IRC log of #schooltool for Tuesday, 2005-08-16

povbot/svn/commits: * srichter committed revision 4707:00:15
povbot/svn/commits: Implemented new pluggable traverser and made initial use of it. The REST traversers were actually already flexible, so I just integrated them. We could convert them all completely to the new way as well, but this would have some implications (not very serious) we might want to discuss first.00:15
povbot/svn/commits: * tvon committed revision 4708:01:41
povbot/svn/commits: keep multiple batches on one page (acl view) from wrapping.01:41
*** jinty has left #schooltool01:51
povbot/svn/commits: * tvon committed revision 4709:07:23
povbot/svn/commits: 'Showing xxx of yyy items' text for containers.07:23
povbot/svn/commits: I put it in the most logical place and tried to style it to look nice, but I think the container UI could look better...  maybe by sliding the search box to the right side... I'm not sure.  Suggestions for something better?07:23
povbot/svn/commits: * tvon committed revision 4710:07:24
povbot/svn/commits: same as r4709 but for schooltool ('showing xxx of yyy' text).07:24
*** pzingg has joined #schooltool07:48
pzinggYikes!  Although schooltool is so cool, I as a sysadmin (and being zope3-ignorant) would like to see some documentation or hints on customization.  F'rinstance:07:50
pzingg1. replacing the zebra with my own logo07:50
pzingg2. replacing the default site-wide "SchoolBell Calendar" title with the name of my own organization.07:51
pzingg3. Leveraging off my OS X server's LDAP directory for persons and group membership.07:52
pzinggI'm sure 1. and 2. can be "easily" done, if I were to be able to figure out how the zope templates and configuration and source repository for resources all work together.  Unfortunately, neither the zope3 nor the schooltool docs are that clear to the outsider.07:53
tvonpzingg: for 2, you can login as manager and go to the managers page... there on the left you should see 'site preferences' which lets you set the name of the site08:04
tvonpzingg: for 1 you could change the schoolbell.png file in src/schoolbell/app/browser/resources08:05
tvoner.. thats for schoolbell, for schooltool it is src/schooltool/browser/resources/logo.png I believe08:05
tvonI don't know anything about getting '3' to work though08:05
pzinggThanks.  I either didn't catch the Site Preferences page in the schooltool docs, or it's new.08:07
tvonIt's new08:08
tvonnow I'm going to bed, more folks will be around in 6-8 hours if you need anything else ;)08:08
*** pzingg has quit IRC08:29
*** ignas has joined #schooltool10:20
*** ignas has quit IRC11:17
*** jinty has joined #schooltool11:56
*** ignas has joined #schooltool12:04
povbot/svn/commits: * jinty committed revision 4711:12:34
povbot/svn/commits: Merged 4962:4964 for issue 328.12:34
povbot/svn/commits: * tvon committed revision 4712:12:37
povbot/svn/commits: passing search results through batch navigation links so we can batch our searches properly.  Also provide a 'clear' button with our search to get back to the normal view.12:37
povbot/svn/commits: Please backport, fixes issue335 and the rest of issue334 (along with the next commit)12:37
povbot/svn/commits: * tvon committed revision 4713:12:38
povbot/svn/commits: Fixes issue335 and the rest of issue334 for schooltool12:38
povbot/svn/commits: Please backport.12:38
povbot/svn/commits: * jinty committed revision 4714:12:41
povbot/svn/commits: Back-ported 4673.12:41
povbot/svn/commits: * jinty committed revision 4715:12:43
povbot/svn/commits: Back-ported 4674.12:43
povbot/svn/commits: * tvon committed revision 4716:12:55
povbot/svn/commits: whoops, tests that should have gone in with 471212:55
povbot/svn/commits: * jinty committed revision 4717:12:56
povbot/svn/commits: Back-port 4678 for
povbot/svn/commits: * tvon committed revision 4718:13:01
povbot/svn/commits: uhm, and the test for 471313:01
*** mgedmin has joined #schooltool13:22
*** gintas has joined #schooltool13:24
gintasjinty, thanks for your help with the release13:26
*** bskahan has joined #schooltool14:24
*** gintas has quit IRC14:25
bskahandoes anyone else get an ftest failure on errors.txt in trunk?14:51
bskahani only get it when I run the tests -uf on schoolbell and schooltool concurrently14:56
*** ignas has quit IRC14:56
mgedminbuildbot doesn't14:56
mgedminwhat sort of failure is it?14:56
mgedminvery interesting14:57
mgedminI'd like to see the traceback14:57
* bskahan nods14:57
mgedminConnectionStateError: Shouldn't load state for 0x07 when the connection is closed14:59
srichteryour DB is hosed14:59
bskahan./ -uf15:00
mgedminwhile providing an adapter in a functional test?15:00
bskahanin schooltool15:00
mgedminwhy does ztapi.browserView even touch ZODB?15:00
bskahanspecifying schoolbell or schooltool doesn't raise it15:00
mgedminit is certainly caused by some test that does not clean up after itself15:01
bskahanso I'm guessing its a test that isn't cleaned up somewhere ...15:01
mgedmindoes it happen if you run ./ -f  (without -u)?15:01
mgedminno, no 'ah'15:09
mgedminI do not understand15:09
mgedminthe test runner runs functional tests first15:09
mgedmin(because unit tests will damage all setup done by ftesting.zcml)15:09
mgedmin./ -fupv works here15:12
mgedminRan 735 tests in 75.921s15:12
mgedminno errors/failures15:12
* mgedmin shrugs15:12
