IRC log of #zope3-dev for Tuesday, 2009-07-28

*** sweh has quit IRC00:02
*** junkafarian has quit IRC00:21
*** dbfrombrc has quit IRC00:21
*** aaronv has joined #zope3-dev00:28
*** pcardune has quit IRC00:29
*** greenman has quit IRC00:29
*** benji has quit IRC00:33
J1msrichter_, ayt?00:51
srichter_J1m: yes00:52
*** srichter_ is now known as srichter00:52
J1mI'm trying to run the kgs tests.00:52
J1musing the trunk of zope.kgs00:52
srichterok00:52
srichteryep00:52
J1mwhen I run bin/generate-buildout, it says it wants a package-cfg-path00:53
srichterreally, that's strange00:53
srichterthe buildout.cfg specifies default parameters00:53
J1mI don't  think it needed that before.00:53
srichterno it does not00:53
srichterRoy Mathew got the tests to run this weekend00:53
srichterhuh00:54
srichterhow did you build? py25 bootstrap.py; ./bin/buildout -N00:54
J1mI used an existing Python 2.5 buildout to bootstrap it.00:55
J1mThen ran bin/buildout.00:55
srichterok, that should still all work as you know00:55
srichtercan you check the generated script whether it contains the arguments00:55
* J1m tries running buildout again00:55
J1mThe main script in buildout.py definitely wants an argument.00:56
J1mI wonder if your checkout is yup to date.00:56
srichterlet me try00:57
srichteryeah, I am all up-todate00:57
srichterlook into buildout.cfg00:57
srichterthe arguments should be listet there00:57
srichteroh00:57
srichteryou need zope.release00:57
srichter :-)00:57
srichterKGS is just the software00:57
srichterzope.kgs that is00:57
J1moh yeah00:57
srichtersorry00:58
J1mnp, I should have remembered that, but your note said zope KGS.00:58
srichterthe note said zope.release :-)00:59
J1mOh yeah, in the body.00:59
srichter(note that I got 12 failures this afternoon because some package does not tear down correctly)00:59
J1mwho reads the body? ;)01:00
srichterLOL01:00
srichterI remember that!01:00
srichternext time I put the entire message in the subject01:00
Theuni1Doesn't matter. Just put it somewhere on the internet. ;)01:00
srichterJ1m: btw, does zope.testing support running only tests of a particular layer?01:01
Theuni1yes01:01
Theuni1--layer01:01
srichterbecause 90% of the tests tun in the non-layer in 10% of the time01:01
Theuni1(No, I'm not trying to be Jim)01:01
srichtercool01:01
Theuni1there is no non-layer anymore01:01
srichterTheuni1: how is the layer called when I do not specify one?01:01
Theuni1it runs all tests in all layers01:02
Theuni1previously the unit-test layer was special but I refactored that about 15 months ago01:02
Theuni1the -f and -u are only command line shims above the --layer argument01:02
srichterbut I only want to run the tests that have no explicit layer specified01:02
Theuni1then they're in the unit test layer01:02
Theuni1as before01:02
srichterok, so -u?01:03
Theuni1yup01:03
srichtercool01:03
Theuni1or using --layer and the really long name that is attached to that layer01:03
srichterI prefer -u :-)01:03
Theuni1That's why I left it there ;)01:04
srichterhaving all these layers really slows down test running01:04
Theuni1only depends on your layers and how much setup they perform01:04
Theuni1it's not the layerism itself01:05
srichteroh, I know01:05
Theuni1As long as you don't have about 50 billion layers01:05
srichterbut most of them bring up a significant part of the app server01:05
Theuni1yup01:05
Theuni1If you do inheritance then some of them will be setup in a combined fashion.01:05
Theuni1IIRC01:05
