IRC log of #zope3-dev for Tuesday, 2007-09-25

timtemgedmin: it's possible to just run the that is in the tarball without installing anything, if you didn't  do that already...00:02
mgedminyes, I found that00:03
*** whit has joined #zope3-dev00:06
*** RaFromBRC|lunch is now known as RaFromBRC00:15
*** timost has quit IRC00:19
*** srichter has quit IRC00:27
*** RaFromBRC has quit IRC00:28
*** RaFromBRC has joined #zope3-dev00:28
ignaswho broke what in zope.catalog? seems like addIntIdSubscriber is not triggering indexing anymore ...00:35
*** dunny has quit IRC00:37
*** harobed has quit IRC00:39
*** junk|afk is now known as junkafarian00:41
philiKONignas: if you really think something broke, svn blame is your friend00:43
ignasphiliKON: eggs don't support that scenario00:44
ignasseems like berndroessl00:44
ignaswas changing something around00:44
philiKONwhat's that go to do with eggs00:44
ignasphiliKON: i don't have a checkout of that module00:44
ignasphiliKON: so i don't have a file i could run svn blame on00:44
philiKONummm... get a checkout?00:44
philiKONalso, svn blame takes an svn url00:45
philiKONsvn blame svn://
ignasi don't have the url ;)00:45
ignasbut thanks00:45
*** pcardune has joined #zope3-dev00:45
philiKONput it in an environment variable in .profile00:46
philiKONmine's called $z00:46
philiKONso i can do svn co $z/
ignasgood itea00:47
*** malthe_ has joined #zope3-dev00:48
*** b52lap has joined #zope3-dev00:50
*** alga has joined #zope3-dev00:51
*** timte has quit IRC00:54
*** benji has quit IRC00:58
*** malthe_ has quit IRC00:59
*** regebro has left #zope3-dev01:01
*** malthe has quit IRC01:02
*** b52laptop has quit IRC01:04
*** junkafarian is now known as junk|afk01:06
*** RaFromBRC has quit IRC01:08
*** RaFromBRC has joined #zope3-dev01:08
ignasphiliKON: still there?01:10
ignasthe thing that changed in catalog is that component.getAllUtilitiesRegisteredFor(ICatalog, context=ob) is used01:10
ignasinstead of component.getAllUtilitiesRegisteredFor(ICatalog)01:11
ignasshould i change registrations of my catalogs in some way?01:11
philiKONi wonder why it didn't show up in the commit messages01:11
philiKONsomebody should tell him to subscribe to checkins@zope.org01:11
philiKONhmm, no01:11
philiKONno idea01:11
philiKONwhat does the commit log say?01:11
philiKONis there a test?01:11
philiKONi would investigate this01:11
philiKONand if it causes further probelms, write an email01:12
philiKONto him and cc the list01:12
ignastests were not changed01:12
philiKONthat's the procedure01:12
philiKONwell, if tests weren't changed, then it means it's untested01:12
philiKONand therefore broken by definition01:12
philiKONi think i know who he is, though01:12
ignasits suposedly a bugfix, but no tests were added (it seems)01:13
philiKONjust write an email01:13
philiKONthere isn't anything i can do for you here...01:14
ignasthere is a test01:14
ignasi'll have to look into it01:14
philiKONok then01:14
philiKONgood luck01:14
ignasit's 1 am here so i'll have to do it tomorrow01:14
ignasso every object has a siteManager associated with it in some way?01:21
ignasor might have that is01:21
ignasif so - how do i "register" my stub object with a specific site manager?01:21
ignasso that getSiteManager(my_object) would return the right site manager01:21
ignasahh, it's __parent__ based01:22
ignasso it's not a bug in zope.catalog, just a change that invalidated my old stubs01:23
ignasa handy tip - if your objects seem to be evading indexing, make sure their __parent__ is set up properly01:26
*** jodok has quit IRC01:26
J1mI'm suspicious of this change myself.01:28
J1mI just don't have time to think about it. :(01:28
J1mThis is a backward incompatible change I suspect.01:29
ignasmost objects have their parents set up01:30
ignasand if they haven't then they were susceptible to bugs when site is ambiguous01:30
J1myeeeees, but I'm not sure that it is a requirement to be a location to be catalogable.01:30
ignasis there another way to have a site?01:31
J1mThere's no obvious reason why this should be a requirement.01:31
ignasso - if there is no parent should a global site, or the active local site be used?01:31
J1mSites are thread-local,01:32
J1mThe thread-loxcal site should be used.01:32
J1mThat was the old behavior.01:32
ignasand if there is parent - should the local be used or the one that object belongs to?01:32
J1mimo, the thread-local one should be used.01:32
ignashmm, i see01:33
ignasi mean - i don't have this problem01:33
ignasi have only 1 site01:33
ignasand my views are in proper places01:33
projekt01J1m, context=ob, should explicit avoid to use local sites01:33
ignasotoh - how would you work around indexing objects that belong to 1 site, while are modified with other site being active01:34
J1mare you saying that that is better?01:34
projekt01I have nothing to do with the catalog changes, but context=ob explicit a way you can get another catalog if the ob is from another site.01:34
ignaswhich might happen with nested sites01:34
ignasprojekt01: yes, the point is should that be done in the default DocIndex subscriber01:35
ignasor not01:35
ignassomeone decided that it would be nice thing to do01:35
projekt01I'm not sure, the catalog indexing has also to do how you setu IIntIds utility.01:35
J1mwas there a proposal to do this?01:36
projekt01Catalog indexing with more then one IntId util can get a real pain, becaue the catalog and the IntId util can't find each other01:36
ignasprojekt01: svn diff -r 79789:79790 svn:// the changeset in question01:37
projekt01I don't think this should be like it is. The use case depends on a havy optimized site setup.01:39
ignasyou mean the bugfix should not have been applied?01:41
projekt01Probably not, but why does this affect your catalog?01:43
ignasunit tests01:43
ignasmy objects had no __parent__ set up01:43
ignasso getSiteManager01:43
ignason them was not returning the local site that i have set with setSite01:43
ignasso my stub objects were not being indexed01:44
projekt01ouch, I see, that changes hurts then01:44
projekt01probably we should at least use: if ILocation.providedBy(ob): context=ob else:  context=None01:45
ignasprojekt01: i think it is possible to set a site manager for a non location object01:45
projekt01hooks.getSite knows about the last traversed site01:46
*** pcardune has quit IRC01:46
ignasi mean - if you can set a specific site manager for objects that is not retrieved using __parent__ then the bugfix will only work for sites that are __parent__ based01:47
ignasand will do something unexpected if your object is not a location01:47
ignasnot sure if i am making myself clear though :/01:47
projekt01providing no __parent__ is not a good idea.01:48
ignaswell - not all objects implement ILocation01:48
projekt01security does also only correct work with ILocation01:48
projekt01such object only depend on global security settings01:49
ignasor custom security policies01:49
ignasfrom what i understand you can have an adapter for IComponentLookup01:49
ignasso anything can have a site attached to them01:50
ignasi might be wrong though01:50
*** rocky has left #zope3-dev01:52
projekt01I didn't used IComponentLookup, I just get allways errors named like this ;-)01:53
projekt01but you could be right01:53
ignasok, i really have to go to sleep01:55
*** dunny has joined #zope3-dev01:55
*** ignas has quit IRC01:55
*** jsadjohnson has quit IRC01:56
*** gstratton has quit IRC01:57
*** junk|afk has quit IRC01:57
*** MrTopf has joined #zope3-dev01:57
*** RaFromBRC has quit IRC02:04
*** gstratton has joined #zope3-dev02:04
*** pcardune has joined #zope3-dev02:05
* mgedmin announces
*** gstratton has quit IRC02:12
*** agroszer has quit IRC02:12
*** ousado_ has quit IRC02:24
*** ousado_ has joined #zope3-dev02:24
*** WebMaven has quit IRC02:27
*** gstratton has joined #zope3-dev02:27
pcardunemgedmin: cool02:28
*** gstratton has quit IRC02:47
*** aclark is now known as aclark|dinner02:50
*** philiKON has quit IRC02:55
*** MrTopf has quit IRC03:06
*** huajie has joined #zope3-dev03:09
*** J1m has quit IRC03:10
*** jsadjohnson has joined #zope3-dev03:17
*** yvl has quit IRC03:20
*** pbugni has quit IRC03:26
*** aclark|dinner is now known as aclark03:37
*** stub has joined #zope3-dev03:48
*** WebMaven has joined #zope3-dev03:49
*** alecm has quit IRC03:51
*** universewisdom has left #zope3-dev04:01
*** jsadjohnson has quit IRC04:02
*** gstratton has joined #zope3-dev04:03
*** niemeyer has quit IRC04:19
*** amit_am has quit IRC04:26
*** jsadjohnson has joined #zope3-dev04:34
*** alga has quit IRC04:36
*** ousado_ has quit IRC04:39
*** ousado_ has joined #zope3-dev04:39
*** gstratton has quit IRC04:51
*** mgedmin has quit IRC04:58
*** b52lap has quit IRC05:02
projekt01Theuni, ayt?05:25
