IRC log of #schooltool for Monday, 2005-10-03

*** srichter has quit IRC00:17
*** srichter has joined #schooltool00:39
*** srichter has quit IRC01:13
*** srichter has joined #schooltool05:04
*** hoffman|edubu has quit IRC06:06
*** Aiste has joined #schooltool09:31
*** tamere has joined #schooltool11:32
*** tamere has left #schooltool11:47
*** ignas has quit IRC12:06
*** mgedmin has joined #schooltool12:15
*** ignas has joined #schooltool12:45
*** vidasp has joined #schooltool13:24
povbot`/svn/commits: * alga committed revision 5144:14:03
povbot`/svn/commits: A view to generate sample data.14:03
*** srichter has quit IRC14:20
povbot`/svn/commits: * ignas committed revision 5145:14:49
povbot`/svn/commits: Bug fix for Issue381 (broken monthly recurrence rules).14:49
*** jinty has joined #schooltool14:58
povbot`/svn/commits: * ignas committed revision 5146:15:09
povbot`/svn/commits: Fixed rest-client system path was not constructed properly.15:09
*** srichter has joined #schooltool15:14
*** thisfred has joined #schooltool15:29
*** ignas has quit IRC16:14
*** ignas has joined #schooltool16:14
*** mgedmin has quit IRC16:16
*** mgedmin has joined #schooltool16:17
* mgedmin clears the mailman queue16:20
*** bskahan has joined #schooltool16:21
mgedminth1a, a long time ago you said you'd take care of pending mailman moderator requests16:21
th1aI do on rare occasions.16:21
mgedminyesterday I noticed and approved a few non-member submissions on the list16:21
mgedmina bunch of them were from mark's canonical address16:22
th1aOh, that's where those came from.16:22
mgedminI thought you'd notice ;-)16:22
th1aYou didn't happen to that address to the list, did you?16:23
mgedminto <what> that address to the list?16:23
th1aadd, that is.16:24
mgedminI added it to the whilelist16:24
mgedminI hope16:24
th1awhitelist?16:24
mgedminfirefox was very unhappy with the huge queue16:24
mgedminit ate 90% of my RAM16:24
mgedminand took ages to post the form16:24
th1aThe whole system is a disaster.16:25
mgedminagreed16:25
mgedminth1a, http://lists.schooltool.org/mailman/admin/schooltool/privacy/sender16:25
mgedminthere's a whitelist of non-members there16:25
th1aAh.  Thanks.16:25
mgedminI do not see mark's address there, though16:26
mgedminperhaps my aborted form submission only let the messages through but didn't change any settings16:26
* mgedmin shrugs16:26
mgedminmailman is ... not nice16:26
mgedmintoday I ssh'ed into the server, played a bit with shell scripts, unplickled the queued messages, piped them to spamassassin, and told mailman to discard those that were spam16:27
mgedminand processed the rest via the web, by looking for word 'school'16:27
th1aThat sounds like a good idea.16:27
mgedminif you could take a look at the pending requests, say, once a week, that would prevent huge browser-killing queues from piling up (hint, hint)16:28
*** alga has joined #schooltool16:28
mgedminperhaps it would be a good idea to move mailman to the new server, and filter all incoming mail through spambayes16:28
mgedminok, the meeting is about to start -- I'll shut up about mailman now16:29
th1aGood morning/evening all.16:30
th1asrichter, bskahan:  ayt?16:31
algaHello16:31
algaFirst let me introduce someone16:32
algavidasp is Vidas Pa┼żusis, our new employee16:32
th1aHello, vidasp.16:32
algaHe's a physics PhD student working for us part time since today :-)16:33
bskahanhi vidasp16:33
th1aMore physicists!16:33
th1aZope 3 seems to make sense to physicists, so I guess that's good :-)16:34
algahe he16:34
th1asrichter is a physicist.16:34
vidasphi everybody16:34
th1aAlthough he doesn't seem to actually be here.16:34
algath1a, maybe you can ping him on the phone?16:35
