IRC log of #zope3-dev for Saturday, 2007-01-20

*** whit has joined #zope3-dev00:01
*** whitmo has quit IRC00:01
*** flox has quit IRC00:05
*** alga has quit IRC00:07
*** benji has quit IRC00:12
*** BjornT has quit IRC00:31
*** d2m has quit IRC00:36
*** romanofs1i has joined #zope3-dev00:48
*** opetznick has quit IRC00:49
*** whit has quit IRC00:58
*** jodok has quit IRC01:00
*** romanofski has quit IRC01:03
*** jodok has joined #zope3-dev01:04
*** alga has joined #zope3-dev01:05
*** J1m has joined #zope3-dev01:15
*** whit has joined #zope3-dev01:22
*** harobed has joined #zope3-dev01:43
*** rmarianski has quit IRC01:50
*** whitmo has joined #zope3-dev01:52
*** whit has quit IRC01:52
*** tomus has quit IRC01:54
*** projekt01 has joined #zope3-dev01:57
*** WebMaven has quit IRC01:58
*** jinty has quit IRC01:58
*** WebMaven has joined #zope3-dev01:59
*** b_52Free has joined #zope3-dev01:59
*** J1m has quit IRC02:02
*** J1m has joined #zope3-dev02:15
*** b_52Centos has quit IRC02:17
*** deo has quit IRC02:21
*** J1m has quit IRC02:35
*** whitmo has quit IRC02:39
*** revpj has joined #zope3-dev02:40
*** norro has quit IRC02:52
*** jodok has quit IRC02:59
*** borax has left #zope3-dev03:15
*** revpj has left #zope3-dev03:15
*** harobed has quit IRC03:18
*** wrobel has quit IRC03:19
WebMavenHmm. I've just run across something odd...03:27
WebMavenIt seems that addMenuItem directives are required to have unique titles...03:27
WebMavenThough I am not sure why...03:28
WebMavenAs far as I can tell, no lookups are done by the title.03:30
WebMavenDoes anyone know?03:34
*** baijum has joined #zope3-dev04:37
baijumgood morning !04:42
*** b_52Free has quit IRC05:04
*** b_52Free has joined #zope3-dev05:04
baijumI am doing some Zope 3 eggification related works today, anyone interested to join ?05:14
* baijum now working on adding test layer to zope.traversing and zope.formlib05:16
*** niemeyer has quit IRC05:27
WebMavensrichter: AYT?05:40
WebMavenOK, would someone please explain to me why the title of an addMenuItem needs to be unique?05:58
WebMavenThis is not making sense.05:58
srichterbecause the title is the name of the adapter05:58
srichterand you get a configuration conflict05:59
srichterthe standard menu implementation is not good in many ways, one of them is that it is really slow05:59
srichtertry to use instead06:00
*** alga has quit IRC06:00
WebMavensrichter: OK, thx.06:52
* baijum is about to commit TraversingLayer to trunk, here is the patch: (can any one review this ? )06:57
* baijum committed TraversingLayer to trunk, r7211207:06
*** projekt01 has quit IRC07:18
*** b_52GM has joined #zope3-dev07:31
*** cursor has joined #zope3-dev07:42
*** b_52Free has quit IRC07:48
* baijum is about to commit FormlibLayer to trunk, here is the patch: (can any one review this ? )08:23
*** pelle_ has joined #zope3-dev08:56
*** pelle_ has quit IRC09:01
*** alecm has quit IRC09:06
*** alecm has joined #zope3-dev09:10
*** ktwilight has joined #zope3-dev09:16
* baijum is away09:30
*** ktwilight_ has quit IRC09:31
* baijum is back10:06
*** mgedmin has joined #zope3-dev10:42
*** opetznick has joined #zope3-dev10:54
WebMavensrichter: are you still around?11:00
WebMavensrichter: it occurs to me that using the title as the name of the adapter is actually a very bad solution.11:02
*** cursor has quit IRC11:02
WebMavensrichter: There should be a separate name parameter.11:02
WebMavensrichter: I suppose, for backward compatibility, it could be made optional, with the current behavior as the default.11:03
*** d2m has joined #zope3-dev11:05
*** dobee has joined #zope3-dev11:07
*** opetznick has quit IRC11:08
*** dunny has quit IRC11:13
*** nuovoM has joined #zope3-dev11:33
*** nuovoM has left #zope3-dev11:33
*** opetznick has joined #zope3-dev11:38
*** opetznick has quit IRC11:43
*** b_52Free has joined #zope3-dev12:02
* baijum hmm..we have to create test layers for many packages, oherwise functional tests of individual packages will not run12:06
* baijum see the list:
