IRC log of #zope3-dev for Friday, 2005-06-03

*** Aiste has joined #zope3-dev00:05
*** Aiste has quit IRC00:26
*** bradb has quit IRC00:41
*** niemeyer has quit IRC01:13
*** SureshZ has left #zope3-dev01:21
*** elbixio has quit IRC01:26
*** yota has quit IRC01:31
*** tarek has quit IRC01:35
*** srichter has quit IRC01:37
*** hazmat_ has quit IRC01:39
*** dagnachew has quit IRC01:40
*** bskahan has joined #zope3-dev02:59
*** srichter has joined #zope3-dev03:08
*** bskahan has quit IRC03:16
*** __odie has joined #zope3-dev04:26
*** projekt01 has quit IRC04:38
*** __odie has quit IRC04:46
*** stub has joined #zope3-dev04:58
*** elbixio has joined #zope3-dev05:19
*** bradb has joined #zope3-dev05:23
*** elbixio has quit IRC05:56
*** elbixio has joined #zope3-dev08:15
*** elbixio has quit IRC08:18
*** guido_g has quit IRC08:23
*** timte has joined #zope3-dev08:25
*** tvon has quit IRC08:49
*** tvon has joined #zope3-dev08:53
*** hdima has joined #zope3-dev09:13
*** zagy__ is now known as zagy09:18
*** guido_g has joined #zope3-dev09:19
*** yota has joined #zope3-dev09:27
*** mooded has joined #zope3-dev09:31
*** Theuni has joined #zope3-dev09:38
*** tarek has joined #zope3-dev09:57
*** sashav has joined #zope3-dev10:00
*** __gotcha_ has quit IRC10:16
*** gintas has joined #zope3-dev10:21
*** mexiKON has joined #zope3-dev10:41
*** philiKON has quit IRC10:49
*** lunatik has joined #zope3-dev10:56
*** lunatik has left #zope3-dev10:59
*** vinsci is now known as vinsci|211:05
*** faassen has joined #zope3-dev11:18
*** ChrisW has joined #zope3-dev12:13
ChrisWanyone here?12:13
*** ChrisW has left #zope3-dev12:16
*** Aiste has joined #zope3-dev12:20
*** gintas has quit IRC12:31
*** projekt01 has joined #zope3-dev12:47
*** vlado has joined #zope3-dev13:03
*** jhauser has joined #zope3-dev13:06
*** Aiste has quit IRC13:15
*** andrew_m has joined #zope3-dev13:26
*** bskahan has joined #zope3-dev13:35
*** dagnachew has joined #zope3-dev13:42
*** dagnachew has quit IRC13:57
*** srichter has quit IRC13:58
*** timte has quit IRC14:01
*** Aiste has joined #zope3-dev14:22
*** srichter has joined #zope3-dev14:52
*** ChanServ sets mode: +o srichter14:52
*** dagnachew has joined #zope3-dev15:04
*** ignas has joined #zope3-dev15:28
*** regebro has joined #zope3-dev15:35
*** alga has joined #zope3-dev15:42
*** dagnachew has quit IRC15:53
*** niemeyer has joined #zope3-dev16:08
*** sashav has quit IRC16:18
*** guido_g has quit IRC16:35
*** lunatik has joined #zope3-dev16:36
*** BjornT has joined #zope3-dev16:38
*** lunatik has left #zope3-dev16:49
*** alga has quit IRC16:52
*** zagy has quit IRC16:52
*** SureshZ has joined #zope3-dev16:58
*** stub has quit IRC17:17
*** tarek has quit IRC17:19
*** tarek has joined #zope3-dev17:20
*** tarek has joined #zope3-dev17:23
*** hdima has quit IRC17:38
*** alga has joined #zope3-dev18:00
*** mooded has quit IRC18:22
*** Aiste has quit IRC18:45
*** elbixio has joined #zope3-dev18:59
*** Theuni has quit IRC19:07
*** elbixio has quit IRC19:26
*** regebro has quit IRC19:34
*** faassen has quit IRC19:44
bradbi'm trying to register a custom display widget. in my browser:schemadisplay directive, i've got:19:52
bradb        <browser:widget19:52
bradb          field="bugdescription"19:52
bradb          class="canonical.launchpad.browser.widgets.BugDescriptionDisplayWidget" />19:52
bradbwhen trying to start my z3 instance though, i get:19:52
bradbConfigurationError: ('Invalid directive', u'widget')19:52
bradbi expected it to not raise an exception. why did it raise an exception instead of simply registering my custom display widget like i asked?19:53
mexiKONno browser: namesapce19:55
mexiKONsimply use <widget ... />19:55
bradbso, for edit forms you use browser: but /not/ for schemadisplays? :)19:55
mexiKON<browser:editform ...><widget ... /></browser:editform>19:56
mexiKONtyhat's what i always use19:57
