IRC log of #schooltool for Friday, 2005-08-19

*** jinty has joined #schooltool01:19
povbot/svn/commits: * srichter committed revision 4819:03:19
povbot/svn/commits: In the noddle somewhere. I need to commit, so I can do some other moves.03:19
povbot/svn/commits: * srichter committed revision 4820:03:20
povbot/svn/commits: Need to commit, so I can move it again.03:20
*** jinty has quit IRC04:28
povbot/svn/commits: * srichter committed revision 4821:04:34
povbot/svn/commits: Delete it now that it is empty04:34
povbot/svn/commits: * srichter committed revision 4822:04:35
povbot/svn/commits: All unit tests in pass. This means that the ZCML loads completely and an application instance can be cleanly created. This, of course, is no indication on the state of the browser code, which is probably still very broken.04:35
*** povbot has joined #schooltool06:04
*** tvon has quit IRC06:18
*** srichter has quit IRC06:18
*** auxesis has quit IRC06:18
*** tvon has joined #schooltool06:18
*** srichter has joined #schooltool06:18
*** auxesis has joined #schooltool06:18
*** maharaja has joined #schooltool09:24
*** jinty has joined #schooltool11:00
*** mgedmin has joined #schooltool14:56
povbot/svn/commits: * gintas committed revision 4823:15:59
povbot/svn/commits: Imported Windows packaging metadata.15:59
*** alga has joined #SchoolTool17:12
povbot/svn/commits: * srichter committed revision 4824:17:39
povbot/svn/commits: Got most tests working now. The two that are still failing and the one I commented out are due to semantic changes from the switch of "ST depending on SB" to "SB depending on ST". I will fix this once the functional tests are passing.17:39
povbot/svn/commits: I really like the new structure; however, most unittests are not really unittests and often depend on other packges and do not use stubs.17:39
*** mgedmin has quit IRC17:40
povbot/svn/commits: * srichter committed revision 4825:17:51
povbot/svn/commits: With a few changed 19 out of 29 functional tests pass.17:51
*** mgedmin has joined #schooltool18:54
*** jinty has quit IRC19:28
povbot/svn/commits: * srichter committed revision 4826:19:30
povbot/svn/commits: Okay, got all SchoolTool functional tests passing. There are only 2 SchoolBell tests failing, because too much is imported.19:30
srichteryipee the rafactor branch can run ST again :-)19:33
srichtermgedmin: do you have a moment?19:34
srichtermgedmin: we need to decide on an order of object hierarchy; basically, what is more "core" than other things19:35
srichteris our application object the most core object?19:35
srichteror is it person, group, resource?19:36
srichteror all four equally core?19:36
mgedminI'd say all four19:36
srichterok, that sounds sane19:36
srichtertimetable is the tricky one19:36
srichterclearly timetable is not important for SB19:37
srichterbut its dependency has crept into many parts of the core components19:37
srichterin other words, none of the core should depend on timetable19:37
srichteralso, I want to provide a package-includes; is that ok?19:40
srichterwe probably want one for SB and one for ST19:41
srichterwell, I should probably have a setup and data dir for both SB and ST, where Data.fs*, conf, package-includes, ftesting.zcml, ... goes19:46
srichtermgedmin: thoughts?19:46
mgedmintimetable: yes, core shouldn't depend on it19:48
mgedminthe rest is fuzzier19:48
srichterwell, person should not depend on course, section, etc19:49
srichterthough it does now19:49
srichterbut only in the view code I think19:49
povbot/svn/commits: * srichter committed revision 4827:20:10
povbot/svn/commits: I just installed reportlab and noticed that the PDF tests were failing.  The fix was easy though.20:10
srichtermgedmin: do you have any ideas on how to structure the root directory properly, so that we can support both schooltool and schoolbell at the same time20:10
srichterone way would be to give different names, like st-package-includes and sb-package-includes20:11
srichteranothe rwould be to have two dirs: schooltool and schoolbell20:11
mgedminfrom the developers' perspective it would be enough to have '' and ''20:11
srichterbut where are all the other files going?20:11
mgedminI defer to the release manager about the rest20:11
srichterwe will have a site.zcml, for example20:11
mgedminone way would be to have qualified names20:12
mgedminschoolbell.conf, schoolbell-site.zcml20:12
mgedminschooltool.conf, schooltool-site.zcml20:12
srichterI guess this might not be too bad after all20:12
mgedminwe used to have something like this a while ago20:13
mgedminthen we split ST and SB20:13
mgedminand I can understand jinty's point of view about going back and forth all the time :-/20:13
srichterI think it is much better now20:14
srichterthere is no duplication anymore and no two repositories to work with20:14
srichterOnce I figure out dependencies, we will see how much interdependent the code is20:15