th1aWell, srichter mentioned that he's going to be traveling (to a sprint) this week?16:36
bskahanyeah16:36
th1aI don't know that I have his phone number.16:36
th1aI'll have to look.16:36
bskahannot sure what the timeframe for that is though16:36
th1aDoes everyone need to talk to him?16:36
mgedminapparently not16:37
th1aI don't think I have his phone number anyhow.16:38
th1aI mean, I can't find it.16:38
th1aSo let's plow ahead.16:38
th1abskahan:  Let's get up to date on the website status first.16:38
bskahanI sent Gus an email on friday, haven't gotten a response yet16:39
bskahanbut he'll probably roll the DNS as soon as he gets that16:39
th1aAnd that'll take a day or so to propogate?16:40
bskahanyes16:40
th1aOK.16:40
bskahanas soon as that rolls I'll set up the cado mailing list16:40
bskahanI want to make sure all the MX dns stuff is still right with the existing lists16:40
th1aOK.16:41
th1aWe should spend a bit more time fixing (simplifying) the documentation and product index pages.16:41
th1aIs there a time today we can do that?16:41
bskahanyeah16:41
bskahanwhat sort of things are you thinking?16:42
th1aJust finishing what we started in the doc index.16:42
th1aDid you work on the chroot problem any further?16:43
bskahannot successfully, but that's my #1 task at the moment16:43
bskahanI'm setting up 216:43
bskahanone for demo and one for your school16:43
th1aIf srichter isn't going to be doing much work this week, are there things you can do to get ahead of him on the viewlet work?16:44
bskahanyes16:44
bskahanI can continue moving existing view to pagelets based on the work he did allready16:45
th1aOK.  After the chroots & demo servers are up, that's probably the biggest priority.16:45
bskahanok16:45
th1aThe rest of the website stuff isn't particularly time-sensitive.16:46
* bskahan nods16:46
th1abskahan:  Anything else?16:47
bskahanno, I think that's it for mew16:47
bskahanme16:47
th1aOK.  POV?16:47
algaThings done last week:16:47
alga* reviewing srichter's refactorings16:47
alga* triaging and fixing bugs: 405 381 379 347 396, found more bugs16:47
alga* missing functional tests for overlay portlet16:47
alga* Started implementing pluggable sample data generators16:47
algaBasically, some of us are still stuck in the bugfixing phase16:48
th1aYou have done a good job of shaking loose bugs, but it is a little depressing.16:48
algaAnd some haven't yet gotten round to reading all of srichter's checkin messages on the refactor branch16:48
* mgedmin blushes16:48
th1amgedmin:  On the whole, you haven't run into many issues with the refactor branch, though, right?16:49
ignaswell16:49
algapersonally, I think it's fine16:49
mgedminI'm the one who hasn't gotten round to looking at it deeply16:50
mgedminwe have encountered issues16:50
mgedminfor example, the restive interface plainly doesn't work in the trunk16:50
mgedmin(probably a Zope3-trunk realted breakage)16:50
ignasit is good enough, some things stoped working though and will have to be fixed16:50
mgedminand there are no functional tests for schoolbell16:50
mgedmin(there are two, but they are disabled, and do not work anyway)16:50
th1aWell, not having functional tests for SchoolBell is acceptable, I suppose, since we don't really know how we're going to build it yet.16:51
th1aI spent a lot of time stressing that we needed to focus on SchoolTool, so that may be the inevitable result.16:52
algath1a, but it feels that there are a lot of unfinished flaky things in the calendaring code16:53
th1aThe REST tests passed until they were broken by the Zope 3 trunk, right?16:53
mgedminthe tests pass now16:53
mgedminthe code is broken16:53
ignasthe server itself16:54
mgedminit's an integration problem somewhere16:54