* baijum committed test layer for (r72118,r72119)12:07
* baijum expected atleast one person to join today's eggification related works :(12:09
mgedminbaijum: I am sort of idle and wondering what to do12:09
mgedminI didn't remember there was eggification scheduled for today12:10
baijumI send a mail to zope3-dev, only Jim responded12:10
mgedminand my experience with eggs is a bit limited12:10
baijummgedmin, familiar with zc.builddout ?12:10
mgedminI've seen the presentation at europython12:10
baijumI am trying to to learn that now..12:11
baijummgedmin, what is your opinion about the out put when running tests in trunk now?12:13
baijumnow I added layers to few packages12:13
baijumso more verbose output is displayed..12:14
baijumfor each layer12:14
* baijum will do more test layer works in another branch (will wait for feedback from others in next week)12:15
*** harobed has joined #zope3-dev12:17
* mgedmin runs tests to see the output12:17
*** romanofs1i is now known as romanofski12:19
*** b_52GM has quit IRC12:20
mgedminhm... -p doesn't work when tests are run in a subprocess?12:22
* baijum runs -p12:27
mgedminprogress indication12:28
mgedminyou don't see it when tests are being run in a subprocess12:28
mgedminor did I just miss it?12:28
baijumhow to run tests are run in a subprocess?12:29
baijumyes, progress indication not coming for tests with layers12:30
baijumonly unit tests progress indication works now12:30
baijum(atleast here)12:31
baijumWhy all functional test run in a subprocess ?12:33
mgedminbecause the functional test layers do not support teardown12:33
* mgedmin is trying to find out if there are good reasons why tearing down the ftest layers is not implemented12:34
baijummgedmin, so any reason for progress indication not working with functional test layers ?12:35
mgedminit works in the first layer, but not in subsequent ones12:36
mgedminbecause the first layer cannot be torn down12:36
mgedminand the test runner is forced to run the next layers in subprocesses12:36
mgedminthe test runner's protocol for communicating with its subprocesses apparently doesn't allow for real-time progress indication12:36
* mgedmin thinks implementing functional test layer teardown is not hard to do12:37
*** zagy_ has joined #zope3-dev12:37
*** dobee has quit IRC12:44
*** zagy has quit IRC12:44
*** romanofski has quit IRC12:45
*** SmokeyD has left #zope3-dev12:53
*** b_52GM has joined #zope3-dev12:59
*** zagy_ has quit IRC13:02
*** zagy has joined #zope3-dev13:02
*** zagy has joined #zope3-dev13:03
*** zagy has quit IRC13:03
*** zagy has joined #zope3-dev13:03
*** b_52Free has quit IRC13:15
* baijum leaving13:25
*** baijum has quit IRC13:26
*** schwendinger has joined #zope3-dev13:42
*** schwendinger_ has joined #zope3-dev13:43
*** dlk has joined #zope3-dev13:51
*** daniele has joined #zope3-dev13:53
*** wrobel has joined #zope3-dev13:55
*** schwendinger has quit IRC14:01
*** schwendinger_ has quit IRC14:15
*** b_52GM has quit IRC14:20
*** b_52GM has joined #zope3-dev14:20
*** philiKON_ is now known as philiKON14:37
*** philiKON is now known as philiKON_14:37
*** lamike has joined #zope3-dev14:38
*** philiKON_ is now known as philiKON14:38
*** philiKON is now known as philiKON_14:39
*** philiKON_ has joined #zope3-dev14:39
*** philiKON_ is now known as philiKON14:40
*** jhauser has joined #zope3-dev14:58
*** jhauser has joined #zope3-dev14:59
*** jinty has joined #zope3-dev15:04
*** schwendinger has joined #zope3-dev15:46
*** schwendinger_ has joined #zope3-dev15:49
mgedminI once got a test failure in zope/app/twisted/tests/test_inputbuffering.py15:56
mgedminit disappeared when I reran the tests15:56
*** oferw has joined #zope3-dev15:58
*** J1m has joined #zope3-dev15:59
*** harobed has quit IRC16:02
mgedminif anyone is interested in the failure, it's here:
mgedminit does not appear to be reproducible16:03
*** lamike has quit IRC16:04
mgedminwhat's the policy on functional tests that change the component registry?16:06