*** jpfarias has left #zope3-dev05:30
*** jsadjohnson has quit IRC05:40
*** ousado_ has quit IRC05:41
*** ousado_ has joined #zope3-dev05:42
*** whit has quit IRC05:49
*** greenman has quit IRC05:53
*** WebMaven has quit IRC05:55
*** projekt01 has quit IRC06:00
edgordonanyone here familiar w/ z3c.authentication that would wanna help me out w/ the special groups? pm me if you don't mind helping for a few minutes06:19
*** rcrafton has joined #zope3-dev06:44
*** greenman has joined #zope3-dev06:51
*** deo has quit IRC07:09
*** aclark is now known as aclark|away07:11
*** chacha_chaudhry has joined #zope3-dev07:19
*** rcrafton has quit IRC07:23
*** rcrafton has joined #zope3-dev07:34
*** dobee has joined #zope3-dev07:42
*** Ariel_Calzada has quit IRC07:44
*** afd__ has joined #zope3-dev07:48
*** rcrafton has quit IRC07:49
*** pcardune has left #zope3-dev07:50
*** jukart has joined #zope3-dev07:57
*** zagy has joined #zope3-dev08:04
*** zagy has quit IRC08:06
*** jukart has quit IRC08:06
*** baijum has joined #zope3-dev08:10
*** naro has joined #zope3-dev08:24
*** sm has joined #zope3-dev08:35
*** zagy has joined #zope3-dev08:36
*** sm is now known as sm-afk08:39
*** instantfoo has quit IRC08:40
*** instantfoo has joined #zope3-dev08:40
*** jukart has joined #zope3-dev08:41
*** romanofski has quit IRC08:42
*** huajie has quit IRC08:45
*** instantfoo has quit IRC08:49
*** dunny has quit IRC08:59
*** instantfoo has joined #zope3-dev09:01
*** afd__ has quit IRC09:05
*** instantfoo has quit IRC09:06
*** jodok has joined #zope3-dev09:10
*** philiKON has joined #zope3-dev09:14
*** chacha_chaudhry has quit IRC09:18
*** hdima has joined #zope3-dev09:23
*** instantfoo has joined #zope3-dev09:26
*** timte has joined #zope3-dev09:30
*** instantfoo has quit IRC09:31
*** chacha_chaudhry has joined #zope3-dev09:33
*** romanofski has joined #zope3-dev09:36
*** davidstryker has joined #zope3-dev09:44
*** davidstryker has left #zope3-dev09:55
*** philiKON has quit IRC09:59
*** instantfoo has joined #zope3-dev10:02
*** baijum changes topic to "Zope 3.4.0b1 is out: | Paste: | Zope 3.4 bugs: | bugs: | FAQ: | wiki: | irc logs:"10:03
*** junkafarian has joined #zope3-dev10:04
*** dunny has joined #zope3-dev10:04
*** philiKON has joined #zope3-dev10:06
*** zagy has quit IRC10:15
*** romanofski has quit IRC10:17
*** pelle_ has joined #zope3-dev10:21
*** bigkevmcd has joined #zope3-dev10:25
*** __gotcha has joined #zope3-dev10:27
*** jukart has quit IRC10:42
*** jukart has joined #zope3-dev10:45
*** MJ has joined #zope3-dev10:46
*** benji has joined #zope3-dev10:47
*** goschtl has joined #zope3-dev11:01
*** philiKON has quit IRC11:10
*** jukart has quit IRC11:13
*** malthe has joined #zope3-dev11:14
*** jodok has quit IRC11:15
*** philiKON has joined #zope3-dev11:30
*** harobed has joined #zope3-dev11:37
*** jfroche has quit IRC11:43
*** ousado_ has quit IRC11:49
*** ousado_ has joined #zope3-dev11:49
*** stu1 has joined #zope3-dev11:50
*** zagy has joined #zope3-dev12:03
*** huajie has joined #zope3-dev12:04
*** zagy has quit IRC12:04
*** zagy has joined #zope3-dev12:04
*** stub has quit IRC12:08
*** stub has joined #zope3-dev12:08
*** romanofski has joined #zope3-dev12:12
*** MrTopf has joined #zope3-dev12:16
*** romanofski has quit IRC12:18
*** romanofski has joined #zope3-dev12:18
*** stu1 has quit IRC12:23
*** __gotcha has quit IRC12:34
*** __gotcha has joined #zope3-dev12:36
*** __gotcha has left #zope3-dev12:36
*** b52laptop has joined #zope3-dev12:38
*** philiKON has quit IRC13:07
*** philiKON has joined #zope3-dev13:11
*** sorin has joined #zope3-dev13:22
*** philiKON has quit IRC13:23
*** ignas has joined #zope3-dev13:26
*** ghendi has joined #zope3-dev13:27
*** regebro has joined #zope3-dev13:31
*** ThomasKarlRichte has joined #zope3-dev13:33
*** philiKON has joined #zope3-dev13:34
*** dunny has quit IRC13:36
*** philiKON has quit IRC13:37
*** MJ has quit IRC13:38
*** mkerrin has joined #zope3-dev13:47
*** gstratton has joined #zope3-dev13:48
*** Ariel_Calzada has joined #zope3-dev13:51
*** cursor has joined #zope3-dev13:53
*** cursor has left #zope3-dev13:53
*** benji has quit IRC13:53
*** cursor has joined #zope3-dev13:54
*** gstratton has quit IRC13:55
*** ThomasKarlRichte has quit IRC14:08
*** ThomasKarlRichte has joined #zope3-dev14:09
*** MJ has joined #zope3-dev14:14
*** jsadjohnson has joined #zope3-dev14:18
*** markusleist has quit IRC14:19
*** junkafarian has quit IRC14:20
*** ThomasKarlRichte has quit IRC14:24
*** philiKON has joined #zope3-dev14:24
*** ThomasKarlRichte has joined #zope3-dev14:24
*** MrTopf has quit IRC14:29
*** aclark|away is now known as aclark14:32
*** rcrafton has joined #zope3-dev14:33
*** rcrafton has quit IRC14:33
*** Aiste has quit IRC14:34
*** lamike has joined #zope3-dev14:37
*** gstratton has joined #zope3-dev14:42
*** yvl has joined #zope3-dev14:49
*** rcrafton has joined #zope3-dev14:51
*** mgedmin has joined #zope3-dev14:52
*** jsadjohnson_ has joined #zope3-dev14:58
*** jukart has joined #zope3-dev14:58
*** jodok has joined #zope3-dev14:58
*** lamike has quit IRC14:58
*** gstratton has quit IRC14:59
*** cursor_ has joined #zope3-dev15:02
*** afd_ has quit IRC15:03
*** cursor has quit IRC15:03
*** pyqwer has joined #zope3-dev15:03
* mgedmin joins #foliage-sprint15:05
*** srichter has joined #zope3-dev15:05
*** ChanServ sets mode: +o srichter15:06
*** WebMaven has joined #zope3-dev15:06
*** jpfarias has joined #zope3-dev15:07
jpfariasgood morning15:07
jpfariasI have this problem15:07
jpfariasI have removed an utility from a local site manager15:08
jpfariasand forgot to unregister it first15:08
jpfariasit was an ICatalog15:08
jpfariasreturns that catalog I removed15:08
jpfariasbut I really don't want it there15:08
jpfariasI tried to unregister it15:09
jpfariasbut seems to not work15:09
jpfariasit keeps registered15:09
mgedminstrange, I thought you couldn't remove a utility while it was registered15:09
jpfariashow can I force the unregistration?15:09
mgedminz3 had the IDependencies framework for ensuring that15:09
mgedminhow are you unregistering it and how doesn't it work (is there an exception, or it just ignores your request?)15:10
jpfariassm.unregisterUtility(utility, ICatalog, name)15:10
jpfariasit returns False15:10
*** projekt01 has joined #zope3-dev15:11
jpfariasif I do catalogs = sm.getAllUtilitiesRegisteredFor(ICatalog)15:11
jpfariascatalogs[0].__name__ is None15:12
jpfariasreturn True15:12
jpfariassame for __parent__15:12
jpfariasthat is just weird15:12
mgedminwhen you remove an object, it's name and parent become None15:13
mgedminbut it is not actually deleted from the ZODB if any other objects have references to it15:13
jpfariasthat the problem15:13
jpfariashow do I get rid of  them now?15:13
* mgedmin reads the source of unregisterUtility15:14
jpfariaswhere is that source?15:15
mgedminare you sure you're unregistering it from the right site manager?15:15
jpfariastried to find it15:15
* mgedmin knows how to use ctags15:15
mgedminwhat if you tried sm.unregisterUtility(provided=ICatalog, name=...) without explicitly passing in component?15:16
jpfariascan I do that?15:17
jpfariaslemme try15:17
mgedminaccording to the source code, you can15:17
jpfariasno luck :-(15:18
jpfariaswhere does getAllUtilitiesRegisteredFor look for the utilities?15:23
jpfariasmaybe that is the place I need to remove it from15:23
*** afd_ has joined #zope3-dev15:24
*** greenman has quit IRC15:24
*** markusleist has joined #zope3-dev15:26
Theunisrichter: hey, any idea about the failing tests on formdemo yet? we have a hard time figuring out what's going on as we're trying to get started with z3c.form15:26
*** stub has quit IRC15:29
mgedminTheuni: I *think* that's about z3c.formdemo depending on features that are only in the latest as-of-yet-unreleased version of z3c.form15:31
*** benji has joined #zope3-dev15:32
* Theuni tries with trunk15:33
Theunithat makes the demo instance come up15:34
Theunithe tests still fail in various ways15:34
*** huajie has quit IRC15:41
projekt01Theuni, can you add me to the packages on PyPi? I need to release the missing packages.15:42
Theuniyes, in a few minutes.15:43
projekt01btw, can we createa a list with pypi usernames and copy all names at once to the packages15:43
projekt01Theuni, thanks, don't hurry15:43
srichterTheuni: I am looking at it now15:44
*** junkafarian has joined #zope3-dev15:44
Theunisrichter: ok thanks.15:44
TheuniI found one error in widgets/README.txt that "can never have worked" (tm) ;)15:45
srichterhe he :-)15:45