srichterrunning all the tests in the KGS takes 50 minutes and with the exception of the ZODB test my feeling is that 50% of the time is spent in layer setup01:06
srichterright01:06
Theuni1but in general, count about 2 seconds for a layer when a new process needs to be spawned01:06
srichterexcept that packages do not do that01:06
Theuni1Hmm.01:06
Theuni1When I wrote the compattest thing I was able to get that time down by using sheer computing power.01:06
Theuni1I can run the compattests in < 2 minutes01:06
srichteryep, you can, absolutely01:06
srichterfor all 16k+ tests in less than 2 mins?01:07
Theuni1I think that was the time i got01:07
Theuni1ran on a 32gb 8 core machine with fast disks01:08
srichterwow, that's impressive01:08
srichtersomeone should sponsor a machine to me to do this sort of work01:08
Theuni1;)01:08
srichterthat would drastically improve my efficiency01:08
Theuni1Yeah, those babies are nice :)01:08
srichterand they are really not that expensive anymore wither01:09
Theuni1Nope01:09
Theuni1Can get them for less than 3k EUR01:09
Theuni1If you build them yourself and don't need support contracts probably even cheaper.01:09
* Theuni1 goes back to selecting a surround system01:09
srichterTheuni1: yeah see you later; are you going to be around next Tuesday?01:10
srichterin Halle that is?01:10
Theuni1It's in Halle. As I said: unfortunately I'm not around. I'll be in Helsinki at that time.01:10
Theuni1But at least three of our people are expecting you.01:10
Theuni1That would be Zagy, Wolfgang and Thomas.01:11
srichtercool, I'll send an E-mail out later this week outlining an agenda and exact times, but it will be on Tuesday01:11
Theuni1fine with me ;)01:12
*** J1m has quit IRC01:12
*** pcardune has joined #zope3-dev01:15
*** aaronv has quit IRC01:19
*** dbfrombrc has joined #zope3-dev01:23
*** Theuni1 has quit IRC01:24
*** dbfrombrc has quit IRC01:25
*** dbfrombrc has joined #zope3-dev01:25
*** Theuni1 has joined #zope3-dev01:27
*** aaronv has joined #zope3-dev01:29
*** junkafarian has joined #zope3-dev01:29
*** jamur2 has quit IRC01:33
*** jamur2 has joined #zope3-dev01:33
*** jamur2_ has joined #zope3-dev01:34
*** jamur2 has quit IRC01:37
*** allisterb_ has quit IRC01:37
*** allisterb has joined #zope3-dev01:39
*** yvl has quit IRC01:39
*** greenman has joined #zope3-dev01:43
*** romanofski has joined #zope3-dev01:44
*** jamur2_ has quit IRC01:50
*** jamur2 has joined #zope3-dev01:50
*** jamur2 has quit IRC01:51
*** allisterb_ has joined #zope3-dev01:53
*** drudi has quit IRC01:57
*** Theuni1 has quit IRC01:58
*** febb has quit IRC02:03
*** allisterb has quit IRC02:04
*** lurkymclurkleton has quit IRC02:06
*** dbfrombrc is now known as dbfrombrc|away02:18
*** aaronv has quit IRC02:24
*** nathany has quit IRC02:25
*** redir has quit IRC02:31
*** hazmat has joined #zope3-dev02:47
*** ChanServ sets mode: +o hazmat02:47
*** jhauser has quit IRC02:51
*** BjornT_ has quit IRC03:03
*** BjornT has joined #zope3-dev03:03
junkafariananyone know if there is an ellipsis option to set the default for running doctests through nose in setup.cfg?03:28
