IRC log of #schooltool for Thursday, 2005-07-21

*** bskahan_ has quit IRC00:09
*** bskahan has quit IRC00:14
*** bskahan has joined #schooltool00:17
*** gml_ is now known as gml00:27
*** SteveA has quit IRC02:32
povbot/svn/commits: * tvon committed revision 4256:02:51
povbot/svn/commits: Base setup for site-wide calendar.02:51
povbot/svn/commits: * tvon committed revision 4257:02:52
povbot/svn/commits: Base setup for site-wide calendar.02:52
th1aIs Zope3 svn wonky?02:53
th1asrichter:  Is there something wrong with Zope3 svn?03:01
srichternot that I know of03:04
srichterTim changed the ZODB versions, so you probably have to recompile03:04
th1aMy connection keeps closing unexpectedly.03:04
tvontis workin here03:05
th1aI meant something wrong with subversion, not the code.03:05
tvoner, no it's not03:05
tvonjust got that on Zope3/src/ZODB03:05
srichterthat would be a good mail to send to zope3-dev and see whether anyone else experiences this03:05
srichteroh, I see03:05
th1aI'm getting it at random points.03:05
srichterhold on, I check03:05
srichtermmh, works for me03:06
tvonI updated again and it worked03:10
th1aI'm trying to do a fresh checkout, and it is just too much I think.03:11
th1aI guess if I do a svn up from the incomplete checkout it gets everything.03:20
tvonyeah, thats working for me03:23
povbot/svn/commits: * tvon committed revision 4258:03:26
povbot/svn/commits: Generation script for site-wide calendar03:26
tvoner, crap03:45
* tvon shrugs03:45
tvonah, nm03:49
povbot/svn/commits: * tvon committed revision 4259:03:58
povbot/svn/commits: Generation for application calendar/timetable03:58
bskahanthey're extending daylight savings time in the US ...04:00
bskahangoing into effect this fall04:01
* tvon cant wait to ditch externals04:10
*** bskahan has quit IRC04:24
povbot/svn/commits: * tvon committed revision 4260:04:51
povbot/svn/commits: app title in page title/footer.04:51
*** gml_ has joined #schooltool06:26
*** gml has quit IRC06:39
*** marcin has joined #schooltool09:35
marcinhi all09:35
marcinanybody here?09:35
th1aI was just getting ready to go to bed...09:36
th1aWhat's up?09:36
marcinth1a, hi09:36
marcinth1a, I got two questions09:36
th1aGo ahead.09:36
marcinth1a, 1. is there something like a sample data for schooltool?09:36
th1aAh... there was, but there isn't currently.  That'll be back soon.09:37
th1aWe went through some major changes, and it is hard to keep all the little details together.09:38
marcinth1a, ok then my second question is ekkhhhm about money09:39
marcinth1a, I've seen bounties on schooltool page (like this
marcinth1a, are there still available?09:39
th1aAh, right.09:39
th1aWell... not bounties as such.09:40
th1aAre you a Python/Zope developer.09:40
marcinwell currently I'm Java developer but I want to switch09:41
marcinand I learn Python now09:41
th1aRight.  We have two small teams of experienced Python developers, and we aren't actively looking for more.09:42
marcinunfrtunately my project for google summer of code was rejected so I'm looking for some other opportunities to hack some code09:42
marcin(and to get some bounty for this maybe ;) )09:42
th1aIf we were, the top priority would be Python/Zope3 development experience.09:42
th1aSo... the short answer is no, we don't really have any active bounties.09:43
marcinwell that's sad09:43
marcin(and a little suggestion - maybe some bounty for uptading your website and removing these old bounty informations ;) ?)09:44
marcin(just a joke)09:44
th1aWell, I've been planning on that.  We're kind of stuck in an endless loop waiting for a server upgrade, at which point I'm going to cull out a lot of the current content.09:45
tvonon the upside, by the time we get a server plone 2.1 will be out09:46
th1aAlso, the bounties were Mark's idea, and since he pays for this, I have to be careful about yanking his stuff.09:46
th1atvon:  It is looking that way.  I think debconf and the LaunchPad launches pushed us back a couple more weeks.09:46
th1aLaunchPad is looking better though, at least.09:47
tvonI'm a bit excited to see that released09:47
th1aI want to move to bazaar.  Sort of.09:47
tvonyeah, I have the same feelings09:48
th1aDid you watch the video of Mark at Debconf?09:48
tvonmhm, no, it's sitting on my desktop09:48
th1aIt is a must-see.09:48
tvonhow long is it?09:48
th1aApparently, he's even richer than I thought.09:48
