IRC log of #zope3-dev for Thursday, 2005-10-27

*** andrew_m has quit IRC00:04
*** jvts has quit IRC00:06
*** andrew_m has joined #zope3-dev00:07
*** niemeyer_ is now known as niemeyer00:08
*** J1m has quit IRC00:13
*** _projekt01 is now known as projekt0100:15
*** GaryPoster has joined #zope3-dev00:26
GaryPostersrichter: hey.  I didn't get your email but Benji mentioned you sent one.  Do you see why you need the _clear call now?00:27
GaryPosterok, cool00:27
srichterI just sent another one to you :-)00:27
GaryPosterOK, cool, hopefully I get that one ;-)00:27
srichterso why do you need clear after BasicTypes are set?00:28
srichterthat makes no sense00:29
srichterclear_ does not mess with BasicTypes00:29
GaryPosterNo, BasicTypes are not the important data structure.  _default_checkers, and _checkers, are the important data structure.00:29
GaryPosterThey are only updated when _clear is called, which is done when the checkers module is imported, and when tests are run.00:30
*** baldtrol has joined #zope3-dev00:31
srichterso, what you really want to do is:00:31
GaryPosterTherefore, without the call to _clear, you were putting the checker setting in for future test runs, but it was not registered for the app itself to run00:31
GaryPosterYes, except that again elevates an _* name (_checkers) to API.00:32
srichterso create an alias to _clear called update(), and we are fine, right? :-)00:32
GaryPosterBut yes, that would accomplish the same thing practically.  Well, I'm OK with that if it is marked as # TODO not API; expect change.00:33
srichterso I guess the right solution would be to make BasicTypes a special dictionary that, when updated, also adds the checker entry00:34
*** whit has quit IRC00:34
srichterthat would be a wuick implementation of an API without adding aPI00:34
GaryPosterGiven that you want a fix (very understandable) and we don't have time to figure out how to do it right (as it had been discussed on the list before, I think) then I'm ok with a to-be-deprecated solution for now.00:34
GaryPosterExcept that BasicTypes itself is new, yes?00:34
srichteris it?00:35
srichterI dunno00:35
*** MJ has quit IRC00:35
GaryPosterHeh, I  thought so00:35
srichtersomething like it has existed for a while, if I remember correctly00:35
GaryPosterIt wasn't there last time I walked through this code.  _default_checkers and _checkers were the only data structures like this back then.  Really don't think it is in 3.1, which I guess is the core question.00:36
GaryPosterLemme look; I think I have a copy of the branch hanging around00:36
srichterok, thanks00:37
GaryPosterYeah, you're right.  It's in 3.1.  Ah well; I don't love it, but yeah, it's there, and you could make that do some of the magic00:38
GaryPosterI'm OK with that or a "TODO"00:38
srichterok, I have it do some magic :-)00:39
GaryPosterok :-)00:39
benjitarek, the tests are still failing  :(00:42
*** GaryPoster has quit IRC00:42
*** Alef has quit IRC00:45
baldtrolhi all...  using apache in front of zope3.1 for the first time, and none of my css, js, etc come through.  pages display, but with no formatting or javascript...  i have the virtual host directives line by line from the zope3 dev book...  zope error logs keep saying "ForbiddenAttribute: ('publishTraverse', ..."00:46
baldtrolgoogled around for it, but other than some other, similar examples quoting stephan or philipp's books, nothing seemed to shine any light00:46
*** zbir has quit IRC00:48
*** benji has quit IRC00:49
*** fdrake has quit IRC00:54
*** baldtrol has quit IRC01:00
*** BjornT_ has joined #zope3-dev01:07
*** nisq has joined #zope3-dev01:16
*** nisq has left #zope3-dev01:17
*** BjornT has quit IRC01:21
*** tarek has quit IRC01:36
*** xenru has joined #zope3-dev01:36
*** natea|busy is now known as natea02:16
*** andres has joined #zope3-dev02:19
andresI have a question regarding persistence.02:20
andresViews dont have persistence, or?02:21
srichterthey are created every time they are used02:21
andresBecause I create an object in a view and this one keeps values.02:22
andresWhich is quite strange i think.02:22
andresAnd that object is just an layer uppon zope.formlib, so nothing special is involved.02:23
projekt01what do you mean with "create an object in a view"?02:24
projekt01srichter, do we need the update method in the viewlet?02:25
andresprojekt01, a = Objectname(context,request, some=other,paremeters=exist, also=...)02:25
srichterprojekt01: yes, at least a dummy implementation02:25
srichterprojekt01: it is part of the content provider API now02:25
