IRC log of #schooltool for Wednesday, 2012-10-10

*** ignas has joined #schooltool00:14
*** th1a has quit IRC01:00
*** replaceafill has quit IRC01:58
*** menesis has quit IRC02:49
*** ignas has quit IRC04:16
*** ignas has joined #schooltool08:50
*** ignas has quit IRC09:18
*** khildin has joined #schooltool09:29
*** menesis has joined #schooltool11:23
*** ignas has joined #schooltool12:49
*** khildin has quit IRC14:16
*** ignas has quit IRC14:45
*** menesis has quit IRC14:59
*** th1a has joined #schooltool15:04
*** ignas has joined #schooltool15:26
*** replaceafill has joined #schooltool16:22
th1ahi replaceafill, aelkner, yvl.16:30
yvlgood morning16:31
replaceafillgood morning/afternoon16:31
th1aOK, replaceafill, how we doing?16:32
replaceafilli'm done with the projects gradebook16:33
replaceafilli'll be done today with the graphical completion report16:33
replaceafillother things i did this week:16:34
replaceafilladded new css for Done links16:34
replaceafillsent email explaining what i did to the dev list16:34
replaceafilland i reset the social skills instance for dwelsh yesterday16:35
replaceafilland extracted the programming skills and added them to jelkner's instance16:35
* replaceafill done16:35
th1aLots of Arlington babysitting.  ;-)16:36
th1aIn the future, they'll pay.16:36
yvlthat sounded like a threat :)16:36
replaceafill"they're going to pay for this!"16:36
th1aSo... we can go over the report and the projects this afternoon.16:36
replaceafillprojects for sure16:37
replaceafilli can set that up now16:37
replaceafilli'll try to have the report ready early in the afternoon too16:38
th1aNow... what we're going to need to do with this is test with some eggs and real databases (skills) before pushing to the PPA, right?16:38
*** menesis has joined #schooltool16:38
replaceafillah yes16:38
th1aNow... what we're going to need to do with this is test with some eggs and real databases (skills) before pushing to the PPA, right?16:38
th1ahi menesis.16:38
replaceafillth1a, i think i can get dwelsh acc Data.fs16:39
replaceafilli'm not sure if it has evaluations already though16:39
replaceafillthat's what we need, right?16:39
replaceafillbut we haven't had issues with jelkner instances so far16:40
replaceafilland they're using bleeding edge16:40
th1aI guess yes, we would want some evaluations.16:41
replaceafilli'll check dwelsh's instance and let you know16:41
replaceafillor we could ask glenda for one16:41
replaceafillyvl, i have a question16:42
replaceafillwhat's the reason for schooltool.cando.course.SectionSkillSet.deployed to be True?16:42
replaceafillthat's why they "appear" as deployed in the skills gradebook :)16:43
yvlno reason16:43
replaceafillis it ok if we change it to False?16:43
yvlas in - an accident16:43
replaceafillok, thanks16:43
aelknerit definitely should not be True, that's for report sheets16:44
* replaceafill done (again) ;)16:44
th1aOK, check dwelsh's instance.16:45
* replaceafill on it16:45
th1amenesis:  What's the status with the release?16:45
menesisI have made releases of all modules16:46
menesisto precise dev ppa16:46
menesisthey were rejected from quantal16:47
th1aah... too late?16:48
menesisbecause of feature freeze that was in effect for several weeks now16:48
menesisI have filed a freeze exception bug16:48
menesisbut it is also past final freeze now16:48
th1aOK, don't worry about it.16:48
menesisdoing all the steps necessary for ffe16:48
menesiswhen I installed the new version of schooltool16:49
th1aSo what version is in Quantal?16:49
menesisit crashed on the new evolve3916:49
menesisso what we have has problems16:50
menesisthere are also two futher bugfixes by yvl16:50
menesisversion in Quantal is only 2.1.116:51
menesisbackported bugfixes16:52
th1aIs that the same as is in Precise?16:52
menesisthat are safe for stable release update (SRU) in precise16:53
menesisif I do that16:53
menesisprecise-updates will have the same as quantal16:53
menesisthe latest release (2.2) is now in precise dev ppa16:53
th1aLet's get precise and quantal synced up.16:54