tvoncould be a nice companion to the exercise bike09:48
tvonheh, oh?09:48
th1aAn hour.09:48
th1aHe talks quite frankly about his life and plans.09:49
th1aThe objective is to reassure the Debian geeks, so he kind of has to lay it all out.09:49
th1aSo it is quite informative.09:49
th1aOK, I'm really going to bed now.09:50
* tvon thinks Mark should do all his speaking in a space suit09:51
marcintvon, so this guy is your sponsor?10:10
*** skaith has joined #schooltool10:11
skaithhello all10:12
skaithhello th1a10:12
skaithth1a : some time to discuss 1à minutes ?10:12
povbot/svn/commits: * tvon committed revision 4261:10:37
povbot/svn/commits: fixing left offset in IE10:37
povbot/svn/commits: * tvon committed revision 4262:11:08
povbot/svn/commits: Removing the footer-at-bottom CSS hack.  This is currently required to avoid a bug in firefox 1.0.x that causes odd stacking to occur when mixing an absolute positioned body (as the hack requires), an absolute positioned element (as the calendar requires) and form fields under the element (our form fields).  The issue appears to be fixed in Deer Park but that doesnt do us much good.11:08
povbot/svn/commits: FWIW, the footer never worked in KHTML anyways (at least not Konq).11:08
*** thisfred has joined #schooltool12:04
povbot/svn/commits: * gintas committed revision 4263:12:38
povbot/svn/commits: Added error checking to SlotEntryStep.12:38
povbot/svn/commits: * gintas committed revision 4264:12:44
povbot/svn/commits: getSessionData takes a view as an argument again.12:44
*** skaith has left #schooltool12:54
povbot/svn/commits: * gintas committed revision 4265:13:10
povbot/svn/commits: Unwedged functional tests, added a branch to test named periods.13:10
povbot/svn/commits: * gintas committed revision 4266:13:34
povbot/svn/commits: Implemented the step that asks if start/end times vary based on day of week or day in cycle.13:34
*** gml_ is now known as gml13:38
*** srichter has quit IRC13:46
*** srichter has joined #schooltool13:51
povbot/svn/commits: * gintas committed revision 4267:14:01
povbot/svn/commits: Unbroke functional test.14:01
povbot/svn/commits: * gintas committed revision 4268:14:21
povbot/svn/commits: Separated RotatingSlotEntryStep and WeeklySlotEntryStep.14:21
*** jinty has joined #schooltool14:23
povbot/svn/commits: * gintas committed revision 4269:14:23
povbot/svn/commits: Removed CycleStep.update which is not really useful any more now that we have WeeklySlotEntryStep.14:23
jintyhoi srichter, re zpkgtools for schooltool14:29
* jinty 's battery dies14:31
povbot/svn/commits: * gintas committed revision 4270:14:34
povbot/svn/commits: Cosmetic fix.14:34
povbot/svn/commits: * gintas committed revision 4271:14:36
povbot/svn/commits: Reverted revision 4269.  It appears that CycleStep.update will come in handy for Albertas.14:36
*** mgedmin has joined #schooltool14:40
povbot/svn/commits: * alga committed revision 4272:14:50
povbot/svn/commits: "Put periods in order if all days same" step.14:50
srichterjinty: I am here14:51
*** jinty has quit IRC14:54
*** SteveA has joined #schooltool14:54
povbot/svn/commits: * gintas committed revision 4273:14:55
povbot/svn/commits: Renamed time_slots{,1,2} to time_slots{_simple,_rotating,_weekly}.  Implemented requiredPeriods properly.14:56
*** matiasV has joined #schooltool14:56
*** eldar has joined #schooltool14:57
*** alga has joined #SchoolTool15:10
povbot/svn/commits: * alga committed revision 4274:15:10
povbot/svn/commits: Made the periods in order view remember the selection when redisplayed due to errors.15:10
*** matiasV has quit IRC15:18
*** matiasV has joined #schooltool15:18
*** gintas has joined #schooltool15:41
povbot/svn/commits: * gintas committed revision 4275:15:42
povbot/svn/commits: Merged time_slots_simple, time_slots_weekly and time_slots_rotating to time_slots, which is now a list of day descriptions(sequences of periods).15:42
povbot/svn/commits: * mg committed revision 4276:16:00
povbot/svn/commits: Cosmetic fixes.16:00
srichterth1a: is elementary school in the US from 1-6 and middle school from 7-8?16:07
povbot/svn/commits: * mg committed revision 4277:16:09
povbot/svn/commits: DRY - Don't Repeat Yourself.16:09
povbot/svn/commits: * mg committed revision 4278:16:10
povbot/svn/commits: Another small DRY refactoring.16:10
povbot/svn/commits: * mg committed revision 4279:16:22