th1aalga:  Oh, so it isn't so much that there aren't SchoolBell ftests as there aren't *calendaring* ftests?16:54
mgedminzope3 style functional tests directly create HTTPRequest objects, and that part works16:54
algath1a: that is not so, sorry to have mislead you16:54
* mgedmin is confused16:54
* th1a is confused.16:55
algacalendaring code is tested, but still has problems16:55
algaschoolbell app does not have ftests16:55
mgedminschooltool/schoolbell restive server doesn't work16:55
mgedminthree unrelated issues16:55
mgedmin(or groups of issues?)16:55
th1aOK, so the policy on SchoolBell as a standalone app in the trunk is that it is a low priority.16:56
mgedminok16:56
th1aAs long as srichter has a plan for getting it to work sometime next spring, that's fine with me.16:57
th1aAre we these issues going into the tracker?16:57
th1aI guess I should read the unread issues in my inbox.16:58
th1aDo we need to make any decisions about outstanding (and recently found) bugs?17:00
th1aWhich ones to fix?17:00
ignasthe amount of bugs is kind of overwhelming, so prioritizing them might be nice ...17:02
mgedminall the problems I've found are in the issue tracker17:02
ignasi allways have the urge to fix something instead of doing the real important work ;)17:02
algaWe'll definitely need REST working to finish this contract :-)17:02
ignass/allways/always17:02
th1aalga:  That's certainly true.17:02
th1aDo you have any sense of why it broke?17:03
mgedminWSGI support in zope 3 trunk, I guess17:03
algaignas has a hacky workaround17:03
mgedminsome interfaces changed17:03
mgedminorder of method arguments17:03
mgedminI'd like to talk to srichter about this17:04
mgedminhe made a couple of WSGI/servertype related changes in the refactor branch17:04
th1aOK.  Can you send him an email?17:04
mgedminI can17:04
mgedminalthough I think he got the issue by mail from the issue tracker17:04
th1aWell, an extra ping might help.  Did you add him to the default nosy list?17:06
th1aWhy do grants persist after a person is deleted?17:06
ignasbecause grants are stored separately17:07
th1aIs that hard to fix or just an oversight?17:08
ignasit is not very hard to fix i guess17:08
th1aWell, it seems like that would be a good one to fix.17:09
th1aOK, before I get too rambly.  What are the goals for this week?17:09
th1aFor POV?17:09
algaI'll go on working on sample data17:11
ignasi'll be working on restclient17:12
ignasthe library17:12
mgedminwe'll continue bugtracking & bugfixing17:12
th1aignas:  Have you started on the restclient?17:13
ignasyep17:13
algath1a: re bugtracking17:13
algathere is a list of issues re. timezones17:13
th1aYes.17:14
algabskahan kind of promised to fix all those on his spare time17:14
th1aThis is the hell that SteveA warned us against.17:14
th1aYes, bskahan did, didn't he?17:14
algahi did promise17:14
algabut he never ever fixed a tz bug17:15
algain the past month or so17:15
th1abskahan?17:15
algahttp://issues.schooltool.org/issue?%3Aaction=search&%3Asearch_text=&title=&%3Acolumns=title&topic=&id=&%3Acolumns=id&creation=&creator=&activity=&%3Acolumns=activity&%3Asort=activity&priority=&%3Agroup=priority&status=-1%2C1%2C2%2C3%2C4%2C5%2C6%2C7&%3Acolumns=status&assignedto=19&%3Acolumns=assignedto&%3Apagesize=50&%3Astartwith=0&%3Asortdir=on&%3Aqueryname=17:16
algaessentially, our approach to timezones is wrong17:17
th1aI was afraid you were going to say something like that.17:17
algathe right thing to do is to store the events *with* timezones17:17
alganot convert everything to UTC17:17
ignaswhich would have to reflect in the UI in some way17:18
th1aHm.17:19
th1aSo just bugfixing is unlikely to solve the problem anyhow?17:19
alganot all problems17:19
algabut some issues are just simple unfinished things17:20