menesisand will copy the same to quantal main ppa16:54
menesiswill do that after quantal release16:54
th1aWell, this is pretty much how this always goes in October.16:56
th1aMy magical thinking; menesis could help by pointing these things out a month or two earlier.16:57
menesisyes we should follow ubuntu release schedule more closely16:57
th1aI then conclude that we should just focus on adding a few more features to this release and get it into the next release cycle at the beginning.16:57
th1aThere's not a big practical impact -- the April release is always the more important one, and we just only seem to be able to do one a year.16:58
menesisi.e. feature freeze means that new features land early16:58
menesis especially those that have interdependencies between modules16:59
th1aThen it comes down to what to call things.16:59
th1aIf we can put out 2.3 whenever, then when should we do it and with what?16:59
menesisstudent's journal for example solves a regression - there was one in old skin. on the other hand it is a brand new page17:00
menesisI think to re-release the same as 2.3.017:01
menesisschooltool has two more bugfixes and a crasher that I am yet to investigate17:01
th1aI think now we should wait to name
th1aWell, include two things:17:02
th1a1) FINISH CanDo;17:02
th1a2) big report push.17:03
th1aIncluding reports using Celery.17:03
menesisit was planned for october release. only last night I decided to call it 2.3.017:03
th1aWell, if it had been in Quantal, whatever we had would have been 2.3.17:04
menesisto have a stable series 2.3 (that should have been in quantal)17:04
menesisfor CanDo in production17:04
menesisso the 1) finish cando is what I accept17:04
th1aLet's assume 2.3 is actually going to be the 13.04 release.17:04
menesisbut celery and friends is a major feature17:05
th1aYeah... we'll have to split that apart somehow.17:06
menesisthat deserves a minor version bump, that is 2.417:06
menesisnow, what is the status of cando/virginia17:07
menesisthey wanted a new release asap17:07
th1aWe have a release which we need to test with their data because it evolves the database and we need to double-check.17:07
th1aI guess we're just going to try to get some dbs from them.17:08
* replaceafill got dwelsh Data.fs, checking if they have grades stored17:08
menesisschooltool 2.2.4 also evolves database and my quantal instance crashes17:09
th1aWhat is the bug?17:11
menesisPOSKeyError: 'No blob file'17:11
menesissomewhere converting people's photos17:12
menesisdid not have time to debug that yet17:12
th1aOK.  That's not the one you pushed anyhow?17:13
th1aI mean, that one isn't in the repository.17:13
menesisit is in the schooltool dev ppa that cando is using17:14
menesisbut I did not have problems with other instances17:14
menesisI assume that if there are no photos, no problem17:15
menesismaybe it's only a permission problem in my installation17:15
menesisyou have to test with some live Data.fs's17:17
th1aOK, so here's what I'm thinking, gentlemen.17:17
th1aMy giant planning post-it is full of reports.17:17
th1areplaceafill:  how long for IEP's?17:18
replaceafillafter completion report?17:18
replaceafillmaybe early next week17:19
th1aOK, so let's say aelkner & yvl keep bugfixing for another week.17:20
th1aI don't think that will be a problem, coming up with bugs.17:20
th1aAnd then we have the report version of last summer's flourish sprint.17:20
th1aEveryone doing reports for a month.17:20
th1aMeeting more frequently, everyone reviewing them so we get the style down.17:21
yvlsounds good17:21
th1aOh, I'm mostly thinging PRINTED reports.17:21
aelknersounds like a plan to me17:21
replaceafillth1a, not errors on evolution for dwelsh instance17:21
th1aWhen we're done with that, we release 2.3 "officially."17:22
th1aTo the PPA.17:22
th1aThen we pivot to getting the 2.4 infrastructure settled (Celery, etc) and getting those packages into the 13.04 repository as quickly as possible.17:23
th1aMake sense menesis?17:23