bskahanthanks for checking15:12
mgedminit might be timing related perhaps?15:14
bskahanI'm going to get a fresh tree and see if I still get it15:15
mgedminI've seen it before15:17
mgedminnot on my machine, but reports about this exception15:17
*** gintas has joined #schooltool15:19
jintyhoi gintas, no worries. currently Im looking at 468215:20
bskahanmgedmin: got the same error on a fresh checkout, must be machine specific15:22
mgedminto fix issue 341 for the release, please merge 4686, [4687], 4688, 4689 to schoolbell's release branch, and 4690 to schooltool's release branch15:24
mgedmin4687 is optional -- it removes a DeprecationWarning15:25
mgedmingintas says he has seen this ConnectionStateError on Windows15:25
* mgedmin is going to resurrect from schooltool-twisted-branch15:25
bskahanmgedmin: what is
gintasit's a collection of functons that check that tests clean up after theirselves15:28
gintasjinty, I'm still not sure I understand what you are proposing with regard to .debs15:28
gintasare you suggesting releasing debs that depend on zope3-lib?15:29
gintasI think it's a better idea to release the current package (with Zope3 inside)15:30
gintaslater we can release 1.2-2 which uses the new scheme15:30
bskahangintas: I think the idea was that we can't get into ubuntu main if we have zope in the package15:31
mgedminoh? why?15:32
jintygintas: you understood correctly. My reasoning is that the tests are run at build time using the installed zope3-lib, if they fail, the build fails.15:32
mgedminoh, security reasons15:32
jintybut I defer to your decision15:33
povbot/svn/commits: * mg committed revision 4719:15:34
povbot/svn/commits: Resurrected from schooltool-twisted-branch (minus old ComponentChecks).  Fixed a unit test that expected sys.stderr to be unwrapped.15:34
* mgedmin suggests merging 4720 -- another DeprecationWarning15:34
povbot/svn/commits: * mg committed revision 4720:15:35
povbot/svn/commits: Fixed another DeprecationWarning.15:35
povbot/svn/commits: Merge candidate.15:35
* jinty will probably upload 1.2-1 (with zope) to unstable and 1.2-2 to experimental (without zope)15:35
bskahanon a fresh checkout, lots of ftests fail because the page headers are getting produced as:15:36
bskahan    - Content-Type: text/xml; charset=UTF-815:36
bskahan    ?                    ^   -        ^^^15:36
bskahan    + Content-Type: text/html;charset=utf-815:36
mgedminbskahan, I don't see those15:37
* bskahan goes to investigate15:38
bskahanmaybe my computer is just hosed15:38
bskahanthat was svn co; cd st-new; make; ./ -pvuf15:39
mgedminbskahan, I suggest you use 'make test ftest'15:40
mgedminI do not entirely trust test -fu15:40
mgedminotoh it would be really good to identify why test -fu sometimes fails when test -f; test -u doesn't15:41
* mgedmin has a zcml i18n question15:41
bskahanstill seeing that error on make test ftest15:42
mgedminoh?  interesting...15:43
mgedminvery interesting15:43
jintybskahan: I think 4680 is also necessary to back port 4682. agree?15:43
bskahanjinty: one sec, I'll look15:44
gintasdon't forget 3723, 8928, 2939, 1239, 4884, 8585 and 7$%^15:44
gintasthese heaps of numbers are getting to me15:44
* jinty is glad there is somebody to share the pain;)15:45
mgedminwe had a process for this15:45
mgedminissue tracker, tags, lists of revisions pertaining for each issue15:46
mgedminmake extract-translations fails15:46
bskahanjinty: 4680 and 4682 conflicted, 4682 wipes out 468015:47
povbot/svn/commits: * srichter committed revision 4721:15:47
povbot/svn/commits: Added a couple tests from schooltool.15:47
jintyif people agree, I am completely prepared to start ignoring back-port requests not made strictly according to the process15:47
bskahanjinty: I think that's a good idea15:48
mgedmincan we ask everyone to look at his checkins since the release branch was made, and make sure that ones that need merging are listed in the issue tracker?15:48
* jinty is +1 for issue tracker + schooltool 0.xx tags15:49
mgedminwhich translation is most complete?15:50
* jinty thinks es_ES for schoolbell might be quite complete15:51
bskahanjinty: I added a bug that referes to 4682/469515:53
bskahanthat's the only one revision I know of that didn't have a specific issue15:54
bskahanmgedmin: do we have the release cycle process written down?15:55
mgedminwe had something like that, I think15:56
* mgedmin doesn't understand15:56
mgedminlang es_ES doesn't seem to work in schoolbell.conf15:57
* jinty goes on 15min typing break15:58
gintaswhat do you think about rev 4654? (generation script to make schooltool's application calendar public and set security policy to the 0.11 defaults)16:00
bskahangintas: I was initially opposed to it being backported16:01
bskahanbut after some thought I think it should16:01
bskahanif it isn't applied, then the top level url will always kick to the login page unless the admin changes ACL on the root/root calendar16:02
bskahandoing that, without going through and explicitly closing the containers, will open the site up to the public16:02
bskahanthe generation at least makes shure /persons/ is denied16:03
gintasOK, I'll backport it then16:03
mgedminis there an issue for deprecation warning cleanups?16:03
povbot/svn/commits: * mg committed revision 4722:16:04
povbot/svn/commits: Fix DeprecationWarnings about missing i18n_domain in SCHOOLBELL_SITE_DEFINITION.16:04