povbot/svn/commits: Refer to steps by their names rather than numbers.  (Besides, the 1st step has nothing to do with weekly/rotating cycles -- the 1st step is where you enter the title).16:22
*** bskahan has joined #schooltool16:29
povbot/svn/commits: * gintas committed revision 4280:16:32
povbot/svn/commits: Split up createSchema, updated it to cope with new semantics (day_names is now always in session).16:32
povbot/svn/commits: * alga committed revision 4281:16:37
povbot/svn/commits: Reorder periods if days are different16:38
povbot/svn/commits: * gintas committed revision 4282:16:42
povbot/svn/commits: Fixed unit test failure.16:42
*** bskahan has quit IRC16:50
povbot/svn/commits: * alga committed revision 4283:16:54
povbot/svn/commits: Hooked up the last question!  Hurray!16:54
povbot/svn/commits: Now all we have to do is to make everything work.16:54
povbot/svn/commits: * mg committed revision 4284:17:00
povbot/svn/commits: Finished my incomplete refactoring in rev 4278.17:00
*** bskahan has joined #schooltool17:03
algabskahan: take a look at the title of http://localhost:7080/terms or whatever :-)17:07
povbot/svn/commits: * mg committed revision 4285:17:08
povbot/svn/commits: Refactoring: untangle FinalStep.dayTemplates a bit.17:08
bskahanalga: thanks17:13
povbot/svn/commits: * gintas committed revision 4286:17:18
povbot/svn/commits: Weekday are now translated properly in CycleStep.update.17:18
povbot/svn/commits: * mg committed revision 4287:17:20
povbot/svn/commits: A dict lookup is more Pythonic, besides it raises an exception rather than returning None when the cycle is something that it should not be.17:20
bskahanth1a: ping17:26
th1aGood morning.17:37
bskahanhey tom17:38
th1aHi Brian.17:38
bskahando you have a written story for overlaying student sections?  tvon isn't around and I can't find a copy17:41
th1aOh, perhaps that didn't get written.  We exchanged some email.  Let me see.17:41
* bskahan is finally getting over jetlag17:42
bskahanpydev plugin for eclipse looks like it improved alot in the last 6 months17:43
povbot/svn/commits: * alga committed revision 4288:17:46
povbot/svn/commits: Fixed the error handling in the simple order selector.17:46
th1abskahan:  Forwarded the updated stories.17:47
bskahanth1a: thanks17:50
th1aIt isn't anything complex.17:51
bskahan"In fact, the17:51
bskahansections should probably always show up as options on the main menu view."17:51
bskahanwhat does that refer to?17:52
bskahanoh, the menu doesn't have "Sections"17:52
bskahanthe reason I did that was because the "Courses" menu item also displays sections17:53
bskahan(grouped by Course)17:53
povbot/svn/commits: * alga committed revision 4289:17:59
povbot/svn/commits: Fixed ftest to handle the last question.17:59
povbot/svn/commits: * mg committed revision 4290:18:00
povbot/svn/commits: Refactored FinalStep.createSchema helper methods to have unit tests and accepts arguments explicitly rather that digging in the session.18:00
povbot/svn/commits: * mg committed revision 4291:18:10
povbot/svn/commits: Updated FinalStep.modelFactory to handle a third case (rotating cycle with time slots based on day in cycle rather than day of week).18:10
*** th1a has quit IRC18:13
*** skaith has joined #schooltool18:23
*** pcardune has joined #schooltool18:23
pcarduneeldar, i'm here18:23
skaithth1A online ?18:24
skaithno ok see you all18:24
skaithmgedmin ?18:24
*** skaith has quit IRC18:28
povbot/svn/commits: * mg committed revision 4292:18:35
povbot/svn/commits: Made FinalStep.createSchema pay attention to period order.  Coincidentally this fixed the failing ftest.18:35
povbot/svn/commits: * gintas committed revision 4293:18:38
povbot/svn/commits: Implemented the restriction in WeeklySlotEntryStep: if a rotating timetable cycle is selected and slots vary based on day of week, each day must have the same number of slots.18:38
pcardunewill there be any addition to the relationship api to allow for one to many relationships?18:40
povbot/svn/commits: * gintas committed revision 4294:18:41
povbot/svn/commits: Cosmetic fix.18:41
srichtermgedmin: how can I run tests of a particular pacakge with the schooltool test runner. the --dir or -s option does not exist18:41
pcarduneand then the path to the package18:42
gintas-s does exist18:42
srichterok, it is not documented18:43
srichtersearch-in is18:43
srichterok, got it18:44
mgedminpcardune, I have no plans to add such an api18:44