menesisI'd say one more week for bugfixes and cando17:24
menesisnext thrusday 2.3.0 and cando releases17:24
th1aYes, but I'd like to get new reports into the release I announce widely.17:24
th1aIt would be worth a couple weeks.17:25
th1aOr perhaps it would be nice to put 2.3.0 out on the Ubuntu release date and then 2.3.1 has reports?17:27
th1aOK, that's fine.17:28
th1amenesis:  Can we move on to yvl for now?17:29
menesisyes of course17:29
menesisI have nothing more to say17:29
yvlnot much to report17:29
yvlfixed a few bugs17:29
yvlexperimented with require-to-reset-password17:30
yvltabled for now17:30
yvlmoved on to timetable scheduling views17:30
yvlbut I do think that password expiring is worth the time17:31
yvlalso - we could do temp password generation17:31
yvl(visible to school administration)17:31
yvl(maybe exportable)17:31
* yvl done17:31
th1aOK.  So you're working on the scheduling multiple terms issue?17:32
yvlnext will be the small issue of photo timestamps17:34
yvlit's a bit weird when you change a photo and get redirected to a page where photo is not changed17:34
yvluntil next request17:34
th1aYes, it is a bug.17:35
th1aOK, thanks yvl.17:36
replaceafillth1a, every seems to work on dwelsh data with latest trunks17:36
replaceafillno error on evolution and grades seem safe17:37
th1aOK, thanks replaceafill.  I'll ask Glenda if she wants us to check her data.17:37
aelknerok, i figured out how to fix the skills order in verso_skills.py17:38
aelkneri want yvl to look at this revision:17:38
aelknerit has to do with the fact that these weird Doodad and Doodads classes are dictionaries17:38
aelknerwhich, obviously, do not preserve the order that the api delivers data17:39
yvllooks good to me17:39
aelknerhowever, luckily, DutyArea had a copy of the raw data in the raw_data attribute17:39
aelkneri was able to verify that the change I made fixes the order of the skills17:40
aelknernow, mind you, Glenda didn't complain about the order of the DutyAreas17:40
yvl(just FYI, all objects keep a copy of raw data - just in case)17:40
aelknerDutyAreas does not17:41
aelknerso we've lost the raw data order in that case17:41
aelknerthat's because Doodads' update method does not keep the raw data around17:41
yvloh, my bad17:42
th1aOh, you're using that to assign numbers as the label?17:42
aelknerso, yvl, should i change the update method in Doodads to save to a new raw_data17:43
yvlyou could add ._order17:44
aelkneri could change update to saw self.raw_items = raw_items17:45
aelknerand use that in the build_xls method as i did with DutyArea17:45
aelknerok, i'll do that next, just wanted to make sure that made sense17:46
yvlit's a bit of overkill, since Doodads is itself a dict, but not that I mind17:46
yvlsomething like...17:46
yvl_order = []17:46
yvlfor item in raw_items:17:47
yvl   if item in self._order:17:47
yvl        self._order.remove(item)17:47
yvljust a suggestion17:48
aelknerwhy was Doodads a dictionary in the first place?  what does it serve to have it that way?17:48
yvleasier to get items by ID maybe?17:49
yvllist Doodads17:49
yvlprobably that17:49
yvlfree methods of a dictionary17:49
aelknernot sure how that helps, seems we need a workaround for a workaround17:50
aelknerand it's a real pain in the ass to maintain17:50
aelkneror is it perhaps dies to the partial update handling?17:50
yvlthat was done in like 10 minutes17:50
yvlit was like a hack, remember?17:50
yvlfor fast downloading17:51
th1aYeah, just change it aelkner.17:51
yvlto whatever you feel is best17:51
yvlbecause you know the data better than me17:51
th1aIt is just a utility script.17:52
th1aOK, so you've got some things in your queue, aelkner?17:53
aelkneri'm not done with this seemingly unimportant utility script that Glenda freeked out about17:54
aelknerand making the change that yvl suggested without testing it would be irresponsible i think17:55
aelknerand testing it requires hours to wait for the downloads, it's not a trivial thing17:55
yvlnobody said it's unimportant17:55
th1aThat it works is important -- just its implementation isn't a big deal.17:56