TheuniE.g. on the float field you add '1234.1234' instead of replacing the value, that borks. a few other things pop up too15:45
*** yvl has quit IRC15:46
srichterTheuni: that could very well be; I have worked on z3c.form a lot and might have not looked at z3c.formdemo to see whether it sitll passes15:46
TheuniToo bad.15:47
srichterwell that happens15:47
srichterit will get much worse15:47
srichtersince it is a direct consequences of eggs15:47
TheuniWe tried using it to get started with a fresh project and that didn't work out too well.15:47
TheuniTrying to get a minimal project running is kind of lots of work.15:48
srichterTheuni: eggs you mean?15:48
Theuniz3c.form i mean :)15:48
srichterTheuni: yes! I think we need a Zope3 meta egg that has all the deps and creates a ZCML file for you15:48
Theuniyikes =)15:49
philiKONsrichter: looked at zopeproject?15:49
srichterTheuni: Roger and I are working on it; Roger is developing a minimal skin that will have all the necessary support15:49
WebMavenphiliKON: does zopeproject handle the library or namespace use-cases yet?15:49
TheuniHmm. Anything visible yet that we could build on?15:49
philiKONWebMaven: no... not yet15:50
philiKONsrichter: btw, i don't buy the "this is a direct consequence of eggs" argument. if you break another package, you introduced a backward-incompatible change15:50
Theuniprojekt01: i just noticed that you wanted access to all zope. and packages.15:50
Theuniin theory i would just give that to you15:50
Theuniin practice i have to do quite some clicking and typing for each individual project15:51
philiKONTheuni: zope.testbrowser is your friend :)15:51
TheuniphiliKON: argh15:51
Theuniyou have anything ready for that?15:51
philiKONsadly no15:51
*** spython has joined #zope3-dev15:53
*** faassen has joined #zope3-dev15:54
projekt01Thenui, should we create a list with all pypi user names from people here at the sprint, so we can add more users at once?15:55
TheuniI'm not sure about that.15:55
WebMavenphiliKON: Hmm. it might be useful if zopeproject added a basic front-page testbrowser test or other doctest...15:55
WebMavenI mean, to the README15:56
projekt01Theuni, you don't think every developer should have pypi permission?15:56
philiKONWebMaven: i'm wary about adding too much boilerplate15:56
srichterphiliKON: yes, but I would prefer not using paster (yet at least)15:56
*** chacha_chaudhry has quit IRC15:57
srichterTheuni: why are you not sure?15:57
philiKONsrichter: it has a zopectl-like script...15:57
WebMavenAh, but a boilerplate test will fail as soon as any development starts...15:57
*** agroszer has joined #zope3-dev15:57
srichterTheuni: the egg process will not work if not all committers can create eggs and upload them15:57
philiKONsrichter, projekt01: releasing stuff takes a certain process, though. i would very much like to make sure that everybody who has the permissions to upload releases is aware of this process15:58
projekt01philiKON, our development process is a f**ing mess right now.15:58
philiKONone bad release can break stuff for everybody15:58
philiKONprojekt01: i tried documenting it15:58
srichterit's a little biit late for enforcing the process agreement15:59
srichterthe egg mess started already15:59
philiKONit's never too late15:59
projekt01philiKON, Theuni, please fix this, we are here at the sprint and don't like to discuss this, we have to do our work, that we discussed with J1m the last couple days.15:59
philiKONfix what?16:00
srichterthis is really frustrating; basically we have an elite that can only upload Zope packagees to pypi; if we trust people with access to, we have to trust them to be able to do releases16:00
philiKONwhat work is it that you're doing? for all i know, i've seen no proposal16:00
*** yvl has joined #zope3-dev16:00
srichtergive Roger (and me as well) access to the zope packages on pypi16:01
projekt01philiKON, fix the development process and give me access to Pypi, or do you like to create eggs for me?16:01
TheuniOk, I thought about that a bit. I think I'd be happy to add committers to pypi. Hwowever. I don't have time to write a script right now and I don't have time to do it  manually. If someone gives me a script and a list of people that are committers that should be added, then I'll run it with my account.16:01
philiKONprojekt01: calm down16:01
philiKONwe're not going to get anywhere by getting cynical or emotional over this16:01
Theuni(I won't fix pypi ;) )16:01
wiggyoh wait, perhaps it's not16:03
mgedminimminent death of zope 3 predicted16:03
spythoni have a clean windows install wanting to install Zope-3, whats _the_ way of doing that nowadays? Seeing that the way of self-contained installer on win is outdated?16:03
philiKONspython: i suggest zopeproject16:04
philiKONspython: make sure you read the section for windows users16:04
*** malthe is now known as malthe|away16:04
srichtermgedmin: unfortunately, not too far from the truth16:05
spythonso i first install python, and then "easy_instal", and then i just run this script to install further things i need?16:05
philiKONspython: right16:06
spythonwill i be able to use zope3 the "old" way of isntances if i install zopeproject?16:06
srichterphiliKON: I am already very emotional about this; first everything gets eggified without a good plan; we end up in a mess; we try to fix some aspects; get blocked; get told to not get pissed off16:06
philiKONspython: no16:07
spythonoh ok16:07
srichterI promised Jim in January I give the approach a try and I did; I at least want to be able to fix the things that mostly disturb me now16:07
philiKONsrichter: well, all i'm going to say is that if you had planned (by advice of J1m or whatever) to make egg releases for a while, then i wonder why you need those access rights in a hurry now. you know pypi as well as the rest of us do16:08
spythonbecause i'm setting up a testing on my machine here, will then comit the code to another machine which runs it live..  that machine does not use zopeproject's way of developing..16:08
philiKONspython: the sandboxes zopeproject creates are pretty self-contained16:08
philiKONsrichter: yes, pypi sucks in that respect.16:08
philiKONsrichter: but what do you want Theuni (or me or J1m) to do right now?16:09
philiKONdrop everything we're doing?16:09
*** elro has joined #zope3-dev16:09
spythonI have to think about this some more then, is Zope 3.3.1 available as eggs for me to easy_install or is it only 3.4 and future versions16:09
WebMavenphiliKON: how much time are we talking about to grant access en-mass to projects on pypi?16:09
*** gstratton has joined #zope3-dev16:10
philiKONsrichter: you're advocating for a plan. i tried to write one down, but apparently that isn't enough either...16:10
philiKONWebMaven: i don't know. there are roughly 100 eggs16:10
philiKONWebMaven: i actually have only access to a dozen or so16:10
philiKON(mostly because i stepped up and released final versions for them)16:11
philiKONgive or take16:11
WebMavenI found 338 zope3-related projects in pypi yesterday.16:11
philiKONi assume projekt01 was talking about the list on
elroHi, I'm trying to backport the Decimal field to 3.3 (in 2.10.4), but I'm getting a ConfigurationError: ('Invalid directive', u'factory') when I try to make the registrations (which I've taken from zope/app/schema/configure.zcml). Any idea what I could be doing wrong?16:12
WebMavenare you just talking about ones in the 'core'?16:12
projekt01philiKON, yes and a concept for all new once.16:12
philiKONelro: use <utility... provides="...IFactory" />16:12
philiKONprojekt01: i don't understand that last bit16:13
* elro wanders if configure.zcml is ever called16:13
projekt01we need a concept if someone adds a new egg. I guess we need to give access to others for that new egg too.16:13
projekt01you never know which egg you need to fix if you do something.16:14
WebMavenphiliKON: Hmm. perhaps those tables need another column for the pypi project owner...16:14
projekt01the -> zope.error separation needs fixes in,, and more...16:15
philiKONprojekt01: about that...16:15
philiKONprojekt01: why does it need fixes?16:15
elrohmm. now I get a ConfigurationError: ('Invalid directive', u'utility'). This is inside a <class> tag16:15
philiKONprojekt01: we said we won't do backward incomaptibilities anymore16:15
srichterTheuni: I fix formdemo first; then I write the pypi script16:16
projekt01Rip out deprecation warnings16:16
philiKONuh huh. why are there even deprecationwarnings?16:16
Theunisrichter: alright16:17
*** MrTopf has joined #zope3-dev16:17
srichter... and once more eggification kills my day of doing real things16:17
*** cursor_ has quit IRC16:18
ignasi so feel your pain16:18