povbot/svn/commits: Fix incorrect i18n translation of principal names -- ZCML does i18n by itself, you don't have to explicitly translate them before specifying them in.ZCML.16:04
povbot/svn/commits: I should have written a unit test for this, but I do not know how.16:04
bskahangintas: can we trigger a debian package install alert that the permissions have changed?16:04
bskahanchanging security policy silently is still pretty rotten16:05
gintasit's possible16:05
gintasbut this will only work for Debian16:05
bskahanit can be in the release notes (for people running from tarballs)16:06
bskahanand I'll ask eric if we can do something similar for RPMs (though I don't recall ever seeing something like that)16:06
mgedminbskahan, can you please svn up and run ./ -fu16:09
gintashmm, doesn't this permission change apply to schoolbell as well?16:09
mgedminI've added -- they will notice if you leave unclean transactions16:10
bskahangintas: no16:10
povbot/svn/commits: * mg committed revision 4723:16:10
povbot/svn/commits: Removed DeprecationWarning, fixed incorrect i18n translation.  The equivalent of rev 4722, but for schooltool.16:10
bskahanmgedmin: no mention of unclean transations16:12
bskahanstill have the errors.txt failure16:13
mgedminI wish there was a way to figure out whether a test crapped in the component architecture configuration16:16
mgedminbut adapter registry's internals are so obfuscated, that you essentially cannot take a snapshot of the state and compare it to another state later16:17
povbot/svn/commits: * jinty committed revision 4724:16:19
povbot/svn/commits: Back-ported 4679:4682 and 4695 for
jintygintas: I'm working on 341 now.16:21
bskahanmgedmin: what revision of Zope are you using with schooltool?16:21
mgedminlet me check16:23
mgedminsvn://, revision 3788516:23
gintashmm, bskahan, that evolve script in rev 4654 does not have tests16:23
bskahangintas: what sort of tests would you like for it?16:23
gintaswell, in other projects we had tests for evolve scripts16:24
gintasbut they're really messy16:24
bskahanI don't think any of our evolve scripts have tests16:24
gintasI assume you have tested the evolution script carefully16:24
gintasbskahan, I don't mean to imply that you should have written a test16:25
bskahanyes, I'd be very happy to add tests for it too16:25
gintasI am just wary of checking in untested code (Python's a dynamic language and all...)16:25
gintasI'll do a manual test and then commit16:26
bskahanI'd like a good way to test generations16:26
bskahanat least functionally16:26
gintasin that other project we had a heap of stubs that we would give the 'evolve' function16:27
gintasbut I think that we don't have time to do this now16:27
gintasthere's a heap of other checkins waiting16:27
*** ignas has joined #schooltool16:27
bskahani like marius' idea of having sample data.fs for each release in svn16:28
* mgedmin does svn up Zope316:28
gintasyes, sounds good16:28
povbot/svn/commits: * srichter committed revision 4725:16:28
povbot/svn/commits: Okay, we cannot use the standard ContainerTraverser, since it also tries to lookup a view. This is really a sign that the default traverser implementation is less than ideal.16:28
mgedminwhoa, massive failures everywhere16:28
mgedminthank you, Zope 3, we needed that on the final release day16:28
mgedminBTW one of the reasons I suffered svn:externals was that it ensured all developers had the same version of Zope 3 (assuming they did svn up)16:28
bskahanmgedmin: oh good16:29
gintasyou mean the Zope-3.1 branch is wedged?16:29
bskahanI thought I was going crazy16:29
mgedminbskahan, no, that is not good16:30
mgedmingintas, the newest Zope 3.1 breaks 12 of our ftests16:30
bskahanyeah, but I though it was just me16:30
bskahanthe html header format changed16:30
mgedminnote that buildbot doesn't update Zope 3 -- it used to rely on svn:externals, but there are no more externals16:31
bskahanand its breaking something fierce in the REST interface16:31
mgedminbskahan, I see 500 internal server errors everywhere16:31
bskahanI think that's the REST pages16:32
gintasso I was right about binding to a specific revision16:32
mgedminand instead of REST pages I see browser pages16:32
povbot/svn/commits: * gintas committed revision 4726:16:36
povbot/svn/commits: Backported revision 4654 (evolution script for new permission policy where the site-wide calendar is public and all containers are closed down explicitly).16:36
povbot/svn/commits: A note about this change in the release notes would be nice.16:36
mgedmin lists revisions that get rid of deprecation warnings16:36
mgedmin(possibly not all of them -- just those that I made)16:38
povbot/svn/commits: * jinty committed revision 4727:16:38
povbot/svn/commits: Back-ported 4685:4689 to release branch for issue 341.16:38
gintasI'm thinking about backporting 4703 together with 4687 and 469516:39
gintasbskahan writes: There is an ftest for this in r4703, but that also includes a security policy change that I don't think should be moved to the release branch this late.16:39
gintasI think that the change is small enough to be backported16:39
gintasand the result is worth it16:40
gintas   <content class=".TimetableDict">16:40
gintas     <require permission="schoolbell.edit"16:40
gintas-             interface="" />16:40
