| *** Pumukel has joined #zope | 00:06 | |
| *** qengho has quit IRC | 00:07 | |
| *** qengho has joined #zope | 00:09 | |
| *** taufik has quit IRC | 00:21 | |
| *** xyld has joined #zope | 00:25 | |
| *** SpankyFromBRC has joined #zope | 00:27 | |
| *** xyld1 has joined #zope | 00:29 | |
| *** xyld has quit IRC | 00:29 | |
| *** Spanktar has quit IRC | 00:29 | |
| *** SpankyFromBRC is now known as Spanktar | 00:29 | |
| *** avoinea has quit IRC | 00:44 | |
| *** avoinea has joined #zope | 00:48 | |
| *** taufik has joined #zope | 00:53 | |
| *** MrTango has joined #zope | 01:00 | |
| *** avoinea has quit IRC | 01:02 | |
| *** supton has quit IRC | 01:16 | |
| *** SpankyFromBRC has joined #zope | 01:17 | |
| *** Spanktar has quit IRC | 01:19 | |
| *** SpankyFromBRC is now known as Spanktar | 01:19 | |
| *** MrTango has quit IRC | 01:31 | |
| *** xyld1 has quit IRC | 01:39 | |
| *** fdrake has quit IRC | 01:41 | |
| *** Pumukel has quit IRC | 01:42 | |
| *** taufik has quit IRC | 01:48 | |
| *** taufik has joined #zope | 01:50 | |
| *** taufik has quit IRC | 01:50 | |
| *** dayne has quit IRC | 02:02 | |
| *** mindthief414 has quit IRC | 02:07 | |
| *** menesis has quit IRC | 02:24 | |
| *** xyld has joined #zope | 02:28 | |
| *** qengho has quit IRC | 03:12 | |
| *** qengho has joined #zope | 03:12 | |
| *** daMaestro has quit IRC | 03:17 | |
| *** Spanktar has quit IRC | 03:18 | |
| *** mindthief414 has joined #zope | 04:08 | |
| *** RichyB has quit IRC | 04:11 | |
| *** kosh has joined #zope | 04:16 | |
| *** xyld has quit IRC | 04:19 | |
| *** ajmitch has quit IRC | 04:22 | |
| *** ajmitch has joined #zope | 04:30 | |
| *** ajmitch has joined #zope | 04:30 | |
| *** mr_jolly has quit IRC | 04:32 | |
| *** RichyB has joined #zope | 04:41 | |
| *** xyld has joined #zope | 04:48 | |
| *** xyld has quit IRC | 04:51 | |
| *** xyld has joined #zope | 04:58 | |
| *** xyld has quit IRC | 05:04 | |
| *** mcdonc has quit IRC | 05:13 | |
| *** mcdonc has joined #zope | 05:13 | |
| *** xyld has joined #zope | 06:03 | |
| *** tiwula has joined #zope | 06:10 | |
| *** xyld1 has joined #zope | 06:24 | |
| *** xyld has quit IRC | 06:27 | |
| *** dayne has joined #zope | 06:44 | |
| *** kosh has quit IRC | 07:08 | |
| *** xyld1 has quit IRC | 07:41 | |
| *** xyld has joined #zope | 07:42 | |
| *** xyld has quit IRC | 07:58 | |
| *** mcdonc has quit IRC | 07:58 | |
| *** mcdonc has joined #zope | 07:58 | |
| *** calvinhp_ has quit IRC | 08:04 | |
| *** zagy has joined #zope | 08:04 | |
| *** zagy has quit IRC | 08:04 | |
| *** tiwula has quit IRC | 08:07 | |
| *** calvinhp has joined #zope | 08:08 | |
| *** prg3 has quit IRC | 08:17 | |
| *** prg3 has joined #zope | 08:20 | |
| *** zagy has joined #zope | 09:06 | |
| *** agroszer has joined #zope | 09:09 | |
| *** __mac__ has joined #zope | 09:26 | |
| *** zagy has quit IRC | 09:27 | |
| *** zagy has joined #zope | 09:27 | |
| *** avoinea has joined #zope | 10:02 | |
| *** goschtl has joined #zope | 10:07 | |
| *** goschtl has joined #zope | 10:08 | |
| *** avoinea has quit IRC | 10:13 | |
| *** avoinea has joined #zope | 10:13 | |
| *** tmog has joined #zope | 10:22 | |
| *** emidal has joined #zope | 10:26 | |
| *** taufik has joined #zope | 10:32 | |
| *** taufik has left #zope | 10:34 | |
| *** mr_jolly has joined #zope | 10:35 | |
| *** thetet has joined #zope | 10:35 | |
| *** giacomos has joined #zope | 10:54 | |
| *** tmog has quit IRC | 11:04 | |
| *** tmog has joined #zope | 11:05 | |
| *** thetet1 has joined #zope | 11:11 | |
| *** maurits has joined #zope | 11:13 | |
| *** thetet has quit IRC | 11:14 | |
| *** mitchell`off is now known as mitchell` | 11:19 | |
