint. Perhaps together with the Neckar sprinters.12:40
projekt01agroszer, are you interested?12:40
agroszersorry, I'm a bit sleepy right now :-)12:40
agroszeryes, of course12:40
*** philiKON has quit IRC12:41
agroszerI'd like to improve my Z3 knowledge12:41
agroszerand worst case latest then we could do the workflow brainstorming12:42
projekt01agroszer, yes, I think that's a good sprint topic.12:42
projekt01agroszer, btw, where are you living?12:43
agroszerdoes Plattensee tell you something?12:43
Theunii wonder whether dispatchToSublocations has to be called explicitly although it is subscribed to the IObjectMovedEvent?12:44
projekt01germany or austria, I'm not sure12:44
agroszerlake Balaton is in Hungary (for the English)12:44
projekt01Ah, ok12:45
projekt01Theuni, it's implicit called as a event subscriber.12:45
Theuniprojekt01:  i see ... so why does the doctest call it explicitly then?12:46
Theunishouldn't it just do event.notify ?12:46
*** alga has joined #zope3-dev12:48
projekt01Theuni, doctests do not use the right setup most the time. (which I really dislike)12:49
Theunido you mean "right" or "real" or "complete" ?12:49
Theunibecause if i make him use notify, the result is slightly different ...12:49
projekt01I mean the same setup like used via the ZCML registration.12:50
*** andres has quit IRC12:50
*** andres has joined #zope3-dev12:51
*** mkerrin has joined #zope3-dev12:54
*** mkerrin has quit IRC12:59
*** faassen has joined #zope3-dev13:04
*** efrerich has joined #zope3-dev13:12
*** philiKON has joined #zope3-dev13:16
Theunithat would be too much for unit tests13:17
Theunithat's what functional tests are for imho13:17
*** elro has joined #zope3-dev13:21
*** mkerrin has joined #zope3-dev13:22
mgedminI think I just found a bug in the <viewlet> directive handler13:22
mgedminif you specify only the template, but do not specify class, extra attributes are not passed to SimpleViewletClass()13:22
*** xenru|clone has joined #zope3-dev13:30
*** philiKON has quit IRC13:30
*** xenru|clone has quit IRC13:32
*** xenru|clone has joined #zope3-dev13:34
projekt01mgedmin, is there a need for extra attributes for the default class?13:34
mgedminprojekt01, there is13:36
mgedminthe viewlet manager may use them13:36
mgedminfor sorting13:36
mgedminthe page template may use them13:36
mgedminmy use case was the sorting one13:37
projekt01Ah, right13:37
projekt01We where thinking of a out-of the box weight support for viewlets, but didn't implement it.13:38
mgedminit would make sense, I think13:39
projekt01Perhaps we can use a different default class with weight support if the weight attribute used.13:39
projekt01...if no custom class is given13:39
mgedminI think relying on __cmp__ is not the answer13:39
mgedminwhat if you get two viewlets, where one defines __cmp__ to look into weight attributes (if they exist)13:40
mgedminand the other defines __cmp__ to look into title attributes (if they exist)13:40
*** ignas has quit IRC13:40
mgedminwhich one gets used for which comparison?13:40
mgedminit's a matter of accident, I think13:40
mgedminI'd prefer if there was a OrderedViewletManager defined in zope.viewlet13:40
projekt01+1, me too13:41
mgedminthere is something very much like it in the boston skin, isn't there?13:41
projekt01and support a weight in the viewlet (required=False) directive13:41
projekt01Yes, I think I implemented this in the Boston skin, but didn't implement a generic pattern for the viewlet directive13:42
mgedminthis is what I wrote for Schooltool:
mgedminactually, 'weight' would be a better name than 'order'13:43
mgedmin"heavier" viewlets would sink below "lighter" ones13:44
projekt01Yeah, cool, this is what we need. I guess we can use such a class if a weight attribute is defined as default if no custom class is given.13:44
mgedminI'm not sure I follow13:45
mgedminthat class is the viewlet manager13:45