*** cursor has joined #zope3-dev16:19
projekt01I'm really get pissed, I'm doing work here at and before the sprint the last 4 days. This work without a egg process could be done in 6 hours.16:20
projekt01I see the benefit of eggs, but the development process is a night mare.16:20
philiKONit's not like we hadn't had eggs before your sprint.16:21
ignashmm, in schooltool we have buildbot that builds the egg and uploads it into our repository if all tests pass16:21
ignasseems like something like that might be quite useful when working on multiple eggs16:21
ignaseven if eggs get uploaded to some developer egg repository16:21
*** cursor has quit IRC16:22
projekt01philiKON; I can't belive that you didn't see the mess we have right now. Or do you really think it's Ok how it is?16:23
philiKONprojekt01: of course i'm aware of the mess16:23
philiKONi'm trying to improve it16:23
philiKONi already brought up some things on the mailinglist16:23
projekt01I trusted you and others and was thinking if they propose a new egg based process it should be ok, but it's clearly not.16:23
philiKONi'm just saying that you folks must've seen the mess too16:23
projekt01But YOU forces us very havy to go that way! I don't forgive you that quick ;-)16:24
philiKONi didn't do anything16:24
* mgedmin decides it is time to read*checkout*/Sandbox/philikon/foundation/maintaining-software.txt and see what the fuss is all about16:25
*** ruda_porto has joined #zope3-dev16:25
baijumphiliKON, Theuni, projekt01, srichter and all : I can spend sometime to give access to srichter and projekt01 manually for whatever packages I owned, for rest of the packages you can ask their owners, is it ok now?16:26
baijumsrichter,projekt01: what's your PyPI id ?16:27
* baijum just reading log16:28
srichterbaijum: thanks a lot!!!!!!!!!!!!!!!!!!16:28
srichterbaijum: srichter -> srichter16:28
srichterbaijum: projekt01 -> projekt0116:29
faassenanyway, on the eggs, one of the main problems with eggs I've been having is that I keep getting different versions for the same buildout, and so do other people.16:30
faassenso I'd like a way to freeze a list of dependencies in a reusable way. there's already such a way using the dependencies thing16:30
faassenbut that locks people into using whatever the person who did the release, even down to bugfix versions.16:31
* mgedmin keeps thinking "you're reinventing Debian"16:31
faassenso I'd like a way to maintain this separately.16:31
faassenmgedmin: there's a crucial difference. we're looking at a development platform.16:31
*** jsadjohnson has quit IRC16:31
wiggythere is very little difference16:31
srichterbaijum: I will still develop a small script (using testbrowser) that will add or at least try to add a user to all packages in the zc.*, zope.*, and z3c.* namespaces16:31
wiggythe package management problems are the exact same16:31
*** jsadjohnson has joined #zope3-dev16:31
faassenwe have different emphasises.16:32
faassenfor isntance, I could install two development projects which use entirely different versions of eggs.16:32
baijumsrichter: ok16:32
faassenwith a linux distrubtion, I'd need to set up two virtual machines or something.16:32
mgedminfaassen: or chroots -- which in the python world translate to virtualenv/workingenv/virtual-python/zc.buildout directories16:33
wiggyfaassen: the package management problems are the same16:33
regebrofaassen: Although I do find it constantly annoying that I can't install eggs with apt-get and I can't install Ubuntu packages with buildout.16:33
wiggyfaassen: only the place where you install them is different16:33
mgedminthe scale is smaller, but the problems are the same16:33
faassenwiggy: the package management problems are the same, but the use cases have different gradations of importance.16:33
faassenwiggy: which might lead to different solutions.16:33
* mgedmin is not getting faassen's point16:33
wiggyso far I still see everyone slowly reinventing the same wheel16:33
faassenso how do you propose we solve this with debian's package management system?16:34
wiggyI mentioned a while ago that someone should factor out the dependency graph logic from dpkg, apt or rpm and reuse that16:34
regebroNot only may you have many instances of the same egg, but only one of a package, there is no apt-get on Windows. :-)16:34
regebrowiggy: That seams like a good idea, yeah.16:35
faassenthis isn't just a dependency management problem. this is a developer project directory management problem.16:35
* wiggy has no time to argue this right now16:35
regebroBut I have to say that I don't think the dependency logic is the difficult bit of this. It seems quite staightforward...16:36
faassenwiggy: I'm not really disagreeing all that much. I'm just saying that our requirements may be exactly the same but our *primary* requirements may diverge. factoring out the stuff from existing packge management systems seems like a, um, significant project, as well. :)16:36
*** jsadjohnson has quit IRC16:36
faassenregebro: one of the main problems I have with it is that I can't reproduce a project reliably.16:37
faassenregebro: we do a release of grok, and one week later, someone does a release of some random egg and poof, grok breaks.16:37
philiKONwith pinned versions in buildout.cfg you can16:37
ignasfaassen: freeze your dependencies16:37
philiKONit's just a lot of work16:37
ignasthe most radical change is not the packaging system, but the fact that everything that was in a sandbox (free to change and refactor) suddenly got released (uploaded to PyPi)16:37
faassenyes, there's no good way to do it yet.16:37
ignasand as with all the released software16:37
ignasit is difficult16:37
ignasto work on it in an agile manner16:38
faassenit's a ton of work right now, it's a pain to maintain it in a framework, etc, etc, it all needs to be worked out and we need patterns to manage it.16:38
benjiphiliKON: why are nailed versions in the buildout config a lot of work?  it takes about 10 seconds to regenerate a totally up to date set of versions16:38
faassenbenji: what recipe do you use?16:38
regebrofaassen: Yeah, that's an annoyance, but that has more to do with the fast speed of change I think.16:38
faassenbenji: you may have a way to do this in 10 seconds, but I certainly didn't know about this. documentation? :)16:38
philiKONfaassen: lovely has a script16:38
benjifaassen: no recipe, just run buildout -vvvvvvv and a little grep/sed (actully vim script)16:38
elroUmm.. it appears that zope.schema.Time has no widgets (in anyway)16:39
regebroAs I mentioned on Europython, after being away from Grok for a month, nothing works anymore, and it tales me a day to get up to speed. But hey, that's the price of progress. :-)16:39
faassenbenji: sorry, that's not a maintainable solution. that's something *you* can do easily. :)16:39
benjiI know my vim skills are ledgendary, but you too can munge text files <wink>16:39
*** malthe|away is now known as malthe16:39
philiKONbenji: we need Fred "Grok developer" Flintstone be able to deal with this too16:39
regebroI got to use virtualenv today, because I installed i18ndude.16:40
faassenbenji: yes, me, call me Fred.16:40
WebMavenSo who is Barney?16:40
regebroAnd it installed loads of zope 3.4 eggs. And hey, presto, I think all my zope-sites broke!16:40
spythonso, no egg zope 3.3 eggs? just wondering16:40
harobedhi, why in apidoc, there aren't __init__ class information ? I read  zope.publisher.browser.TestRequest and I don't see def __init__(self, path, body_instream=None, environ=None) in APIDOC16:40
spython*"so no zope 3.3 eggs"16:41
faassenbenji: anyway, pinning doesn't work aut of the box if the framework developers want to do the pinning.16:41
faassenbenji: we don't want this to be something all the separate application developers are *required* to care about.16:41
benjiphiliKON: I'd hope that anyone that can type can use grep; even so; a simple buildout-output-to-versions script would be trivial16:41
regebrospython:  ???16:41
philiKONbenji: yeah, or a recipe16:41
spythonregebro: just wondering if i can install zope-3.3 via eggs (easy_install)16:42
benjifaassen: I think that's where philiKON's idea of using HTTP-gettable versions comes in; I don't have that need, so I haven't given it much thought16:42
philiKONspython: no16:42
spythonok danke16:42
faassenbenji: saying that the solution is possible or that it's easy or that you have done it doesn't mean it's community standard practice. that takes communication, possibly code release, and such things. otherwise nobody actually *knows* what the hell to do. philipp and I only knew about version pinning 3 weeks ago!16:42