mexiKONsame with addform19:57
bradbi use browser:widget19:57
mexiKONi c now, you want schemadisplay19:57
mexiKONdoes schemadisplay even support the widget subdirective?19:57
bradbit's documented that it does, yes19:57
mexiKONhmm, it does, indeed19:57
mexiKONbradb, will you be at EP?19:58
bradbnot this year, unfortunately. :) too much work to do + moving into a new place right now.19:59
*** elbixio has joined #zope3-dev19:59
mexiKONtoo bad20:00
elbixiohi, are pluggableauth and pau related or do they provide completly different functionality?20:03
mexiKONpluggableauth is x3.020:04
mexiKONpau is 3.120:04
mexiKONthose numbers are zope versions in case you're wondering20:04
mexiKONso, yes, they provide the same functionaity20:04
mexiKONpluggableauth will be deprecated with zope 3.120:05
elbixiobut if i try to use Zope trunk i could use pau20:05
mexiKONyou really should20:08
elbixiook, i'll read the apidoc, thanks philipp20:08
*** J1m has joined #zope3-dev20:13
*** __odie has joined #zope3-dev20:13
*** dagnachew has joined #zope3-dev20:15
mexiKONhi J1m20:16
mexiKONdo you have a minute? i would like to talk to you about a testing issue i have20:16
J1mmexiKON, what's up?20:19
*** bradb is now known as bradb|lunch20:21
mexiKONJ1m, so20:21
mexiKONJ1m, it's about Five20:21
mexiKONFive has >70 tests now20:21
mexiKONwe used to do testing like this: there was a "FiveTest" product that had all the mock objects; tests would play around with those20:22
mexiKONof course, this makes tests neither atomic nor is it easy to refactor. it's also harder and harder to add more tests because "FiveTest" grew20:22
mexiKONso got rid of FiveTest20:22
mexiKONproblem is this:20:22
mexiKONmore than 7020:23
J1mThat's not very many20:23
mexiKONit isn't20:23
mexiKONbut the number will increase20:23
mexiKONi hope20:23
mexiKONthat's my goal at least20:23
J1mSo what is wrong with unit tests?20:23
mexiKONso i want to make sure we do it right20:23
mexiKONproblem is this:20:23
mexiKONit's hard to not make a unit test an integration test in Zope 220:24
mexiKONthat's just the way tests are in Zope 220:24
mexiKONor, rather, ZopeTestCase-based tests20:24
mexiKONso, most tests rely on Five's ZCML being executed20:25
mexiKONbecause they need the services, etc. etc.20:25
J1mI'm not sure I believe that.  I was able to make unit tests for ZClasses.20:25
J1mwell, that's not entirely true20:25
J1manyway, go ahead with your question.20:25
mexiKONwe have *some* unit tests20:25
mexiKONbut very few20:26
mexiKONif you want test stuff like Five-based traversal, it's hard to test w/o the object actually in the whole Zope 2 context20:26
mexiKONso, most tests need Five ZCML being executed, also becaus ethey themselves would like to execute some ZCML so they want the metadirectives to be active20:27
mexiKONof course, by loading custom ZCML, each test case also loads stuff20:27
mexiKONwe want this to be cleaned up20:27
mexiKONcleanup after every test and reexecuting Five's ZCML + some custom ZCML is very very very slow20:27
mexiKONi thought of doing setUp/tearDown before adn after every test *suite*20:28
J1mYou should look at the layer support in my new test runner20:28
mexiKONbut, alas, the test runner grabs the individual test cases and puts them into itsown suite20:28
mexiKONlayer support... hmm. how does that work and help me?20:28
J1m(My new test runner is being used for ZODB 3.4, and will be used for Zope 2.9 and 3.2)20:28
J1mLook at the documentation first. :)20:29
mexiKONnote that this is Zope 2.7 in the wrost case20:29
mexiKONare you sure that the zope 2.7 test runner has this?20:29
J1mNo, Zope 2.7 doesn't have this.20:30
J1mNow does Zope 2.8 or 3.120:30
J1mNor does Zope 2.8 or 3.120:30
J1mIs Five limited to the Zope 2.7 test runner?20:30
mexiKONwell, we need a Zope 2.7 solution20:30
mexiKONwell, Five is at least included in Zope 2.8, so we're limited to what Zope 2.8 has20:31