algalike boundaries of dates17:20
mgedminlet's split this into two parts17:20
mgedmin1) we can support many use cases by storing times in UTC17:20
mgedmin2) we can support more use cases if we store time + timezone17:20
mgedminthere are a number of bugs in our implementation of (1)17:21
mgedminlet's fix those17:21
algayep17:21
th1aOK.17:21
th1aI'll follow up with bskahan.17:22
algaany feedback would be nice17:22
alga'cause we assign issues to bskahan and nothing happens17:23
th1aUnderstood.17:23
algawe don't know whether he agrees with them or not17:23
algawhether he even notices17:23
th1aApparently not noticing at the moment.17:23
mgedminwe included a couple of timezone-related bugs in our contract, as part of the bugfixing story17:24
mgedminI don't see why we shouldn't just go ahead and fix those17:24
th1aIf they're in the contract, definitely do that.17:24
th1aAnything else before we wrap up?17:26
mgedminI don't think so17:26
th1aOK.  I'll talk to bskahan about the timezone bugs.17:27
th1aWelcome aboard vidasp.17:28
* th1a bangs the virtual gavel.17:28
srichterdarn, too late :-(17:30
srichterI was so into my grading today :-(17:30
th1aPhysicists!17:30
srichteranything else I should know?17:30
th1aThe main point seems to be that REST was broken by the Z3 trunk.17:31
srichteroh, probably the result stuff17:31
th1aPOV has some REST related stories to complete, so they need it to work.17:32
srichterI will catch up as soon as I have my stuff graded17:35
srichterI will also work from home thise afternoon17:35
th1asrichter:  OK.17:36
algasrichter: there's an issue assigned to you in the tracker17:48
*** alga has quit IRC17:48
mgedminsrichter, http://issues.schooltool.org/issue41117:48
mgedminwe though you might know what is going on from a glance17:49
mgedminwe suspect that you know everything about Zope 3 WSGI support and server types17:49
srichterthat is a true assumption :-)18:14
srichterbasically get rid of passing in any output stream :-)18:14
srichterI'll fix the bug this afternoon, or at least have a detailed look at it18:14
* srichter got a new 24'' 1920x1200 plasma screen this weekend and hopes to increase his productivity greatly once he gets it working properly ;-)18:16
mgedminwoohoo18:27
*** Aiste has quit IRC18:31
*** Aiste has joined #schooltool18:34
povbot`/svn/commits: * mg committed revision 5147:18:38
povbot`/svn/commits: Tiny refactoring.18:38
bskahanmgedmin, I missed alga earlier, I will tackle the bugs assigned to me in the issue tracker18:40
ignasmgedmin, we had a way to get all possible Relationship URI's out of schooltool through RESTive interface, should i reimplement it or just hardcode the list of relationships in the client code ?19:02
mgedminignas, yes19:03
*** srichter has quit IRC19:08
*** jinty has quit IRC19:44
*** vidasp has quit IRC19:52
povbot`/svn/commits: * mg committed revision 5148:19:53
povbot`/svn/commits: Calendar overlay API enhancement: `overlaid_calendars.add()` now returns the ICalendarOverlayInfo object.19:53
povbot`/svn/commits: This is necessary if you want to disable timetable overlay on the new calendar:19:53
povbot`/svn/commits: >>> info = person.overlaid_calendars.add(other.calendar) >>> IShowTimetables(info).showTimetables = False19:53
*** pcardune has joined #schooltool20:00
*** Aiste has quit IRC20:04
povbot`/svn/commits: * mg committed revision 5149:20:07
povbot`/svn/commits: Oops.  Forgot to add the file.20:07
mgedminth1a, http://issues.schooltool.org/issue367 (duplicate events in calendar overlay)20:10
mgedminthe proposed solution is to disable the Timetable checkbox by default20:10
mgedminfor groups20:10
mgedminand, I assume, sections20:10