srichterprojekt01: that was the agreement I made with Gary02:26
projekt01I guess it's needed for subviews? right?02:26
projekt01Ok, I see02:26
srichterI think this is a good agreement actually02:26
srichterbecause it makes content providers two-stage views, which gives them a better meaning02:26
srichterand it adresses a very important use case02:27
projekt01Yes, I can live with that ,-)02:27
projekt01I think two interfaces where better then all in one, but that's ok.02:28
srichterI thought so first too, but while talking to Gary I convinced myself otherwise02:29
projekt01andreas, that can you do in the view.02:29
andresprojekt01, do you mean me?02:29
srichterI really want to make the two-phase process a core content provider implementation02:29
projekt01srichter, if you say that's Ok, I'm fine, then I know you have allways good arguments ;-)02:30
projekt01andreas, yes02:30
srichterbtw, I agree with Jim and ZC now, that the two-phase process is very JSR-168 friendly02:30
projekt01srichter, sounds interesting this two-phase process02:30
andresprojekt01, i just are called andres ;-)02:30
projekt01andres, Ok02:31
projekt01andres, after you created this object, where do you store them?02:31
projekt01andres, I guess you call this view on a "content type instance", then you create a object, ad now you have to store the object somewhere02:33
andresprojekt01, no, i dont want to store something somewhere (actually, i do, but thats not the point).02:34
projekt01what do you do with object "a"02:35
andresprojekt01, i let it generate some text for me, nothing more.02:35
andresbut, a has an array as one of its attributes, and that still contains values from the last request.02:36
projekt01use sessions for share request data02:36
projekt01over more then one view02:36
andresprojekt01, i dont want that, it is so at the moment, thats my problem.02:37
andresand the strange thing.02:37
projekt01can you explain the problem a little more. I think I didn't understand it right.02:38
andresI will try so.02:38
andresI have a View in which i create an instance of my class Form. Then i do some things like Form.add(x). Which saves x in an array. On the first request, the array contains only one item (which is normal). On the second request, it contains _two_ elements.02:42
andresThe one from the request before (and thats the strange thing) and the newly added (which is normal)02:42
andresWhich i dont understand.02:43
projekt01do you use a "input field" for show the value of the first request in the form?02:44
andresNo, just text.02:44
*** zbir has joined #zope3-dev02:46
projekt01can you trace the request?02:46
projekt01I mean do you have a tool for see the request variables?02:46
andresyes, i have that, and its not in there.02:47
andresARG. My Python unawareness shoot me in my foot. The array was declared outside of __init__() (but still in the class).02:48
andresafter moving that into __init__ it works.02:48
andresBut is that normal?02:49
projekt01I guess you defined a class variable and set a value there, so all classes provide this attribute as default02:50
projekt01so they first value get stored there02:51
andresBut, that those are persistent over multiple request seems strange to me.02:51
andresAnyway, its soled and working...02:52
projekt01that isn't persistent, after a server restart all class variable will have the initial values02:53
andresYes, stupid naming from my side.02:54
projekt01but in the runtime you can change class variables and thy get shared on all class instances, except a class overrides this value in a self.xy02:54
andresEven if i only access them via self.xy they are shared? So, i should never declare variables outside of __init__?02:55
andresif i did classname.attribute = 'xyz' i would have expected this...02:55
projekt01It depends how the attribute "attribute" is declared in the class02:57
andresclass xyz:02:57
andres....attributename = []02:57
projekt01but if you call self.attribute you call "as a kind of fallback" the class variable if no private variable is defined.02:58
projekt01self.attribute is private02:58
andresOk. Strange that didnt hit me earlier.02:59
projekt01your attributename list is a class variable. You have to define "self.attributename" in the __init__ or in a method in your case03:00
andresprojekt01, yes, thats what i do now.03:02
projekt01srichter, there is no viewlet base class for use/inherit in custom viewlets. Or I have to say the metaconfigure mixin the class ViewletBase if I whould use one.03:03