| *** MrTango has joined #zope | 11:42 | |
| *** menesis has joined #zope | 11:48 | |
| *** evilbungle has quit IRC | 11:49 | |
| *** evilbungle has joined #zope | 12:00 | |
| *** mindthief414 has quit IRC | 12:17 | |
| *** fredvd has joined #zope | 13:00 | |
| *** zagy has quit IRC | 13:04 | |
| *** phimic has joined #zope | 13:08 | |
| phimic | does zope intereprete a webdav GET on 8091 different than a GET on default http port 8081? | 13:09 |
|---|---|---|
| *** menesis has quit IRC | 13:14 | |
| *** tmog has quit IRC | 13:15 | |
| *** crazydiamond has joined #zope | 13:19 | |
| *** KageSenshi has quit IRC | 13:21 | |
| *** KageSenshi has joined #zope | 13:22 | |
| *** tmog has joined #zope | 13:36 | |
| *** phimic has left #zope | 13:37 | |
| *** mwilkes|away is now known as MatthewWilkes | 13:38 | |
| mgedmin | agroszer, winbot question | 13:59 |
| mgedmin | do you have setuptools installed into the system pythons? | 13:59 |
| mgedmin | as in the real setuptools, rather than distribute? | 13:59 |
| agroszer | depends on the python folder | 13:59 |
| agroszer | there's *_clean which never have setuptools | 13:59 |
| agroszer | but package tests mostly run with setuptools | 14:00 |
| mgedmin | would it be hard to switch to distribute? | 14:00 |
| agroszer | where setuptools is setuptools | 14:00 |
| mgedmin | because https://bitbucket.org/pypa/setuptools/issue/5 breaks everything :-) | 14:00 |
| mgedmin | we can wait for that issue to be resolved, or we can install distribute ourselves, since buildout is unable to do so | 14:01 |
| *** menesis has joined #zope | 14:01 | |
| *** crazydiamond has quit IRC | 14:01 | |
| agroszer | setuptools vs distribute s*cks | 14:01 |
| agroszer | I guess switching to distribute isn't hard | 14:02 |
| mgedmin | yeah, python packaging | 14:03 |
| mgedmin | hpk42 on twitter suggested a python packaging conference | 14:04 |
| mgedmin | mcdonc dubbed it #sadconf :-) | 14:04 |
| mgedmin | eventually setuptools 0.7 will be out, and distribute will die | 14:04 |
| mgedmin | (and perhaps distlib will replace setuptools in the far far future, in a galaxy far far away) | 14:04 |
| agroszer | my basic problem is that e.g. distribute guys ignore the winbot tests | 14:05 |
| agroszer | which never ever passed | 14:05 |
| agroszer | anyway, which way do we want to go? | 14:07 |
| mgedmin | installing distribute (instead of setuptools) into the pythons is the quickest fix, IMHO | 14:08 |
| agroszer | okay... soon | 14:10 |
| agroszer | mgedmin, can I disconnect you from winbot? | 14:13 |
| agroszer | or logoff | 14:13 |
| mgedmin | oh, am I connected again? sorry | 14:15 |
| mgedmin | I keep forgetting that closing rdesktop doesn't log me off | 14:16 |
| agroszer | please logoff | 14:16 |
| *** yvl has quit IRC | 14:16 | |
| agroszer | winbot is low on RAM and one shell takes quite a lot of RAM | 14:16 |
| mgedmin | ouch | 14:17 |
| mgedmin | is it possible to enable auto-logoff after a timeout? | 14:17 |
| agroszer | dunno | 14:17 |
| agroszer | ouch setuptools supports 64bit? | 14:19 |
| agroszer | okay, mgedmin 25_32, 26_32, 26_64, 27_32, 27_64 have now distribute 0.6.44 | 14:27 |
| agroszer | where I think package tests are run with 26_32 | 14:27 |
| betabug | hey agroszer! 45km/h wind this morning, with 75km/h wind gusts | 14:36 |
| agroszer | heya | 14:37 |
| agroszer | meeeh | 14:37 |
| betabug | south wind, so it's quite warm too | 14:38 |
| mgedmin | whee, travis job for ZODB runs 22 test from a test suite that has more than a thousand | 14:51 |
| mgedmin | looks like zope-testrunner quits after running the first test | 14:51 |
| *** xyld has joined #zope | 15:11 | |