gintas+             interface="" />16:40
gintas+    <require permission="schoolbell.view"16:40
gintas+             interface="" />16:40
gintas   </content>16:40
bskahangintas: the change is small, but it is a feature addition with a security change between RC3 and final16:41
bskahanif you and jinty agree on it though16:42
gintaswell, thing is that we will not be doing releases for a while16:43
* jinty is not the release manager for this release, meerly helping out with the obvious16:43
gintasthis is a once-a-year shot, I'd like to have it as polished as possible16:43
mgedmindo I fix the ftests now, or do I revert to an older Z3 snapshot and help with merging/etc?16:45
bskahanmgedmin: this implies that we can't depend on zope 3.1 branch being stable enough not to break schooltool16:46
bskahando we want to move back to a particular zope revision?16:47
* mgedmin wishes srichter said something16:47
gintaswe don't have an external16:47
srichterok, I did not follow16:47
srichterI am having deep issue with my code16:47
gintasI say just grab a several-day old revision that works and let's put this aside for now16:47
srichterhold on, I'll read on16:48
povbot/svn/commits: * jinty committed revision 4728:16:48
povbot/svn/commits: Back-ported 4689:4690 to release branch for issue 341.16:48
bskahanI really would like to get zope 3.1 bug fixes for the next year16:49
srichterwhat failures do you guys get?16:49
mgedminsrichter, briefly, some recent change in Zope 3.1 branch broke something16:49
mgedminone of the failures is16:49
srichter(btw, I think stuart's HTTP header fix is correct)16:49
mgedmin    TypeError: mapply() can not call < object at 0xb580b8ec>16:49
mgedminrev 37885 worked, rev 37961 broke16:50
srichterso we have to look at stuart's changes more carefully?16:50
mgedminwhat was changed in Z3.1 recently?16:50
mgedminI saw that was updated16:50
mgedminand some http code16:50
srichterStuart fixed the behavior when having multiple HTTP headers16:50
srichterx-foo: 116:50
srichterx-foo: 216:50
povbot/svn/commits: * gintas committed revision 4729:16:50
povbot/svn/commits: Cosmetic fixes, added svn:keywords.16:50
srichterbecomes x-foo: 1, 216:51
srichterI am updating ST right now to check out the failures16:52
bskahanI'll be back in 30 minutes16:52
*** bskahan has quit IRC16:53
mgedminI suspect
srichterDAMN, I wish I would not have approved this16:54
srichterI think I actually did not16:55
mgedminI'll try to run tests with that checkin reverted and see if it helps16:55
mgedminwe need change the request class for our RESTive requests16:55
mgedminsrichter, if you have a moment, take a look at src/schoolbell/rest/ftests/__init__.py16:59
mgedminall we need to do is to make GET requests in functional doctests, but use HTTPRequest rather than BrowserRequest16:59
srichter(BTW, I was just about to discover this problem myself, since this is exactely the same deep issue I was looking at ;-)17:00
gintasI'm backporting 470317:02
povbot/svn/commits: * gintas committed revision 4730:17:07
povbot/svn/commits: Backported 4703.17:07
povbot/svn/commits: This is a controversial change because it affects the security policy a little bit (see change in timetable/configure.zcml), but I think that it would be very nice to have this in 0.11.17:07
srichterwhee, I hope benji is done soon with this17:13
*** bskahan has joined #schooltool17:16
srichterok, that seems to fix the problem17:17
srichterI have tried this on my refactor branch and the tests go on for a while longer17:18
mgedminas a workaround I did cd Zope3 && svn merge 37948:37947 .17:19
mgedminall tests pass17:19
srichterwe just need to verify that the fix benji is checking in still has all the tests pass17:19
gintaswhat can we do about issue 323?
gintashoffman says: Actually, the way it is now (rc3) kicks people back to the front page WAY to much.  It is super confusing.17:20
srichterbtw, eventually we should create an IRestRequest interface and RestRequest class17:20
gintasI'll take tvon's new checkins: 4693, 4696, 4697, 4698, 4699, 4700, 4701, 4708, 4709 and 471017:21
srichterthis way views registered for REST (pure HTTP) don't accidentily crop into the browser17:21
*** tvon has quit IRC17:22
ignaswould have done this in the first place yet we had a bug in our skin code that made me try a few different approaches, and HTTP request was the one that was implemented when gintas found the bug17:23
mgedminwe had a IRestRequest once, I think17:23
mgedminsomebody decided it wasn't necessary and refactored it out17:24
povbot/svn/commits: * mg committed revision 4731:17:24
povbot/svn/commits: Spacing17:24
srichter(I think this was a bad decision ;-)17:25
ignassrichter, you should have been there when we were debuging the code ;)17:26
srichterI would think that IRestRequest would make it easier to debug17:26
povbot/svn/commits: * mg committed revision 4732:17:27
povbot/svn/commits: Fix nondeterministic test failures (issue 340).17:27
povbot/svn/commits: Note that the placement of sections that happen simultaneously (i.e. scheduling conflicts) remains nondeterministic in the real user interface.  I think that's not a problem.17:27
ignaswell - the syptom of the bug was - in some cases we had rest specific traversers working and in some cases  - not17:28
ignaschanging the request to HTTP one was making it work in more places ;)17:28
mgedminignas, it is irrelevant17:28