mgedminthe order/weight attribute would be defined in a viewlet13:45
*** philiKON has joined #zope3-dev13:45
mgedminyou cannot choose a default class for the manager by looking at viewlets registered for that manager13:45
projekt01Ah, and the weight (additional attribute) isn't dispatched to the manager right now?13:46
*** xenru has quit IRC13:47
*** tav has joined #zope3-dev13:47
projekt01sorry, if I get it right, the viewlet defines a weight, and the manager sorts on this attribute. How do you make sure that the weight get added to the viewlet?13:47
mgedminlook at the code13:47
*** J1m has joined #zope3-dev13:47
mgedminif there is no weight attribute, I sort by title13:47
mgedminif there is neither weight nor title, I raise an exception and tell the developer the name of the viewlet13:48
mgedminso that the developer can fix his zcml13:48
projekt01Ah, I got it, the viewlet directive supports extra (custom) attributes13:48
mgedminonly that support is a little bit broken13:49
mgedminand I just fixed it13:49
mgedminand wrote a test13:49
*** graham has joined #zope3-dev13:49
mgedminshould I mention the bugfix in CHANGES.txt as well?13:49
*** alga has quit IRC13:50
projekt01Cool, yes a note about this would be good, perhaps srichter is looking for this since we didn't implement it.13:50
mgedminpatch here:
mgedminI have to run now13:51
projekt01mgedmin, Thanks a lot13:51
mgedminI can commit this after I come back13:51
mgedmincan I backport this to the 3.2 branch as well?13:51
projekt01I guess so13:51
*** mgedmin has quit IRC13:52
*** ruda_porto has joined #zope3-dev13:53
*** xenru|clone has quit IRC13:53
*** xenru has joined #zope3-dev13:54
*** elro has quit IRC14:01
*** philiKON has quit IRC14:09
TheuniJ1m: what is the idiom "break glass"?14:28
J1mFor example, here many public buildings have manual fire alarms that require breaking a glass cover before they are pulled.14:30
Theuniah right.14:30
J1mDo you have these in Germany?14:31
Theunii don't remember a special name for the glass itself14:32
J1mHere, these thinsg typically have signs with them that say: "in case of emergency, break class".14:34
J1mThus the idiom.14:34
* J1m really shouldn't use idioms in these sorts of discussions.14:34
tarek"in case of emergency, break class" nice quote to insert in a module header :-)14:41
*** benji has joined #zope3-dev14:41
Theuniah. J1m, we have this idiom too, but we don't use it as a noun :)14:41
TheuniJ1m: Hmm. Is it possible you want to solve two issues with the ZODB permission storage?14:45
* J1m tries never to be unduly constrained by limits of grammer, spelling or vocabulary. :)14:45
*** BjornT has quit IRC14:46
J1mNot in practice. Just dreaming out loud.14:46
TheuniJ1m:  i see that point one and two would work for a while. Not beeing constrained by vocabulary would be ... kind of very isolating.14:47
*** BjornT has joined #zope3-dev14:47
J1mNo, it's liberating. ;)14:48
Theunitwo sides of one coin ;)14:49
projekt01Jim, is the global principal registry queryable?14:52
*** dobee has quit IRC14:54
projekt01does the pluggableauth package get removed sometime?14:54
J1mI'm not sure what you men.  Any authentication system is optional.14:55
* Theuni needs to get packaging stuff for the office move next week ...14:55
projekt01J1m, what's the generic way to get principal's information (for build a vocabulary) which will work for all IAuthentication utilities14:56
J1mThe generic way, is to use a search-and-select UI.14:57
J1mThe most generic way, if to delegate to each principal source to provide it's own search-and-select, as I did. This tends to provide an unusable UI.14:58
J1mIMO, the best way, is to assume that there is a simple string-bases searching API, providing adapters for eacg principal source.14:59
J1mThen you can provide a simple string-based search-and-select UI.14:59
J1mIMO, this is best if it works,14:59