philiKONbenji: it works as long as you're not on a train or plane wanting to do some development...16:42
faassenbenji: yes, but maintaining a HTTP resource separate to egg releases sucks. :)16:42
benjifaassen: <shrug>16:42
faassenbenji: yes, that's a typical response on this channel, shrug. :)16:43
philiKONfaassen: you should read what tres has to say about developer convenience in this :)16:43
regebrospython: Well, there aren't that many of them, and you would have to specify that you want the 3.3 version and so on. It's probably possible, but I wouldn't be able to tell you how. And besides, wy would you want to?16:43
faassenbenji: but it's not my response anymore. anyway, you shrug and we'll solve it. :)16:43
benjiphiliKON: true; in that case giving the end-developer a versions file they can use (and change if they wish) sounds like a good idea16:43
spythonphiliKON: so everyone's that are using z3c.form and other neat stuff are running zope-3.4b ?16:43
faassenphiliKON: where does he say what?16:43
regebrospython: Oh, now I see that you said *via* eggs. No you can't.16:43
philiKONspython: yup16:43
faassenphiliKON: I mean, which list, when, where?16:44
philiKONfaassen: zope3-dev16:44
faassenphiliKON: recently?16:44
philiKON(originated on zope-cmf, i started cc'ing zope3-dev)16:44
spythonalright, thanks for your time in this busy moment it seems, i'll step back now16:44
*** pelle__ has joined #zope3-dev16:44
elroHas anyone ever used a Time field as a form field? I have a suspicion that if there is no widget in then there is not widget anywhere16:44
faassenphiliKON: Tres seems to be specifying a ton of complications? I just want to release a grok that continues to work for a week. :)16:46
philiKONfaassen: i'm just encouraging you to share your views on this matter16:46
philiKONi think we can only solve this well if we all pull on the same rope16:46
ignasfaassen: get a buildbot ;)16:47
ignasfaassen: works for me, i have to fix things time to time, but schooltool has a pretty good "uptime"16:47
ignasbuildout with pinned versions + buildbot are the most reliable way to do it i have found yet16:48
mgedminbuildbot and buildout have names that are too much alike16:48
benjiignas: yep; I also like the idea of a non-nailed buildbot run to see how things are on the bleeding edge16:49
ignasthough we (schooltool) were expecting Zope3.4 egg set to get packaged for debian16:49
ignasand would have used that as a stable base set of eggs16:49
ignasit didn't happen though, but someone will have to do a lot of work to assemble a stable set of eggs for a release into any linux distribution ...16:50
ignasbecause I am afraid it will not "emerge" from the chaos by itself16:50
faassenI'm not talking about *testing* whether it works.16:50
faassenI can do this.16:50
faassenI'm talking about *releasing* a framework that continues to work after a few days. The only way I can do this is if I can specify explicitly what versions it requires and that all applications that reuse this framework take this into account.16:51
faassenthere should be a way to disable this feature and do your own dependencies if you want to, but that should be the default way it works.16:51
faassenotherwise we cannot release a framework and expect it to work after 2 days.16:51
faassenbecause someone might've released an egg that breaks it. this is not theory, this is what has been happening.16:51
ignasoh, you mean "package"16:52
*** hdima has quit IRC16:52
ignasso that people would use the same grok egg instead of downloading a new one every 2 days16:53
faassenignas: yes, if they use grok 0.11, they use the same set of eggs.16:53
faassenignas: including all eggs grok depends on.16:53
ignaspin them in buildout.cfg you are generating, and maybe even go as far as to download all the eggs16:53
faassenignas: if they have applications that uses grok 0.11, those applications also use the same eggs.16:53
ignasand put them in grok website16:53
ignasso they would always be available16:53
faassenignas: you can't pin it. I can't expect every user of grok to copy a pinned list!16:53
ignasand put the find_link16:53
ignasgrokproject ?16:54
ignasand you can pin version in setup.py16:54
ignascan't you?16:54
faassenignas: no, my individual grok application needs to say: I am using grok 0.11, please use all eggs defined by such, exact versions.16:54
faassenignas: yes, you can pin in, but that locks people in dramatically into exact versions. it may be what we end up doing for now, though. :(16:54
mgedminfaassen: it would help if we had periodic releases (say, every 6 months) of a set of eggs that have been tested to work together16:54
faassenignas: you want to pin people down, but you don't want to lock them in forever.16:54
faassenmgedmin: yes, that's the testing issue. but we can take care of this with grok for now.16:55
faassenmgedmin: it's not the thing that is hurting. we can run the grok tests with a set of eggs and pin them down.16:55
ignasfaassen: so you release grok 0.1216:55
ignasand ti has an updated list of eggs16:55
faassenignas: yes, I proposed this, but philipp doesn't agree. :)16:55
faassenignas: but it may be what we need to be doing.16:55
ignasif Zope3 has no release managers for sets of eggs, you will have to release your own sets of eggs16:56
faassenignas: philipp makes the argument that this locks people in when they develop their own app. if they know what they're doing this should be able to user a newer version of a particular dependency explicitly.16:56
faassenignas: but yes, that's what I have been advocating. :)16:56
ignasi'd just enforce buildout on them ;)16:57
*** jsadjohnson has joined #zope3-dev16:57
ignasand make grokproject the single entry point16:57
ignasit depends on 2-3 eggs only16:57
*** rocky has joined #zope3-dev16:58
faassenignas: we can't put this list of dependencies in grokproject?16:58
faassenignas: we need to put it in grok's, right?16:58
ignasyes you can16:58
ignasyou can put it into a template16:58
ignasthat gets generated by grokproject16:59
ignasand then people can unpin whatever they like16:59
ignasand learning how to pin dependencies is good for them anyway ;)16:59
mgedminand poor people will be forced to maintain that version list forever?16:59
mgedminor do you propose a tool that can upgrade to a newer stable tested set of versions?16:59
ignasmgedmin: yes, because it's their application ;)16:59
faassenmgedmin: I agree with your objection.16:59
faassenthe poor people should have a way to upgrade to grok 0.1117:00
faassenby just saying, I wanna use grok 0.11 now, please.17:00
faassenit's also important in communication.17:00
faassenif you tell me, my application breaks, help!17:00
faassenand I say, which version of grok are you using.17:00
faassenand you say, grok 0.1017:00
faassenI know a lot.17:00
faassenright now, I dno't.17:00
faassenas you might use any mix of dependencies there.17:00
faassenand if people are forced to maintain their own lists of dependencies, you'll have the same problem.17:00
faassenyou'll have to ask: give me your complete list of dependencies.17:00
ignasso you want to know precise set of dependencies, and want to allow people to change that set if they want?17:00
faassenthat's not very easy to talk about, so maintenance becomes harder.17:01
faassenignas: yes. obviously the first is more important than the latter, but philipp and others have advocated the latter a lot.17:01
ignasyou know that you can't have both?17:01
faassenignas: and if you have a layered tree of dependencies, you might want to let some zope 3 egg determine *its* exact dependencies, and have grok depend on it, instead of grok having to maintain its own full list.17:01
*** pelle_ has quit IRC17:01
faassenignas: actually we are pretty close to having both, as you can have a version section in buildout.cfg and refer to a list of versions over http.17:02
faassenignas: but that's not ideal.17:02
mgedmininstead of an egg with a long list of versions you could have several repositories for eggs17:02
faassenmgedmin: that's an idea, though that sort of breaks if you want to use the cheeseshop, right?17:02
*** pelle__ has quit IRC17:02
mgedminyou could upload the latest stable release to cheeseshop17:03
faassenmgedmin: plus it requires all kinds of ways to manage these repositories. I prefer to maintain lists in svn. :)17:03
faassenmgedmin: I think the lists should be maintained close to the things they talk about, i.e. the packages.17:03
baijummgedmin, btw, sm has fixed wiki indexing problem17:04