mgedminthat's a conceptual decision -- is a RESTive request something new, or is that just a plain old HTTP request/17:29
gintasdoesn't really matter17:29
gintasright now at least17:29
mgedminalga argued that it is just an HTTP request17:29
mgedminI had no strong opinion one way or another17:29
srichterignas: yeah, and that is exactely the problem17:29
srichternow you use browser traversers for HTTP :-(17:29
mgedminI do not believe anybody raised srichter's argument -- preventing RESTive views from cropping up in Browser UI17:29
srichterbut I fixed most of this already ;-)17:29
mgedmingintas, is issue 340 (rev 4732) worth backporting?  I leave the decision to you17:30
srichterI think it is more than a plain HTTP request, because we have definitely expected semantics17:31
mgedminsrichter, the whole philosophy of REST is that you use plain HTTP requests rather than building something complicated on top17:31
mgedminthis might translate into Zope3-land as "IRestRequest is an HTTP request that is used for RESTive API"17:32
srichteryeah, but it becomes more complicated, because now the lines are blurred17:32
mgedminbut the temptation of just using HTTP requests ("because that's what REST means!") is great17:32
gintasjinty, I'm backporting 4713 right now17:32
srichterpretty much all HTTP protocols now have REST views17:32
* jinty urges anyone who checked in depreciation warning fixes to add them to issue 344 for back-porting17:33
* mgedmin ran out of things to do17:33
jintygintas: anything you want me to take a look at?17:33
gintaslet's see17:33
mgedmingintas, $same_question17:33
srichteralso, most REST views are registered for IPublishTraverse, which is even worse17:34
srichternow RESTive views will be available even in FTP17:34
srichter(I fixed that on the way as much as I could)17:35
gintashmm, issue330 is for me probably17:35
gintasjinty, you can merge fixes in, I think17:36
jintygintas: ok, i'll have a look17:37
mgedminam I the only one who thinks issue numbers ought to have a checksum digit?17:38
gintashmm, I'm starting to think that we will be able to release today after all17:38
jintymgedmin: no, some debian developers also want it for debian bugs, they have 6 digits to get right...17:40
gintasis 4706 backported?17:40
gintaslooks like it is not17:43
gintasby the way, jinty, have you backported 4673 & 4674?17:45
gintasI haven't looked into them deeply, if you know they're simple enough, feel free17:46
povbot/svn/commits: * srichter committed revision 4733:17:48
povbot/svn/commits: Update schooltool to the new packaging structure and traverser code in schoolbell. The pluggable traverser worked great and did not cause any troubles.17:48
jintygintas: yep, back-ported them this morning.17:49
mgedminignas, take a look at
jintygintas: looks like issue 334 & 335 are going to need a huge swathe of back-ports, starting from at least 4696. If you want it I can make a non-conflicting patch (and revision numbers) and submit it to the issue tracker for review first.17:59
gintasI'm still wrestling with 4712 ;(18:02
gintasumm, tvon writes: As of r4699 next/previous now show appropriate numbers (they indicate the batch18:03
gintassize on most pages, or the remainder when there are fewer items than the batch18:03
gintashave we backported changes "up to 4699"?18:04
jintythats the problem18:04
jintywe are also missing quite a few tests, some committed in 4696. (README.txt).18:05
gintasI think we have everything up to 469618:06
gintas(excluding that one)18:06
gintas4699 touches membership.txt which we don't have in the branch at all18:08
jintythat's what I was about to say.18:08
mgedminit's just an ftest18:09
mgedminotoh... I see your point18:09
mgedmin9how do you verify that the backport works?)18:09
gintasit was added in r468818:10
gintasI'll fix this18:10
srichterwhich objects can have timetables again? person. group, resource, term, course, section?18:11
srichtertheir containers too?18:11
mgedminITimetabled objects18:11
mgedminI think containers have no timetables18:12
srichterwell, my branch does not have this directly implemented anymore18:12
mgedmin(unless people who added a schoolwide calendar added a schoolwide timetable as well)18:12
mgedmincan apidoc list all known classes that implement a given interface?18:12
* jinty waits for membership.txt before attacking 334 again18:12
srichterit is part of the interface details view18:13
mgedminmembership.txt appeared in rev 468818:14
mgedminoh, gintas is already on it18:14
povbot/svn/commits: * gintas committed revision 4734:18:15
povbot/svn/commits: Backported rev 4688, which adds the functional test membership.txt.  Later changes modify this doctest.18:15
gintasjinty, you can then merge the changes outlined in issue 334, I'll examine your commit and then see what's missing18:18
gintaswe want that stuff merged anyway18:18
jintycool, I'll get to it after my typing break18:18
gintasI'll check out 330 then18:19
gintasjinty, how's it going?18:26
*** tvon has joined #schooltool18:32
gintashi tvon18:34
tvonhey gintas18:35
tvongintas: whats up with 334?18:37
jintygintas: just starting18:37
gintastvon, jinty's working on backporting all the stuff18:38
tvonah, okay18:38
gintasbut I just found a supposedly major problem (see issue345)18:38
gintasI don't think we'll be able to release today, I really want to test this whole thing manually18:39