srichterI think the most pressing issue is getting page templates going20:16
srichterI think the most pressing issue is getting pagelets working, so we can componentize the views20:16
povbot/svn/commits: * srichter committed revision 4828:20:40
povbot/svn/commits: A couple name renamings.20:40
*** jinty has joined #schooltool21:24
povbot/svn/commits: * srichter committed revision 4829:21:31
povbot/svn/commits: Started to make the startup a bit more Zope 3 like and thus more pluggable.21:31
srichtermgedmin: do you have any idea how I could load two different ftesting.zcml files depending on whether I run tests in sbapp (schoolbell app) versus app (schooltool app)21:58
mgedminyou can pass the filename of ftesting.zcml to FunctionalTestSetup22:00
mgedminbut you can only do this once22:00
mgedminthat is, you can only pass one file name in one test run, but you can actually do that several times22:00
srichtermmh, so it needs to sample the files for sbapp differently22:01
srichterI have no clue how I can do that22:04
povbot/svn/commits: * srichter committed revision 4830:22:09
povbot/svn/commits: Fix tests after my recent startup changes.22:09
srichtermgedmin: what is the -level option in
*** kobold has joined #schooltool22:27
jintyhoi hoi kobold22:27
koboldjinty: ok, let's move here.22:28
jintysrichter: kobold and I have a couple of questions.22:28
srichteryep, go ahead22:28
jintykobold is working on debian support of zope22:28
srichterthough I might be afk in few minutes22:28
jintyok, basically the question is how best to handle product-include dependencies22:29
jintyA debian specific solution is possible, but probably not best22:29
srichterpackages-include you mean?22:29
jintyyep, oops:)22:29
jintyis there a way of handling them right now22:30
srichterwell, basically, if I install a third party package, it should automatically add an entry to package-includes22:30
srichterI think zpkgtools knows how to handle them22:30
srichterit is part of the setup meta-data22:30
koboldbut where the dependencies are stored?22:31
srichterthe package-includes in the trunk is not the one we use for the release22:31
srichtereach sub-package is itself responsible for installing it22:31
srichterdependencies should be in DEPENDENCIES.cfg22:31
koboldIf I understand correctly, that file is then parsed by zpkgtools ...22:32
koboldwhich should automagically handle them.22:32
jintybut then when a sub package is installed, the package is automatically included for all instances??22:32
srichterwell, it depends where you install the package22:33
srichterif you install it in Zope 3's installation path, then yes, it is available everywhere22:33
* jinty wants to install all packages on the default python path22:33
srichterif you install it on the instance, ...22:33
srichterthat's a bad idea22:34
srichterbecause not every instance wants a certain package22:34
* kobold agrees22:34
jintybut it is ignored if the package include is not installed?22:34
srichter(btw, I think you really should have this discussion with fdrake; he knows the system much, much better than I do)22:35
srichterhe and Jim know all the use cases22:35
srichterI wonder whether this is the purpose of package includes22:36
jintyindeed, but also perhaps another day22:36
koboldI've looked inside zope3 sources, but I couldn't find zpkgtools22:36
* jinty needs to go to the fiesta de gracia soon22:36
srichterkobold: you have to check it out separately22:36
koboldisn't there a release somewhere on
srichteryeah, fred mentioned a release, but I do not know where it is22:37
koboldI'll look on the website.22:37
srichteryou can always get the tag22:37
koboldthanks for the discussion.22:37
koboldjinty: happy party!22:37
srichterI still have some time :-)22:37
jintykobold: I am also still trying to figure out zpkgtools, so could you mail the list with what you find?22:38
* jinty bids adeu22:38
koboldjinty: sure, I'll do22:38
*** jinty has quit IRC22:38
koboldanother question, what zope3 packages worth shipping inside debian/ubuntu?22:38
srichteronly the stuff that we distribute int he Zope 3.1 release22:39
srichtermaybe wiki and bugtracker, though it has not been tested22:39
srichterbasically the release is the API we are willing to support22:40
koboldI really find myself uncomfortable with zope3 ...22:40
mgedminsrichter, -level? do we have it?22:41
koboldI've used zope2, and zope3 is quite different ...22:41
srichtermgedmin: yeah, and it works great22:41
mgedminif so, it should do more or less the same as the corresponding option in Zope's test.py22:41
srichterI just make the sbapp ftests level 222:41
srichterso they are commonly not run22:41
* kobold found
koboldI have to go, thanks for your help srichter.22:45
*** kobold has left #schooltool22:46
*** mgedmin has quit IRC23:03
*** alga has quit IRC23:48

Generated by 2.15.1 by Marius Gedminas - find it at!