| *** Spec has quit IRC | 00:02 | |
| *** vidasp has quit IRC | 01:38 | |
| *** RockyBurt has joined #schooltool | 04:40 | |
| *** wrobel has quit IRC | 05:04 | |
| *** wrobel` has joined #schooltool | 05:04 | |
| *** didymo has quit IRC | 06:36 | |
| *** tiredbones has quit IRC | 09:52 | |
| *** tiredbones has joined #schooltool | 09:52 | |
| *** alga has joined #SchoolTool | 12:04 | |
| *** shapr` has joined #schooltool | 12:20 | |
| *** shapr has quit IRC | 12:20 | |
| *** alga has quit IRC | 12:42 | |
| *** Aiste has quit IRC | 12:45 | |
| *** shapr` is now known as shapr | 12:54 | |
| *** mgedmin has joined #schooltool | 13:00 | |
| *** vidasp has joined #schooltool | 13:08 | |
| *** alga has joined #SchoolTool | 13:14 | |
| *** Aiste has joined #schooltool | 13:14 | |
| *** vidasp has quit IRC | 14:25 | |
| alga | tiredbones: here's an interesting snippet for you: | 15:33 |
|---|---|---|
| alga | Troubleshooting | 15:33 |
| alga | * We've seen segfaults when running the test suite. This was | 15:33 |
| alga | invariably caused by having parts of Zope 2 or ZODB 3 (which | 15:33 |
| alga | goes with Zope 2) installed in site-packages or elsewhere on | 15:33 |
| alga | Python's module search path (sys.path). So make sure you have | 15:33 |
| alga | no parts of Zope 2 or ZODB 3 on sys.path when running Zope 3! | 15:33 |
| alga | Zope3/doc/INSTALL.txt | 15:34 |
| tiredbones | alga, thanks very much! I hope that's my problem. I was just getting ready to order a motherboard. | 15:43 |
| *** mgedmin has quit IRC | 15:48 | |
| alga | ignas said that your gcc was crashing, and that python has brouhgt the whole system down | 15:58 |
| *** ignas has joined #schooltool | 15:59 | |
| alga | wrong ZODB version on pythonpath does not affect such things | 15:59 |
| tiredbones | alga, the system wasn't crashing, but I couldn't run 'make test'.I was getting seg faults on build for BTree. | 17:01 |
| tiredbones | alga, what do you mean here - wrong ZODB version on pythonpath does not affect such things | 17:02 |
| alga | I had an impression that it was not just the tests that were failing, that the compilations were also | 17:40 |
| alga | sorry, my English is not at its straightest today | 17:41 |
| tiredbones | 'make test' ran now I'm having prblems with 'make ftest' any help -> http://pastebin.com/556054. I removed an old copy of Zope I had around. | 17:41 |
| tiredbones | ls | 17:42 |
| alga | what if you start python2.4 and type in "import ZODB"? | 17:45 |
| alga | if it succeeds, ZODB.__file__ | 17:45 |
| tiredbones | alga, looky, looky. It succeed - '/usr/local/Zope-3.2/zope3/src/ZODB/__init__.pyc' | 17:47 |
| tiredbones | alga, can I/should I just dump ZODB? | 17:48 |
| *** RockyBurt has quit IRC | 18:00 | |
| *** th1a has joined #schooltool | 18:18 | |
| alga | OK | 18:27 |
| alga | tiredbones: it looks like the correct version of ZODB... | 18:27 |
| tiredbones | alga, So this is not my problem/ | 18:28 |
| *** Aiste has quit IRC | 18:28 | |
| tiredbones | when I run 'make build update-translations following http://pastebin.com/556146. | 18:35 |
| tiredbones | when I run 'make build update-translations' i get the following http://pastebin.com/556146. | 18:36 |
| povbot | /svn/commits: * ignas committed revision 5759: | 19:16 |
| povbot | /svn/commits: Refactored DailyCalendarRowsView so it would be easier to implement a timezone timetable interaction bugs. | 19:16 |
| ignas | srichter, ping | 20:08 |
| srichter | pong | 20:12 |
| ignas | i am going through schooltool trying to fix some TZ bugs | 20:13 |
| ignas | and you are using "self.date = datetime.date.today() | 20:13 |
| ignas | " in commendation | 20:13 |
| ignas | is it intentional ? | 20:13 |
| ignas | most of the time the date should be taken with the users timezone or app timezone taken into account | 20:14 |
| ignas | i can't recall exactly now, but i am afraid that date.today() is in the server timezone | 20:14 |
| ignas | not utc event | 20:14 |
| ignas | even | 20:14 |
| srichter | feel free to change it | 20:15 |
| srichter | I probably did not even think about it when I wrote it | 20:15 |