th1aYou can do whatever you want aelkner.17:56
yvlI apologize for making a suggestion17:56
th1aYou just have to learn to predict when a change will raise yvl's hackles.17:56
* replaceafill assumes every changes he makes, raises yvl's hackles ;)17:57
th1aOK, so I think aelkner has tasks for now.17:58
aelkneryeah, i'll work out some change to make like that suggested by yvl17:58
aelknerthing is, how to deliver to Glenda?17:59
aelknerand with new version, what is the result of having a xls file with different order?17:59
th1aOh, wait.17:59
aelknercan they reimport it without breaking things17:59
th1aThat won't work at all, will it?17:59
aelkneryeah, that's what i'm thinking18:00
th1aAll they need is new labels though.18:00
th1aSo will the importer add labels without changing anything else?18:00
*** magespawn_ has joined #schooltool18:00
th1aAlso, this doesn't do any good until we confirm that we sort by label everywhere we should.18:00
*** magespawn has quit IRC18:00
aelknerthe new verso_skills.xls file cannot be reimported into an existing Data.fs18:01
aelknerthat's because the ids are different, a different order is established18:02
th1aWhy are the ids different?18:02
aelknerthe Skills sheet has Skillset Id, Skill Id, other things, then Label18:03
aelknerSkill Id and Label are determined by the order of processing the Doodads18:03
aelknerso having changed the order of the Doodads, the skills themselves have different labels AND Skill Ids18:04
th1aSo to actually fix this, you need to generate the skill id's the old way and the labels the new way?18:04
aelkneri mean, that's the only way the script can run, dynamically creating the Skill Ids as it processes the downloaded data18:05
aelknerhm, that's an idea18:05
aelknerthe problem is, the old way of creating the skill ids was using the dictionary order, which is unpredictable18:06
aelknerso, i'm not sure how safe it's going to be to just change the script to change the labels18:06
aelknerand reimport, but i guess we could hold our breath and hope that the dictionary order is the same18:07
th1aMaybe you need to just process their current import sheet.18:07
th1aAdd the new labels to that.18:07
aelknerlet's think about the users and how they use the script18:08
aelknerthey don't have an existing verso_skills.xls necessarily18:09
aelknermy understanding was that we delivered the script to them a whilw ago18:09
aelknerand they came up with a set of instructions for each sysadmin to use18:10
aelknerthat involved running the script to create the verso_skills.xls file18:10
th1aOK, so basically every site is running this and getting potentially different results?18:10
aelknerin theory, yes18:11
replaceafillpage 318:11
replaceafillthey make clear not to mess with verso_skills.xls18:11
th1aAre they downloading it or generating it?18:12
th1aIt doesn't make sense that they'd all generate it.18:12
replaceafill"Download the following files"18:12
aelknerwell, if they are downloading it, then we are in better shape18:12
aelknerin that case, we could talk to Lee Capps about creating a new version18:12
th1aMaybe you can just download that aelkner and add the labels to it.18:12
aelknerat this point, if we can establish that the user does NOT run getverso.py18:13
aelknerand we can just patch the verso_skills.xls to have the right labels18:13
th1aI believe that is the case.18:13
replaceafilli don't think they do18:13
aelknerand the whole state just uses the same verso_skills.py18:14
aelknerthat could resolve the situation18:14
th1aJust make sure nothing weird happens when you reimport with the new labels.18:15
aelknerso, th1a, as you suggested, i could modify the script to take a verso_skills.xls file as an argument18:15
th1aaelkner, I think you can just hack together something for this case.18:15
aelknerand if that argument is passed, make sure to only update xls rows in plave18:15
th1aIt can be a one shot deal and then fix the main script as well.18:16