*** junkafarian has quit IRC03:45
*** gary_poster has joined #zope3-dev03:45
*** redir has joined #zope3-dev03:50
*** pcardune has quit IRC03:52
*** skriv1 has joined #zope3-dev04:09
*** davisagli has left #zope3-dev04:09
*** skriv has quit IRC04:26
*** greenman has quit IRC04:55
*** drudi has joined #zope3-dev05:00
*** gary_poster has quit IRC05:01
*** alecm has quit IRC05:02
*** BjornT has quit IRC05:02
*** markusleist has joined #zope3-dev05:15
*** davisagli has joined #zope3-dev05:22
*** stub has joined #zope3-dev06:04
*** drudi has quit IRC06:06
*** pcardune has joined #zope3-dev06:07
*** pcardune has quit IRC06:10
*** afd__ has joined #zope3-dev06:16
*** alecm has joined #zope3-dev06:33
*** BjornT has joined #zope3-dev06:36
*** redir has quit IRC06:43
*** pcardune has joined #zope3-dev06:48
*** jpcw has quit IRC06:55
*** pcardune has quit IRC07:03
*** febb has joined #zope3-dev07:09
*** alecm has quit IRC07:18
*** greenman has joined #zope3-dev07:40
*** sunoano has quit IRC07:40
*** sunoano has joined #zope3-dev07:42
*** hdima has joined #zope3-dev08:17
*** hdima has quit IRC08:18
*** skriv1 has quit IRC08:18
*** Theuni1 has joined #zope3-dev08:22
*** JaRoel|4D has quit IRC08:25
*** jukart has joined #zope3-dev08:42
*** afd__ has quit IRC08:42
*** agroszer has joined #zope3-dev08:59
*** __mac__ has joined #zope3-dev09:05
*** febb has quit IRC09:23
*** JaRoel|4D has joined #zope3-dev09:23
*** yvl has joined #zope3-dev09:32
*** yvl has quit IRC09:34
*** kaeru has joined #zope3-dev09:34
*** linmyat has joined #zope3-dev09:50
*** goschtl has joined #zope3-dev10:01
*** jhauser has joined #zope3-dev10:07
*** iham has joined #zope3-dev10:09
*** __mac__ has quit IRC10:09
*** __mac__ has joined #zope3-dev10:10
*** sweh has joined #zope3-dev10:23
*** romanofski has quit IRC10:26
*** davisagli has left #zope3-dev10:35
*** iham has quit IRC10:40
*** junkafarian has joined #zope3-dev10:43
*** __mac__ has quit IRC10:48
*** ktwilight_ has joined #zope3-dev10:53
*** ktwilight has quit IRC10:55
*** ktwilight__ has joined #zope3-dev10:58
*** agroszer has quit IRC11:02
*** fcorrea has quit IRC11:02
*** mcdonc has quit IRC11:02
*** VladDrac has quit IRC11:02
*** agroszer has joined #zope3-dev11:02
*** fcorrea has joined #zope3-dev11:02
*** mcdonc has joined #zope3-dev11:02
*** VladDrac has joined #zope3-dev11:02
*** junkafarian_ has joined #zope3-dev11:07
*** vaab has joined #zope3-dev11:08
*** ignas has joined #zope3-dev11:09
vaabhi everyone, I've found a bug in ZConfig and I have a patch, what is the best way to get it reviewed (and integrated) ? (I've already sent it on launchpad, but it seem not really alive)11:10
*** ktwilight_ has quit IRC11:12
*** __mac__ has joined #zope3-dev11:14
*** iham has joined #zope3-dev11:14
*** romanofski has joined #zope3-dev11:16
*** pelle__ has joined #zope3-dev11:17
Theuni1vaab: launchpad is a reasonable place. if it languishes for long, you might consider posting a reference to the launchpad entry on the developer list.11:22
*** pelle__ has quit IRC11:25
*** pelle__ has joined #zope3-dev11:26
*** junkafarian_ has quit IRC11:29
*** pelle__ has quit IRC11:29
*** greenman has quit IRC11:31
vaabTheunil: thx for the tip11:37
*** iham has quit IRC11:37
*** markusleist has left #zope3-dev11:47
*** markusleist has joined #zope3-dev11:48
*** jpcw has joined #zope3-dev11:57
*** kaeru has quit IRC12:21
Theuni1np12:27