*** schwendinger has quit IRC16:06
mgedminI'd like them to clean up after themselves16:06
mgedmintoo bad you cannot use placlessTearDown in a functional test16:06
*** b_52Free has joined #zope3-dev16:12
J1mmgedmin, I'm in favor of tests cleaning up any changes they make.16:21
J1mmgedmin, ayt?16:22
* mgedmin is16:22
J1mIf you feel like some r&d :)16:22
J1mThe stage is set for doing thigs better.16:23
J1mBut it will take some work from somebody.16:23
J1mThe component registries are more flexible than they used to be.16:23
J1mFor example, it should be possible, somehow, to play with subregistry or temporary hacks to component __bases__ to allow new registries to be inserted during tests and removed after tests.16:24
J1mThis would allow things like placelessTearDown to be useful in this context.16:24
*** oferw has quit IRC16:25
J1mI'm waving my hands but I think something clever could be done with some effort.16:25
J1mFor example, I'm pretty sure something like this would allow teardown of zcml layers.16:25
J1mUnfortunately, I don't have tome to do anything with this myself.16:25
J1mThis is something I had in mind when I did the last wave of component registry refactoring.16:26
mgedminI've just implemented teardown of zcml layers :)16:27
J1mah, cool16:27
J1mhow did you manage that?16:27
mgedminplacefulTearDown() + reset two FunctionalTestSetup attributes to allow it to load a different zcml file next time16:28
*** b_52GM has quit IRC16:28
J1mah, hm, so this works for top-level layers bit not for sublayers.16:29
J1mso registry manipulation would allow tear down of sublayers.16:29
mgedminwell, depends on the sublayers, I guess16:29
mgedminit would be really nice if zcml layers could be nested16:30
J1mand, of course, none of this deals with changes to globals -- including security declaratations.16:30
J1mwell, they can be16:30
mgedminsecurity declarations are cleared by CleanUp().cleanUp(), I think16:30
J1mbut you have to tear down the whole thing after each layer...16:30
mgedminthere might be problems with changes to globals that are not registered with CleanUp()16:31
J1mand global changes are usually pretty minor16:31
J1mMost of those are pretty benign.16:31
mgedminif you'll take a look at my last checkin to a new zope.testing branch16:31
mgedminthere's a way to catch uncleaned-up changes of any specific global value16:32
mgedminI used it to test whether layer teardown cleans everything I could think og16:32
mgedminwhich was: transaction contents, component registry contents, python loggers, zope.event subscribers16:32
mgedmin(and incidentally found a bunch of unit tests that had incomplete teardowns)16:34
J1mThat's the --checkers option?16:34
J1ma generalization of the mechinery that checks for left-over threads.16:35
J1mMakes me want to add a component registry to the test runner. :)16:36
mgedminessentially, yes16:36
J1mIf the runner had a component reg of it's own, then these could just be event subs.16:36
mgedminthey have to store state somewhere16:37
J1mNow really :)16:37
*** jodok has joined #zope3-dev16:42
J1moops, I meant "not really" in response to the tricky remark.16:43
J1mI don't see that state management really matters.16:43
J1mSubscribers can csertainly keep track of state if they wish.16:44
mgedminthat's true, I suppose16:44
mgedminfoo = SomeObj(); provideSubscriber(foo.startTest, for=ITestStarted); provideSubscriber(foo.stopTest, for=ITestStopped)16:45
*** projekt01 has joined #zope3-dev16:51
*** jinty has quit IRC16:52
*** benji has joined #zope3-dev16:57
*** Londo has quit IRC17:16
d2manyone please hit the server (mail archives stopped archiving yesterday, is deliveing mails though)17:21
J1mI'll put in a request to reboot mailman17:24
*** jfroche_ has joined #zope3-dev17:34
mgedminhm, buildbot thinks I broke unit tests17:41
mgedminbut testHandlingSystemErrors passes here17:41
mgedminis it necessary to run the tests with -vv?  it makes it rather difficult to see the errors in buildbot logs17:42
mgedmina single -v would suffice17:42