aelknerok, so the deliverable is a new verso_skills.xls file with same skill ids but new labels18:16
aelknerand assume that noone is running themselves18:17
aelknerok, i'll download their posted xls file and patch it for them18:18
aelknerand yes, changing the views to sort by label will be part of the fix18:18
aelknerwish me luck :)18:18
th1aWell, let us know how that is coming, because it would make sense to get it in the "big release" which is immanent.18:19
yvlbest of luck, aelkner ;)18:19
aelkneri'll write the developer list when i have news18:20
th1aI guess I'll let menesis know when I think we're ready to push the CanDo changes to the PPA.18:20
replaceafillouch, cando's data model change broke virginia's section reports... :(18:20
th1aAh, good catch.18:21
th1aYeah, I thought we discussed that.18:21
th1aThanks aelkner.18:21
* replaceafill goes to fix that first...18:21
th1aI worked on a new example of standards for the docs, and I thought I found a straightforward one, but it ended up being non-hierarchical.18:22
th1aIt is organized by grade level and the same standards have to appear in multiple grades.18:22
th1aSo... there you go.  Good thing we allow that.18:22
th1aOK, thank you gentlemen.18:22
th1aHave a good week and weekend.  Keep me posted on your progress.18:23
* th1a drops the bag of gravel.18:23
yvlthanks guys18:23
replaceafillthanks everybody18:23
yvlsee you Mon18:23
aelknerthanks guys, cya mon18:26
replaceafilldoes anyone knows if it's possible to "inherit" buildout sections?18:34
replaceafillinherit everything from foo18:35
replaceafilli remeber i read about it some time :)18:35
replaceafilland i'm getting tired of adding all those lines to my buildouts for adding headless functionality18:36
replaceafillah! <=18:38
replaceafillif we create a section after the skills have been assigned to the course, the section doesn't get the skills?!?!?!19:10
replaceafillyep, confirmed19:13
replaceafillth1a, can i file a bug about this and give it to yvl19:14
replaceafillhe knows better how to fire events19:14
th1aSet it as critical.19:16
*** magespawn_ is now known as magespawn19:16
*** ignas has quit IRC19:53
*** menesis has quit IRC20:08
replaceafillth1a, to me dwelsh is a bug20:52
replaceafilldwelsh problem :D20:52
replaceafillwe have a copySection function that copies courses, instructors and members20:53
replaceafillbut no locations20:53
th1aYes, that's a bug.20:53
th1areplaceafill:  Maybe you should put dwelsh's bug next in your queue.21:33
replaceafillth1a, ah ok21:34
replaceafillth1a, i was thinking21:35
replaceafillfor locations, should we use the same "update" logic in linked sections21:35
replaceafillthat we use for instructors/members21:35
th1aI think it is slightly less likely to be the correct guess.21:36
th1aIf you withdraw from a term as a student, you're almost always not coming back intentionall.21:36
replaceafilli'll limit it the fix to the add form then21:36
th1aHowever, it seems more likely to me that you would move rooms around per term.21:36
replaceafillmakes sense21:37
replaceafillth1a, for the release you want both types of completion reports, right?21:37
replaceafillcharts + tables21:37
replaceafilltables = the one you don't want to do :D21:38
th1aRight now I consider the tables to be va only.21:38
replaceafillth1a, before we tabled the completion report, you told me we should probably do it by term21:59
replaceafilldoes that still apply?22:00
replaceafillyour "(maybe)" in the log made me wonder :)22:02
th1aThe skills are copied to the section but that is separate for each term?22:03
replaceafilli haven't checked yet if equivalency works across linked sections22:04
replaceafillbut your point about the kid leaving in the middle of the year makes sense to me22:04
th1aBasically, it should be per term, but it should include scores from earlier terms in the linked section for kids in the current term.22:05
th1aThat shouldn't be TOO crazy.22:06
replaceafillkk, that solves my question22:06
*** ignas has joined #schooltool23:36

Generated by 2.15.1 by Marius Gedminas - find it at!