*** ccomb has joined #zope3-dev13:14
*** linmyat has quit IRC13:17
*** MJ has joined #zope3-dev13:21
*** pelle__ has joined #zope3-dev13:46
*** aaronv has joined #zope3-dev13:49
*** redir has joined #zope3-dev13:54
*** drudi has joined #zope3-dev14:03
*** junkafarian_ has joined #zope3-dev14:08
*** junkafarian_ has quit IRC14:09
*** drudi has quit IRC14:15
*** iham has joined #zope3-dev14:16
*** dunny has quit IRC14:21
*** iham has quit IRC14:27
*** projekt01 has joined #zope3-dev14:29
*** aaronv has quit IRC14:33
*** davisagli has joined #zope3-dev14:53
*** redir has quit IRC14:55
*** iham has joined #zope3-dev15:06
*** malthe is now known as malthe|away15:21
*** gary_poster has joined #zope3-dev15:29
*** drudi has joined #zope3-dev15:32
*** redir has joined #zope3-dev15:38
*** jamur2 has joined #zope3-dev15:41
*** sunoano has quit IRC15:41
*** jamur2 has quit IRC15:41
*** sunoano has joined #zope3-dev15:42
*** benji has joined #zope3-dev15:43
*** redir has quit IRC15:55
*** J1m has joined #zope3-dev15:59
*** redir has joined #zope3-dev16:00
*** redir has quit IRC16:01
*** romanofski has quit IRC16:02
*** jamur2 has joined #zope3-dev16:02
*** iham has quit IRC16:03
*** hexsprite has joined #zope3-dev16:04
*** lurkymclurkleton has joined #zope3-dev16:05
*** stub has quit IRC16:06
*** aaronv has joined #zope3-dev16:07
*** hexsprite has quit IRC16:09
*** Theuni1 has quit IRC16:12
*** Theuni2 has joined #zope3-dev16:12
*** __mac__ has quit IRC16:56
*** jdobrien has joined #zope3-dev17:09
jdobriencan anyone point me to information using zope's get/query utilities for dependency injection....specifically I want to change the provider for unit tests17:10
jdobrienjdobrien, preferably without a monkey patch hack17:11
*** ccomb has quit IRC17:18
srichterjdobrien: in your test setup you simply need to do a zope.component.provideUtility() call with your testing component17:19
*** sweh has quit IRC17:19
srichtermany packages do that17:19
srichterfor example zope.keyreference17:19
*** ccomb has joined #zope3-dev17:21
jdobriensrichter, thanks...i guess it looked to easy :)17:21
*** ktwilight__ has quit IRC17:38
*** ktwilight__ has joined #zope3-dev17:38
*** davisagli has quit IRC17:39
*** kaeru has joined #zope3-dev17:50
*** ktwilight__ has quit IRC17:53
*** drudi has quit IRC17:53
*** drudi has joined #zope3-dev17:53
*** redir has joined #zope3-dev17:56
*** ktwilight__ has joined #zope3-dev17:57
*** ktwilight__ has joined #zope3-dev17:58
*** projekt01 has quit IRC18:07
*** iham has joined #zope3-dev18:11
*** kursor has joined #zope3-dev18:19
*** iham has quit IRC18:20
*** skriv has joined #zope3-dev18:23
*** nathany has joined #zope3-dev18:30
J1msrichter, ayt?18:33
srichterJ1m: yeah18:33
J1mI'm getting lots of errors running tests of zope.release.18:33
J1mw python 2.5 on a 32-bit ubuntu machine18:34
srichterhow many?18:34
srichterI get 12 right now due to the "provider" tales expression type not being registered18:34
srichtersomehow some tearing down does not work right18:35
J1m17 failures, 4 errors out of 16465 tests.18:35
srichterI have my suspicions, since something similar happens with the vocabulary registry18:37
srichtermmh, that's definitely more than I am seeing, though some tests might depend on an external script or so18:37
srichterlet's start with the errors. which tests/packages cause them?18:37
J1mTests with errors:18:37
J1m   /home/jim/.buildout/eggs/lovely.recipe-1.0.0b5-py2.5.egg/lovely/recipe/egg/README.txt18:37
J1m   /home/jim/.buildout/eggs/lovely.recipe-1.0.0b5-py2.5.egg/lovely/recipe/fs/README.txt18:37
J1m   /home/jim/.buildout/eggs/lovely.recipe-1.0.0b5-py2.5.egg/lovely/recipe/i18n/README.txt18:37
J1m   /home/jim/.buildout/eggs/lovely.recipe-1.0.0b5-py2.5.egg/lovely/recipe/importchecker/README.txt18:37
J1mTests with failures:18:37
J1m   /home/jim/.buildout/eggs/keas.profile-0.1.0-py2.5.egg/keas/profile/README.txt18:37
J1m   /home/jim/.buildout/eggs/keas.unit-1.0.0-py2.5.egg/keas/unit/README.txt18:37