*** th1a has joined #schooltool18:45
srichteryou guys should port the coloring stuff to Zope 3; it is nice ;-)18:46
srichterof the test output18:47
pcardunethere's coloring?  what option is that?18:47
pcardunemgedmin, very well then, i guess we will extend it ourselves then...18:48
mgedminI was thinking about that too18:51
gintasuse -C on terminals with light background18:51
srichterit really makes the output more readable; it would be a great improvement18:52
povbot/svn/commits: * gintas committed revision 4295:18:58
povbot/svn/commits: Added a file type hint for vim.18:58
*** alga has quit IRC19:01
povbot/svn/commits: * mg committed revision 4296:19:08
povbot/svn/commits: Updated FinalStep.createSchema to deal with different period orders on different days, at least with the weekly cycle.19:08
th1amgedmin:  Have you thought about how that will need to be represented in XML?19:09
*** bska|mobile has joined #schooltool19:09
th1aThose cycles?19:09
mgedminwhat cycles?19:10
th1aSchedules with different sequences of periods within the cycle.19:11
mgedminthose are all representable with our current xml, since yesterday19:11
th1aYou mean, you changed the XML yesterday.19:12
mgedminalga did19:12
th1aOK.  That's what I wanted to know.19:12
mgedminth1a, I've a question for you19:13
th1aI should send some new examples to Andy@ReportLab then.19:13
mgedminone of the three penultimate steps in the timetable workflow has a comment19:13
mgedmin"the layout of this form should be[sic] reflect the schema created on the slot/time form."19:14
mgedminwhat if the user said that classes begin and end at the same time each day?19:14
mgedminthen the slot/time form was a single list of start/end times19:14
mgedminthere's no grid19:14
mgedmincan the sequence of periods each day be different in that case?19:15
mgedminand if so, will it depend on the day in cycle or on the day of week?19:15
th1aLet me summon my diagram.19:15
*** bskahan has quit IRC19:15
* mgedmin nods19:15
th1aThat screen will either need the custom grid created on the slot/time form, or it will need to generate a simple one from the list of start/end times.19:18
mgedminwas that your answer?19:18
th1aI hope so.19:18
mgedminok, then19:19
mgedminimagine this:19:19
mgedminI choose "rotating cycle", enter 3 day names19:19
mgedminsay that classed begin and end at the same time each day19:19
mgedminenter 4 slots19:19
mgedminsay that periods are named, and enter 6 period names19:19
mgedminthen I say the sequence of periods each day is different19:19
mgedminnow I have to specify the order in the grid19:20
mgedminmy question: should the grid have 3 columns (for 3 days in the cycle), or 5 columns (for week days)19:20
mgedminif I had chosen to say that classes begin on different times on different days19:20
mgedminI would have given an extra choice between the two grids19:21
mgedmineven if I then would have entered the same times for each day19:21
* mgedmin pauses19:22
mgedminperhaps I'm looking too deep19:22
mgedmindoes the weekly grid even make sense here?19:22
th1aI think the key here is that you have to say whether or not the schedule varies b the day of the week or day in the cycle.19:22
mgedminI think it doesn't19:22
th1aThat's in the diagram.19:22
mgedminyes, but that step is skipped if I say classes begin at the same time each day19:23
mgedminand, actually, if I say that19:23
mgedminand then if I say that period order varies by the day of the week19:23
mgedminthen I do not have a rotating cycle at all19:23
mgedminboth times and period names depend only on the day of week19:24
mgedminand the day name doesn't indicate anything19:24
mgedminso maybe I can just assume that in this specific case (rotating cycle, same slots, different period order) the grid will have days in cycle, and not days of week19:24
th1aAs my friend Chris Lehmann wrote yesterday:  Scheduling is pedagogy. In fact, it is one of the most powerful pedagogical tools we have at our disposal. How we structure our school day... what choices we make in the ways in which our students move through their day... affects everything else we do.19:25
mgedminthat's solved then19:26
th1a;-)  Good.19:26
mgedminuh oh19:26
mgedminI think I see another problem19:26
mgedminremember that weird example?  4 days in cycle, periods go A,B,C,D in the first day, B,C,D,E in the second one, etc.19:26
mgedminbut slots depend on the time of day19:26
