*** Joe_Violanti has left #SchoolTool | 00:25 | |
*** didymo has joined #schooltool | 01:10 | |
*** th1a has quit IRC | 04:55 | |
*** th1a has joined #schooltool | 04:55 | |
*** th1a has quit IRC | 05:01 | |
*** th1a has joined #schooltool | 05:01 | |
*** th1a_ has joined #schooltool | 06:16 | |
*** th1a has quit IRC | 06:29 | |
*** didymo has quit IRC | 10:08 | |
*** ignas has joined #schooltool | 10:44 | |
*** srichter has joined #schooltool | 10:47 | |
*** th1a has joined #schooltool | 11:22 | |
ignas | th1a: ping :) | 11:24 |
---|---|---|
*** Aiste has quit IRC | 11:28 | |
*** faassen has joined #schooltool | 11:35 | |
*** th1a_ has quit IRC | 11:39 | |
*** tiredbones has quit IRC | 11:44 | |
*** tiredbones has joined #schooltool | 11:45 | |
*** alga has joined #SchoolTool | 11:50 | |
*** Aiste has joined #schooltool | 12:35 | |
*** alga has quit IRC | 12:42 | |
*** mgedmin has joined #schooltool | 12:57 | |
*** vidasp has joined #schooltool | 13:02 | |
*** alga has joined #SchoolTool | 13:02 | |
*** ignas has quit IRC | 13:18 | |
*** ignas has joined #schooltool | 13:18 | |
tiredbones | th1a, I was looking at your permissions chart and was wondering what the colors represent? | 13:58 |
tiredbones | sorry th1a, I should read my email more carefully. | 14:14 |
*** jinty has joined #schooltool | 14:29 | |
*** mgedmin has quit IRC | 15:17 | |
faassen | jinty: I don't think the same trick would work. | 15:30 |
faassen | jinty: I mean, I tried at your suggestion | 15:30 |
faassen | jinty: but it just won't find the zc.resourcelibrary directory at all, I think, as there's only an egg link. | 15:30 |
jinty | Strange, I use it in sqlos and it seems to work | 15:30 |
faassen | really? | 15:30 |
faassen | maybe I'm doing something stupid wrong. | 15:30 |
* jinty has a look at his sqlos package again | 15:31 | |
faassen | what happens is that it doesn't complain; it does find zc.resourcelibrary, I guess. | 15:31 |
faassen | but it doesn't find the tests. | 15:31 |
faassen | oh wait, perhaps those aren't installed. | 15:31 |
faassen | would that be possible? | 15:31 |
faassen | or wait.. | 15:32 |
faassen | no. | 15:32 |
faassen | I mean, these are functional tests. | 15:32 |
jinty | faassen: http://pastebin.com/682835 | 15:33 |
faassen | yeah, I added taht to Zope's test runner. | 15:34 |
jinty | then I use the Zope3 test runner to run the sqlos tests | 15:34 |
faassen | but doesn't seem to have any effect. | 15:34 |
faassen | in your case, is it picking up tests in development mode/ | 15:35 |
faassen | ? | 15:35 |
faassen | or in an installed egg? | 15:35 |
jinty | yes | 15:35 |
faassen | ok. | 15:35 |
jinty | no development mode | 15:35 |
faassen | that's weird. | 15:35 |
faassen | going to try this with zc.table now, see whether that works. | 15:35 |
jinty | indeed. | 15:35 |
jinty | but note that I also have an easy-install.pth file in Zope3/src | 15:36 |
faassen | there is one too here. | 15:37 |
faassen | I don't know how to read those. :) | 15:37 |
jinty | perhaps I can duplicate your setup somehow | 15:37 |
faassen | shouldn't be too hard, the schooltool-eggs branch. | 15:38 |
faassen | I shall check in my notes. | 15:38 |
faassen | in README. | 15:38 |
faassen | ok, there's now a section in README.txt | 15:39 |
faassen | Egg dependencies and checkouts | 15:39 |
* jinty starts to get a totally new checkout and might be a while | 15:39 | |
faassen | which contains the instructions on how to replicate this situation. | 15:39 |
povbot` | /svn/commits: * faassen committed revision 5972: | 15:40 |
povbot` | /svn/commits: Initial notes on how to: | 15:40 |
povbot` | /svn/commits: * install egg dependencies into a Schooltool tarball. | 15:40 |