| *** xyld1 has joined #zope | 15:16 | |
| *** fdrake has joined #zope | 15:17 | |
| *** xyld has quit IRC | 15:18 | |
| *** xyld has joined #zope | 15:25 | |
| *** xyld1 has quit IRC | 15:25 | |
| *** xyld has quit IRC | 15:25 | |
| *** goschtl has quit IRC | 15:34 | |
| *** yvl has joined #zope | 15:35 | |
| *** thetet1 has quit IRC | 15:45 | |
| *** kosh has joined #zope | 15:51 | |
| *** tmog has quit IRC | 15:54 | |
| *** tmog has joined #zope | 15:55 | |
| *** foodoo has quit IRC | 16:05 | |
| *** mcdonc has quit IRC | 16:15 | |
| *** mcdonc has joined #zope | 16:18 | |
| *** tmog has quit IRC | 16:34 | |
| mgedmin | I think you can break zope.testrunner's -j mode if you do os.chdir() in a test | 16:39 |
| mgedmin | if that's true, I'll be laughing all the way to the madhouse | 16:40 |
| fdrake | mgedmin: You can break just about anything relying on process state with os.chdir(). That's a mighty hammer. | 16:41 |
| benji | yep; when parallelizing tests the pernicious influence of global state becomes very apparent | 16:44 |
| benji | I recently spent weeks (or maybe months) squeezing global state dependencies out of a very large test suite. | 16:45 |
| fdrake | benji: Yeah; we're still nailing crap in a certain large test suite you may be familiar with. :-) | 16:48 |
| benji | heh :) | 16:48 |
| fdrake | The zope.testrunner subprocess communication problem looks like it isn't all that hard to fix, though. | 16:49 |
| mgedmin | I'm trying to do that right now :) | 16:50 |
| fdrake | mgedmin: Yay! | 16:50 |
| fdrake | Another nice tool would be something that hooks os.chdir and test setup/teardown, and reports what test failed to restore state appropriately. | 16:51 |
| mgedmin | oh yes | 16:52 |
| mgedmin | I had a pluggable architecture for state verifiers | 16:52 |
| mgedmin | I miss it | 16:52 |
| fdrake | But that's pretty heavy for the general case of tests that don't touch things like os.chdir. | 16:52 |
| fdrake | Yeah. | 16:52 |
| mgedmin | that's why I think the test suite should define what things to check | 16:52 |
| mgedmin | e.g. an uncommitted zope transaction can wreak havoc on other tests | 16:52 |
| fdrake | I'm more likely to add things like that for functional tests, where everything is heavy anyway. :-/ | 16:52 |
| fdrake | mgedmin: Yeah; I've got checks that no interaction is left behind. | 16:53 |
| mgedmin | if layers had pluggable testSetUp/testTearDown hooks, this could be done per-layer | 16:53 |
| fdrake | That caught a *lot* of junk. | 16:53 |
| * mgedmin wonders if https://bugs.launchpad.net/zope.testrunner/+bug/1026576 hasn't been fixed in trunk yet | 16:53 | |
| _mup_ | Bug #1026576: Count test import errors as errors <zope.testrunner:New> <https://launchpad.net/bugs/1026576> | 16:53 |
| fdrake | mgedmin: Like testSetUp and testTearDown? | 16:55 |
| mgedmin | ? | 16:56 |
| mgedmin | sorry, lost context | 16:56 |
| mgedmin | what is like testSetUp? | 16:56 |
| mgedmin | wait, oh, cool! it's been implemented! | 16:57 |
| mgedmin | I never noticed | 16:57 |
| fdrake | Those are methods on the layer object. :-) | 16:57 |
| fdrake | Time machines are wunnerful! | 16:57 |
| mgedmin | what worries me a bit is that https://travis-ci.org/zopefoundation/ZODB/jobs/7595879 skips all the layers other than the first, but doesn't emit *** subprocess communication error *** messages | 17:00 |
| mgedmin | but one thing at a time | 17:01 |
| fdrake | Perhaps travis isn't showing stderr? | 17:02 |
| fdrake | Dunno; don't normally deal with travis myself. | 17:02 |
| mgedmin | no; "could not communicate with subprocess" is written to stdout | 17:03 |
| fdrake | Well, dang. | 17:03 |
