IRC log of #schooltool for Monday, 2005-10-24

*** ignas has quit IRC00:43
*** ignas has joined #schooltool00:51
*** ignas has quit IRC01:18
*** ignas has joined #schooltool01:29
*** strichter has joined #schooltool02:13
*** srichter has quit IRC02:21
povbot/svn/commits: * ignas committed revision 5310:02:28
povbot/svn/commits: Fixed the test for AtomCalendarView. The test was taking UTC monday, yet the code was using datetime.date.today() which yelded non UTC today (the bug only surfaces atound 1-2am on Sunday in Europe/Vilnius). I am not 100% sure that we want the local current week not the UTC one in the code.02:28
ignasoops :) it's Monday 1-2 am02:28
*** didymo has joined #schooltool02:33
*** strichter has quit IRC02:50
*** ignas has quit IRC02:54
*** tiredbones has left #schooltool03:31
*** srichter has joined #schooltool03:43
*** tvon has joined #schooltool03:46
*** didymo has quit IRC07:03
*** didymo has joined #schooltool07:04
*** Aiste has joined #schooltool09:07
*** Mattias has quit IRC09:11
*** SteveA has quit IRC09:33
*** auxesis has quit IRC10:40
*** auxesis has joined #schooltool10:42
*** ignas has joined #schooltool11:52
*** mgedmin has joined #schooltool11:54
*** thisfred has joined #schooltool13:14
*** ignas has quit IRC13:27
*** didymo has quit IRC14:20
*** ignas has joined #schooltool14:24
*** alga has joined #SchoolTool14:25
*** th1a has quit IRC14:25
*** bskahan has joined #schooltool14:54
*** tiredbones has joined #schooltool15:10
*** dwelsh has joined #schooltool15:51
*** dwelsh has left #schooltool15:52
*** dwelsh has joined #schooltool15:53
*** dwelsh has left #schooltool15:53
*** th1a has joined #schooltool16:09
*** vidasp has joined #schooltool16:11
*** vidasp has joined #schooltool16:12
srichtergood morning (I am here ;-)16:20
*** erchache has joined #schooltool16:23
erchachehi16:23
erchachethis night i travel to merida for a conference16:23
erchacheanybody knows Jeff Waugh of Ubuntu project?16:24
th1aerchache:  I met him a couple weeks ago.16:27
erchachehe will going to participate on a dissertation of open source development projects16:28
erchacheth1a: for schoolbell reasons?16:29
erchachethe meet16:29
th1aSocial, essentially.16:30
erchacheahhh is he your friend?16:30
th1aWell, I just met him the one time.  Seems like a nice guy.16:30
erchacheahhh ok16:30
th1aNow it is time to start the weekly meeting.16:31
* th1a gavels in the meeting.16:31
th1aEveryone awake?16:31
srichterI am here16:32
th1abskahan?16:32
erchachewell...im here too...16:33
erchache;)16:33
th1aSince POV isn't working on SchoolTool right now I don't expect all of your undivided attention.16:35
* mgedmin is here16:35
th1aThat's better.16:35
algamost of are :-)16:35
ignasNobody expected the Spanish inquisition!16:35
algamost of us are :-)16:35
erchacheim going to promote schoolbell on this conference....16:35
bskahanhi16:36
th1aAh.16:36
th1aGood.16:36
erchachelinex and guadalinex are using on schools and universities16:36
th1aOK, I want to thank ignas for coming up with the list of outstanding issues.16:36
erchacheand we are making of our guadalinex distribution for university of sevilla on Spain16:36
th1aIt has been helpful.16:36
erchachei think we can integrate schoolbell/schooltool on our repositorie16:37
ignasthank albert for translating it into human language :)16:37
th1aAnything else POV wants to report?16:38
erchachePOV? what is this?16:39
mgedminits us -- www.pov.lt16:39
erchacheahhhh ok :-P16:39
algath1a: we're waiting for stories for the next contract16:40
th1aRight.16:40
erchacheok...16:40
th1aWorking on it.16:40
th1aOK.  bskahan:  what's up?16:40
bskahanclosed the 2 most reported TZ bugs16:41
bskahanreaminging TZ bugs, ical import/export issues and recurrence rules16:41
bskahanthen back to viewletifying calendar view in trunk16:42
bskahanGus just emailed me today about the forums.st.o url so our forums should be up after a DNS roll16:43
th1aWhat's the status of fhs.schooltool.org?16:43
bskahanyou can start using it tomorrow, I'll give you the login and pass via email16:43
th1aOK.16:43