projekt01srichter, the problem is, the custom viewlets have to call the super class constructor an they are only based on "object"03:06
projekt01I don't like this, because we have to write mixin classes03:07
andresbtw, the combination zope3 + postgres + sqlos is quite neat.03:10
*** xenru has joined #zope3-dev03:37
*** stub has joined #zope3-dev03:52
*** srichter has quit IRC04:18
*** srichter has joined #zope3-dev04:18
*** ChanServ sets mode: +o srichter04:18
*** sashav has joined #zope3-dev04:45
*** projekt01 has quit IRC04:46
*** benji has joined #zope3-dev05:00
benjisrichter, ayt?05:02
srichterbenji: yeah05:04
benjihave you checked out lately?  :)05:04
srichterno, but I just did as soon as you pinged me05:05
*** niemeyer has quit IRC05:10
*** gdsgdsgvdd has joined #zope3-dev05:19
gdsgdsgvddi have zope 3.0.0 installed and was tring to use ISession05:20
gdsgdsgvddimport ISession05:21
gdsgdsgvddbut there is no session.interfaces, in what version can i find this session interface05:22
srichterI think we did not ship the session package in 3.005:25
gdsgdsgvddis it there in 3.105:26
srichteryeah, it is in 3.105:31
srichterbenji: are you still there?05:32
srichterbenji: we really need to get the urllib2 404 error problem addressed05:33
srichterhave you looked into that already?05:33
srichteralso, I really would like to have a testbrowser method that can return all links in a page05:33
srichterthis way I can check for broken links05:34
benjiI think something can already check for broken links, but can't remember05:34
benjiis there a helper function to do that in the functional testing stuff?05:34
srichteryeah, but the mechanize browser class has a very efficient link retriever05:35
benjiI'd be in favor of reworking that function to use mechanize05:36
srichterbtw, I built today the static apidoc maker using mechanize and your publisher handler and it is a really nice interface05:37
srichterexcept that it is not tested at all05:37
srichterbtw, if I check in fixes to the Zope 3 version of mechanize, will you guys send those fixes upstream?05:38
benjiwe can, it'd probably be best if you do though (in case he has any questions it'd be easier for you to answer directly)05:43
srichterok, though I really don;t want to join yet another mailing list05:49
srichterdarn contributing must be easier ;-\05:49
*** gdsgdsgvdd has quit IRC06:01
*** lucia12345 has quit IRC06:06
*** strichter has joined #zope3-dev06:14
*** srichter has quit IRC06:23
*** deo has quit IRC07:19
*** d2m_ has joined #zope3-dev07:29
*** d2m has quit IRC07:29
*** jhauser has quit IRC07:29
*** d2m_ is now known as d2m07:29
*** benji has quit IRC07:30
*** natea has quit IRC07:30
*** bradb has quit IRC07:30
*** gnosis has quit IRC07:30
*** mp has quit IRC07:30
*** tiredbones has quit IRC07:30
*** vinsci has quit IRC07:30
*** benji has joined #zope3-dev07:31
*** bradb has joined #zope3-dev07:31
*** natea has joined #zope3-dev07:31
*** gnosis has joined #zope3-dev07:31
*** mp has joined #zope3-dev07:31
*** tiredbones has joined #zope3-dev07:31
*** vinsci has joined #zope3-dev07:31
*** xenru has quit IRC07:38
*** gnosis has quit IRC07:45
*** tiredbones has quit IRC07:45
*** benji has quit IRC07:45
*** vinsci has quit IRC07:45
*** mp has quit IRC07:45
*** bradb has quit IRC07:45
*** natea has quit IRC07:45
*** benji has joined #zope3-dev07:45
*** bradb has joined #zope3-dev07:45
*** natea has joined #zope3-dev07:45
*** gnosis has joined #zope3-dev07:45
*** mp has joined #zope3-dev07:45
*** tiredbones has joined #zope3-dev07:45
*** vinsci has joined #zope3-dev07:45
*** dobee has joined #zope3-dev08:02
*** lucia12345 has joined #zope3-dev08:04
*** BjornT_ is now known as BjornT08:08
*** strichter has quit IRC08:23
*** dlk has joined #zope3-dev08:41
*** sashav has quit IRC08:44
*** dobee has quit IRC08:47
*** MJ has joined #zope3-dev09:08
*** dobee has joined #zope3-dev09:09
*** hdima has joined #zope3-dev09:33
*** mp has quit IRC09:39
d2mhow do i POST to a URL using a python module like httplib ? are there functional tests for POST ?09:44