projekt01I agree on the unusable UI, I like to implement a different UI, but it's to hard to get the right principal information right now.14:59
projekt01Can we support a simpler API for get the principal information?15:00
J1mWe've had customers that wanted a much richer UI for serachingg their specific database. Then we've provides a specific UI for them.15:00
J1mprojekt01, I don't think we need one.15:01
J1mI've been able to provide a simple UI based on what's there now.15:01
projekt01Why do we not provide a method getPrincipalInformation at the IAuthentication level like PrincipalFolder is implementing?15:04
projekt01This would make lookup prncipal information much easier15:04
J1mWhat information?15:05
J1mThis cannot be standardized.15:05
*** MJ has quit IRC15:06
J1mThe best way, IMO, is to provide access to a searchable string.15:06
J1mThis can be done via adaptation.15:06
projekt01what do you mean with a searchable string?15:07
*** MJ has joined #zope3-dev15:07
J1mA string that could be matched by searches.15:07
projekt01which searches?15:08
J1mSearches for principals.15:08
projekt01Ah, I have a different use case15:09
projekt01I like to provide a vocabulary containing all principals from a IAuthentication utility. There is no need for search because the result has to contain all principals15:10
J1mOK, I don't have that use case.15:10
projekt01This uscase is only to implement via the search queryable part right now. Which is very bad for performanc15:11
J1mIn any production application I dal with, there are too many principals to present alll of them.15:11
J1mIn any production application I deal with, there are too many principals to present all of them.15:11
projekt01We provide only a searchable API right now for IAuthentication. What's about if I like to implement a different search on full results?15:12
J1mThen you are missguided.15:12
projekt01Do you mean I should not do this?15:13
J1mIt certainly should not be part of IAuthentication.15:13
J1mOr IPrincipalSource.15:14
*** philiKON has joined #zope3-dev15:14
J1mI would not object to IIterableAuthentication or IIterablePrincipalSource, but I would absolutely not require or expect all authentication or principal source utilities to also provide this interface.15:15
projekt01I guess I will implement a adapter where is able to get all principal informations like title, login and id where every future principal must support.15:17
projekt01And use this adapter for my need. Then I can implement this adapter for future IAuthentication implementations15:17
projekt01So then I can implement a tree like UI for IAuthentication15:18
*** philiKON has quit IRC15:18
*** andres has quit IRC15:18
projekt01It's really to complex the built in search concept for straight forward concepts.15:19
J1mIt's actually not that complex at all if you go with a simple string search.15:19
J1mThe user has a single text area to type in a string.15:20
J1mNote that if you don't have many users, you can even arrange that typing in an empty strings gets all users.15:20
J1mI agree that the UI I provided in Z3 sucks.15:22
J1mI was badgered (by someone at ZC) into overgeneralization.15:23
J1mNote that I plan to release the simple text-bases searaching facility.15:23
J1mWe've been planning to release our "Sharing" authorization system, which includes this, for some time.15:24
projekt01Note the batching is also a concept for break down large results. Search is not the only way.15:25
*** dobee has joined #zope3-dev15:25
J1m(It's been hard to find time to get this released, especially since there have been a number of dependency issues to sort out and we've continued making enhabcements to it.)15:25
faassenhm, I am getting this error and I'm at a loss on how to debug happens somewhere deep when I call a page template...15:26
faassen  File "/home/faassen/working/doclib/z31/lib/python/zope/interface/", line 261, in unsubscribe15:26
faassen    del self.dependents[dependent]15:26