| _mup_ | Bug #98250 was filed: Some doctests can't be run twice <bug> <bugday20100424> <core> <zope.testing:Won't Fix> <zope.testrunner:New> <Zope 3:Won't Fix> <Zope 3 3.4:Won't Fix> <https://launchpad.net/bugs/98250> | 17:05 |
| mgedmin | oh, _mup_, it wasn't _filed_, it was _marked as a duplicate_ | 17:07 |
| mgedmin | there's a difference | 17:07 |
| mgedmin | bots | 17:07 |
| *** prg3 has quit IRC | 17:08 | |
| *** prg3 has joined #zope | 17:10 | |
| *** MatthewWilkes is now known as mwilkes|away | 17:11 | |
| *** dayne has quit IRC | 17:19 | |
| *** kosh has quit IRC | 17:24 | |
| mgedmin | oooookaaaaay | 17:29 |
| *** mitchell` is now known as mitchell`off | 17:30 | |
| mgedmin | I fixed the subprocess communication error (yay) and now zope-testrunner does the same thing on my laptop that it did on that travis job: nothing! | 17:30 |
| mgedmin | oh, of course, --test-path is relative, so it doesn't find any tests | 17:31 |
| mgedmin | this is bad | 17:32 |
| *** nueces has joined #zope | 17:38 | |
| fdrake | mgedmin: Doesn't the first layer run in the manager process? | 17:44 |
| mgedmin | yes | 17:44 |
| *** avoinea has quit IRC | 17:45 | |
| fdrake | Changing it so the manager only dispatches seems reasonable, and takes care of this problem for both communications and test-path. | 17:45 |
| fdrake | We'd also win in that -j > 1 could have more win. | 17:45 |
| mgedmin | okay, now zope-testrunner -j99 reports errors when the subprocess is unable to find the right layers | 17:49 |
| mgedmin | no more silent failure | 17:49 |
| mgedmin | now all I need to do is capture the original os.getcwd() and pass it to subprocess.Popen() | 17:49 |
| *** mwilkes|away is now known as MatthewWilkes | 17:51 | |
| *** avoinea has joined #zope | 17:52 | |
| benji | I'm not sure changing the directory back to the original for each subprocess is the right route. Making global state changes in any test (whether across layer boundries or not) seems like a bug and papering over it with the test runner doesn't seem like the right approach. | 17:52 |
| mgedmin | it's what zc.recipe.testrunner does | 17:53 |
| benji | hrm | 17:53 |
| fdrake | I think there are two distinct problems: | 17:53 |
| fdrake | 1. The runner should not be subject to problems with the tests themselves, as much as possible. | 17:54 |
| fdrake | 2. Some tests don't clean up correctly, and should reported & fixed. | 17:54 |
| fdrake | mgedmin is dealing with 1. | 17:54 |
| fdrake | The second needs to be handled in a more fine-grained approach, and probably requires a number of specific solutions for different kinds of process-global state. | 17:56 |
| mgedmin | and 3. Errors should never pass silently | 17:56 |
| mgedmin | testrunner has a bunch of sanity checks | 17:56 |
| mgedmin | like "test left threads behind" etc. | 17:56 |
| mgedmin | "test changed os.getcwd()" is maybe a useful check | 17:56 |
| benji | ah! that makes sense | 17:57 |
| mgedmin | but what about things like doing a os.chdir(some_temp_dir) in a layer setUp? | 17:57 |
| fdrake | Right. That should probably end up allowing additional checks to be plugged in. | 17:57 |
| mgedmin | I had a branch in svn for pluggable checks | 17:59 |
| mgedmin | I never merged it to trunk because doctests aaargh | 17:59 |
| *** prg3 has quit IRC | 17:59 | |
| mgedmin | hm, no, I didn't | 18:00 |
| mgedmin | http://zope3.pov.lt/trac/log/zope.testrunner/branches/mgedmin-transaction-nanny is something completely different (and useless, since I forgot to svn add one file) | 18:00 |
| *** alecm has quit IRC | 18:00 | |
| mgedmin | no, I did! 6 years ago! | 18:00 |