mgedminI do not think you can represent this example with your diagram19:27
mgedminsince the slot grid will be based on weekdays19:27
mgedminbut period order should depend on days in cycle19:27
* mgedmin thinks of a possible solution19:28
mgedminif we changed the last question to19:28
th1aThere is a note that says "for a rotating cycle that varies by day of the week, each day must have the same number of slots."19:28
th1aI think that covers the problem.19:29
mgedminthat covers a different problem19:29
th1aIf you can enforce that restriction.19:29
th1aWhich form is the problem here?  "Put the periods in order for each day:"?19:30
th1aThat form doesn't have to indicate the start and end times of the slot.19:30
th1aSo if that varies by day of the week, it doesn't matter.19:30
mgedminthe problem is when I want the period time to depend on the day of week, but the period order to depend on the day in the cycle19:30
srichtermmh, schooltool's test runner cannot count ;-)19:31
mgedminsrichter, can't it?19:31
mgedminthe zope 3 one doesn't19:31
srichterJim fixed the doctest counting so that every code snippet in a doctest is a test19:31
srichter(its more rewarding ;-)19:31
mgedminI fixed that in a different way19:31
srichterhow so?19:32
mgedminone doctest is 1 test19:32
mgedminth1a, do you understand what I'm driving at?19:32
srichtermmh, ok19:32
th1aI think so, but I don't think it is a problem.19:33
mgedminif you say so19:34
mgedminit's just that this schedule was one of the examples that we based our initial design on19:34
th1aOh, we can do that schedule.19:34
mgedminif it's ok when it's not representable, then that's fine19:34
* mgedmin pauses19:34
mgedmincan you give me a walkthrough?19:35
th1aAre we worrying about the forms or the underlying data model?19:35
mgedminthe forms19:35
mgedminthe model can represent that example19:35
mgedmindo you remember the example?19:36
mgedminall lessons shortened to 30 minutes on Wednesdays because of baseball19:36
mgedminor something like that19:36
th1aThat's the example from our school, so I know it well.19:36
povbot/svn/commits: * gintas committed revision 4297:19:36
povbot/svn/commits: Cosmetic fixes.19:36
th1aI'm saying the "Put the periods in order for each day" form19:36
th1aonly has to set the sequence each day.19:36
th1aIn the cycle.19:37
th1aIt doesn't need to know what the start and end times in each day of the week are.19:37
mgedminuh oh19:37
th1aNow do we have a problem with the data model?19:37
mgedminin a minute19:38
mgedminlet's establish that I should disregard the comment about the layout matching the slot/time form layout19:38
mgedminand always have a grid of days in the cycle19:38
mgedminI think there might be a problem with the data model...19:40
mgedminthe data model doesn't know about slots19:41
mgedminit knows about period names and times19:41
mgedminI'm confused19:42
th1aI had a feeling it was going to come to this.19:42
mgedminI'll chew on it and come back here later19:42
mgedminperhaps I remember the example incorrectly, because I'm sure it used to be representable19:42
th1aYes, I'm confused about that too, because I know we at least thought we could do it, but now I can't see how it is possible.19:43
*** erchache has joined #schooltool19:43
erchacheth1a: netboot runs on ubuntu?19:43
th1anetboot?  I have no idea.19:44
erchacheim triying to install the other server via netboot because doesnt have floppy/cdrom bay19:44
erchachei install dhcp3 an tftpd on my old laptop19:44
erchacheand runs19:44
erchachebut when are on mirror select to download packages doesnt run19:45
bska|mobileerchache: its possible to install ubuntu that way, that's how I did it19:45
bska|mobileI used the debian documentation to do it, iirc19:45
erchachei do too19:45
erchachebut when i select mirror fails19:46
erchacheask for proxy or similar19:46
bska|mobileprobably ask in #ubuntu19:46
erchachei think is a bug19:46
erchacheim doing now19:46
erchacheis a opteron and im using amd-6419:47
erchachebut at first kernel load good19:47
povbot/svn/commits: * mg committed revision 4298:19:47
povbot/svn/commits: Another test for different period order.19:47
*** SteveA has quit IRC19:52
*** gintas has quit IRC19:55
*** SteveA has joined #schooltool19:56
*** alga has joined #SchoolTool19:57
*** bska|mobile has quit IRC19:58
povbot/svn/commits: * alga committed revision 4299:20:06