*** MJ has quit IRC09:46
*** jvts has joined #zope3-dev10:15
*** sashav has joined #zope3-dev10:22
*** alienoid has joined #zope3-dev10:27
*** faassen has joined #zope3-dev10:32
*** vlado has joined #zope3-dev10:35
*** j-w has joined #zope3-dev10:59
*** Aiste has joined #zope3-dev11:01
*** andres has quit IRC11:15
*** andres_ is now known as andres11:15
*** povbot` has joined #zope3-dev11:24
*** ignas has joined #zope3-dev11:26
*** povbot has quit IRC11:40
*** MJ has joined #zope3-dev11:40
*** projekt01 has joined #zope3-dev11:48
*** mkerrin has joined #zope3-dev12:16
*** mgedmin has joined #zope3-dev12:22
*** regebro has joined #zope3-dev12:40
*** j-w has quit IRC12:42
*** tarek has joined #zope3-dev12:45
*** j-w has joined #zope3-dev13:21
*** jvts has quit IRC13:43
*** philiKON has joined #zope3-dev13:50
*** strichter has joined #zope3-dev13:54
*** efge has joined #zope3-dev13:57
*** niemeyer has joined #zope3-dev14:03
*** philiKON has quit IRC14:04
projekt01srichter, ayt?14:05
*** benji has quit IRC14:22
*** strichter has quit IRC14:23
*** jinty has joined #zope3-dev14:48
*** dlk has quit IRC14:48
*** MrTopf has joined #zope3-dev15:05
*** srichter has joined #zope3-dev15:10
*** jvts has joined #zope3-dev15:12
*** srichter has quit IRC15:12
*** jvts has left #zope3-dev15:13
*** srichter has joined #zope3-dev15:15
*** srichter has quit IRC15:19
*** srichter has joined #zope3-dev15:19
*** jvts has joined #zope3-dev15:22
efgehas anyone profiled zcml loading to find out what's so slow ? adding the load of 4 little zcml files in a unit test makes things slower by at least 1.5 s per test... :(15:27
*** ChanServ sets mode: +o srichter15:28
srichterefge: never add ZCML files to unit test setup; it's a recipe for disaster15:28
srichterin ZCML, the schema field conversion is what's taking the time15:29
efgesrichter: well I need to load some Five stuff before anything can be tested15:29
srichterit loads a lot of Python too of course15:29
srichterso if your FS is slow, then you have your bottle neck15:29
srichterthen create a FiveSetupBase class or helper function15:29
srichterthe new test runner will encourage this15:30
efgefor some tests, every test will need the setup15:30
*** povbot has joined #zope3-dev15:38
*** andres has quit IRC15:38
*** benji has joined #zope3-dev15:40
srichterefge: right, that's the reason you should write a base class or helper function15:40
srichterusing the new test runner, all those tests would be in the same layer15:40
*** MrTopf_ has joined #zope3-dev15:40
*** MrTopf has quit IRC15:41
*** povbot` has quit IRC15:41
*** hdima has quit IRC15:41
*** dobee has quit IRC15:41
*** hdima has joined #zope3-dev15:44
*** dobee has joined #zope3-dev15:44
*** philiKON has joined #zope3-dev15:44
*** Aiste has quit IRC15:45
*** zbir has quit IRC15:47
*** Faelin has joined #zope3-dev15:50
efgesrichter: you mean the zcml would be loaded once for all the tests ? that's what layers are ?15:57
efgeif that's the case, cool15:57
srichteryeah, I think this is the idea15:59
srichteryou load some common setup once15:59
srichter(I am not sure how well cleanup will be15:59
srichterbecause using the placelesssetup.tearDown() is very simple and powerful, and that would certainly not work anymore15:59
benjisrichter, currently tearing down is not done, and a new clean subprocess is spawned that loads the new ZCML16:00
efgeah yes16:00
srichterI'll wait for Jim to make the switch16:00
srichterbenji: I see16:00
* philiKON is waiting for Jim, too... his testrunner change on the zope 2 trunk made running tests in instance products impossible16:01
srichterbenji: oh, so the setup is loaded and then a process is spawned16:01
srichterbenji: this way you keep your setup base line and you always have a clear tear down16:01
srichterbenji: that's clever16:01
benjisrichter, it'd be more clever if we could undo ZCML :)16:02
efgebetter have fast process spawn though, I wonder how windows fares :)16:02
*** zbir has joined #zope3-dev16:02
jvtshi, I'm integrating Zope 3's zope.component in Gaphor ( The problem I'm having is that I just want zope.interface, zope.component and zope.exceptions. But the packages somehow also depend on zope.testing, zope.proxy and zope.deprecation.16:02
srichterbenji: undoing ZCML is very tough, simply because you would need to define anti-actions for each action to make it work16:03
alienoidbenji: hi, you dislike zcml?16:03
srichterbenji: I have thought about this many times :-)16:03
benjialienoid, nope.  I think ZCML is great for the things it's good at16:04
alienoidoh, ok :)16:04
srichterjvts: it should not depend on zope.proxy16:04
srichterjvts: zope.testing and zope.deprecation sound right though16:04
jvtstestins is indeed listed in the dependencies16:05
srichterjvts: zope.testing is only required, if you want to run the tests of course16:05
benjialienoid, I also think it really, really bad for other things, let's not do those  :)16:05
philiKONzcml is great for turning things on and off. that's it :)16:05