faassen  File "/usr/lib/python2.4/", line 216, in __delitem__15:26
faassen    del[ref(key)]15:26
faassenKeyError: <weakref at 0xb4374b1c; to 'LocalSurrogate' at 0xb431764c>15:26
J1mprojekt01, no, batching doesn't solve the problem at all.15:26
faassenthat't the end of a long traceback. anyone have a clue? I just extracted a large piece of code from the document library (which doesn't ahve this)..15:26
faassenbut it seems to have something to do with authentication.15:26
faassenhm, wait, no, it seems to have to do something with language support.15:26
faassen    target_language = negotiator.getLanguage(langs, context)15:27
faassen  File "/home/faassen/working/doclib/z31/lib/python/zope/i18n/", line 40, in getLanguage15:27
faassen    envadapter = IUserPreferredLanguages(env)15:27
faassenis somewhat higher up..15:27
faassenthat gets triggered by the ZPT i18n engine, but I don't do any i18n.15:27
faassenin fact, there's no teven a page template registered by my app for what I try to view. ugh.15:27
J1mfaassen, this seems vaguely familiar.15:27
J1mWhat version of Zope?15:27
faassenwell, that's what I was hoping for. 3.115:28
faassenlet me check the exact version, I think 3.1 proper though.15:28
projekt01J1m, ok, thanks for the hints15:28
faassenyup, 3.1.0 proper.15:28
faassenso somehow it's not doing proper in some kind of fallback default page tempalte.15:29
faassenas it works fine in case I do provide my own templates. no big deal for me right now, it's just rather scary.15:29
faassenlocalsurrogates? weak references? ugh. :)15:29
faassengoes back to the theme that the page template engine should not do any calling into the Zope api.15:30
faassenJ1m: hey, you should go beyond one entry in your blog. :)15:31
faassenJ1m: philipp is outracing you!15:31
faassenJ1m: did you see the recent back and forth between me and Paul and Ian Bicking and Jeff Shell and such?15:33
*** GaryPoster has joined #zope3-dev15:33
*** zbir has quit IRC15:36
benjifaassen, I did, interesting discussion15:36
faassenJ1m: I'll dump some links on you.15:38
faassenPaul's fairy tale ('in the beginning, the world was nice')
faassenmy followup15:38
faassenIan Bicking's followup to mine15:38
faassenJeff Shell's followup to the discussion15:39
*** MrTopf has joined #zope3-dev15:39
faassenand my followup to myself :)15:39
faassenbenji: any thoughts?15:40
benjistrangely enough, I pretty much agree with everyone :)15:40
faassenbenji: well, me too. :)15:41
faassenbenji: I don't think we're all on different pages.15:41
faassenbenji: just different perspectives.15:41
*** GaryPoster has quit IRC15:41
benjiZ3 is /so/ nice for "pure Python" development, OTOH, I think it's very important to empower "power users" to extend systems we build in well defined, but useful ways15:41
faassenbenji: I just thought this was a nice opportunity to create some community awareness of this issue. it sort of got off the radar, then recently Plone people wanted to know where customization was gone in Five.15:41
faassenand I thought we'd better step back for a bit.15:41
faassensince my blog is now also on lots of plone people are hopefully seeing this stuff.15:42
benjiI feel like there is some mechanism that'll enable this, but not sure exactly what it is15:43
faassenI've been vaguely considering the following simple mechanism.. (in the context of five, but applies to z3 as well)15:43
faassen(include lots of handwaving please in this discussion)15:43
faassenlike, someone registers a page template through some zcml directive. or a resource like a css file.15:44
faassennow, customizer comes in, and says, this is great, I just want to tweak it here and there.15:44
faassenso they go to this web interface.15:44
faassenand there's a list of all the templates registered in the package..handwave how we find this out.15:44
faassenand he clicks on it, and he sees the thing in a textarea.15:44