povbot` | /svn/commits: * install egg dependencies in development mode in an SVN checkout. | 15:40 |
faassen | okay, I have the same issue with zc.table tests. | 15:40 |
faassen | it finds the package, just doesn't seem to pick up tests. | 15:40 |
faassen | I say: python2.4 test.py -vpf1 -s zc.table | 15:40 |
faassen | or -vpu1 | 15:40 |
jinty | is the package a zipfile? | 15:41 |
jinty | er no, obviously not | 15:41 |
faassen | well, development packages aren't zipfile. | 15:41 |
faassen | but I turned zipfile stuff off. | 15:41 |
* jinty urges on his internet connection | 15:41 | |
faassen | I suppose that even though the egg may not ship with tests.. | 15:41 |
faassen | the development mode package will find tests. | 15:42 |
faassen | perhaps that assumption is wrong. | 15:42 |
jinty | hmm, eggs should ship with tests, shouldn't they? | 15:42 |
faassen | I don't know. | 15:42 |
jinty | it's not celar from your notes where I am supposed to checkout zc.resourcelibrary | 15:44 |
jinty | ah, somewhere | 15:44 |
* jinty thinks eggs should definitiely ship with tests | 15:44 | |
faassen | ok. | 15:44 |
faassen | I shall try to adjust them that they have tests included. | 15:45 |
jinty | but that's just my opinion I formed myself | 15:45 |
faassen | anyway, I don't see how that could be the cause. | 15:45 |
jinty | otherwise how do you test installations? | 15:45 |
faassen | unless there's a mechanism that makes sure exactly only those packages listed in setup.py are installed. | 15:45 |
faassen | somewhere in setuptools. | 15:45 |
faassen | I was assuming that typically one doesn't test eggs unless you're co-developing them. | 15:46 |
faassen | it depends on your view of tests. | 15:46 |
faassen | like, are tests for developers or are they also for end users to test whether their install works? | 15:46 |
jinty | yeah, but if I install a server I like to run the tests on the installation | 15:47 |
jinty | to make sure I have all dependencies and everything is working fine | 15:48 |
jinty | (from a sysadmin's point of view) | 15:49 |
jinty | from a debian packagers point of view, it is nice to run the tests on an installed egg at package build time | 15:50 |
faassen | ok. | 15:50 |
faassen | well, I'll look into somehow bundling tests with eggs and having them run. | 15:50 |
faassen | but I don't know how to run tests in eggs. | 15:50 |
faassen | is that possible? | 15:50 |
jinty | I don't know about zipfiles, but it certainly seems possible for non-zipfile eggs | 15:52 |
faassen | by the way, you'll have to put in the 'import site' thing into the test runner, if this is really necessary. | 15:58 |
faassen | I get the same behavior (no error, but no tests found in zc.table or zc.resourcelibrary) without or without that addsitedir. | 16:00 |
jinty | I see your problem | 16:00 |
jinty | moment | 16:00 |
faassen | ok you can reproduce it not picking up tests, right? | 16:01 |
jinty | yep, it's because the test runner ignores tests found outside certain paths | 16:01 |
jinty | there's a option to add them... | 16:02 |
jinty | .PHONY: testall | 16:02 |
jinty | test: develop | 16:02 |
jinty | cd Zope3 && $(PYTHON) test.py --test-path=../src -s sqlos | 16:02 |
jinty | that's the makefile rule for sqlos | 16:03 |
faassen | aah. | 16:03 |
faassen | but how can I add.. | 16:03 |
faassen | something taht's not a directory? :) | 16:04 |
faassen | II mean, I just tried.. | 16:04 |
jinty | PYTHONPATH=~/pylib:src ./test.py -s zc.resourcelibrary --test-path=../zc.resourcelibrary/src | 16:04 |
jinty | that works, without patching test.py | 16:04 |
faassen | oh, but then we're actually just putting the checkout on the test path. | 16:04 |
faassen | right? | 16:04 |