th1aThat needs to be a trunk version.16:44
bskahan?16:44
bskahaneek16:44
bskahanok16:44
th1aI will deal with the consequences.16:44
bskahanI'll switch it16:44
th1aSorry I must have forgotten to tell you.16:44
bskahanwe talked about switching it after the first snapshot release16:45
th1aWell, the timetable model was updated in the trunk to handle FHS's schedule.16:45
bskahanah16:45
th1aBut we didn't backport it.16:45
bskahanok16:45
th1aOK.  srichter?16:47
th1aStephan and I discussed rearranging the order of some of his work on Friday.16:47
srichterright, so I updated you already on Friday ;-)16:48
th1aWell, it may be of interest to everyone else.16:48
srichterI just finish addressing the issues raised by Marius' comments16:48
srichterthen I finish doc phase I and do phase II16:49
srichterI'll push back the REST fixes till later16:49
srichtersince the doc stuff is time sensitive16:49
srichter<done>16:49
th1aWell, specifically, this means writing a tutorial for SchoolTool that includes using viewlets16:50
th1aso that POV has some guidance when they start attendance.16:50
th1aAnd doing that prior to viewletizing the current views.16:50
srichterright, I will do viewletization as much is necessary for the documentation though16:51
erchachesend me a mail when finish first preview of that......16:51
erchachei can do documentation.....16:52
srichterI will also switch to the new viewlet implementation this week16:52
th1asrichter and bskahan need to coordinate on the shared task they have in the viewlet conversion.16:52
srichterright16:52
th1aSo...16:53
srichterbtw, ZC, Projekt01 and us now agreed on the content provider API, so that it can make it into the core for 3.216:53
srichterth1a: I think the overlap is small right now and not acute until I am done with the docs16:53
bskahansrichter, I'm going to have to hit you up with some questions tomorrow when I start working on it agin16:54
th1aOK.16:54
srichtersure, so I am not available all day tomorrow, since I am in school16:54
bskahanok16:54
th1aAs long as you two work it out among yourselves.16:54
th1aI'm going to be traveling the next three days, unless I get socked in by a hurricane.16:55
th1aI'm going to an ed-tech conference in Colorado.  They gave me a booth in this exhibition, which is nice, except it is a lot harder to come up with stuff to put in a booth than do a powerpoint.16:56
th1aSo my sister and I are whipping up some posters.16:56
th1aIt is a bit stressful.16:56
th1aHm... no jinty today.16:57
ignashe told he'll be away on sunday16:58
th1aignas: right.  Thanks.16:58
th1aI guess everything seems in order for the release.16:58
th1aI got the RPM's already.16:59
th1aI sent Gintas a plea to make more Windows .exe's.16:59
th1aI don't think I'm going to have Mac packages...16:59
th1asome work is still needed there.16:59
erchachedaemon for windows are made?17:00
th1aWe're getting closer though.17:00
th1aWe've got a services script for Windows which might work.17:00
erchacheok....17:00
erchacheincludes on tarballs?17:00
th1aNot at this point.17:01
th1aIt is on the website.17:01
th1aMoving on, the response to the testing guidelines was less violent than I expected.17:02
bskahan"As much as possible, browser view classes should not perform complex computations.  These should be handled by the Python API code.  "17:03
th1aI should be able to make some minor changes to satisfy some of mgedmin's concerns.17:03
*** dwelsh has joined #schooltool17:04
algamy rule of thumb: if there are if statements in the code, there needs to be a test17:04
*** dwelsh has quit IRC17:04
algaif there is a loop longer than a oneliner list comprehension, it also needs a test17:04
th1aI think the question is, when is it sufficient to cover a view class with a functional test.17:05
th1aEverything is still tested.17:05
algaha!17:05
bskahanI'm not sure I agree with that part of the guideline, there's reasonably complex logix in view code that doesn't make sense to put anywhere else because its browser specific17:05
algawhat about edge cases?17:05
algawhat happens when we press this, this and this button?17:06