faassenand he can edit it. if he saves, the page template is overridden, that is, for that site..15:44
faassenthe page template is just registered in memory like the ones are now.15:45
faassenno need to store the compiled version in the ZODB or anything.15:45
faassenso the registration is in fact the same.15:45
faassenit's just finding the code in some other place.15:45
faassenI think that limits all kinds of growths, as you could present this as 'you are now modifying component X'15:46
faassenand if you want to move this to filesystem development, you simply dump that text to the filesystem and register that instead.15:46
faassenbut there's already a lot of tweakability.15:46
faassenanyway, it's not a very great idea, but it might be controllable while still giving enough power.15:46
faassenat least not whole skin directories are being copied, no python code is copied, etc.15:47
faassenand the page template code is the same as it'd be on the filesystem.15:47
faassenso it's easy to move it, then refactor it, etc.15:47
faassenthen again, it might not be too different from the CMF skin model after all.15:48
benjiit sounds like the basic mechanism would be TTW changes of component registration, with some nice UI to help specific tasks like ZPT editing15:48
J1mI love this discussion, this topic, and those links.15:48
J1mUnfortunately, I don't have time to have this discussion or read all of these links now.15:48
J1mI fear there is going to be a lot of work necessary to get the next beta out this weekend.15:49
J1mThe most important thing facing us ATM is succeeding at time-based releases.15:49
J1mI'm working on twisted issues.15:50
benjiJ1m, is it preferred that bug fixes be applied to the 3.2 branch first and the merged to the trunk, or the other way 'round?15:50
*** _anguenot has joined #zope3-dev15:51
J1mIt would be helpful if someone would go through the new issues and decide if any are critical, so that we know what we have to work on.15:51
faassenJ1m: sure, read the stuff later.15:51
J1mbenji, the right answer is probably 3.2 first, but I don't think it matters at this point.15:51
J1mfaassen, will do.15:52
*** zbir has joined #zope3-dev15:52
J1mbenji, for some reason, I prefer working on the trunk and porting from there.15:52
J1mI should probably get over that.15:52
*** mkerrin has left #zope3-dev15:53
J1mIf we every start using svnmerge, then it will be compelling to do bug fixes on the release branch and use svnmerge to bring them to the trunk.15:53
*** J1m changes topic to "Bug day today! | logs available at"15:53
J1mAny volunteers to classify the issues?15:54
faassenhm, how would I do that?15:55
faassenI'm a bit scared of the issue tracker.15:55
J1mGo through the issues posted since the beta.15:56
J1mMark the onces that need to be fixed as critical.15:56
faassenis there an easy way to get the issues posted since the beta?15:56
faassenalso, I'm not sure my user (faassen) can mark anything.15:57
faassenI don't know whether I have the rights.15:57
J1mI would jump to the end and work my way backward.15:58
faassenokay, I'll do that.15:58
J1mYou are now a supporter. Congratulations!15:58
faassenJ1m: whee. :)15:58
*** mkerrin has joined #zope3-dev15:58
faassenafter all these years, I support Zope! :)15:58
* faassen tries to find the little twiddly buttons to mark issues that should appear.15:59
J1mWhen viewing an issue, to change it's importance, use the edit link.15:59
* faassen tries to find the 'edit' link. hm..16:00
J1mat the top of the page, next to follow up etc.16:00
faassenyeah, didn't find it. logging out and in, see whether that helps.16:00
J1mMaybe I need to make you a manager, but I didn't think so.16:00
J1mshoudkn't matter.16:00
J1mYou have to be looking at an issue.16:01
faassentrying again.16:01
faassenlogging out and in shouldn't matter, but somehow it did matter. dunno, some caching going on.16:02