jinty | you might be | 16:05 |
faassen | well, are you? | 16:05 |
faassen | by the way, that doesn't work for me. | 16:05 |
faassen | I get a no such file or directory. | 16:05 |
jinty | --test-path=TEST_PATH | 16:05 |
jinty | Specify a path to be searched for tests, but not added | 16:05 |
jinty | to the Python search path. This option can be used | 16:05 |
faassen | IOError: [Errno 2] No such file or directory: '/home/faassen/working/schooltool/eggs/schooltool/Zope3/zopeskel/etc/securitypolicy.zcml' | 16:06 |
faassen | don't you get that? | 16:06 |
jinty | I have that file. but I did a make inplace first | 16:07 |
faassen | I'm in the directory Zope3 | 16:07 |
jinty | so am I | 16:07 |
faassen | hm, perhaps I didn't get made yet. | 16:07 |
faassen | in this particular checkout. | 16:07 |
faassen | I thought I had.. | 16:07 |
faassen | no, that wasn't it. | 16:07 |
faassen | hm, but I can run the normal z3 tests. | 16:08 |
faassen | oh wait, i can't. | 16:08 |
faassen | that's weird, I thought I did before. | 16:08 |
faassen | hm, I wonder what triggers that error. | 16:08 |
faassen | python test.py -vp | 16:09 |
faassen | seems to work. | 16:09 |
faassen | at least my tests are being executed. | 16:09 |
faassen | ok appears to be only the functional tests. | 16:11 |
faassen | so that's not related to eggs. | 16:11 |
faassen | jinty: did you actually get those zc.resourcelibrary tests to run? | 16:12 |
faassen | the zc.table tests run | 16:12 |
jinty | yep. they failed beautifully for lack of a configure-zcml file | 16:13 |
jinty | ZopeXMLConfigurationError: File "<string>", line 6.2 | 16:13 |
jinty | ConfigurationError: ('Unknown directive', u'http://namespaces.zope.org/zope', u'resourceLibrary') | 16:13 |
faassen | right.. | 16:13 |
faassen | well, that makes sense. | 16:13 |
faassen | ok. | 16:13 |
faassen | I still think the Zope3/src is better than schooltool/src | 16:13 |
faassen | I mean, to put them in. | 16:13 |
* jinty hopes that when zope becomes a collection of eggs, Zope3/src will dissapear | 16:15 | |
jinty | so I slightly disagree, but not much | 16:16 |
faassen | you'd put them all in src? | 16:16 |
faassen | it's just that you wouldn't want your test runner to pick up all the Zope 3 tests. | 16:16 |
faassen | when you're testing schooltool. | 16:16 |
faassen | putting all your eggs into that single space might mean all the tests are run. | 16:17 |
jinty | sorry, gotta eat the pasta provided by my roommates;) | 16:17 |
jinty | see you a bit later! | 16:18 |
faassen | ok, thanks for the help! | 16:18 |
faassen | see you later! | 16:18 |
jinty | just a quick note, the zope test runner ignores the ZODB tests by default. we might want something similar. | 16:19 |
faassen | we still have the issue of where this site.zcml is coming from. :) | 16:19 |
faassen | ok. | 16:19 |
faassen | your tarballdoesn't contain the .zcml that I adjusted in the brnch for some reason. | 16:19 |
faassen | branch. | 16:19 |
* jinty thought he had posted a mail to the list about the missing zcml | 16:44 | |
* jinty decides he very much prefers the PYTHONPATH approach to the import site; site.addsitedir(src) approach | 16:48 | |
povbot` | /svn/commits: * faassen committed revision 5973: | 16:57 |
povbot` | /svn/commits: Add some more notes about development installs of eggs. | 16:57 |
*** mgedmin has joined #schooltool | 17:08 | |
faassen | jinty: I hadn't seen that reply.. | 17:21 |
faassen | jinty: what do you mean the PYTHONPATH approach compared to the impot site story? | 17:21 |
* faassen looks for the reply in possible spam places. | 17:21 | |
jinty | they are equivalent. If you set PYTHONPATH, you don't need to do the import site dance | 17:22 |