*** ghendi has quit IRC17:05
*** junkafarian has quit IRC17:06
*** junkafarian has joined #zope3-dev17:08
*** whit has joined #zope3-dev17:10
*** whit has joined #zope3-dev17:11
*** davidstryker has joined #zope3-dev17:15
*** junkafarian has quit IRC17:22
*** junkafarian has joined #zope3-dev17:25
baijumsrichter,projekt01: Done ! Now you must be owners of more that 90 packages in PyPI ;)17:25
*** davidstryker has left #zope3-dev17:26
* baijum leaving, good night all !17:28
*** baijum has quit IRC17:28
*** jsadjohnson_ has quit IRC17:32
*** niemeyer has joined #zope3-dev17:36
*** agroszer has quit IRC17:38
*** pcardune has joined #zope3-dev17:39
*** pelle_ has joined #zope3-dev17:43
*** J1m has joined #zope3-dev17:58
*** timte has quit IRC18:14
*** malthe is now known as malthe|dinner18:20
*** niemeyer has quit IRC18:21
*** pcardune has quit IRC18:21
sm-afkmorning all.. mgedmin, I wondered how did you notice the problem ? did you expect google to index some page quicker ?18:23
*** sm-afk is now known as sm18:23
mgedminyes, the foliage sprint wiki page18:23
mgedminI was really surprised not to find it on the front page on google18:23
mgedminwhen my blog entry about it appeared in 8th position five minutes after I blogged18:24
*** jukart has quit IRC18:24
mgedminthen I started looking for robots.txt and decided to take a look at the <head>18:24
*** RaFromBRC has joined #zope3-dev18:24
smthanks for noticing.. I posted about it at
smI expect when google hits us again active pages should start appearing there quicker18:26
*** niemeyer has joined #zope3-dev18:26
mgedminah, it was a spam-prevention feature18:28
mgedminI thought the zope servers were being hammered by bots and someone hacked in those headers as a relief18:28
smthat may have been part of it too.. but I think google fetches the page either way18:29
smif there's no robots.txt18:29
mgedminwhat about the <a rel="nofollow" href=...> solution for spam?18:29
smI never heard it was very effective18:30
* sm reads up on it18:33
smthanks, and
*** timte has joined #zope3-dev18:38
*** zagy has quit IRC18:38
*** jpfarias has left #zope3-dev18:38
*** MrTopf has quit IRC18:38
WebMavennofollow prevents spammers from getting most of the *benefit* of spam... but you can't assume they care enough about any individual site to not spam it.18:42
mgedminyes, but the same applies to <meta robots noindex>18:43
WebMavenit is easier to spam everything they can, rather than focus on the sites that give them benefit.18:43
WebMavenno, because they can more easily tell that spamming that site is pointless.18:44
*** jodok has quit IRC18:44
WebMavenit won't stop a spammer from crawling the site, of course.18:44
*** ruda_porto has quit IRC18:44
* sm opens
*** MrTopf has joined #zope3-dev18:44
smwe've been *hushed voice* blessedly free from spam on the zope wikis recently18:46
smthe username requirement stops most of it18:47
sm"all edits" subscribers catch the rest18:47
smnow they'll have to respond a bit quicker to keep spam links out of google, that's all18:48
*** romanofski has quit IRC18:48
smany other zope3 wiki needs, while we're talking about it ?18:49
WebMaven'all edits by'?18:50
benjism: I've threatend to delete about 80% of the content (that which is out of date); you can do that if you want <wink>18:50
smWebMaven: I meant wiki subscribers who select "all edits", therefore see the spam edits18:51
smbenji: no time, unless you want to work on it together.. that could go fast18:52
WebMavenyes, I know. I was answering the Q about needed functionality.18:52
smbenji: a cleanup would be great18:52
WebMavena link (on a username) that showed all edits made by that user would be hlpful.18:53
smincluding past edits ? not just pages for which they are the last editor ?18:54
*** febb has joined #zope3-dev18:56
*** romanofski has joined #zope3-dev18:59
*** pbugni has joined #zope3-dev19:03
*** srichter has quit IRC19:03
*** harobed has quit IRC19:03
WebMavensm: yes19:04
*** srichter has joined #zope3-dev19:04
*** ChanServ sets mode: +o srichter19:04
smthanks WebMaven.. can you say what is difficult right now without it ?19:05
smthat would help me prioritise it19:05
WebMaven"What is user X involved in?"19:05
WebMaven"User X just mae a spam edit, what other edits have they done?"19:06
smwe do have a /expungeEditsEverywhereBy?username=... , which managers can use for cleanup19:06
smbut not a user-visible listing19:06
srichterJ1m: We have a working version of static APIDOC; whom do I have to contact to get this uploaded to let's say apidoc.zope3.org19:06
WebMavenouch, that's dangerous.19:06
srichtersm: ^ Are you one of the web masters for
smthat's why it's managers only19:07
smhi srichter.. no19:07
WebMavensm: since you can fake the name.19:07
smsrichter: actually, there is no ?19:07
smif you mean the zope3 wiki ? then yes19:07
*** yvl has quit IRC19:07
WebMavensm: there are also some forms of introspection into a wiki based on IP addresses.19:09
J1msrichter, cool!19:09
WebMavenie. IP addresses from User, Users from IP, edits from IP.19:09
J1msrichter, I could give you a folder on
smWebMaven: good idea19:10
smso far we haven't needed it enough to prioritise that19:11
WebMavensm: I'd consider those lower priority unless concerted endemic spamming is ocuring.19:11
J1msrichter, is there any chance that using the tool to generate api documentation for just a package that could be incorporated into the package's documentation on pypi?19:11
WebMavenbut edits by user is more generally useful.19:11
*** timte has quit IRC19:12
ignassomeone should upgrade the search on PyPI19:13
ignaszope is literally hijacking it at the moment19:13
ignasas it is substring based19:13
ignasso the longer the documentation19:13
ignasthe higher chances of showing up in search results19:14
ignasbuildout shows up as one of the top hits for table, because it has "executable" mentioned a lot in it's readme.txt19:14
ignasadding more information about pypi will make most zope3 eggs show up no matter what you search for :/19:15
ignass/about pypi/to pypi19:15
*** b52laptop has quit IRC19:15
philiKONsrichter: you can contact me to have it on :)19:16
philiKONsrichter: i wasn't aware there were plans for zope3.org19:16
*** MJ has quit IRC19:16
elroCould someone take a look at - it's a patch for with widgets for Time fields19:19
* elro should probably get round to signing a contributor agreement19:22
*** sm is now known as sm-afk19:24
philiKONelro: +1 :)19:25
*** goschtl has quit IRC19:26
ignas+1 if you add a unit test for parsing code19:27
philiKON-1 to the parsing code19:27
* philiKON responds via email19:27
*** pyqwer has quit IRC19:27
mgedminthe launchpad bug is unreadable :/19:28
elrolaunchpad has no option to attach a file19:29
elroso I had to paste inline19:29
ignaselro: really?19:29
ignasbut it has19:29
elrolooks fine on safari to me19:29
philiKONthe launchpad UI just sucks19:29
elroah yes, but only once you have posted the bug19:29
philiKONmgedmin: the email looks just fine :)19:29
*** RaFromBRC has quit IRC19:29
philiKONelro: replied via email. might take some time before it reaches you19:30
elrophiliKON: what should the parsing code be instead?19:30
philiKONsee my email.19:30
philiKONit has shown up on the bug page already19:30
philiKONelro: basically, don't re-invent the wheel19:31
philiKONespecially if your wheel is a clunky, untested one :)19:31
elrodidn't know that existed19:31
philiKONyou could've looked at the datetime widget19:31
philiKONwhich could've lead you to zope.datetime19:31
elrohmm. datetime.time has no utcfromtimestamp method19:33
elrodatetime.datetime.time() wil do though19:34
ignasfrom pytz import utc; utc.localize(timestamp)19:35
ignasif you want to "utc" it19:35
elromissed it because the DatetimeWidget does it's conversion on the field19:35
ignasyou better be sure that the time that was entered is utc19:36
elrotimes should not be treated as being in any particular time zope19:36
elroyou don't know what time of year it applies to19:36
ignaswhy would you need utcfromtimestamp then?19:36
philiKONprojekt01: ping19:37
elrofromtimestamp puts it in the local machines time19:37
*** pelle_ has quit IRC19:38
ignasso you want to convert a timestamp, that you know is a utctimestamp into a datetime object that is utc19:39
elrono, I want a timezone naive time19:39
elroI don't have that info, I don't want to claim I have it19:40
elroI want a datetime.time object, not a datetime.datetime19:40
ignasand what is the input?19:40
elrotext from the user. a zope.schema.Time field19:41
elroI need to record the time of a trade to satisfy reporting requirements, and put this into a database with a time column19:42
projekt01philiKON, yes19:44
philiKONprojekt01: btw, a tip19:44
* projekt01 is fixing my broken egg19:44
philiKONprojekt01: i suggest changing and CHANGES.txt after a release so that it says the *next* version19:44