gintasmaybe that bug is bogus18:39
gintasotoh, it's not18:40
tvon'tomorrows events' portlet maybe18:43
bskahangintas: I can't recreate that bug18:48
tvonI can18:48
* tvon tries wiping data.fs18:48
tvonI still getit18:50
tvonget it18:50
tvonIts only with recurring events, if you turn that off again it's okay18:51
bskahanwhat Recurrence are you using?18:51
tvonjust 'repeat until'18:51
tvon(this is in SB, havent tried ST)18:52
bskahantesting in schooltool18:52
bskahanI'll try in SB18:52
* mgedmin trying to reproduce it in SB too18:53
mgedminUnauthorized: (< from /home/mg/src/schoolbell/src/schoolbell/app/browser/templates/ object at 0xb12fd78c>, 'browserDefault', 'schoolbell.viewCalendar')18:53
mgedminwhaaa? this makes no sense!18:53
tvonmgedmin: I was getting an error on 'allday' access18:53
mgedmintvon, aha!18:53
tvonUnauthorized: (<schoolbell.calendar.mixins.ExpandedCalendarEvent object at 0xb14a056c>, 'allday', 'schoolbell.view')18:53
mgedmingo to src/schoolbell/app/configure.zcml18:54
mgedminline 25818:54
mgedminchange attributes="resources" to attributes="resources allday", try again18:54
mgedminno, don't18:54
tvonbskahan: you fixed a site-wide calendar permissions bug in ST, right?18:56
* tvon thinks the permissions are set differently in SB (and probably wrongly)18:57
mgedminmissing __parent__ somewhere?18:57
bskahanjust reproduced it in SB trunk18:57
gintas    *  Module, line 550, in getEvents18:57
gintas      if (removeSecurityProxy(event.__parent__) is removeSecurityProxy(self.context) and18:57
gintasUnauthorized: (<schoolbell.calendar.mixins.ExpandedCalendarEvent object at 0xb123216c>, 'allday', 'schoolbell.view')18:57
tvonno, its the ACL18:59
tvonin SB I set the ACL on app.calendar, not app18:59
mgedminno, no argh18:59
mgedminthe ACL indicates that Unauthenticated users have no permissions at all18:59
mgedminwhy can they see nonrecurring events then?18:59
mgedminshould I rm my Data.fs?18:59
* mgedmin looking at http://localhost:7080/acl.html19:00
tvonmgedmin: http://localhost:7080/calendar/acl.html19:00
mgedminalso empty19:00
mgedminmust be my data.fs19:00
tvonoh? thats peculiar19:00
bskahantvon: they have viewCalendar on the calendar, but not 'view'19:01
mgedminwhy does ExpandedCalendarEvent require schoolbell.view, and not schoolbell.viewCalendar?19:02
mgedminsrc/schoolbell/app/configure.zcml line 25719:02
mgedminalso, look around19:02
tvonthat fixes it19:04
mgedminImmutableCalendar a few pages down also wants schoolbell.view rather than viewCalendar19:04
gintasI really want a functional test for this19:05
tvonI've started19:07
tvonadding recur event to site-wide and checking that it can be viewed anon19:07
gintasI would like to leave in a few minutes19:10
tvonalmost there19:10
gintasI will probably be back online in a few hours and will try to solve some of the remaining problems19:10
gintastvon, that's OK, it's obvious that I won't release today19:10
gintasthere aren't too many revisions to be backported; hopefully all of these can be covered today19:11
gintasI still want to do some clickety-click testing after everything is backported19:11
jintyaargh the patches for 334 break schoolbell/app/browser/ftests/app.txt19:11
gintasto catch the likes of issue34519:12
gintasjinty, if you have a reasonable cumulative patch, you can upload it to the issue tracker19:12
gintasand I'll have a closer look and figure out the gory details19:12
bskahanjinty: can you the app.txt failure?19:13
* mgedmin is feeling very disoriented19:13
gintasmgedmin, you're not the only one19:13
bskahanjinty: thanks19:14
jintyI am probably missing a revision somewhere...19:15
*** gintas has quit IRC19:25
jintybskahan: now it passes... huh... guess its time to stop for the day19:28
jintygintas: I compiled a list of revisions for 344 (safer than patches as files are being added). they merge cleanly and all tests pass, I think. Possibly the list is too complete.19:31
ignasgintas has quit19:31
jintyoops, hope he reads the logs.19:32
*** jinty has quit IRC19:32
mgedminI home jinty added that list to the issue itself19:33
mgedminnope :(19:34
povbot/svn/commits: * tvon committed revision 4735:19:38
povbot/svn/commits: fix for ExpandedCalendarEvent permissions in issue34519:38
ignasemm cr*p19:40
ignastarball release of ST does not allow one to look at calendars of resources19:40
ignascan someone please try and reproduce it ?19:40
ignasrelease tarball19:41
mgedmincannot repro with current trunk19:46
ignasoh, found the problem - we didn't backport it19:46
ignasit is fixed in the trun rev4040 and rev404119:46
ignasbut it was not backported, because we marked it as resolved ...19:47
* mgedmin hasn't eaten today19:47
mgedminis it present in 0.11 release branch?19:47
ignasguess not, it was fixed while releasing 0.1019:48
povbot/svn/commits: * srichter committed revision 4736:19:53
povbot/svn/commits: Okay, I would have been surprised, if I would have gotten the traverser stuff right the first time. We have to be very careful that none of the traverser plugins tries to lookup a view. Views should only be looked up after all the plugins failed. Otherwise the semantics of traversal change in very unexpeced ways -- in other words I just spent 2 hours debugging this. ;-)19:53
povbot/svn/commits: * mg committed revision 4737:20:01