| ignas | that's why i am asking you to think of which date you want in there | 20:16 |
| ignas | as i don't know what that date is used for | 20:16 |
| ignas | the only test is checking that it is read only | 20:16 |
| ignas | and yes - date.today() returns a date in the server timezone | 20:16 |
| srichter | it is the date at which the commendation was made | 20:16 |
| ignas | not really the date you want to use in a web application | 20:16 |
| srichter | the timezone for that date is not important in my opinion | 20:17 |
| srichter | for example, let's say you created the commendation at 1:00 am on February 2, 2006 | 20:18 |
| srichter | then in my timezone it would be February 1, 2006 | 20:18 |
| srichter | I think it would be ok, if the commendation always shows it was created on February 2, 2006 | 20:19 |
| srichter | also, we are talking about small schools here, so server tz == user tz | 20:19 |
| ignas | err the problem is - that more likely scenario is - i am making a commendation 2006-02-02 and seeing it listed under 2006-02-01 in the list | 20:19 |
| srichter | and commendations is a demo package anyways, though it should document the right way of course | 20:19 |
| ignas | because iirc many unix/linux servers are using UTC as the default timezone | 20:20 |
| srichter | as I said, I did not even think about it, so feel free to change it to whatever is right | 20:20 |
| srichter | then I at least know where to look the next time I am dealing with TZs | 20:21 |
| ignas | so which one is *righter* UTC or *application timezone* ? | 20:21 |
| ignas | guess should talk to th1a about these | 20:22 |
| ignas | :/ | 20:22 |
| srichter | user timezone is more correct | 20:24 |
| ignas | user as in - the currently logged in user | 20:24 |
| ignas | the principal | 20:25 |
| srichter | mmh, I guess it should see that it is Feb 2 for the user that did input the date and just always display Feb 2 | 20:25 |
| srichter | yes | 20:25 |
| srichter | well, I take that back with the user; I think application timezone is pretty good | 20:25 |
| ignas | i guess the *right way* is storing whole UTC timestamp, and displaying it according to user preferences | 20:26 |
| ignas | only the date part | 20:26 |
| ignas | so no info is lost, and it is easy to fix the code ;) | 20:26 |
| srichter | ok | 20:29 |
| th1a | OK... back from my shower. | 20:33 |
| ignas | i guess i should write up some guidelines | 20:33 |
| ignas | for datetime/date/timezone usage | 20:33 |
| ignas | in schooltool | 20:33 |
| th1a | ignas: That's exactly what I was thinking in the shower. | 20:34 |
| th1a | Commendations is tough in part because it is a trivial case. | 20:34 |
| th1a | So nobody ultimately cares. | 20:34 |
| th1a | Let's think about something more important. | 20:34 |
| ignas | at the moment one guideline - "Store timezone('utc').localize(datetime.datetime.utcnow()) not the date in ZODB" | 20:35 |
| th1a | If I'm a teacher, and I mis-set my personal timezone, when I enter attendance, it should be stored in the school's timezone. | 20:35 |
| th1a | Although, if I'm not mistaken, we can't explicitly set the timezone of the school at this point? | 20:36 |
| ignas | we have "application timezone" | 20:36 |
| ignas | though - you can set timezone to anything you want | 20:36 |
| ignas | you will not change the "real time" of the event | 20:36 |
| ignas | you might mess up the datetime.date | 20:37 |
| ignas | that's why we should never store it in the ZODB | 20:37 |
| th1a | OK. Where do you set application timezone? | 20:37 |
| ignas | click on Manager in the breadcrumbs | 20:38 |
| ignas | click Site prefferences | 20:38 |
| ignas | iirc | 20:38 |
| ignas | it is the application timezone setting, which means - untimezoned things like timetables, term end/start will act according to it, or are already doing that | 20:39 |
| th1a | OK. That's in the trunk but not the releases. | 20:39 |
| ignas | yes | 20:39 |
| th1a | So basically, the tradeoff is this: | 20:40 |
| th1a | If a person is working outside the school's timezone, from there perspective, sometimes unexpected things will happen. | 20:40 |