| mgedmin | http://zope3.pov.lt/trac/log/zope.testing/branches/test-checkers | 18:00 |
| fdrake | mgedmin: Luckily, the entire zuniverse has been refactored since then. :-) | 18:01 |
| *** alecm has joined #zope | 18:01 | |
| *** alecm has joined #zope | 18:01 | |
| mgedmin | IT WORKS! MY FIX WORKS! MUAHAHAHAHAAaa | 18:02 |
| mgedmin | ahem | 18:02 |
| mgedmin | long day | 18:03 |
| *** prg3 has joined #zope | 18:04 | |
| *** mcdonc has quit IRC | 18:07 | |
| *** daMaestro has joined #zope | 18:08 | |
| *** prg3 has quit IRC | 18:11 | |
| *** prg3 has joined #zope | 18:11 | |
| mgedmin | pushed a branch with fixes for https://github.com/zopefoundation/zope.testrunner/issues/6 and the related "silently skips ALL THE LAYERS but one" issue I haven' t bothered filing | 18:17 |
| mgedmin | I'd merge to trunk, but the changes aren't unit-tested | 18:18 |
| mgedmin | does anybody here enjoy writing zope.testrunner-style doctests? ;) | 18:18 |
| *** dayne has joined #zope | 18:18 | |
| *** tiwula has joined #zope | 18:19 | |
| fdrake | Crickets. | 18:26 |
| * mgedmin still hopes that somebody-not-him will one day walk up to zope testrunner, say "this is bull*", and rewrite it to use unittest assertions | 18:28 | |
| mgedmin | well, maybe not all of the tests | 18:28 |
| mgedmin | doctests seem to work well as integration tests | 18:29 |
| mgedmin | (aka system tests, aka functional tests) | 18:29 |
| fdrake | doctests work well for a lot, if we restrain ourselves from loading up each test file with too many things. | 18:30 |
| fdrake | We've definitely trained ourselves to ignore other tools, though. | 18:30 |
| *** dayne has quit IRC | 18:36 | |
| *** tmog has joined #zope | 18:37 | |
| *** agroszer has quit IRC | 18:39 | |
| mgedmin | I'll try to write some kind of test | 18:40 |
| mgedmin | (tomorrow) | 18:40 |
| * fdrake appreciates mgedmin's dedication. | 18:40 | |
| * fdrake also believes mgedmin is a masochist. | 18:41 | |
| *** __mac__ has quit IRC | 18:50 | |
| *** fredvd has quit IRC | 19:02 | |
| *** MrTango has quit IRC | 19:03 | |
| *** giacomos has quit IRC | 19:06 | |
| *** maurits has quit IRC | 19:07 | |
| *** __mac__ has joined #zope | 19:29 | |
| *** emidal has quit IRC | 19:41 | |
| *** crazydiamond has joined #zope | 19:45 | |
| *** zagy has joined #zope | 19:52 | |
| *** tmog has quit IRC | 20:02 | |
| *** TresEquis has joined #zope | 20:05 | |
| *** evilbungle has quit IRC | 20:05 | |
| *** agroszer has joined #zope | 20:06 | |
| *** MrTango has joined #zope | 20:15 | |
| *** tmog has joined #zope | 20:15 | |
| *** dayne has joined #zope | 20:19 | |
| *** evilbungle has joined #zope | 20:25 | |
| *** Pumukel has joined #zope | 20:28 | |
| *** __mac__ has quit IRC | 20:31 | |
| *** mindthief414 has joined #zope | 20:41 | |
| *** tmog has quit IRC | 20:46 | |
| *** menesis has quit IRC | 21:09 | |
| *** MatthewWilkes is now known as mwilkes|away | 21:34 | |
| *** Pumukel has quit IRC | 21:36 | |
| *** tiwula has quit IRC | 21:43 | |
| *** zagy has quit IRC | 22:26 | |
| *** mwilkes|away is now known as MatthewWilkes | 22:27 | |
| *** zagy has joined #zope | 22:27 | |
| *** avoinea has quit IRC | 22:29 | |
| *** MatthewWilkes is now known as mwilkes|away | 22:33 | |
| *** tiwula has joined #zope | 22:35 | |
| *** evilbungle has quit IRC | 22:54 | |
| *** MrTango has quit IRC | 23:08 | |
| *** mwilkes|away is now known as MatthewWilkes | 23:21 | |
| *** evilbungle has joined #zope | 23:23 | |
| *** menesis has joined #zope | 23:38 | |
| *** MrTango has joined #zope | 23:47 | |
| *** MatthewWilkes is now known as mwilkes|away | 23:53 | |
| *** sp0cksbeard has quit IRC | 23:55 | |
Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!