*** alga has joined #zope3-dev17:49
*** jfroche has quit IRC17:50
* mgedmin can reproduce17:50
*** alecm has joined #zope3-dev17:55
mgedminokay, that error is caused by the garbage collector kicking it during an inopportunate moment18:15
mgedminthe extra error message that gets logged is produced from a __del__ method18:15
mgedminI need to find the earlier test that leaves that object uncleaned18:15
*** opetznick has joined #zope3-dev18:23
*** dobee has joined #zope3-dev18:29
*** projekt01 has quit IRC18:31
*** yota has quit IRC18:36
*** dobee_ has joined #zope3-dev18:37
*** opetznick has quit IRC18:38
*** yota has joined #zope3-dev18:39
*** opetznick has joined #zope3-dev19:01
*** dobee has quit IRC19:03
mgedmintests always take longer when you;19:04
mgedmin're waiting for them19:04
*** softdevr has joined #zope3-dev19:12
*** J1m has quit IRC19:15
*** harobed has joined #zope3-dev19:16
mgedminis this another non-deterministic failure, or what?19:20
mgedmin^^ Failure in test testThreading (zope.server.http.tests.test_httpserver.Tests)19:20
* mgedmin suddenly realizes why tests are taking so very very long: he left an extra gc.collect() after *each* test19:21
philiKONmgedmin: way to go on the test cleanup, btw19:31
mgedminI have a patch for transaction, and one for ZODB19:32
mgedminhow do I go about pushing those in?19:32
philiKONyou mean checkers?19:32
mgedminno, a couple of tests are missing transaction.abort() at the end19:33
philiKONi see19:33
philiKONmgedmin: well, the 'transaction' and 'ZODB' packages are maintained in the 'ZODB' project19:36
philiKONyou commit them there and update the svn:external in Zope 319:36
* mgedmin will try to send an email to the list instead19:37
*** dobee_ has quit IRC19:43
softdevrHi, can someone quickly point me in the right direction?  How can I create a web app without using the TTW interface to create each folder, page, image and so on?19:51
mgedminby writing Python code in a text editor?19:53
* mgedmin is not quite sure what softdevr wants19:53
philiKONsoftdevr: you can start by reading one of the numerous tutorials:
softdevryeah, that's look useful - thanks20:02
*** povbot` has joined #zope3-dev20:15
*** povbot_ has joined #zope3-dev20:23
*** povbot has quit IRC20:25
*** alga has quit IRC20:29
*** povbot` has quit IRC20:35
* mgedmin is tired and needs to stop20:57
*** dunny has joined #zope3-dev21:17
*** b_52GM has joined #zope3-dev21:24
*** b_52Free has quit IRC21:27
*** povbot has joined #zope3-dev21:33
*** povbot_ has quit IRC21:33
*** povbot has joined #zope3-dev21:37
*** tomus has joined #zope3-dev21:39
*** povbot_ has joined #zope3-dev21:58
*** povbot has joined #zope3-dev22:06
*** vbachs has joined #zope3-dev22:18
*** povbot_ has quit IRC22:18
*** povbot` has joined #zope3-dev22:24
*** alecm has quit IRC22:30
*** whit has joined #zope3-dev22:34
*** whit has quit IRC22:34
*** yota has quit IRC22:48
*** vbachs has quit IRC22:50
*** philiKON_ has joined #zope3-dev22:52
*** cursor has quit IRC22:58
*** ChrisW has joined #zope3-dev22:59
ChrisWthere some problem with the lists?23:00
*** philiKON has quit IRC23:00
mgedminapparently so23:00
mgedmin<d2m> anyone please hit the server (mail archives stopped archiving yesterday, is deliveing mails though)23:00
mgedmin<J1m> I'll put in a request to reboot mailman23:00
mgedmin<d2m> thanks23:00
ChrisWand just when a topic was getting interesting23:00
mgedminthat was about 5 hours ago23:00
ChrisWhow's .li?23:01
*** dlk has quit IRC23:01
WebMavenChrisW: I noticed my last two checkins didn't get posted.23:01
ChrisWI got some replies this afternoon23:01
ChrisWwas just wondering where my replies to the replies went23:01
mgedminI saw my email23:02
ChrisWI didn't get mine back, and I didn't see them in pipermail23:02
*** alecm has joined #zope3-dev23:32
d2mjust looked, archive seems to work again23:36
ChrisWand i have my replies, cool ;-)23:38
*** ChrisW has left #zope3-dev23:38
*** jhauser has quit IRC23:47
*** jhauser has joined #zope3-dev23:48

Generated by 2.15.1 by Marius Gedminas - find it at!