philiKONprojekt01: so let's say you just released 3.4.0, then change to say version="3.4.1dev"19:45
philiKONand add a new entry in CHANGES.txt19:45
philiKONwith "unreleased" as date19:45
philiKONthis should prevent skipping releases (like you accidentally did today) or doing them double (like J1m accidentally did today)19:45
philiKONall documented here:*checkout*/Sandbox/philikon/foundation/maintaining-software.txt19:45
elrocool, neither datetime.datetime.utcfromtimestamp or fromtimestamp add tzinfo for a plain time19:46
projekt01philiKON, agreed, just need to get comfortable with the process ;-)19:46
philiKONprojekt01: sure.19:46
philiKONprojekt01: that's why i'm saying it :)19:46
projekt01philiKON, np, btw, I'm happy again19:47
philiKONphew ;)19:47
projekt01I hacked my IPhone19:47
philiKONjodok has one too :)19:47
projekt01And installed pyhotn 2.5 on it19:47
projekt01Hope to get Zope3 running on it ;-)19:48
philiKONprojekt01: anyway, i can understand the frustration. but that's no reason to yell at each other :). it's not like the eggs are news from yesterday. we've had them for a while now19:48
philiKONand some of us are actually working on trying to improve the siutation19:48
*** Rolando has joined #zope3-dev19:48
*** markusleist has quit IRC19:48
ignaso snap, that code might be buggy19:48
ignasnot 100% guarantee though19:49
projekt01yes, sorry about that, you did a good job with helping with documentation etc.19:49
ignasit is saying that it will return utc timestamp, but is using time.localtime() to get the date19:49
ignaselro: it is not returning a UTC timestamp19:50
ignaselro: the parsing function gives you a timestamp that when parsed with fromtimestamp gives you the same date that is on the server, and the time entered19:50
*** pelle_ has joined #zope3-dev19:51
elroyep, so I should use fromtimestamp.19:52
ignasuntil someone will fix the function ;)19:53
ignasto return a timestamp in UTC19:53
ignasi might be wrong though19:53
ignasand it might be that fromtimestamp takes the timezone i am in into account19:54
philiKONprojekt01: i'm waiting for the iphone to come to germany officially. i can wait a couple of months19:54
projekt01Yes, I heard they will sell it shortly in germany19:55
projekt01but not in switzerland19:55
*** pcardune has joined #zope3-dev19:57
* J1m likes seeing people buy iPhones. In fact, they should buy 2. Or 3.19:58
*** rcrafton has quit IRC19:59
philiKONJ1m: so you can't? or so they get cheaper? :)19:59
J1mI have apple stock. :)20:00
projekt01J1m ;-)20:00
*** rcrafton has joined #zope3-dev20:01
*** sm-afk is now known as sm20:05
spythonhi, in my simple AddForm i want to add the object i create in a utility instead of the folder the view was accessed from. Any ideas?20:06
pcardunespython: just override the add method20:10
spythonI'm curious as to how you programmatically add to a container?  I don't want to care about giving the object i'm adding a name20:11
spythonpcardune: yes, so in add something like "getUtility(IMyUtil).add(myobj)" ?20:12
spythonthe utility is a folder20:12
*** jfroche has joined #zope3-dev20:12
*** rcrafton has quit IRC20:14
pcardunespython: you can use a namechooser20:14
pcarduneI mean, every object that is added to a container *must* have a name associated with it20:15
pcarduneit doesn't matter what the name is as long as it is unique20:15
pcarduneyou can adapt your container to INameChooser, and then use the chooseName method to generate a unique name20:15
pcarduneI think the default name chooser uses the name of the class of the object and then append "-2" or "-3" to make it unique in the container20:16
spythonok, how can i access the default namechooser? the one that adds an incrementing number at the end for occupied names (foo1 foo2, foo3, etc..)20:16
*** fcorrea has joined #zope3-dev20:16
spythonmeh, i just think i'll read some source.. :)20:17
pcardunespython: you simply adapt the container to it.  i.e. name=INameChooser(folder).chooseName(obj); folder[name]=obj20:17
*** ignas has quit IRC20:19
*** pcardune_ has joined #zope3-dev20:21
spythonpcardune great, works20:24
*** pcardune_ has quit IRC20:25
*** rcrafton has joined #zope3-dev20:26
*** b52laptop has joined #zope3-dev20:27
*** dobee has quit IRC20:29
*** spython has quit IRC20:31
*** alecm has joined #zope3-dev20:31
elroright, it appears that zope.datetime.time is buggered20:34
elroit can't handle a time with seconds20:34
elroi.e. what datetime.time.isoformat returns20:35
* elro is off20:41
philiKONprojekt01: piong20:50
philiKONprojekt01: ping20:50
philiKONsrichter: ping20:54
philiKONmgedmin: ping?20:55
philiKONsomebody from the sprint please?20:56
mgedminphiliKON: pong?20:56
philiKONcan you please get projekt01 to irc20:57
*** pcardune has quit IRC20:57
* mgedmin will try20:57
projekt01philiKON, yes20:57
*** schwendinger has joined #zope3-dev20:58
philiKONprojekt01: the 3.4.0 egg you released is faulty20:58
philiKONit's missing meta.zcml20:58
philiKONall of our biuldouts are breaking20:58
*** agroszer has joined #zope3-dev20:58
projekt01what did I miss?20:59
philiKON no idea20:59
philiKONit's just breaking20:59
philiKONprojekt01: and the 3.4.0 egg is also broken20:59
philiKONthe ZIP file is missing CHANGES.txt20:59
philiKONbut needs it20:59
*** pbugni has quit IRC20:59
philiKONplease fix this too20:59
projekt01this is probably because we can't test it.21:00
projekt01How do you test this?21:00
philiKONi created a new project21:00
philiKONand loads the 3.4.0 egg21:00
philiKONand it won't start21:00 includes
philiKONbut it's not there21:00
projekt01hm, is there a way to test this, I have no project which uses eggs?21:01
philiKONyou can test it with zopeproject, for example21:01
philiKONprojekt01: if you make me an owner of the PyPI page, i can upload a new tarball21:02
projekt01can we write a test which will test the basics? like the egg installation e.g. nissing files?21:02
philiKONit might be a windows issue?21:02
projekt01I think it's my fault, I probably was running sdist and didn't delete the generated sdist folder if I was running it again21:02
philiKONyou always need to clear the 'build' directory21:03
projekt01Does the sdist folder get deleted and created again if I run .. setup . sdist21:03
mgedminwaaaah :(21:03
philiKONyes, i ran into that problem too21:03
projekt01argh, Ok, then it's my fault not windows21:03
philiKONprojekt01: also, don't forget
philiKONthis one is actually not even installable21:04
philiKONbecause CHANGES.txt is missing from the egg21:04
mgedminis there a way to make sdist do the right thing always?21:04
projekt01Yes, how can I grant you access on pypi?21:04
philiKONmgedmin: don't know21:04
philiKONprojekt01: by going to PyPI and adding me to the owner role21:04
* mgedmin looks for a wall to bang head against21:04
philiKONprojekt01: my name is philikon21:04
philiKONprojekt01: so, about the missing CHANGES.txt file. You always need to first check in, *THEN* create the egg21:06
philiKONprojekt01: looks only for files that are in subversion21:06
philiKONyou shoudl really create the sdist from the tag21:06
philiKONa fresh checkout21:06
philiKONthen neither the 'build' stuff nor the missing CHANGES.txt will happen21:07
philiKONargh21:08 is also missing configure.zcml21:08
philiKONprojekt01: what did you do?!?21:08
* mgedmin suggests a sprint topic: buildbot for zope 3 eggs21:08
philiKONprojekt01: i suspect all the eggs you created today are missing files21:09
philiKONthey should be recreated21:09
schwendingeri have nailed = 3.4.121:09 = 3.4.0a1_221:09 = 3.4.0a121:09 = 3.4.0a121:09 = 3.4_dev_r7371521:09
schwendingernow my app is working again21:09
*** elro has quit IRC21:09
*** whit has quit IRC21:09
* philiKON suggests that this gets fixed in a timely fashion. there's already complaints on zope3-users!21:10
philiKONmgedmin: Theuni has a buildbot setup21:10
projekt01just sdist a couple time, did recognize that I have to delete the old dist folder, so the packages will contain early files and not my changes.21:10
projekt01did/did not21:10
philiKONmgedmin: that won't help if you don't create the eggs properly21:10
mgedminphiliKON: my idea was that it will email you and tell you that you haven't created the egg properly21:10
philiKONmgedmin: well, that'd be good21:11
philiKONor everybody just read my guide21:11
mgedminwon't work in practice21:11
*** mkerrin has quit IRC21:11
mgedminif you don't automate it in some way21:11
philiKONi suppose21:11
*** ignas has joined #zope3-dev21:11
mgedmin(you could "automate" by having a person keep track of stuff and shout at people, if you find someone who has the energy for that)21:12