mgedminbut keep the checkbox enabled by default for persons and resources20:11
mgedminis that right?20:11
*** ignas has quit IRC20:15
*** jinty has joined #schooltool20:16
*** Aiste has joined #schooltool20:22
*** pcardune has quit IRC20:27
*** srichter has joined #schooltool20:41
th1amgedmin:  Only leave it enabled for persons.20:43
mgedminok20:46
bskahanmgedmin, do we have a use case where a person would have a timetable and another user would subscribe to it?20:49
mgedminreal-world use case?20:50
mgedmin"When does my buddy Paulius have lessons?"20:50
th1abskahan: Certainly the main office often has to locate a student.20:51
* mgedmin edits src/schooltool/app/browser/overlay.py20:53
* mgedmin adds a print statement, runs the tests, and is mystified when none fail20:54
* mgedmin adds raise NotImplementedError and gets the failure he wants20:54
* mgedmin discovers that the debug print was eaten by a wildcard in the doctest20:54
*** pcardune has joined #schooltool20:57
*** pcardune has quit IRC20:58
*** pcardune has joined #schooltool20:59
pcardunejinty, ayt?21:18
jintyindeed21:25
jintypcardune: time for an update to cando?21:25
povbot`/svn/commits: * mg committed revision 5150:21:28
povbot`/svn/commits: Refactoring: extracted method CalendarSelectionView._updateSelection, added unit test.  Added a TODO comment where the code will have to be modified to fix issue 367.21:28
povbot`/svn/commits: * srichter committed revision 5151:21:40
povbot`/svn/commits: Implemented ignas' suggestion of the fix for issue 411. I agree that this is the best way of doing it.21:40
srichtermgedmin: I fixed 41121:40
mgedminsrichter, joy!21:40
mgedminsrichter, have you seen how the developer mode menu interacts with the daily calendar view?21:41
srichterno21:41
mgedmintry it, it is fun21:41
srichterok, will do21:41
mgedminthe menu hides behind every second hourly bar21:41
mgedminand behind all events21:41
srichteroh yeah, I have seen that21:42
srichtertoo much CSS woodoo for me21:42
mgedminalso, the absolutely positioned daily/weekly/monthly/yearly icons21:42
mgedminyeah21:42
srichterI refer to our CSS guys to fix that :-)21:42
mgedminperhaps .devmode { z-order: 999; } or something would help21:42
mgedminI should create an issue for this21:42
mgedminsrichter, could you take a look at http://issues.schooltool.org/issue39121:44
mgedminits related to one of your refactorings21:44
mgedminpermissions for the IShowTimetables adapter21:45
mgedminperhaps it can be fixed by making it a trusted adapter?21:45
* mgedmin runs make ftest21:45
* mgedmin discovers that he cannot commit a fix for issue 367 because he just triggered issue 391 in functional tests21:46
srichterI have a look21:47
* mgedmin tries to add trusted="true" to the IShowTimetables adapter21:48
* mgedmin runs make ftest and gets a bunch of HTTPError: HTTP Error 500: Internal Server Error21:48
pcardunejinty, it's time for a new update21:48
mgedminaha!     ForbiddenAttribute: ('showTimetables', <schooltool.app.app.ShowTimetables object at 0xb5d1c5ac>)21:49
* mgedmin adds security declarations for the adapter21:50
* mgedmin gets different ftest failures21:51
mgedmin    Unauthorized: (<schooltool.app.app.ShowTimetables object at 0xb5d638ac>, 'showTimetables', 'schooltool.view')21:51
* mgedmin adds goes to ShowTimetables.__init__21:52
* mgedmin adds self.__parent__ = context.__parent__ # for local security grants21:52
* mgedmin runs ftests21:52
mgedminsrichter, I've a fix for issue 39121:54
mgedminpatch: http://issues.schooltool.org/file72/fix-for-issue391.patch21:54
jintypcardune: looking at the changelog now,21:54
mgedminall it needs is a functional test21:54
jintypcardune: in future could you add which rvisions have been merged to the log? "merge from trunk" tells me nothing21:55