povbot/svn/commits: Fixed the number of slots on the periods order (complex) form.20:06
algaalga to office, alga to office20:06
algaanything else to do?20:06
th1aYou might need to re-write the tiimetabling data model.20:07
mgedminalga, read the backlog20:08
mgedminalga, do you now enforce unique day and period names?20:13
povbot/svn/commits: * mg committed revision 4300:20:15
povbot/svn/commits: Updated FinalStep.createSchema to handle different time slots.20:15
mgedminthe answer is "no"20:15
pcarduneanyone, with functional tests, if the server returns that fun little page that says "A system error has occured" how can i see the error without running zope and looking through the web interface20:15
mgedmin>>> print http("""...""", handle_errors=False)20:16
mgedminI don't know why that isn't the default20:16
pcarduneoh, very very good20:16
*** erchache has quit IRC20:16
povbot/svn/commits: * mg committed revision 4301:20:17
povbot/svn/commits: Fixed ftest.20:17
th1aWe generally need to get SchoolTool to spew its errors more visibly.20:17
eldarthere was Falsh20:18
eldaroops, wrong window20:18
mgedminth1a, agreed20:19
pcarduneif arguments are passed through the url, like foo.html?bar=w00t... how do you access bar from the browser view class?20:21
eldarshouldn't they be in request?20:23
algath1a: presentation of errors is just a view in the skin.  easy20:25
mgedmineldar, they are20:26
pcardunemgedmin, as in like self.request['bar']?20:26
th1aI think everyone would be very happy if the traceback showed up on the web interface when you have an error.  It isn't perfect security practice, but I'll take easier debugging right now.20:26
pcarduneth1a, amen to that!20:27
mgedminI was very happy with the custom error view in schooltool 0.920:27
mgedminor 0.820:27
mgedminthe one that was twisted based20:27
algamgedmin: yep.  periods and slots are tied in our model20:28
*** thisfred has quit IRC20:28
mgedminalga, I finished createSchema20:29
mgedmin(I hope)20:29
mgedminwhat else is left?20:30
mgedmin- enforce unique day/period names20:30
mgedmin- functionally test at least 4 paths through the graph20:30
mgedmin- play with the view manually and try to break it20:30
algayep, pretty much20:31
alga  - check slots for sanity20:31
alga12:50-1:35 is an error :-)20:31
eldarumm does anyone know where the variable will be stored after being passing by url20:32
eldarlike ?eval=test20:32
eldarwhere do i get the eval?20:32
mgedminthe request20:32
mgedminself.request['eval'] also works20:33
eldarrequest['eval'] gives key error20:33
pcarduneok, so in a functional test, how do you pass stuff like that? "POST /blah.html?bar=foo20:33
mgedminof course not20:33
mgedminGET /blah.html?bar=foo20:33
pcarduneoh... i can't believe i didn't see that20:33
pcardunethat really is embarrassing20:34
mgedminif you use POST, you have to specify form values in the request body20:34
algaeldar: see posts on our ftests20:34
pcardune(that is what i get for copy/past)20:34
mgedminwhat happens if I have <form action="somehtml?bar=1" method="POST">?20:34
pcardunethat would create a problem presumably20:34
algacan you do that?20:34
mgedminI'd expect to see both bar and other form values...20:34
tvonTestRequest ?20:34
mgedminalthough that's bad style20:35
*** munkee has quit IRC20:38
pcarduneoff-topic: is schooltool compatible with lynx?20:39
th1amgedmin:  How much of a problem do we have with the data model?20:40
th1apcardune:  The intent is that SchoolTool is compatible with Lynx.  I haven't looked at our calendar layout under Lynx, however.20:40
mgedminthat depends on the requirements20:41
mgedminthe data model cannot vary period times and order independently20:42
mgedminthey are either both tied to day of week, or to the day in cycle20:42
mgedminso the yellow comment in the pdf was correct20:43
th1aSo we couldn't model the example we were discussing all along.20:43
th1aThat point has caused me much confusion over the past year.20:49
mgedminme too20:54
*** munkee has joined #schooltool20:55
th1aActually, maybe I've confused myself a bit...20:55
th1aSo the only case we can't do is a rotating cycle that doesn't match the days of thew week going over a weekly pattern of changing start/end times.20:57
th1aAnd even if we did that, it would be really hard to display it as a table, for example.20:57
th1aThe completed timetable.  You'd need two tables.20:58