J1m   /home/jim/.buildout/eggs/lovely.remotetask-0.4-py2.5.egg/lovely/remotetask/processor.txt18:37
J1m   /home/jim/.buildout/eggs/z3c.coverage-1.1.3-py2.5.egg/z3c/coverage/README.txt18:37
J1m   /home/jim/.buildout/eggs/z3c.formui-2.0.0-py2.5.egg/z3c/formui/README.txt18:37
J1m   /home/jim/.buildout/eggs/z3c.macro-1.2.1-py2.5.egg/z3c/macro/README.txt18:37
J1m   /home/jim/.buildout/eggs/z3c.macroviewlet-1.0.0-py2.5.egg/z3c/macroviewlet/README.txt18:37
J1m   /home/jim/.buildout/eggs/z3c.menu.ready2go-0.6.0-py2.5.egg/z3c/menu/ready2go/README.txt18:37
J1m   /home/jim/.buildout/eggs/z3c.menu.ready2go-0.6.0-py2.5.egg/z3c/menu/ready2go/zcml.txt18:37
J1m   /home/jim/.buildout/eggs/z3c.pagelet-1.1.0-py2.5.egg/z3c/pagelet/README.txt18:37
J1m   /home/jim/.buildout/eggs/z3c.recipe.egg-0.2-py2.5.egg/z3c/recipe/egg/tests/../setup.txt18:37
J1m   /home/jim/.buildout/eggs/z3c.wizard-0.6.0-py2.5.egg/z3c/wizard/README.txt18:37
J1m   /home/jim/.buildout/eggs/zc.loggermonitor-0.1-py2.5.egg/zc/loggermonitor/README.txt18:37
J1m   /home/jim/.buildout/eggs/zc.recipe.testrunner-1.2.0-py2.5.egg/zc/recipe/testrunner/README.txt18:37
J1m   /home/jim/.buildout/eggs/zc.recipe.testrunner-1.2.0-py2.5.egg/zc/recipe/testrunner/bugfixes.txt18:37
J1m   /home/jim/.buildout/eggs/zope.contentprovider-3.5.0-py2.5.egg/zope/contentprovider/README.txt18:37
J1m   /home/jim/.buildout/eggs/zope.viewlet-3.5.0-py2.5.egg/zope/viewlet/README.txt18:37
J1mTotal: 16465 tests, 17 failures, 4 errors in 131 minutes 38.403 seconds.18:38
J1m:)18:38
J1m(benji hates me)18:38
benjimy eyes!18:38
* J1m tries running tests individually.18:38
benjiheh18:38
srichterJ1m: ok, keas.unit is due to the fact that you do not have the units script installed; we should make those tests optional18:38
J1mwhy are keas tests in the zope.release tests?18:39
srichterJ1m: zope.contentprovider, z3c.macroviewlet, z3c.menu.*, zope.viewlet are most likely all due to the missing provider registration18:39
J1mThe lovely recipe tests fail for me even when run individually.18:39
srichterkeas.unit is a svn.zope.org package18:39
srichterok, we should take lovely.recipe out of the KGS then18:39
J1mI thought you had all of the tests passing.18:39
J1mwhy does it matter whether it is a zope.org package?18:40
srichteryeah, then I added some more, which switched some test orders around that broke things and I did not investigate yet18:40
*** JaRoel|4D has quit IRC18:40
J1mdang18:40
srichterpretty much all tests should pass individually18:41
srichterthat's how I tested it18:41
srichterthe full run just takes too long; though the "provider" problem is visible when running "-u" which takes only 5 mins on my machine18:41
J1mbtw, I see no good reason to run the zodb tests.18:42
srichterright, I usually comment them out too :-)18:42
srichterwe can just set them to tested=False18:42
J1mzc.recipe.testrunner tests fail for me. :(18:43
*** jukart has quit IRC18:43
J1mError: Couldn't find a distribution for 'zope.exceptions'.18:43
J1mI wonder what the heck it;s doing.18:44
srichterah, where?18:44
srichterthis is probably due to the fact that zope.exception is not listed as a dependency in the setUp()18:44
srichterI had to fix that before in one of the recipes18:44
J1mREADME.txt line 150.18:44
srichterwhich package?18:44
J1mzc.recipe.testrunner18:45
srichtermmh, I thought I had fixed that18:45
J1mI get a failure for z3c.recipe.egg.18:46
J1mSomehow, I don't think we should be testing recipes here.18:46
J1mBut it depends on what the goals are. :)18:47