| th1a | From their perspective. | 20:41 |
| ignas | not really | 20:41 |
| th1a | But everything will look right to the user in the regular timezone? | 20:41 |
| ignas | i don't think we have any glitches | 20:41 |
| ignas | i mean - i am working on bugs that app timezone enables fixes for | 20:42 |
| ignas | they were flaged already because they were triggered by users not being in UTC | 20:42 |
| th1a | If I can stop trying to think about this, I will happily do so. | 20:42 |
| th1a | ignas: Are there any design decisions to be made then? | 20:43 |
| th1a | That you need my opinion on? | 20:43 |
| ignas | not at the moment, there might appear some ... like - when i should use the user timezone and when i should use app timezone | 20:43 |
| ignas | but not any i am aware of at the moment | 20:43 |
| th1a | When you say "use" what do you mean? | 20:43 |
| ignas | an example from the past: | 20:44 |
| ignas | timetable events had no timezone | 20:44 |
| ignas | question was - when should they be ? | 20:44 |
| ignas | answer was - application timezone should be used for timetable events | 20:44 |
| ignas | as timezone support is flaky in some parts, i might find more places that are not using timezones or are using wrong ones | 20:45 |
| th1a | OK. | 20:46 |
| *** shapr has quit IRC | 20:56 | |
| *** mgedmin has joined #schooltool | 21:00 | |
| ignas | th1a, just found such a place | 21:05 |
| th1a | Go ahead. | 21:05 |
| ignas | th1a, when one is adding an emergency timetable change | 21:05 |
| ignas | one should see the today of his own or app ? | 21:06 |
| th1a | Which day is displayed? | 21:06 |
| ignas | at the moment - the server date today | 21:06 |
| ignas | user should see the today of the application i guess | 21:07 |
| th1a | Yes. | 21:07 |
| ignas | this stuff is confusing :/ | 21:08 |
| ignas | poorly configured schooltool might have usability problems ... | 21:08 |
| ignas | as probably any application ... | 21:08 |
| ignas | maybe we should display the application timezone in the user preferrence dialog, to give them a hint | 21:09 |
| th1a | Do we have anything at this point that sets new users to the application timezone by default? | 21:10 |
| ignas | no | 21:10 |
| ignas | all users start with UTC iirc | 21:10 |
| th1a | I think that's a major bug. | 21:10 |
| th1a | Why don't you implement that. | 21:11 |
| ignas | as soon as i'll finish with the timezone mess etria left | 21:11 |
| th1a | :-( | 21:12 |
| th1a | Lessons learned, at least. | 21:12 |
| ignas | timezones are difficult, and tricky even if they don't look like that | 21:12 |
| ignas | datetime has misleading API | 21:12 |
| th1a | Ultimately, timezone support was a mistake. | 21:13 |
| th1a | But it was kind of inevitable once we decided to create SchoolBell. | 21:13 |
| ignas | well - it is difficult to implement/fix something you don't understand, now that i think that i *do* understand, it seems fixable | 21:14 |
| ignas | in a day or two | 21:15 |
| ignas | or three | 21:15 |
| ignas | ;) | 21:15 |
| ignas | though at the moment i am making it technically _correct_ not _good_ | 21:16 |
| th1a | The difference is? | 21:16 |
| ignas | if users will ask - "Why are events jumping when DST kicks in?" we will be able to answer - "Works as designed", not "It's a bug that was reported 5 times already, sorry." | 21:17 |
| ignas | ;) | 21:17 |
| ignas | there is a difference between a code that was intentionally written to behave in an acceptable manner, and the code that has small unintentional bugs left in there because of ignorance | 21:18 |
| th1a | I see. | 21:22 |
| *** alga has quit IRC | 21:26 | |
| *** ignas has quit IRC | 21:53 | |
| *** auxesis has quit IRC | 22:33 | |
| *** auxesis has joined #schooltool | 22:39 | |
| *** mgedmin has quit IRC | 23:08 | |
| *** wrobel` has quit IRC | 23:10 | |
| *** jinty has joined #schooltool | 23:10 | |
| *** Aiste has joined #schooltool | 23:38 | |
| *** didymo has joined #schooltool | 23:58 | |
Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!