faassennow I have an 'edit' link.16:02
* J1m really away for a while, breakfast16:03
*** zbir has quit IRC16:03
*** BjornT has quit IRC16:17
J1mah, faassen, wrt to the traceback16:23
J1mThis is fixed in 3.2. :)16:23
J1mIt happens when the same specification appears more than once in a specification's bases.16:23
J1mfor example, using alsoProvides more than ince with the same interface.16:24
benjijust so I don't step on anyone's toes, I'm working on XXX cleanup16:42
*** mgedmin has joined #zope3-dev16:46
benjiJ1m, there is an XXX note on IFormAPI's validate method that it should be renamed something like "sumbitted"; here's your chance16:46
J1mHey, you are trying to credit me with a non-existent misspelling,16:48
J1mI don't care.16:49
benjijust asking if you really want it changed or not16:49
J1mI don't remember typing that xxx16:49
benjioh, I guess that answers that16:49
*** dobee has quit IRC16:53
*** fermigier has quit IRC16:55
faassenJ1m: sorry, I just got phone call etc, so haven't been able to do much with issuse so far.17:17
J1manything you do would be appreciated.17:17
J1mAre you going to PyCon? :)17:17
faassenJ1m: no, unlikely.17:18
faassenwhy is everyone asking me whether I got o PyCon, SteveA was too.17:18
*** romanofski has quit IRC17:18
faassenI just noticed on the silva channel, an hour before.17:18
J1mBecause we're all going to sprint and wish you could join us.17:19
faassenwell, I'll think it over.17:19
J1mAnd Dallas is so close to Rotterdam17:19
faassenwasn't planning on going.17:19
faassentrue, very close.17:20
*** BjornT has joined #zope3-dev17:20
faassenI've actually been in dallas. :)17:20
*** ignas has joined #zope3-dev17:20
faassenI've been all over Texas. Houston, Austin, Dallas.17:20
J1mI wasn't planning to go either, but people kept asking me about sprinting there. :)17:20
* faassen grins.17:20
mgedmindoes anyone else see a failure in test testNotFound (
benjinot me, mgedmin17:22
mgedminI'll revert my bugfix and rerun the tests then17:22
*** zagy has quit IRC17:28
mgedminI see that failure in a clean Zope3 trunk checkout17:28
J1mdoes the test fail when run by itself?17:30
mgedminpython2.4 ./ -f -pv -s
*** zagy has joined #zope3-dev17:31
J1mI'll try in a minute17:31
mgedminwhoa, never mind17:32
mgedminlooks like I have strange pathnames in .pyc files17:32
*** GaryPoster has joined #zope3-dev17:32
mgedminyep, that's the cause17:32
mgedminthat test wants to see a line of source code in a traceback in a html page17:33
mgedminbut because of invalid paths in my .pyc files, the traceback generator is unable to find the source file17:33
* mgedmin runs find -name '*.pyc'|xargs rm && python2.4 ./ -f -pv -s
J1mmgedmin, should consider using the -exec option in find.17:34
mgedminxargs is easier to type, and faster17:35
mgedminfind|grep|xargs is a flexible pattern17:35
SteveAnot many char 0 in .pyc file names17:48
*** yota has joined #zope3-dev17:49
*** zagy has quit IRC17:52
projekt01 J1m, are you really sure that the PrincipalRegistry is queryable?18:01
J1msee getPrincipals18:02
projekt01I found the ISourceQueryView view but the ISourceQueriables isn't implemented.18:03
projekt01It seems only to work because there is a ISourceQueryView view registered, but that doesn't mean it's implemented.18:05
J1mIf you are going to ignore my answers, I won't bother providing them.18:05
projekt01Sorry ,I can't follow you, getPrincipals isn't a method of the ISorceQueriables interface ???18:07
J1mno, it's a method of a principal registry.18:08
J1mYou can write an adapter using it.18:08
projekt01Yes, but you told the PrinciplaRegistry is queryable (perhaps my question was incorrect) I was looking for a the ISourceQueriable adapter by asking "is queryable".18:11
*** Hellfrie2 has quit IRC18:11
projekt01Ok, I see the PrincipalRegistry doesn't provide ISourceQueryable but we register a view providing ISourceQueryView and this will make it working and let's the PrincipalRegistry act as it would be ISourceQueryable18:15