srichterI rather test more, sub-KGSs can then be derived from it18:47
J1mMy goal is to try to make sure the packages commonly used to build web applications actually work together.18:47
J1mWe need people who modify core packages to run core tests.18:47
J1mThese ancillary failures will discourage that.18:48
srichterMy explicit goal is to be able to say: Mmh, that package is useful right now I just want to be able to use it18:48
srichteryou have to define core18:48
J1mfair enough.18:48
J1mBut we should do that.18:48
srichterI bet you that our commonly used packages vary a bit18:48
srichterI think though that it would be useful to write a script that allows you to extract sub-lists18:49
srichterfor example for a definition of "core"18:49
J1mRight now, we, ZC, are in a tough spot. We need to update packages but it's hard finding a set that work together.18:49
J1mSome sort of KGS could help a lot.18:49
srichterI hear you18:49
srichterI was in the same spot too several times in the last few months18:50
J1mso I'm most concerned about the packages that actually interact.18:50
srichterI could not even start a new project because I could not find a working set18:50
srichterright, and those pacakges are all passing now18:50
*** iham has joined #zope3-dev18:50
srichtermaybe we could make use of layers18:50
*** goschtl has left #zope3-dev18:50
srichterI mean levels18:50
srichterso the most basic tests run in level 1, then level 2, etc18:51
srichterthis way we can tell core developers to only worry about level 1 or 2 and release managers worry about higher-levels18:51
J1mI suggest that:18:52
J1m1. It would be easier to implement this as separate sets of tests.  Changing test levels would be to invasive. Levels were a bad idea.18:53
J1m2. We should automate running tests individually.18:53
srichterI thought Theuni had set up individual testing before18:54
J1mBTW, I don't necessarily have a problem with running application level tests to catch backward-compatibility breakage, as long as the application tests only due to library breakage.18:54
J1mHe did, but I think the thing they came up with didn't have a way to lock down version. IOW, no known good set.18:55
srichterso what would be the best way to set up individual test runs that do not take forever to complete?18:55
J1malso, it was rather ad hoc.  It was appropriate for a sprint, but not as a development check.18:55
J1mToo bad he's not around atm.18:56
srichterTheuni2: ping18:56
J1mWell, the thing that generates a test script now could generate set of scripts, or it could generate a master script that just invoked the script we have now.18:57
J1mIt would take just as long unless the paster script ran the individual scripts in separate threads.18:58
J1mIt might be cool to have an option in the test runner to automate this.18:58
J1mPerhaps something that puts the unit packages for a package in a separate layer.18:59
srichterunit pacakges?18:59
srichterthere is already a layer for unit tests that gets invoked when setting "-u"18:59
J1mThen we could use the -j switch to run each test/layer in a separate subprocess.18:59
srichterright19:00
J1mright. Maybe that could be hacked to create multiple unit layers.19:00
srichterbtw, I think we can reuse some z3c.recipe.compattest code to make this work19:00
srichterit basically executes every package in its own sub-process19:01
J1mThat sounds good.19:01
*** MJ has quit IRC19:01
srichterso basically, we could generate a "test-<PACKAGE>" script for each pacakge in the KGS19:02