J1mI wonder if that is strictly necessary20:31
mexiKONyou want to be able to run all the tests in a Zope 2.8 release with its rest runner, right?20:31
J1mIt would be nice20:31
J1malternatively you could use the same trick the Z3 functional tests currently use.20:32
*** __odie has quit IRC20:32
mexiKONwhat's that?20:32
J1mThey have a separate setup thing that they run once.20:32
J1mIt's a bit of a hack20:32
J1mBut the set up runs the first time a functional test runs20:32
J1mThat may not work for you though20:33
mexiKONwell, that's what  sorta have now20:33
mexiKONi want tesets to be more atomic, though20:33
J1mZ3 knows to run the functional tests after the unit tests.20:33
J1mI thought you wanted to share setup?20:33
mexiKONbetween the tests of one test suite20:33
mexiKONtake a doctest, for example20:33
mexiKONlots of tests, one setUp/tearDown20:33
mexiKON(I want to convert more stuff to doctests, too)20:34
J1mAnother problem is that ZCML doesn't provide any facilities for tearing down.20:34
J1mIt might be nice if it did.20:34
mexiKONwell, what i do is this:20:34
mexiKONi load zcml, but don't execute actions20:35
mexiKONno, wait, i used to do that20:35
mexiKONnever mind20:35
mexiKONwhat i do is:20:35
mexiKONto setUp, i load the zcml20:35
mexiKONto tearDown, i use placelesssetup.tearDown20:35
mexiKONall the services gone -> all zcml is gone20:35
J1msome z3 zcml modified module globals.20:36
J1msome z3 zcml modifies module globals.20:36
mexiKONyes, and some <five: /> directives monkey patch classes20:36
mexiKONbut that's a minor glitch20:36
mexiKONa glitch that even Zope 3 doesnt' cope with20:36
mexiKONunless you have explicit tearDown20:36
*** elbixio has quit IRC20:36
J1mIt does in the case of functional tests by running  them last.20:37
J1mdo you have a question? :)20:38
mexiKONi'm pondering20:38
mexiKONit's not obvious for me what to do20:38
mexiKONi guess i should a) convert as much to doctests20:38
mexiKONand b) try to minimize the amount of config needed for a test case so that individual and atomic setup/teardown isn't so expensive20:39
J1mYou could probably design a test suite that has it's own suite-based setUp and tearDown.20:39
mexiKONi did that already20:39
mexiKONit works when running each test module individually20:39
mexiKONpython2.3 tests/test_foobar.py20:39
mexiKONhowever, the zope test runner ignores the custom test suite, it uses a bulk one and inserts the tests into that20:40
J1mI guess that's true. So does mine, now that you mention it.20:42
J1mI don't have any ideas for you off the top of my head.20:43
J1mAnd I need to get back to thinking about multi-database zope 3 apps :)20:43
mexiKONok :)20:49
*** __odie has joined #zope3-dev20:52
*** srichter has quit IRC21:02
*** dagnachew has quit IRC21:06
*** bradb|lunch is now known as bradb21:09
*** bskahan has quit IRC21:21
*** mgedmin has joined #zope3-dev21:32
*** srichter has joined #zope3-dev21:32
*** ChanServ sets mode: +o srichter21:33
*** __gotcha_ has joined #zope3-dev21:59
*** __gotcha_ is now known as __gotcha21:59
*** dagnachew has joined #zope3-dev22:18
*** mgedmin has quit IRC22:28
projekt01Hi together, I'll working on the nested menu implementation for the Boston skin right now.22:28
projekt01Does anybody has a better idea then using the javasript "IE7" for build nested dropdown menus22:29
projekt01Otherwise I will implement it like here:22:30
projekt01Ok, that was the wrong copy/paste22:30
*** ignas has quit IRC23:01
*** __odie has quit IRC23:11
*** __odie has joined #zope3-dev23:12
*** alga has quit IRC23:21
*** __odie has quit IRC23:22
*** _projekt01 has joined #zope3-dev23:37
*** bradb is now known as bradb|afk23:39
*** projekt01 has quit IRC23:39
*** nederhoed has joined #zope3-dev23:53
*** nederhoed has left #zope3-dev23:57
*** nederhoed_ has joined #zope3-dev23:59

Generated by 2.15.1 by Marius Gedminas - find it at!