projekt01Ok, important to know for a testing setup where includes the PrincipalRegistry and a local PAU ;-)18:16
*** projekt01 has quit IRC18:20
*** yota has quit IRC18:32
*** stub has quit IRC18:32
*** Theuni has quit IRC18:33
*** stub has joined #zope3-dev18:39
*** yota has joined #zope3-dev18:44
*** tahara has quit IRC18:45
*** stub has left #zope3-dev18:48
*** MrTopf has quit IRC19:10
*** MJ has quit IRC19:15
*** agroszer has quit IRC20:01
*** BjornT has quit IRC20:06
*** sashav has quit IRC20:09
J1mmkerrin, ayt?20:27
mkerrinJ1m: hi20:28
J1mI was about to look at the code, but you are here. :)20:29
J1mI'm guessing/hoping that the wsgi gets output from the app and queues it for deleivery.20:29
J1mI'm guessing that this code doesn't do the I/O.20:30
J1mand thus isn't blocked by it.20:30
J1mIf this is the case, then the wsgi output processing should happen very fast and should not be held up by the network I/O.20:31
mkerrinI am not following - so the output isn't an issue??20:32
J1mI'm trying to figure that out.20:33
J1mUnfortunately, I'm not good at reading twisted code.20:33
J1mIf what I said above is true, then I don't think that output is an issue.20:33
J1mWe don't want the app thread to spend time doing I/O.20:34
J1mIf write in WSGIHandler queues data that is actually send by the reactor, then it doesn't have to wait for I/O.20:35
J1mDo you undertand what the write method is doing?20:36
J1m(This raises another issue ...)20:38
mkerrinkind off - the callFromThread method calls seem to tell me that we are still going to be in the same thread has the application (even though Zope has finished processing) until all the data has being written to the socket stream. Although I will still have to write some kind of code to test for this20:38
J1mI have a feeling that this is too big to deal with before the beta.20:40
J1mI really can't help much because this code really doesn't mean much to me.20:41
* mgedmin fixing issue 49720:41
* mgedmin found a commented out functional test for it in
* mgedmin happy20:42
mkerrinWhen is the beta planned20:42
J1mwe hope to make it this weekend, or Monday latest.20:43
J1mWe are hoping that this will be the last beta.20:43
J1mWe really want to release 3.2 this month.20:43
mkerrinright - so this isn't looking good for twisted at the moment20:44
J1mOK, I suggest that flailing would be bad.20:45
J1mSo let's not flail.20:45
benji-1 on flailing20:46
benjiI've done enough of that for all of us today20:46
*** _anguenot has quit IRC20:47
mkerrinBut wouldn't what I said in the email work - if we had a threadless wsgi handler then the glue between twisted and zope can handle the threads20:47
J1mDid you see the discussion on the web sig yesterday?20:47
J1mIt isn't possible to write thread-management WSGI middleware.20:48
mkerrinah - looking that up now20:49
J1mI think the best thing is to step back, state the requirements clearly, and hope I can convince others.20:49
J1mThat will take time.20:49
J1mTime we don't have for this release.20:49
*** graham has quit IRC20:51
J1m(I hope I don't come to regret endorsing wsgi.)20:52
mkerrinJ1m: I am going to have to think more about over the weekend - 7pm on a Friday is a bad time for this21:14
*** fcorrea has joined #zope3-dev21:14
J1mFair enough21:14
J1mNext week, I'm going to pursue this issue on the web-sig list.21:14
J1mI'm going to try to set out the following requiremets for WSGI servers:21:15
J1m- don't make applications wait for network I/O.21:15
J1m- Don't hold large inputs or outputs in memory.21:15
J1m(I fear that the second is actually prohibited by the PEP.)21:16
*** ruda_porto has quit IRC21:17
J1mI don't really want to spend any more time on this now.  I need to concentrate on the beta.21:17
*** ignas has quit IRC21:18
*** xenru has quit IRC21:19
mkerrinfair enough21:20