povbot/svn/commits: Backported Makefile changes from rev 4659: check out Zope3 if it is not there already.20:01
povbot/svn/commits: * mg committed revision 4738:20:07
povbot/svn/commits: Backported Makefile changes from rev 4659: check out Zope3 if it is not there already.20:07
mgedminok, I'm going to backport deprecation warning changes20:09
mgedminum, how do you merge stuff?20:13
mgedminsvn merge -r 4686:4687 svn+ssh:// .20:13
mgedminoh, it is already applied20:14
mgedminand so is 472020:18
povbot/svn/commits: * mg committed revision 4739:20:19
povbot/svn/commits: Merged rev 4722 -- killing deprecation warnings on startup.20:19
ignastvon, and in case of heh - migration scripts do not migrate personal preferences properly ...20:20
mgedminbug: schooltool-0.11.x footer says " SchoolTool Server Running SchoolTool 1.2rc "20:21
povbot/svn/commits: * mg committed revision 4740:20:22
povbot/svn/commits: Merged rev 4723 -- killing deprecation warnings on startup.20:22
tvonin case of heh?20:23
mgedmintvon, bskahan: the revision number in the footer -- where does it come from?20:23
ignastvon, is buggy, debugging it  now20:25
tvonmgedmin: erm, thats for schoolbell anyways20:25
tvonone sec20:25
tvonschooltool should come from schooltool/__init__20:26
mgedminSchoolBell 1.2 footer says " SchoolBell Server Running SchoolBell 1.2rc " -- which is fine20:26
mgedminbut SchoolTool says its version 1.2, which is not fine, as it is version 0.1120:26 is not migrating prefs.dateformat correctly when given a 0.10 data.fs20:26
tvonmgedmin: I have it, making a ftest20:27
tvonhrm, ideas on how I could change the version in an ftest?20:29
srichterI fixed that footer problem on the refactor branch too20:29
mgedminevolve6 bug is clear20:30
ignas        elif prefs.dateformat == 'YYYY-DD-MM':20:30
ignas            prefs.dateformat = '%Y-%m-%d'20:30
mgedmin'YYYY-DD-MM' is something exotic20:30
mgedminwe used to have 'YYYY-MM-DD'20:30
mgedminwhich is ISO-860120:30
mgedminignas, create an issue please20:31
mgedmintvon, where does the version number come from?20:31
mgedminand why do you want to change it in an ftest?20:31
tvonignas: whats wrong with that?20:32
tvonYYYY-DD-MM does not exist afaik, it changes it to yyyy-mm-dd20:32
tvonmgedmin: so I can test that the ftest shows the schooltool version and not the schoolbell version20:32
mgedminand you do not want to hardcode a specific version in the test?20:33
mgedminI suggest >>> 'SchoolTool %s' % VERSION in str(response)20:33
tvonthat wont prove anything, in svn it will be 'SVN' for both schoolbell and schooltool20:34
tvon(afk 1 min)20:34
mgedmintvon, you can do20:34
mgedmin>>> import schooltool20:35
mgedmin>>> old_version = schooltool.VERSION20:35
mgedmin>>> schooltool.VERSION = '(the correct version number)'20:35
mgedmin>>> print http("""...""")20:35
mgedmin... SchoolTool Server Running SchoolTool (the correct version number) ...20:35
mgedmin>>> schooltool.VERSION = old_version20:36
povbot/svn/commits: * ignas committed revision 4741:20:41
povbot/svn/commits: Fix the issue349 a bug when evolving preferences that had 'YYYY-MM-DD' set as dateformat.20:41
povbot/svn/commits: * srichter committed revision 4742:20:42
povbot/svn/commits: Changed ITimetabled to ITimetables. Also added a new IHaveTimetables interface that the content objects can implement to declare that they have timetables.20:42
povbot/svn/commits: * mg committed revision 4743:20:48
povbot/svn/commits: Backported rev 4741: fix issue 349 -- bug in database upgrades (ISO 8601 date format is not migrated correctly).20:48
povbot/svn/commits: It is not good that evolution scripts have no unit tests.  :(20:48
ignastvon, who did the section table ?20:53
tvonignas: bskahan20:53
tvonignas: that generation was intentional20:53
ignaswhich generation ?20:54
tvon4064:4065 on, personpreferences dateformat was wrongly stored as YYYY-DD-MM, the generation fixes it by changing it to %y-%m-%d20:54
tvonignas: I might be misunderstanding what you did20:55
ignastvon, well what I did was get the 0.10 tarball, create some stuff20:56
ignasand put the data.fs for current release as a base...20:56
mgedminignas, tvon: perhaps you need to convert both 'YYYY-MM-DD' and 'YYYY-DD-MM' into '%Y-%m-%d'20:56
ignasand well - apparently default preferences are stored as YYY-MM-DD20:56
mgedminreopening the issue20:57
tvonso a new generation to do YYYY-MM-DD -> %Y-%m-%d then?20:57
ignasnext time please put a comment near the code that looks outright buggy ...20:58
mgedminit's OK to change existing evolution scripts, if they were not present in any previous release20:58
tvonignas: nod20:58
mgedminI guess was present in rc2, so there's a strong argument to not touch it20:58
mgedmintvon, I created for the version issue20:59
mgedminI assumed you were working on it20:59
ignaswell, imho we should change it because it was buggy after all - the default pref for all the users was YYYY-MM-DD, yet the one you could select when modifying your preferences is YYYY-DD-MM20:59
povbot/svn/commits: * mg committed revision 4744:21:00
povbot/svn/commits: Reverting incorrect backport.  See for more info.21:00
mgedminignas, create evolve${N+1}.py that converts YYYY-MM-DD21:00
mgedminso that people who upgrade 0.10 -> 0.11rc2 -> 0.11 final will not be screwed21:00
ignasi still get it :D21:01
tvonmgedmin: yeah, thanks21:01