srichterjvts: zope.deprecation is a simply package to make deprecation warnings much easier. It allows us to evolve the code16:06
*** alga has joined #zope3-dev16:07
jvtssrichter: ok16:07
benjiefge, it works well enough on Windows; you generally only spawn a new process fairly rairly, not for every test or anything16:08
philiKONzope.depecration is quite useful. the only thing that i'm missing from it is the functionality to deprecate a whole module and still play nice during certain imports16:09
benjiphiliKON, I wish we had a dedicated evangelist to promote individual packages (like zope.deprecation, zope.wfmc, etc.) to non Zope 3 users16:12
benjimaybe I need to figure out ZPackage and do it myself16:12
philiKONwe'd need an inside man ;)16:13
philiKONlike, someone who's actually respected in the pythonc ommunity16:13
philiKONthe twisted guys seem like a good group of evangelists16:13
philiKONwe already got them to use zope.interface16:13
philiKONzope.component should be next :)16:13
*** SteveA has joined #zope3-dev16:16
*** sashav has quit IRC16:18
srichternote that for 3.3 we will be individual packages for 3.316:20
srichterFred and I will work on that16:20
philiKONs/packages/eggs/ ?16:21
srichterI don't know what eggs are16:21
srichterwe will distribute many small zpkgtools-created packages16:21
philiKONeggs are zipfiles containing python modulees16:21
philiKONerr, packages16:21
philiKONthey are like JAR files16:22
faassensrichter: you should learn what eggs are, and easy_install, and setuptools. :)16:22
philiKONa whole package inside a zip file16:22
srichteryeah, when I need it16:22
faassensrichter: I think Zope 3 should be packaged as an egg as soon as possible.16:22
philiKONthe beauty of eggs is that in addition to putting everything in a zip file (which python already supports) is the dependency management16:22
faassensrichter: it's an essential requirement for Zope 3 marketing in my opinion. part of the reason TurboGears is hyped to death is that you can install it in a few minutes.16:22
philiKONfaassen, i agree!16:22
faassensrichter: even though it's a humongous package with many dependencies.16:22
philiKONnot only from a marketing perspective16:23
philiKONit has so many other advantages16:23
faassenof course. :)16:23
faassenanyway, I think it's going to be the Python standard way to do packaging, and Zope 3 should do it the Python standard way.16:23
srichterif eggs fulfill our use cases then sure16:23
*** J1m has joined #zope3-dev16:23
srichterI'll defer to Jima nd Fred to make the decision16:23
philiKONmorning J1m16:23
faassensrichter: the tutorial even has zope 3 examples!16:23
philiKONfaassen, it does?16:23
* philiKON doesn't remember16:23
faassensrichter: I just wanted to make sure eggs settled in your mind as something potentially Important. :)16:24
faassenphiliKON: yes, it does, one of the tutorials talks about namespace packages, for instance 'zope'16:24
srichterfaassen: ok :-)16:24
faassenphiliKON: and now we install a sub package into it, etc.16:24
philiKONJ1m, your changing of on the Zope 2 trunk made it impossible to execute something like "bin/zopectl test --dir Products/MyProduct" in an instance (to run the tests of MyProduct16:24
faassenit's the combo setuptools/eggs/easy_install that's the important bit.16:24
philiKONfaassen, ah, cool16:24
J1mphiliKON, ise -s instead og --dir16:25
J1mphiliKON, use -s instead og --dir16:25
J1mphiliKON, use -s instead of --dir16:25
* philiKON tries, but doesn't think that that changes anything16:25
*** fdrake has joined #zope3-dev16:25
J1mI think the goldegg people are leading the way on eggs in zope16:26
srichternote that I am totally indifferent about the packaging solution; I am just following what I am told and zpkgtools is very powerful; it handles ZCML, which is a big feature for us16:26
srichterthe thing is that we do not just install simple packages16:26
J1mfdrake, will be participating in a packaging sprint next month.16:26
fdrakewoohoo, San Jose, here I come!16:26
srichter(as I said I just do what Fred will tell me I have to do to make a release)16:26
J1mphiliKON, hm, --dir should work too.16:27
philiKONJ1m, when i comment out the automatic generation of the --libdir flag that's causing the trouble, i get different errors16:27
philiKONJ1m, really, i don't think that the new test runner behaves like the old at all when it comes to bootstrapping zope (setting up Products, for intsance)16:28
J1mphiliKON, that's odd wrt --libdir.16:28