algawhat if there is no data?17:06
algafunctional tests rarely go into such details17:06
algawhat happens if there are these error conditions?17:06
th1abskahan:  That's why it says "as much as possible."17:07
th1aSometimes you need more complex logic in view classes, and then you need to test those classes.17:07
srichteralga: how likely are you to do a mistake in the test setup that is soemtimes in excess of 100 lines?17:08
ignassrichter, mistakes in test setup surface easily17:08
srichterthe point is that with a functional test you can cover all the edge cases that can occur17:08
algayep17:08
ignasif you do a mistake in 100 line test setup - your test doesn't work17:08
srichterignas: no, not necessarily17:08
srichterthat is not true17:09
ignasnot absolutely true17:09
ignasbut i haven't seen that many subtle test failures because of some small mistake in the test setup17:09
th1aOK, there are a couple questions here.17:09
srichterignas: I have spent hours getting the right test setup going to make a trivial test pass, which is a sign that the test setup is much more complex than the benefit of the test will ever be17:10
srichteralso, coming back to the browser functional test edge case17:10
ignassrichter, if your test setup is lower than some margin - you can't trust any of the tests anymore :17:11
srichteraccessing those views via Python more or less tests a use case that can never occur; i.e. use the Python API of a view17:11
ignass/setup/coverage/17:11
srichterexactely17:11
srichterI do not trust many ST tests for this reason17:12
srichternote that Zope 3 itself has  a bunch of bad tests too, but this is mainly historical of course17:12
ignasdid you try looking at the real coverage of our unit tests17:12
ignas?17:12
srichteryeah, and they are mostly integration tests17:13
ignasno17:13
ignasi mean --coverage17:13
srichteroh I see, no I have not looked at this17:13
srichteris that an option on the test runner?17:13
ignasnot sure it still works17:13
ignassome time ago it worked17:13
ignastakes a looong time though :/17:14
ignasmost of the untested code of ours are in places that throw exceptions ...17:14
ignascorner cases ...17:15
th1aOne thing that would be helpful to wrap this up is a better definition or some examples of a "simple" view class.17:16
th1aAny suggestions?17:17
srichterfor me a simple view class is one that only collects data from other unit tested Python APIs and prepares them in a PT-friendly format17:17
algaI offered mine17:18
mgedminI agree with srichter's definition -- assuming there are no complicated conditions/aggregations17:18
bskahansome of the container views fit alga's definition, iirc17:18
algaif your methods are mere list comprehension one liners, you can rely on ftests17:18
th1aalga: I misread yours earlier.17:18
bskahanbut combine that with marius' email comment about a minimum render test and you probably haven't changed from the current policy17:19
mgedminminimum render test?  what's that?17:19
srichterthere are also trivial try/excepts that are easily covered in a ftest17:19
* mgedmin is reading http://c2.com/cgi/wiki?CanFunctionalTestsReplaceUnitTests now17:19
bskahanmgedmin, you mentioned in an email that view's should have a minimum test (as a placeholder for when they get complicated later, if nothing else)17:20
mgedminyes, but I didn't say anything about rendering!17:21
th1aI don't think I agree with requiring placeholder tests, since setup seems to be timeconsuming.17:21
mgedminrendering views in unit tests requires complicated setup that srichter objects to17:21
bskahanmgedmin, /me nods17:21
bskahanok17:21
mgedminand I sometimes consider to be a mistake17:21
bskahanmisread17:21
srichterright17:21
srichterthis is exactely what I am saying17:22
mgedminan empty test module that just imports a view class requires no expensive setup17:22
srichterif we want to do unit tests right, we have to write many more stubs17:22
srichterunfortunately, I do not think this is practical17:22
mgedminit requires the tedious setup -- creating a test_foo.py module, putting the boilerplate inside, svn add, etc.17:22