J1mThat would be one way.19:02
srichterthen use a master script to control the run19:02
srichterit would also allow you to run a package individually19:03
J1mOr we could arrange that the unit tests for each package end up in a different layer.19:03
srichterright19:03
J1mwe already have that with test selection.19:03
J1mWhat would be nice is to have a "core" test script that just tests libraries commonly used to actually run zope apps.19:04
srichterif ./bin/test -m <PACKAGENAME> would work reliably, we just need to have an additional script19:04
Theuni2srichter: pong19:04
srichterTheuni2: see discussion above19:04
* Theuni2 reads19:04
J1msrichter, you mean bin/test -s package_name, which works reliably for me.19:04
srichterJ1m: I would implement that by having a script that can produce a sub-list of the large KGS19:05
srichterJ1m: oh, I might have used the wrong option then19:05
J1m-m selects test modules.19:05
J1mBut it accepts regexes, so can be used to select packages sometimes.19:06
srichter[srichter@localhost test]$ ./bin/test -vpc1 -s zc.recipe.testrunner19:06
srichterRunning tests at level 119:06
srichterTotal: 0 tests, 0 failures, 0 errors in 0.000 seconds.19:06
Theuni2actually the compattests  should support your usecase19:07
J1mIt appears that the tests for zc.recipe.testrunner actually install things and may depend on what you have installed.19:07
Theuni2I'm not sure it's obvious. I'll have to re-investigate.19:08
J1msrichter, although I don't actually care about zc.recipe.testrunner in this context. :)19:08
srichteryeah19:08
Theuni2I'm on a run currently and I'm not sure what the actual question is after reading your discussion.19:08
srichterTheuni2: we just try to figure out the best way to run all tests, but running the tests of each package in its own process19:09
srichterTheuni2: looking at compattest, it looks as you are doing just that19:09
J1mIMO, I think we want to define a set of packages that are commonly used in a running zope(3) application.  We need to make sure those packages work together.19:09
Theuni2yeah, that's what the compat tests do and they also allow starting them in parallel. I think it should also be possible to specify versions.19:09
J1mThat's the goal. :)19:09
Theuni2right. the compattests are there so you can work on a package and have the tests of other packages which will be future consumers of your changes checked right away.19:10
J1mI think we need to be able to test these packages in a known configuration that we update only after passing the tests on a specific set of platforms.19:10
srichterright19:11
Theuni2Even if I'm wrong and it currently might only support testing other packages trunks I think the compattest recipe can be modified in a simple manner to consume e.g. a version section.19:11
J1mwhere the platforms include: windows, linux, max, Python 2.619:11
J1mwe need to make it as easy as people who work on these packages to do the right thing.19:12
Theuni2jop19:12
*** ignas has quit IRC19:12
Theuni2I need to leave now, though :)19:12
srichterTheuni2: see ya19:12
J1mI gotta go too.19:12
J1mThanks for the great work srichter !19:12
srichterJ1m: note that the KGS software also support extensions, like buildout19:12
srichterso we could have several small ones, like for ZODB, Zope 3, zc.buildout and extras and then one large one that pulls them together19:13
srichterno problem; I wish I would not be the only one doing this relatively boring work ;-)19:14