J1mI changed the code that generates that.16:28
philiKONoh, wait, maybe i didn't update that16:28
* philiKON has a weird sandbox16:28
philiKON(because of the zope 3.2 integration)16:29
* philiKON svn ups16:29
philiKONJ1m, works now16:30
philiKONJ1m, sorry about that. was my bad16:30
philiKONooooooh. i get lots of dots now...16:30
* philiKON loves the new testrunner's interpretation of doctests16:30
srichterbtw, I made some really good headway with static apidoc16:31
philiKONsrichter, saw that, nice16:31
philiKONfor zope 3.3 i'll look at generating docbook16:32
srichterthere are some issues left with images16:32
philiKONimages? what do we need images for?16:32
srichterand recently it fails to find some links using pullparser, which I have to investigate16:32
srichterphiliKON: opening and closing sections in the interface details16:32
srichterthe URLs of those images are not referenced in the usual way and the URL is hard coded16:33
srichterbut by making the JS a bit nicer, we should be able to get it working16:33
srichterbut the missing links really kill me right now, because I know they worked before, but pullparser seems not to be able to find them16:34
srichterand of course, we have massive amounts of ReST complaints16:34
srichterwe need to attack fixing those problems16:34
philiKONbasically making docstrings rest compliant16:36
*** anguenot has joined #zope3-dev16:36
*** jinty_ has joined #zope3-dev16:38
*** dlk has joined #zope3-dev16:39
*** jvts has left #zope3-dev16:41
*** SteveA has quit IRC16:43
*** JoaoJoao has joined #zope3-dev16:43
JoaoJoaogood morning16:43
*** jinty_ has quit IRC16:45
JoaoJoaoIs there a way to make a Zope3 server always do HTTP Basic Auth16:46
srichterwhat do you mean? This question does not make sense.16:47
JoaoJoaosrichter: I have a PDA app that do POST requests to a Zope3 server and I'd like to make it use HTTP Basic Auth16:54
*** jinty has quit IRC16:54
philiKONJoaoJoao, protect it with a permission that the anonymous principal doesn't have16:54
philiKONthat will make zope automatically request login information16:55
philiKONthrough whichever method you prefer16:55
philiKONhttp basic auth is just one16:55
srichteryeah, and the default in Basic Auth16:55
*** baldtrol has joined #zope3-dev16:55
JoaoJoaookay, thanks a lot17:01
*** andres has joined #zope3-dev17:08
*** hdima has quit IRC17:12
baldtrolhey all...  i tossed this out to the zope3-users list (and in here way toward the end of the day yesterday), but if no one minds i'd like to ask again, just because i'm stumped...17:15
baldtrolrunning zope3.1 behind apache, i keep getting "ForbiddenAttribute: publishTraverse in traverseName" whenever a request looks for anything configured as a resource in browser:resource...17:17
VladDracnot if you access it directly?17:18
baldtrolnosir...  on 8080 i get everything fine17:18
*** SteveA has joined #zope3-dev17:19
VladDraccheck your zope's accesslog and see if apache somehow mangles the requests17:19
baldtrolyessir...  i have, and i'm pretty sure it has to do with that, but i'm confused about what to do about it.  everything i want to redirect is under a site named "agency_application", so my RewriteRule is "^/agency_application(/?.*)"17:22
benjibaldtrol, that's the match part of the rewrite, what is the redirect part say?17:23
baldtrolbut then all the traversals are looking for things at agency_application/@@/zope3.css (as an example) as opposed to /@@/zope3.css...  but still, if I change the rewrite rule to /.* i still have the same issue...17:24
baldtrolthe rest of the rewrite is:17:24
baldtrolhttp://localhost:8080/agency_application/++vh++http:localhost:80/agency_application/++$1 \17:24
baldtrolwhich, per the zope3dev book and what i can find online, i think that's right (assuming, as i am, that i'm running zope and apache on the same box)17:25
benjiand what does the URL look like when Zope sees it?17:25
baldtrolrequest url that is coupled with the error are like the following: http://localhost/agency_application/@@/zope3.js17:26
baldtrolthanks for helping me out (or at least chattin' with me about it... sometimes talking it over helps me understand it better)17:27
benjiand if you put that URL in your browser it works?17:28
*** dobee has quit IRC17:28
*** dlk has quit IRC17:28
baldtrolnosir, that's the frustrating bit.  I have a url that works... http://localhost/agency_application/1130357904.01075245452/agency_applicant/index.html (numbers are tmp while the object is created and worked through, before archiving), which returns the html based on the object views.17:29