* jinty actually tests that | 17:22 | |
jinty | yep it's true | 17:22 |
faassen | I think that's true, but I can't tell everybody to set their PYTHONPATH all the time. | 17:22 |
faassen | the whole point of those schooltool-server and test.py site path manipulation is to avoid having to do that. | 17:23 |
jinty | but PYTHONPATH feels more correct | 17:23 |
faassen | well, it's a shift in philosophy that I'm not going to propose, but feel free. | 17:23 |
faassen | concerning those missing zcml files. | 17:23 |
faassen | I can't find your post. | 17:23 |
jinty | anyway, it's just my opinion | 17:23 |
*** Aiste has quit IRC | 17:23 | |
faassen | what did you say about it? | 17:23 |
jinty | moment | 17:23 |
jinty | ok, re-sent it, but CCed the list this time | 17:27 |
jinty | fassen, the crux is, look in releases/SchoolTool/PACKAGE.cfg | 17:27 |
jinty | faassen, rather | 17:28 |
faassen | hm, where is this Schooltool? | 17:28 |
jinty | I think this is a rather hoinking big bug | 17:28 |
jinty | yep | 17:28 |
faassen | svn+ssh://source.schooltool.org/svn/schooltool/trunk/releases | 17:28 |
faassen | doesn't have a subdirectory. | 17:28 |
faassen | should I look somewhere else? | 17:28 |
faassen | ok, did I run into a bug? :) | 17:28 |
jinty | no in the schooltool source tree | 17:29 |
faassen | ok, looking there.. | 17:29 |
jinty | there is a releses directory there | 17:29 |
faassen | yes, looking in that PACKAGE.cfgf file. | 17:29 |
faassen | found it. | 17:29 |
faassen | so the bug is we should load the stuff from the branch in there? | 17:29 |
faassen | that isn't a bug though, just a neglect on the branch? | 17:30 |
faassen | or what is this bug? | 17:31 |
jinty | it's a bug somewhere (perhaps in my branch tagging script), I am just not sure exactly where yet. | 17:31 |
faassen | but why is it a bug? I mean, that file refers to the trunk. | 17:31 |
faassen | so if it pulls that stuff in to make a tarball.. | 17:31 |
faassen | then it's going to get trunk stuff. | 17:31 |
faassen | or am I missing something? | 17:31 |
faassen | I don't know whether it pulls stuff that's listed there. | 17:32 |
jinty | yes, but if you build a tarball from a tag in three months and include this stuff from the trunk you end up with a horribly broken tarball | 17:32 |
faassen | yes. | 17:32 |
faassen | I mean, so it's a design flaw? | 17:32 |
faassen | I mean, clearly you if you're branching a stable branch or have a tag. | 17:33 |
faassen | you'd like to pull in the stuff local to that tag or branch. | 17:33 |
faassen | not from the trunk. | 17:33 |
faassen | the simple fix would be to adjust that file. | 17:33 |
faassen | each time you branch or tag. | 17:33 |
faassen | but that'd be suboptimal. | 17:33 |
jinty | I think I remember there was some zpkg magic to get around changing that file | 17:34 |
jinty | but I could have my tagging script edit it as well | 17:34 |
jinty | but don't worry about it for merging your branch, this is a different issue | 17:35 |
faassen | well, I could merge my branch then. | 17:35 |
faassen | I mean, otherwise I think I'm ready now. | 17:36 |
jinty | +1 | 17:36 |
faassen | except I couldn't test the tarball generation. | 17:36 |
faassen | completely. | 17:36 |
faassen | I mean, the tarball works, I *think* I successfully can install eggs into it. | 17:36 |
faassen | but I cannot 100% verify they're really loadable as I didn't load them. | 17:36 |
faassen | anyway, I'll merge and we'll deal with that when it's time. | 17:36 |
faassen | then at least I can proceed using those dependencies, which my point was in the first place. :) | 17:36 |
jinty | I am comfortable that we can resolve the remaining issues before release, thus there is little risk in merging. | 17:38 |