ignasok, it seems that zope.datetime.time() is buggy21:12
ignasit returns timestamp thinking that the clock of your PC is set to UTC21:12
ignasso if you will run it with "20:15" as a parameter on PC's with different timezones set21:12
philiKONmgedmin: well, i'm doing that once in a while :)21:12
ignasyou will get different timestamps, though unix timestamps should be identical among all the machines21:12
ignasis it a bug in documentation or implementation?21:12
philiKONprojekt01: and while i'm at it, produces a deprecation warning talking about "removal in zope 3.6". first of all, i don't think there'll be a zope 3.6 (at least it seems that nobody will want to continue making one big zope 3 releae), and second of all i thought we weren't going to remove stuff anymore21:13
projekt01philiKON, I added you to a couple other packages on pypi21:15
philiKONprojekt01: i'm not going to have time to do anything right now. it would've been fine it was just 1 package21:15
philiKONbut i don't have time for 4 packages21:15
philiKONi think you should fix them...21:15
philiKONin fact, i need to get out of here21:16
projekt01We are working on using zope without any app packages, J1m suggested that this should be the way to go.21:16
philiKONbut not deprecdation with removal21:16
philiKONi'm ok with deprecation21:16
philiKONbut we won't remove stuff21:16
projekt01This means app contains all the views and zmi relevant parts.21:16
philiKONyes, yes, i understand21:16
philiKONi'm talking about the message21:16
philiKON"This reference will be gone in Zope 3.6"21:16
*** MrTopf has quit IRC21:17
projekt01we need a way to tell people that something is changing21:17
philiKONagain, i said i'm fine the deprecation warning21:17
philiKONbut we won't remove stuff21:17
philiKONthere'll be no zope 3.6 and we won't remove it even if there were21:17
philiKONbecause we no longer break BBB21:17
philiKON(unless we really really have to, apparently()21:17
philiKONplease fix the eggs soon. it's ok if we discover broken eggs, but not everybody is as buildout-savvy as schwendinger21:18
philiKONand can nail versions21:18
projekt01what should i do if I remove things like move classes from to zope.error?21:18
philiKONadd deprecation warnings21:18
projekt01I need to support this and give poeple time to following21:18
philiKON"This isn't the recommended location anymore. Use XYZ"21:18
*** MrTopf has joined #zope3-dev21:19
philiKONJust don't say "This will be removed by XXX"21:19
projekt01No nail version will get us in trouble21:19
philiKONsigh. i'm going to say this in german now21:19
projekt01why should we not support it for a longer time e.g. a year?21:19
philiKONbecause that's the policy now21:19
philiKONJ1m wrote an email to zope3-dev21:19
projekt01I should say it will get removed in 12 month21:19
J1mphiliKON, we should say: "We may remove this someday.". Keep em guessing. :)21:20
philiKONJ1m: ok21:20
J1mprojekt01, remove what?21:20
*** sm is now known as sm-away21:20 classes21:20
projekt01they are now in zope.error21:20
philiKONanyway, projekt01, this isn't the priority. the priority is getting the eggs fixed. they're breaking everybody's buildouts21:20
* philiKON needs to leave21:20
*** philiKON has quit IRC21:20
J1mI'm not really paying attension, so I'll say this abstractly.21:20
J1mnever mind, I'm not sure what I want to say. :)21:21
projekt01I see21:21
ignasok, zope.datetime.time is not outright buggy21:22
J1mNo, Now I do....21:22
ignasit's let's say documentation challenged21:22
J1mso must either:21:22
ignasand test challenged a little bit too21:22
J1m- not have future releases, or21:22
J1m- be backward compatible.21:22
J1miow future releases must be backward compatible.21:23
ignasJ1m: does that include interegg backwards compatibility?21:23
J1mIf there are new releases, they should build on
ignasso if i need to change function signature to fix a bug21:23
J1mThis is my jusdgement, not an edict. :)21:23
projekt01J1m, yes, that's what I do21:24
ignasand that function is only needed in another egg21:24
projekt01I agree21:24
*** lamike has joined #zope3-dev21:24
J1mignas, the external api needs to be BC.21:24
*** lamike has quit IRC21:24
J1mLeaving aside the fact that the api is not always well defined. :)21:25
ignasJ1m: hmm, and what is external API? i have seen code that philikon said is not API, but it had deprecation warnings ...21:25
J1mUnfortunately, that is often gray.21:26
ignasso it's more of a "use common sense" and "don't break things too much" thing ;)21:26
J1mfor example, people don't think of class names as part of an external api, but f the classes get pickled, then the names are externally visible.21:26
J1mThis is a dark side of pickling.21:26
J1mignas, we should try very hard not to break things at all.21:27
*** sm-away is now known as sm21:27
J1mwithin reason --and if we screw up, we should fix it.21:27
* ignas is trying not to touch Zope code at all ;)21:27
ignasi mean going through committee, keeping it backwards compatible, checking out all the eggs, and uploading it back21:28
J1mIf you want people to use your stuff, it shouldn't break.  Almost makes you not want people to use your stuff. :)21:28
ignasis just too much for small bugfixes21:28
benjiJ1m: I look forward to your motivational speaking tour.21:29
ignasJ1m: having to "release" something if you want to use the egg yourself is the biggest problem21:29
ignasif it's not "released" (Zope3 trunk checkout in good old days)21:29
ignasyou fix it, and use it21:29
ignasnow you fix it, and then release it, and as soon as your changes get released21:29
ignasit's a totally different thing21:29
*** markusleist has joined #zope3-dev21:30
J1mIn the good old days the zope3 trunk became a bit of a dumping ground.21:30
ignasbut you could experiment with backwards incompatible changes for at least some time21:31
ignasand then make it backwards compatible before release21:31
ignasnow there is no grace period21:31
*** Aiste has joined #zope3-dev21:31
ignasit's either not available to anyone21:31
ignasor it's already a release21:31
J1mignas, it's pretty easy to set up private repos.21:32
J1mor to use externals21:32
ignaswell - it just decreases the ammount of testers, i mean - lot21:32
ignaslot's of people used zope trunk21:32
ignasbut only 2-3 developers will use your private repo21:32
ignasi hope this will stabilize, with all the beta alpha and dev releases, with sane dependencies21:33
*** lamike has joined #zope3-dev21:34
J1mI guess I don't see your point.21:35
*** lamike has quit IRC21:36
*** MrTopf has quit IRC21:38
ignaswell - during the timespan between Zope3.2 and Zope3.3 at least some of the things got broken, same for Zope3.3 -> Zope3.4, and most of these breakages got fixed only before the release21:43
ignasso even if the switch required some effort21:43
ignasyou could chose when it will happen, and it was bearable21:43
ignasnow we will have a lot of small upgrades21:43
ignaswithout a stable checkpoint21:44
*** pelle_ has quit IRC21:44
*** febb has quit IRC21:44
*** pelle_ has joined #zope3-dev21:46
*** yvl has joined #zope3-dev21:48
ignasanyway, just a small warning - zope.datetime.time() assumes server local timezone if no timezone is passed21:50
ignaswhich might cause trouble, just like
*** philiKON has joined #zope3-dev21:55
*** ignas has quit IRC21:58
*** sm has quit IRC22:09
*** timte has joined #zope3-dev22:12
*** sm has joined #zope3-dev22:14
*** faassen has quit IRC22:17
*** naro has quit IRC22:18
*** Ariel_Calzada has quit IRC22:21
*** pcardune has joined #zope3-dev22:22
*** pcardune has quit IRC22:25
*** sm has quit IRC22:27
*** Aiste has quit IRC22:27
*** philiKON has quit IRC22:28
*** philiKON has joined #zope3-dev22:29
*** ousado_ has quit IRC22:31
*** ousado_ has joined #zope3-dev22:31
*** Ariel_Calzada has joined #zope3-dev22:34
*** MrTopf has joined #zope3-dev22:50
*** Ariel_Calzada has joined #zope3-dev22:52
*** lucielejard has joined #zope3-dev22:56
*** lucielejard has quit IRC22:56
*** lucielejard has joined #zope3-dev22:57
*** malthe|dinner is now known as malthe23:01
*** dunny has joined #zope3-dev23:03
*** ThomasKarlRichte has quit IRC23:16
*** nathany has joined #zope3-dev23:16
*** hazmat has joined #zope3-dev23:20
*** ChanServ sets mode: +o hazmat23:20
*** greenman has joined #zope3-dev23:27
*** schwendinger has quit IRC23:30
*** wreutz has joined #zope3-dev23:36
*** ktwilight has quit IRC23:36
*** ktwilight has joined #zope3-dev23:38
*** ousado_ has quit IRC23:51
*** ousado_ has joined #zope3-dev23:52
*** junkafarian_ has joined #zope3-dev23:56
*** sm has joined #zope3-dev23:57

Generated by 2.15.1 by Marius Gedminas - find it at!