benjiwhat is the working form of the zope3.js URL you gave above?17:29
baldtrolbut the url for @@/whatever.css or whatever.js, or .png...  none of them work.17:29
baldtrolworking form is: src="http://localhost:8080/@@/zope3.js"17:30
benjiis "agency_application" a site?17:30
baldtrolit is...  weird thing though (i had assumed this was a tangential issue, but it's a subclass of Folder, which lets me turn it into a site, but not manage it...)17:32
*** alga has quit IRC17:32
benjiI think the problem is that when agency_application is asked for the zope3.js resource, it can't find it17:33
benjithis is only a work-around, but add another rewrite so JS, CSS, and images are rewritten without the "agency_application" part and see what happens17:34
baldtroli agree...  the error and the way it's not returning my images, zope images, my css, and zope css, makes me think it's a traversal issue.  and ok...  i'll get on that.  (brb, workplace issues calling my name)17:34
benjiglad to help; what little I did :)17:35
srichterwhat do you guys think: Should the static apidoc generator use a config file or command line options for configurability?17:46
andressrichter, the best would be both, with commandline overwriting the configfile, or?17:47
srichteryeah, but it takes twice the work right now17:47
J1msrichter, look at the approach I took for the new test runner.17:49
J1mBasic idea:17:49
J1m- most of logic is in a module. The module defines default defaults17:49
J1m- A small script uses the module. This small script overrides the default defaults with defaults. Basically the script is the config file.17:50
J1m- User-supplied options override the defaults.17:50
J1mLow rent and flexibly enough imo.17:50
srichterok, I have a look17:51
J1mIt could be easily extended to load defaults from a config file.17:51
J1mbut I doubt that is necessary,17:51
srichterme too17:51
SteveAit is nice if that happens in the snall script17:51
SteveAnot in the library17:51
srichterI do not want to overdesign but remove some of the spghetti code17:51
SteveAbecause having the config file policy in the library makes problems if you don't want that17:52
srichterbtw, how is the Authentcation token for HTTP Auth calculated?17:52
srichtera md5 hash of user:pw?17:52
SteveAbase64 digest17:52
SteveAof user:pw17:53
SteveAi think17:53
srichterthanks, I'll try17:53
SteveAit is in the spec for Basic auth17:53
SteveAdigest auth is of course better ;-)17:53
srichteroh, I just need it for the static apidoc generator, so I don't care17:53
baldtrolbenji: that worked.  made my first rewrite rule:18:04
baldtrol    RewriteRule ^/agency_application/(.*css|.*png|.*js) \18:04
baldtrolhttp://localhost:8080/++vh++http:localhost:80/++/$1 \18:04
benjibaldtrol, well at least that's somethign :/18:05
baldtroli mean, i can run it that way in production.  it wouldn't be the most tangled apache config i've ever decided was "eh, good enough"18:05
benjiyeah, me too  :)18:05
baldtroli just wasn't sure if there was something on the zope side that i'm doing wrong...18:06
benjiI don't think so, I think it's in the rewrite, but I'd have to think about it harder18:06
baldtrolcool.  thanks much man.18:07
baldtroli'll post a reply to my own zope3-users question just so your sol'n is archived in these logs and those18:07
*** j-w has quit IRC18:08
SteveAresources should be rooted based on virtual hosting stuff18:11
SteveAnot based on sites18:11
*** xenru has joined #zope3-dev18:15
baldtrolmy pagetemplate usage of resources is always context/++resource++<resourcename> that not right?18:17
*** mohsen has joined #zope3-dev18:22
philiKONthat's correct18:26
philiKONthe point is how that is rendered18:26
*** mohsen has quit IRC18:27
*** stub has quit IRC18:33
*** alga has joined #zope3-dev18:33
d2mi posted to zope-users mailing list about the problems to POST data to a zope3 server using httplib -- problems seem to be httplib related; it turned out that POSTs through urllib work just fine18:36
d2mwhat remains is the question if posting to /@@edit.html is necessary -- or better: how a default view for the POST method can be created/configured18:36
philiKONsrichter, is it me or is the attribute nowhere documented in an interface???18:39
philiKONsite managers are just assumed to have this attribute, or what?18:40
JoaoJoaohow do I create a new zope role in my package's configure.zcml?18:40
philiKONJoaoJoao, <role />18:41
philiKONJoaoJoao, look at etc/securitypolciy.zcml for examples18:41