faassen | ok, thanks for the go ahead. | 17:40 |
povbot` | /svn/commits: * faassen committed revision 5974: | 18:11 |
povbot` | /svn/commits: merge 5865:5973 svn+ssh://source.schooltool.org/svn/schooltool/branches/schooltool-eggs | 18:11 |
faassen | hm, I think I eliminated history for setup.py as I moved it around.. | 18:11 |
povbot` | /svn/commits: * pull in eggs (zc.table in particular) automatically when doing a 'Make' | 18:11 |
povbot` | /svn/commits: * explicit version.txt to maintain schooltool version so that this can | 18:11 |
povbot` | /svn/commits: be retrieved without having to import schooltool. | 18:11 |
povbot` | /svn/commits: * remove unused MANIFEST.in | 18:11 |
faassen | even though I didn't change anything. | 18:11 |
faassen | I tried to fix this by reverting setup.py before checking in, but that completely screwed up the checkout. | 18:11 |
faassen | so I hope the erasure of history from that file isn't too bad. | 18:12 |
faassen | it's a short file. | 18:12 |
faassen | :) | 18:12 |
*** alga has quit IRC | 18:17 | |
th1a | ignas: pong. | 18:32 |
ignas | th1a: i am ripping away timetables for persons/resources/groups (they will still have their composite timetables) | 18:33 |
th1a | ignas: Thank you. | 18:33 |
ignas | faassen: err, what magic must i do to run schooltool ? as dapper is kind of missing some python packages | 18:51 |
jinty | ignas: try installing the latest setuptools, probably in /usr/local/lib/python2.4/site-packages | 18:56 |
jinty | http://peak.telecommunity.com/DevCenter/EasyInstall#installing-easy-install | 18:57 |
mgedmin | uh oh, I suppose I will have to do that on the buildbot machine too | 18:57 |
ignas | jinty: err i kind of would like to avoid that :/ as i don't like installing packages anywhere out of my homedir | 18:58 |
ignas | jinty: imho messing with apt managed files is allways a bad idea, especially if one is running unreleased ubuntu version (dapper) | 18:58 |
jinty | ignas: you can also install into a directory in your home directory | 18:58 |
jinty | but it needs to be a site somehow, and I don't know the exactly right way to do that | 18:59 |
* jinty is using ~/pylib | 19:00 | |
ignas | will i find such info in schooltool's README.txt some day ? | 19:00 |
jinty | then you would probably need to add PYTHONPATH=~/pylib to your environment | 19:01 |
mgedmin | and the makefile should take care to add to $PYTONPATH instead of overriding it like I saw on faasen's branch | 19:01 |
jinty | and change the scholltool Makefile PYTHONPATH declaration to be PYTHONPATH=~/pylib:src:Zope3/src | 19:01 |
ignas | jinty: as i am often connecting to remote PC's and kind of trying to develop a distributed unit test runner sometimes, it hurts a lot :/ | 19:02 |
jinty | mgedmin: feel free, I investigated the make manual for a short while but didn't find the "right" way | 19:03 |
jinty | ignas: this is just until setuptools 0.6a11 makes it into ubuntu | 19:03 |
ignas | jinty: Edgy Eft ? | 19:03 |
jinty | they are on 0.6a9 right now | 19:03 |
jinty | ignas: definitely | 19:04 |
ignas | jinty: is python ./ez_setup.py --install-dir /usr/local/lib/python2.4/site-packages/ enough to make it not polute my /usr/lib ? | 19:15 |
ignas | as the dry run is broken | 19:15 |
ignas | :/ | 19:15 |
jinty | ignas: why don't you run that as your own user and change the ownership of that directory temporarily? | 19:16 |
jinty | then change the ownership back to root | 19:17 |
ignas | an idea | 19:17 |
jinty | that way you can be sure | 19:17 |
* jinty notes that leaving that directory owned by an un-priviliged user is a big security risk | 19:19 | |