mkerrinby the way - issue 521 in the collector should now be fixed. James Knight committed a fix to the Twisted code so read(-1) now returns all of the stream21:22
*** hazmat has joined #zope3-dev21:24
J1mmkerrin, did he commit the releases/2.1.x branch?21:24
J1mmkerrin, did he commit the fix on the releases/2.1.x branch?21:25
J1mah, he did :)21:26
J1mI'll update the external to that.21:26
* J1m adds a new mkzopeinstance test21:27
* J1m hates hates hates inscrutable classic unittests. Blech.21:28
* benji does too.21:28
*** deo has joined #zope3-dev21:29
* J1m debugs a test to find out wtf it does21:31
J1m(wtf is a tla for "what the fudge")21:32
*** faassen has quit IRC21:33
* benji wants some fudge21:34
* benji goes to get some fudge21:35
* benji returns with chips and dip; all out of fudge :(21:36
*** jukart has joined #zope3-dev22:13
*** mkerrin has quit IRC22:25
*** lunatik has joined #zope3-dev22:26
*** jukart has quit IRC22:39
*** J1m has quit IRC22:40
benjisrichter, ping22:56
*** J1m has joined #zope3-dev22:57
benjiJ1m, bbwin seems to be exibiting the flaky network behavior, just like fred-win22:59
J1mI guess my switch is golden. :)22:59
benjiI guess we can't make fun of your switch any more22:59
*** lunatik has left #zope3-dev22:59
benjiJ1m, do you consider lack of I18N coverage a blocker?  (there is an XXX about some messages not being I18Ned)23:08
*** mgedmin has quit IRC23:29
J1mwhat is using
benjiok, all XXXs are fixed except for four, one I need to work on, I sent a message about the other three to zope3-dev23:39
*** jinty has joined #zope3-dev23:44
*** zbir has joined #zope3-dev23:45
dman13regarding mkzopeinstance ... I became aware last night that doesn't properly xml-quote values that are put in xml files (eg admin password in principals.zcml)23:46
dman13are there two duplicate mkzopeinstance's now?  (one for twisted and one for zserver)23:47
J1mseem to be.23:47
dman13oh, while I'm on the topic, is there a convenient function available in python to properly escape characters for XML?  If someone can point me to it I can make an easy fix for mkzopeinstance23:48
dman13does there need to be two mkzopeinstances?  isn't the only difference the zope.conf parameter for using twisted? (or was that parameter not implemented yet?)23:48
benjidman13, xml.sax.saxutils.escape and/or quoteattr23:49
J1mI have no idea wht there is a twisted mkzopeinstance.23:49
benjiI have no idea what you said.23:49
dman13benji: sweet!  thanks.23:49
* dman13 hasn't done enough XML work in python to have learned the APIs23:50
* benji hasn't either, he just can't forget anything :)23:51
J1mI think srichter just copied to make
J1mBut I suspect that not all if it is used.23:52
benjioh, BTW I'm glad you got your WSGI bug fixed23:52
J1mNow I guess I'll spend time trying to find out.23:52
benjiis that the one Gary was working on?23:52
J1mNo, it wasn't fixed afaik23:53
J1mwell, one was.23:53
J1mIt was pretty minor.23:53
J1mNot the big one.23:53
benjinot the bug Gary had (with the streams/bodies/consumption/etc.)?23:54
J1mYeah, that one was fixed.23:54
*** ClaesBas has joined #zope3-dev23:54
J1mYesterday, we also realized that Twisted's wsgi strategy has inadequate thread-management.23:55
J1mWhich means that no one in theit right mind would use it in production.23:55
J1mThat's why today I modified to add a --zserver option.23:56
benjipfft, that's great!23:56
ClaesBasHi, have a question about "Zope3 Instances", how to update after a "svn update" of main source?23:56
J1mI brought this up on the Z3 list.23:56
benjihmm, I must have missed it23:56
J1mTres made a very funny response.23:56
benjiahh, OK, the electing-people-who-think-goverment-can't-work joke, got'cha23:58
benjiI guess that precludes me from running for office, of course the +70 IQ already disqualified me23:59
J1mYou would make a great President.23:59

Generated by 2.15.1 by Marius Gedminas - find it at!