th1aMaybe I've forgotten and we used a weekly rotation.20:59
th1aI think the case we can't do is extremely uncommon, since it is so difficult to even describe as a table.21:02
povbot/svn/commits: * alga committed revision 4302:21:03
povbot/svn/commits: Post events to schoolwide calendar when an emergency day is replaced.21:03
povbot/svn/commits: * mg committed revision 4303:21:08
povbot/svn/commits: Another walkthrough in the ftest: different period order in different days.21:08
*** bskahan has joined #schooltool21:17
th1aI'm going over to school to see if my memory is deceiving me.21:20
*** th1a is now known as th1a|school21:20
povbot/svn/commits: * alga committed revision 4304:21:24
povbot/svn/commits: Make sure period and day names are unique.21:24
pcardunehow do we get testing to stop on the first error?21:30
*** alga has quit IRC21:31
mgedmintry passing -1 to the test runner21:32
mgedminthen, when it doesn't work, find the place where you create your DocTestSuite21:32
mgedminand make sure its optionflags argument includes doctest.REPORT_ONLY_FIRST_FAILURE21:33
povbot/svn/commits: * mg committed revision 4305:21:33
povbot/svn/commits: Another walkthrough: different time slots.21:33
pcardunewhat if that doesn't work?  -l isn't a recognized command21:35
pcarduneoh wait... it's -121:36
*** th1a|school is now known as th1a21:41
th1aNope, our school uses the schedule we can't model :-(21:42
th1aI'm going to poke around the code and see if I can figure out some options.21:43
*** marcin has quit IRC21:46
povbot/svn/commits: * mg committed revision 4306:21:58
povbot/svn/commits: Illustrate the tests, otherwise it's impossible to keep them straight.21:59
srichterth1a: Steve just suggested having a Zope 3 bug day. I guess since ST benefits from the release, would it be possible that the ST team could participate in a bug day?22:04
th1aWell, I think your question is will we pay them to participate in a Zope 3 bug day.22:05
srichteryes ;-)22:05
povbot/svn/commits: * mg committed revision 4307:22:13
povbot/svn/commits: Another walkthrough.22:13
*** munkee has quit IRC22:32
povbot/svn/commits: * mg committed revision 4308:22:33
povbot/svn/commits: Sixth and final walkthrough.22:33
povbot/svn/commits: Now all the nodes and all the edges in the workflow graph have been traversed at least once in functional tests.22:33
mgedminth1a, it appears that we're done22:36
mgedminit's a first22:36
mgedminI don't remember any other proposals that we managed to finish before the deadline22:37
* mgedmin pauses22:37
mgedminotoh there may be bugs...22:38
th1aWell, we're done aside from the fact that we realized that one use case we thought was covered hasn't worked for a year and a half and isn't an easy fix, right?22:38
mgedminyes, there is that too22:38
th1aBut I should have nailed down that problem a long time ago, too.22:39
povbot/svn/commits: * mg committed revision 4309:22:46
povbot/svn/commits: Cleaned up outdated TODOs and XXXes.22:46
povbot/svn/commits: * mg committed revision 4310:22:49
povbot/svn/commits: Another obsolete TODO (implemented elsewhere)22:49
srichterth1a: btw, here is the final workflow I came up with:22:54
srichter                                           +----------+     +----------+22:54
srichter                            +------------> |   Pass   | --> | Graduate | --> E22:54
srichter                            |              +----------+     +----------+22:54
srichter                            V22:54
srichter        +--------+     +----------+        +----------+22:54
srichter  S --> | Enroll | --> | Complete | <----> |   Fail   |22:54
srichter        +--------+     +----------+        +----------+22:55
srichter                            |22:55
srichter                            |              +----------+22:55
srichter                            +------------> | Withdraw | --> E22:55
srichter                                           +----------+22:55
srichter(switch temporarly to monospace font)22:55
* mgedmin whispers, "why temporarily?"22:55
th1aThat's what I was just thinking...22:55
povbot/svn/commits: * mg committed revision 4311:22:55
povbot/svn/commits: Renamed the old "wizard" view name to complexadd.html.  Updated links and surrounding text.22:55
srichtermgedmin: I assume that most people do not use monospace font all the time; I don't! :-)22:56
srichterbtw, you can think of activities like Zope 3 events22:57