pcarduneyes, i can do that21:57
pcardunewhen i did the merge, i just said merge -r 390:HEAD blah blah blah21:57
srichtermgedmin: an ftest should be easy using test browser :-)22:00
jintypcardune: yeah, it's a fundamental problem with svn. the standard workaround is something like: svn ci -m 'merged from trunk, used command svn merge -r 390:HEAD blah'22:00
srichterI wonder whether we should have a directory in which we can put tests for issues22:00
pcardunejinty, that makes sense, i'll be sure to do that next time22:00
srichterI would prefer not having those test be around the documentation driven ones22:00
mgedminsrichter, do you want to write that easy ftest?22:01
mgedminotoh, perhaps I can do it22:01
mgedminif I'm careful enough22:01
srichterwell, you just follow your own recipe from the original report22:01
pcardunejinty, going to lunch now, will be back in 30 minutes or so, email me if there are any problems22:02
pcardunejinty, actually, i won't be back for 2 hours, as I have class.22:03
jintyno worries, I'll mail you when I'm finished22:03
povbot`/svn/commits: * alga committed revision 5152:22:04
povbot`/svn/commits: Set svn:keywrods.22:04
mgedminalga, I hope you didn't really set svn:keywrods22:06
mgedminoh, screw the separate ftest for 391, my change that fixes 367 requires the fix for 391 anyway22:07
mgedminso I've got an (indirect) regression test22:07
* mgedmin waves hands vigorously22:07
srichterhe he22:08
srichterfine with me :-)22:08
*** srichter has quit IRC22:09
povbot`/svn/commits: * mg committed revision 5153:22:10
povbot`/svn/commits: Fix issue 391: traceback when trying to uncheck timetable overlay checkbox.22:10
povbot`/svn/commits: (This fix should come with a functional test, but, as it happens, my subsequent fix for issue 367 triggers this bug in our existing functional test suite.)22:10
mgedminaargh, my fix for 391 fixes functional tests, but breaks unit tests22:13
mgedminthis is what happens when you try to fix multiple issues in a single dirty sandbox22:13
povbot`/svn/commits: * mg committed revision 5154:22:15
povbot`/svn/commits: Oops.  My fix for issue 391 broke unit tests, and I didn't notice.22:15
povbot`/svn/commits: * mg committed revision 5155:22:17
povbot`/svn/commits: Fix issue 367: duplicate timetable events showing up in calendars.22:17
povbot`/svn/commits: The implemented fix is the one suggested in the issue tracker: when you include calendars of sections/groups/resources, disable the timetable checkbox by default.22:17
mgedminsrichter is gone :-/22:20
mgedminI wanted to ask a question about restive cookies22:20
mgedminbskahan, you're the CSS guru22:45
mgedminsrichter would be happy if you looked at http://issues.schooltool.org/issue41422:46
th1aActually, I think tvon is responsible for the devmode CSS.22:46
th1aWell, more precisely, he volunteered to help srichter, iirc.22:47
mgedminah, ok22:47
th1aIt wasn't part of a formal contract.22:47
tvonsrichter is the one that fixed positioned it... I can take a look at it if you like22:47
th1aHowever, it seems like we should be able to clean it up as part of the overall UI rearrangement which is underway.22:48
th1aThat is, I'm not sure that there is much point in fixing it this week.22:48
mgedminI hope that it can be fixed by adding a single CSS rule22:49
mgedmin.devmode {z-order:999;} or something22:49
tvonthats what I was thinking22:49
th1aWell, I presume this will become some kind of "viewlet."22:49
mgedmingood night, all22:54
*** mgedmin has quit IRC22:54
*** Aiste has quit IRC22:54
*** bskahan has quit IRC23:11
*** srichter has joined #schooltool23:15

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