*** Aiste has quit IRC17:22
mgedminand I want to invest in that tedious but simple and inexpensive setup earlier17:22
th1aSo by alga's standard, in http://source.schooltool.org/svn/trunk/schooltool/src/schooltool/level/browser/level.py17:23
* mgedmin reads http://c2.com/cgi/wiki?DoBothUnitAndAcceptanceTests17:23
th1aLevelContainerView doesn't need a unit test.17:24
th1aLevelValidationView does?17:24
th1aIEditLevelSchema doesn't.17:24
th1aEditLevelAdapter does by anyone's definition.17:24
th1aLevelAddView does17:25
th1aand LevelEditView doesn't.17:25
* mgedmin nods17:25
th1aIt seems to me that LevelAddView can be tested perfectly well with just a functional test.17:26
th1aOr tests.17:26
mgedminthats.... true, technically17:26
algawhat happens when they grow?17:26
* mgedmin coughs17:26
* mgedmin notices that he omitted an apostrophe in his last uttering, and is flabbergasted17:27
th1aIf they grow, you have to write unit tests.17:27
* mgedmin wonders why the lack of capital letters doesn't annoy him17:27
th1aGives you some incentive to keep the view classes simple.17:28
srichteralga: then you can decide to: (a) put the complicated code into another code segment and unit test that, or (b) write a unit test for the view17:28
th1aWell, I think we're within a range of opinion that we can all coexist within.17:29
th1aCertainly I'm not trying to force POV to test less.17:29
th1aAnd with that, our time is up.17:30
algaAre we still formally responsible for the quality of the codebase?17:30
th1aI still intend to write time for code reviews into your contracts.17:31
ignaswhat do you expect as a result of the "code reviews" ?17:33
th1aPretty much what you've done before.17:33
th1aPoint out bugs, problems, missing tests, style issues.17:33
ignasbut well - what is the point of code review if well - no one listens to our comments ?17:34
mgedminignas, people do listen17:34
mgedminpeople just don't agree with everything17:34
srichterright17:35
* mgedmin mutters "when I will be an evil dictator, I'll remember this"17:35
th1aI certainly don't think you should feel like I'm making a blanket rejection of your point of view.17:35
* th1a bangs the virtual gavel.17:37
* mgedmin nods17:37
*** alga has quit IRC18:14
erchachecan make a zope question?18:37
bskahanth1a, what program did you use to do the flash demo for ST?18:39
erchachei hate flash....18:39
th1aI didn't do one with flash, but if I did, I would use vnc2swf.18:40
erchachehow i can connect a ms-sql server to zope on the most easy and fast way?18:41
th1aI don't know.18:43
srichteralso look at wink18:43
srichter"wink" is a very cool screen capture program to make nice demos18:43
bskahanthanks18:45
povbot/svn/commits: * srichter committed revision 5311:19:06
povbot/svn/commits: Implemented several suggestions from Marius:19:06
povbot/svn/commits: - Wrote unit tests as required by Tom. Note that this was prior to19:06
povbot/svn/commits: the testing requirements discussion.19:06
povbot/svn/commits: - Completed IManagerWorkItems interface.19:06
povbot/svn/commits: - Work items are now a special dictionary and not a list anymore.19:07
povbot/svn/commits: - Ids for work items are now generated within the custom dictionary, thus19:07
povbot/svn/commits: making it much more efficient.19:07
*** dwelsh has joined #schooltool19:27
dwelshth1a:  you there?19:27
*** erchache has quit IRC19:42
*** alga has joined #SchoolTool19:42
*** srichter has quit IRC19:56
*** JohnnyST has joined #schooltool20:09
*** vidasp has quit IRC20:23
*** srichter has joined #schooltool20:24
*** thisfred has quit IRC20:32
th1adwelsh:  I21:06
th1aam here.21:06
*** mgedmin has quit IRC21:14
*** bskahan has quit IRC21:36
*** ignas has quit IRC21:46
*** alga has quit IRC21:47
*** pcardune has quit IRC21:51
povbot/svn/commits: * srichter committed revision 5312:22:15
povbot/svn/commits: Added missing subscriber unit tests.22:15
dwelshth1a:  can you hear me now?23:04
dwelshth1a:  yo tom.23:40
*** ignas has joined #schooltool23:44

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