th1aIf you don't use monospace, you might as well be using Windows!22:57
th1aAnyhow... so "Complete" is sort of a neutral state.22:57
srichteryou then register workitems (like event listeners) that need to be completed before going to the next event22:57
srichter"Complete" is the state that needs manual interaction22:58
th1aEach block is called an "activity?"22:58
srichterso "complate" is a verb in the sense of "completing"22:58
th1aDo the transitions have a special name in this scheme?22:58
srichterit is the activity of the student completing the level22:59
srichteryou can give transitions an id, but it is not that useful22:59
srichtertransitions only have conditions22:59
srichterall the work is really being done in an activity22:59
mgedminth1a, could you look at the new timetable wizard?23:01
mgedminI'm not entirely sure the case with a rotating cycle that has slots defined by weekdays works as expected23:01
mgedminI'll try to look at it again tomorrow, when my brain works again23:02
th1aI think something like "pending" is better than "complete"23:02
th1amgedmin:  I'll take a look.23:02
*** bska|mobile has joined #schooltool23:03
srichterth1a: but pending is a state not an activity23:04
th1aOh... ok.  I have yet to make the conceptual shift here...23:05
srichterMAybe I should call it "Completing" or "Executing" or something like that23:05
th1aI'm thinking this works for a manual process, but you'd need another state to start using it dynamically.23:06
srichterno, you would simply make "Complete" a sub-workflow23:07
srichterthat contains the complexity that goes into it23:07
srichterwfmc allows for this23:07
srichter(our implementation might not, but it should be fairly straightforward to add)23:08
th1aMy brain wants a state machine.23:08
srichteryep :-)23:08
srichteronce you get it, it very obvious ;-)23:08
th1aSo I shouldn't be trying to picture "where" the student is between when they pass one grade and complete the next one?23:08
th1aThat's just not relevant?23:09
srichterthat's not relevant23:09
srichterwhen the manager says: Student X passed, the workflow will progress to the "Pass" activity23:09
srichterthe "Pass activity has several23:10
srichterthe "Pass" activity has several automatic work items:23:10
th1aOne other issue is that after you withdraw, you may re-enter the system, so it isn't necessarily an end state.23:10
srichter1. Make an entry in the history23:10
srichter2. Set the current level to the next one23:10
*** bskahan has quit IRC23:10
srichter3. If no next level is available forward to the graduate activity23:10
srichter4. set the student status "Enrolled" (if necesary)23:11
*** bskahan has joined #schooltool23:11
srichterif you do not want a student to be forwarded to the next level automatically:23:11
srichter5. Wait for user input okaying progress to next level23:11
srichter1-5 are work items that are completed during the pass activity23:12
srichternote that I can attach any amount of workitems to an activity23:13
*** bska|mobile has quit IRC23:14
th1aThat seems fine, with the possible exception of my last comment.23:15
srichterwell, noone said you cannot enroll again23:16
srichterbut withdraw usually means you are out of the scool23:16
srichterah, I just noticed:23:18
th1aI guess the way it works now is fine for the moment.23:18
srichterthe "Enroll" activity will also require a manual input.23:19
srichterit will let you set the initial level you are at23:19
th1aOnce we have a stronger sense of what "enrolling" means in the system we might change it a bit.23:19
srichterAgain, then "Enroll" just becomes a sub-process, which WfMC supports nicely23:19
th1aRight.  I'm getting that.23:21
srichterI think this is the true strength of WfMC23:22
srichterIt allows you to evolve a process without necessarily conflicting with what you have23:22
th1aI'm glad we're trying this route.23:23
th1aI think it is going to be really powerful.23:23
srichterI think it will be very cool23:23
srichterI'll send you the README.txt I am developing right now later this week, so you can see how far I am23:24
srichteroh, I also forgot to tell you: the level container has a validate() method that verifies the correct setup of level graphs23:24
mgedminsee ya23:25
*** mgedmin has quit IRC23:26
*** bskahan has quit IRC23:31
*** munkee has joined #schooltool23:50

Generated by 2.15.1 by Marius Gedminas - find it at!