philiKONsrichter, never mind about the documentation of next18:48
philiKONsee it now18:48
*** sashav has joined #zope3-dev18:51
*** fdrake has quit IRC18:59
srichterokay, I did my third closed run of the API doc generation and it seems to stabilize around 260 MB19:04
srichteron my laptop it takes 1 hour 20 mins to generate19:05
srichterthat's not bad19:06
srichterI hope you were not hoping for something smaller19:06
philiKONdn't care so much about size19:07
philiKONthe time it takes to generate is just enormous19:07
*** faassen has quit IRC19:07
philiKONi wonder how big a tar.bz2 tarball of the whole thing would be19:07
srichterit is probably very very compressible, because a lot of information is duplicated19:08
srichterIt tries to load up about 22000 links btw19:09
srichtersome fail, which you could consider a ftest failure of course ;-)19:09
srichterparsing an HTML file takes also a long time19:12
philiKONah, you're parsing each file?19:13
srichterof course, you need to get all the links19:14
tareksrichter, i was wondering something about the tests we made yesterday19:31
tarekfor instance: this was'nt failing on my box:
tarekin this code19:33
tarekthere's a missing bracket (</param>)19:33
tarekit's not catched on my box19:33
srichterno clue why that would be19:34
*** alga has quit IRC19:34
*** deo has joined #zope3-dev19:37
*** tarek has quit IRC19:40
*** natea is now known as natea|out19:43
*** vlado has quit IRC19:49
*** benji has quit IRC19:53
*** MrTopf_ has quit IRC20:06
*** natea|out is now known as natea20:08
*** philiKON has quit IRC20:17
*** benji has joined #zope3-dev20:18
*** regebro has quit IRC20:26
*** alga has joined #zope3-dev20:27
*** Alef has joined #zope3-dev20:56
JoaoJoaoHow do I grant permissions to users authenticated with a specific principal source21:01
*** zbir has quit IRC21:10
*** andres has quit IRC21:13
benjisrichter, ayt?21:32
*** natea has quit IRC21:52
JoaoJoaoI'm a little confused with PAU21:59
JoaoJoaoI have built a principal source, and I want to grant all users authenticated from this principal source a special role22:00
*** efge has quit IRC22:14
*** alienoid has quit IRC22:20
baldtrolJoaoJoao, i've run into that same issue...  i've tried it myself, and hadn't found a way to do it.22:21
JoaoJoao that's bad news =(22:28
JoaoJoaoI can grant permissions for roles and principals22:29
JoaoJoaoPAU is damn complex22:31
J1mJoaoJoao, you don't want to use a role for that.22:32
J1mUse a group.22:32
J1mIt should be easy to arrange that all users from a principal source are in a group.22:32
J1mThen you can grant a permission to the group or assign a role to the group.22:33
JoaoJoaooh that's cool, thanks22:38
J1mBTW, I agree that pau is complex.  Many situations are well served by that.22:39
J1mBut maybe it would be word while to have a really simple ZODB based auth service that let you manage users and groups.22:39
J1mMight be nice for people getting started and for simpler sites.22:39
JoaoJoaoJ1m: Maybe a doc explaining such simple use cases would be enough22:40
J1mYou mean a recipe for how to set it up.  I agree.22:41
J1mOf course, pau could probably benefit a log from some UI work.22:41
J1mVolunteerrs would be much appreciates.22:41
J1mGarrett Smith done a ton of great work a few months ago.22:41
J1mGarrett Smith did a ton of great work a few months ago.22:42
srichterbenji: what's up? (I am about to go though)22:47
*** srichter has quit IRC22:53
*** mgedmin has quit IRC22:53
JoaoJoaoJ1m: So, I only need to know how to put all users from a principal source in a group ;)22:57
J1mYes, and you can do that with a subscriber on the principal creation events.23:00
*** MJ has joined #zope3-dev23:00
*** JoaoJoao has quit IRC23:09
*** alga has quit IRC23:11
*** ignas has quit IRC23:13
*** anguenot has quit IRC23:23
*** benji has quit IRC23:31
*** baldtrol has left #zope3-dev23:48
*** fdrake has joined #zope3-dev23:51
fdrakewould anyone have a problem if I change zope.schema.interfaces.IIterableSource to extend zope.schema.interfaces.ISource?23:54
fdrakemaking the change does not cause any tests to break, and this only affects the trunk (3.2)23:55
fdrakethe issue w/ the current arrangement is that a source that implements both depends on the order they're specified to select widgets registered for IIterableSource23:56
fdrakeI don't think that's acceptable.23:56
fdrake(especially since I'm adding such a widget for 3.2)23:56
fdrakeok, silence indicates consent!  :-)23:59

Generated by 2.15.1 by Marius Gedminas - find it at!