*** alga has joined #SchoolTool | 19:24 | |
ignas | jinty: err, well now i have setuptools-0.6a11-py2.4.egg in /usr/local/lib/python2.4/site-packages, but schooltool still is not buildable | 19:34 |
ignas | ImportError: No module named setuptools | 19:35 |
faassen | hm, I need to put in anote. | 19:38 |
faassen | my apologies | 19:39 |
faassen | concerning how to set up setuptools. | 19:39 |
faassen | ignas: that one is weird. | 19:39 |
*** vidasp has quit IRC | 19:41 | |
povbot` | /svn/commits: * faassen committed revision 5975: | 19:42 |
povbot` | /svn/commits: Some fixes to the shell commands. | 19:42 |
ignas | faassen: fixed that :/ apparently -s was not the key i should have used when installing | 19:42 |
ignas | --prefix tries to install easy_install to /usr/local/bin though | 19:44 |
ignas | so i still don't know what was the proper way of installing setuptools | 19:44 |
ignas | but it works now | 19:44 |
*** srichter has quit IRC | 19:44 | |
faassen | okay, that's a relief. | 19:45 |
povbot` | /svn/commits: * faassen committed revision 5976: | 19:46 |
povbot` | /svn/commits: Added some notes about installing setuptools first. | 19:46 |
ignas | but i am begining to dislike eggs/setuptools | 19:47 |
jinty | ignas: Soonish apt-get will work | 19:48 |
mgedmin | jinty: define soonish | 19:48 |
mgedmin | I thought dapper was very deeply frozen now? | 19:48 |
mgedmin | and I'd be afraid to jump to edgy as soon as it is branched | 19:48 |
mgedmin | people are talking about breaking everything in it to get the latest crackful candy | 19:49 |
mgedmin | to balance the very stable dapper release | 19:49 |
* jinty thinks that it might have been a good idea to provide an updated setuptools deb for schooltool developers | 19:49 | |
jinty | mgedmin: apt-pinning? | 19:52 |
jinty | mgedmin: apt pinning? | 19:52 |
* mgedmin does not like apt pinning | 19:53 | |
* mgedmin thinks the apt preferences file format was designed by a drunken monkey on drugs | 19:53 | |
faassen | anyway, we'll see how eggs work out. | 19:55 |
faassen | it's complicated having two package formats, one distribution based and one language based. | 19:55 |
faassen | then again a language based packaging system was something long missing in python. | 19:55 |
faassen | it gets hairiest if you want to install the packaging system itself as a package. :) | 19:55 |
faassen | out of here for today. | 19:56 |
faassen | jinty: thanks again for all the help. | 19:56 |
faassen | see you! | 19:56 |
*** faassen has quit IRC | 19:57 | |
jinty | faassen: caio | 19:57 |
povbot` | /svn/commits: * ignas committed revision 5977: | 20:02 |
povbot` | /svn/commits: Fix a broken test. | 20:02 |
povbot` | /svn/commits: * ignas committed revision 5978: | 20:02 |
povbot` | /svn/commits: Abolish tripple single quotes. | 20:02 |
mgedmin | is setuptools brainded, or is this its way to tell me "everything is now all right"? | 20:14 |
mgedmin | http://schooltool.pastebin.com/683351 | 20:14 |
mgedmin | at least '>>> import setuptools' works now | 20:19 |
mgedmin | let's see if buildbot becomes green again | 20:19 |
mgedmin | all green! | 20:49 |
*** jinty has quit IRC | 20:54 | |
mgedmin | woohoo, buildbot is finally packaged for Debian! | 20:56 |
mgedmin | or at least for Dapper | 20:56 |
*** srichter has joined #schooltool | 21:11 | |
ignas | th1a: ayt | 21:52 |
ignas | ? | 21:52 |
*** ignas has quit IRC | 22:11 | |
*** mgedmin has quit IRC | 22:18 | |
*** Hwyvar has joined #schooltool | 22:34 | |
*** alga has quit IRC | 22:38 | |
*** srichter has quit IRC | 22:51 | |
*** srichter has joined #schooltool | 22:52 | |
*** Aiste has joined #schooltool | 23:27 | |
*** jinty has joined #schooltool | 23:52 |
Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!