*** ktwilight__ has quit IRC19:23
*** jpcw has quit IRC19:32
*** jdobrien has quit IRC19:37
*** jdobrien has joined #zope3-dev19:41
*** jpcw has joined #zope3-dev19:50
*** alecm has joined #zope3-dev19:51
*** aaronv has quit IRC19:55
*** aaronv has joined #zope3-dev19:56
*** sweh has joined #zope3-dev20:03
*** redir has quit IRC20:09
*** redir has joined #zope3-dev20:12
*** pcardune has joined #zope3-dev20:20
*** jukart has joined #zope3-dev20:28
*** iham has quit IRC20:32
*** pcardune has quit IRC20:33
*** jukart has quit IRC20:43
*** kapil__ has joined #zope3-dev20:47
*** jukart has joined #zope3-dev21:02
*** sweh has quit IRC21:04
*** junkafarian has quit IRC21:07
*** aaronv has quit IRC21:13
*** aaronv has joined #zope3-dev21:14
*** junkafarian has joined #zope3-dev21:14
*** __mac__ has joined #zope3-dev21:15
*** Theuni2 has quit IRC21:18
*** Theuni1 has joined #zope3-dev21:18
*** kursor_ has joined #zope3-dev21:19
*** kursor has quit IRC21:19
*** Theuni1 has quit IRC21:24
*** ktwilight has joined #zope3-dev21:26
*** kursor_ has quit IRC21:31
*** kursor has joined #zope3-dev21:31
*** jukart has quit IRC21:46
*** sweh has joined #zope3-dev21:56
*** kursor has quit IRC21:57
*** ktwilight has quit IRC22:06
*** agroszer has quit IRC22:09
*** Theuni1 has joined #zope3-dev22:24
*** kapil__ has quit IRC22:28
*** __mac__ has quit IRC22:57
*** mintsauce has joined #zope3-dev23:00
mintsauceI'm having problems with zope.app.renderer23:06
*** dunny has joined #zope3-dev23:07
mintsauceThe exact error is: IOError: [Errno 2] No such file or directory: '/home/liveweb/site/eggs/zope.app.renderer-3.5.0-py2.4.egg/zope/app/renderer/meta.zml'23:07
mintsaucehah ... and pasting that, I've just noticed the problem23:08
mintsauceNope, thought the mis-spelt zcml link (as zml) was the problem, but still get the same error23:09
mintsauceI note one of the latest updates to the package says 'Removed deprecated meta.zcml stuff and zpkg stuff.'23:09
mintsauceRemoving the meta.zcml reference i now get23:15
mintsauceConfigurationError: ('Invalid value for', 'package', 'ImportError: Module z3c has no global macro')23:15
mintsaucefor: <include package="z3c.macro" file="meta.zcml" />23:16
*** dbfrombrc has joined #zope3-dev23:32
*** dbfrombrc has quit IRC23:33
*** junkafarian has quit IRC23:34
*** junkafarian has joined #zope3-dev23:37
*** jpcw has quit IRC23:40
*** malthe|away is now known as malthe23:42
*** sunoano has quit IRC23:43
*** ccomb has quit IRC23:43
*** ccomb has joined #zope3-dev23:44
*** sweh has quit IRC23:46
*** sunoano has joined #zope3-dev23:48
*** pcardune has joined #zope3-dev23:50
pcarduneanyone know why an OOBTree would enter into an inconsistent state where [tree[key] for key on tree.keys()] would yield a KeyError?23:53
pcarduneMaybe Theuni1 or srichter know?23:53
Theuni1ist that all the code that is there?23:53
Theuni1you aren't modifying it while iterating?23:53
Theuni1from a different thread?23:53
pcarduneTheuni1: pretty sure I'm not modifying it from a different thread23:54
Theuni1so presumably it's already inconsistent when the transaction starts?23:54
pcarduneoh, no, modified before in that thread23:55
pcarduneso potentially put into an inconsistent state in that thread23:56
pcarduneTheuni1: one thing I did notice is a None key23:57
Theuni1Wouldn't think that would make a problem.23:58
Theuni1What are your other keys?23:58
pcarduneTheuni1: all persistent objects23:58
*** greenman has joined #zope3-dev23:59

Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!