mgedminignas, can you revert 4741, and also add a comment in evolve6 that YYYY-DD-MM is intentional there21:02
mgedminnah, I'll do that myself21:02
povbot/svn/commits: * mg committed revision 4745:21:06
povbot/svn/commits: Reverted 4741, added informative comments.  See for more info.21:06
povbot/svn/commits: * tvon committed revision 4746:21:06
povbot/svn/commits: fix for issue350 (schoolbell version showing up in schooltool footer)21:06
*** tvon|x31 has joined #schooltool21:06
tvon|x31argh wifi21:08
tvon|x31ignas: are you doing evolve8 or shall I?21:09
* mgedmin loves svn commit announcements on irc21:09
tvon|x31it's nice21:09
ignasi am already testing it21:09
* tvon|x31 nods21:09
tvon|x31tvon: go away21:09
*** tvon has quit IRC21:10
*** tvon|x31 is now known as tvon21:10
povbot/svn/commits: * mg committed revision 4747:21:10
mgedminwhat's left to merge?21:10
povbot/svn/commits: Backported rev 4746: fix for issue350 (schoolbell version showing up in schooltool footer).21:10
mgedminis anyone keeping track?21:10
tvonmgedmin: ever toy with CIA?21:12
mgedminlooked at it once or twice21:12
mgedmincan't say I've ever toyed with it21:13
povbot/svn/commits: * ignas committed revision 4748:21:13
povbot/svn/commits: The real fix for the issue349.21:13
* tvon thinks it's pretty nice21:14
povbot/svn/commits: * mg committed revision 4749:21:16
povbot/svn/commits: Backported rev 4748: The real fix for the issue349 (YYYY-MM-DD date format not migrated).21:16
mgedmintime for me to go home21:21
mgedminI guess issue 345 needs to be backported, and perhaps investigated21:23
mgedminI don't understand the status of issue 33021:23
bskahanmgedmin: I can reproduce 33021:28
bskahanignas: what was the question about the section index table?21:29
mgedminbskahan, can you write up the steps to reproduce it?21:31
bskahanmgedmin: new data.fs, front page (/calendar), click the printer icon21:34
bskahanthis is in schoolbell21:34
mgedminbskahan, in the issue tracker, please21:34
ignasbskahan, why search only searches in section ids ?21:34
ignaswhy so many spaces before commas ?21:34
ignaswhy when deleting sections you still see only their ids ?21:35
*** mgedmin has quit IRC21:36
bskahanignas: not sure about the search, didn't change its behavior, commas is an html artifact, good point about deleting21:37
* tvon did searching21:37
tvonI know what 330 is I believe... where should I see a print icon?21:38
tvonah, missing timesbd.ttf21:39
* tvon has Times_New_Roman_Bold.ttf21:39
bskahantvon: 330 is caused by ISchoolBellApplication not having a title attribute21:39
* tvon nods21:40
tvonit's more directly caused by access to title not being specified in zcml.  we can just add it there21:41
bskahanshouldn't the interface not being correct be fixed?21:42
tvonsrichter: did you create an ITitleOwner or something along those lines in your branch?21:42
* tvon shrugs21:42
tvonjust because it's in the class doesnt mean it has to be in the interface21:42
tvonI'm not sure what the right answer is21:42
bskahanshouldn't it be in an interface somewhere?21:42
tvonISchoolBellApplication inherits ICalendarOwner which has title21:43
*** pcardune has joined #schooltool21:43
tvonzcml doesnt follow subclasses21:43
pcarduneeldafar: ayt?21:44
bskahantvon:  add     <allow interface=".interfaces.ICalendarOwner" />21:45
eldafarpcardune: yes21:45
tvonbskahan: yeah, that sounds right to me21:45
pcarduneHas anyone ever gotten a Bus Error?  as in, i push a button and suddenly zope crashes and all it says is "Bus error"21:45
tvonwait a sec...21:45
bskahani realize that's not the exact module21:45
tvonicalendarowner doenst have title, I think it should though21:45
tvonno, I guess it should go on ISBApp21:46
tvonI thought ICalendarOwner had it, my mistake21:46
tvonbskahan: are you working on that one?21:55
bskahantvon: is there a default title?21:58
tvonbskahan: yeah, SchoolBell/SchoolTool21:59
*** ignas has quit IRC21:59
bskahanwhere is that default set?22:00 in ApplicationPreferences22:00
povbot/svn/commits: * srichter committed revision 4750:22:13
povbot/svn/commits: Made CalendarOverlayInfo's showTimetables available via an adapter instead of implementing a derived class. This allowed me to remove the second implementation of IPerson. This simplified code a bit.22:13
bskahanfor some reason the PDF view doesn't display the site name correctly22:14
srichterI wonder whether courses and sections should be together in a package22:14
bskahansrichter: yes, I think they should22:14
srichterok, will do that then :-)22:14
povbot/svn/commits: * bskahan committed revision 4751:22:20
povbot/svn/commits: added a title TextLine attribute to the ISchoolBellApplication interface.  This allows the current security definition to work properly.22:20
povbot/svn/commits: There is no attached doctest because verifyObject is allready tested in test_app.  There is an ftest for regressions though.22:21
povbot/svn/commits: This should close issue330 and is a candidate for backporting.  It exposes a new bug however, the pdf of the frontpage calendar does not properly print the site title in the header.  creating a seperate issue351 for that bug.22:21
*** tvon has quit IRC22:28
*** tvon has joined #schooltool22:50
*** pcardune_ has joined #schooltool23:39
*** pcardune has quit IRC23:39
*** bskahan has quit IRC23:58

Generated by 2.15.1 by Marius Gedminas - find it at!