IRC log of #schooltool for Monday, 2013-03-11

*** menesis has quit IRC00:45
*** veloutin has quit IRC02:50
*** Lumiere has quit IRC02:50
*** th1a has quit IRC02:50
*** paulproteus has quit IRC02:50
*** paulproteus has joined #schooltool09:26
*** th1a has joined #schooltool09:26
*** Lumiere has joined #schooltool09:26
*** veloutin has joined #schooltool09:26
*** yvl has joined #schooltool10:29
*** replaceafill has joined #schooltool16:11
th1ahi replaceafill, yvl.16:30
yvlgood morning16:31
replaceafillgood morning/afternoon16:31
yvlit's bank holiday here so I guess menesis won't be around16:32
th1aOK.  It is the second day of daylight savings time here.16:32
th1aShould we shift an hour earlier next week?16:33
replaceafillfine with me16:35
*** ChanServ sets mode: +o yvl16:35
*** yvl changes topic to "SchoolTool development | IRC logs at | SchoolTool Dev meetings Mon, 13:30 UTC (15:30 EET, 9:30 EST) | Use for pasting."16:35
*** yvl changes topic to "SchoolTool development | IRC logs at | SchoolTool Dev meetings Mon, 13:30 UTC (15:30 EET, 9:30 EDT) | Use for pasting."16:35
th1aOK, yvl, how are reports coming?16:36
yvlwell :)16:37
yvlI did get few reports to render in celery16:38
yvlwithout reporting back to ST16:38
yvland I'm somewhat happy with "message sending"16:39
yvlso on Wed I expect to have at least one report available for demonstration16:39
yvlwriting new reports will be essentially the same as the old ones16:40
yvljust change the report request dialog class name16:41
yvland specify which request parameters you want to pass on to the report16:41
yvlpictures will be a bit different16:42
yvl(because I'd like to do without traversal if possible)16:42
th1aFor ID's?16:43
yvlso, for example, url to a static resource like schooltool logo16:43
yvlis rendered as data: instead of http:16:43
yvlI think we should avoid getting files with urllib where possible in general16:44
yvl(for reports)16:44
yvlwell, anyway16:45
yvlthat was just FYI16:45
th1aWell, keep it moving forward.16:46
*** ChanServ sets mode: -o yvl16:46
th1aThanks yvl.16:47
replaceafilllast week i basically gave up on the journal bug16:48
replaceafillsorry about that16:48
replaceafillbut i wasn't able to reproduce at all16:48
replaceafill(from scratch)16:48
replaceafilli did identify the problem (i think)16:48
replaceafillthe section journal adapter sometimes retrieves the wrong meetings16:49
th1aWell... I don't know that we've given up on it.  Set it aside for the moment...16:49
replaceafilli just switched ;)16:50
replaceafillso i started on glenda's changes to the VA skills reports16:50
replaceafilli've finished the pdf of the completion one16:51
replaceafillthe progress one doesn't need a pdf16:51
replaceafilli added some JS to set the header of the table fixed16:51
replaceafillsame way we do it in the book with the navbar16:51
replaceafilland now i need to work on the totals of the columns16:52
replaceafillyvl, do you have a recommendation on how to create a total row for a column in a table formatter?16:52
replaceafillprobably at the bottom of the table16:53
yvladd an extra item?16:53
replaceafilli thought of that, but how do we prevent sorting on it16:53
th1aWe should probably show Glenda what you have and double check what she means there anyhow.16:53
replaceafillth1a, ah ok16:53
replaceafilli think the rest of her requests are easy enough16:55
replaceafilli'll make the report not to render initially16:55
replaceafillth1a, i was thinking we need a submit button after all for that16:55
replaceafillor showing the filters with '-------' values maybe?16:55
replaceafilli like the submit button idea better, you can set several filters and then run once16:56
th1aTo keep it from timing out?16:56
replaceafillwell, at least the user will get to the report generation page16:56
replaceafillit may timeout with the spinner showing :)16:57
replaceafillso, i'll work on this and have it ready for showing it to glenda tomorrow16:57
replaceafillah btw16:57
replaceafilli'm flying on wednesday to santa clara to meet with jelkner16:57
replaceafilland paul and david16:57
replaceafilland work on quiz16:57
replaceafilland attend pycon :)16:58
th1aWhen will you be back?16:58
replaceafillbut i plan to work normally the rest of this week16:58
replaceafillnext thursday 2116:58
replaceafilli also plan to attend the dev meetings next week16:58
yvlbtw, replaceafill, a quick hack would be adding an invisible column that has False for all items, True for totals and sort by it16:58
replaceafillyvl, in addition to the current sorting?16:59
replaceafillthat will work! :D16:59
yvlbut that's more of a hack16:59
replaceafillmy kind of solution ;)17:00
replaceafilli'll give it a try17:00
replaceafillthanks yvl17:00
* replaceafill done17:00
th1ayvl:  Have you worked at all on this area: "section journal adapter sometimes retrieves the wrong meetings"17:02
th1aI think the problematic code my reach back to ignas' days.17:02
yvlnot recently17:02
th1aDo you know if you ever touched it at all?17:03
yvlI did17:03
yvlwhen rewriting timetables17:03
yvlwell, after that17:03
th1aPerhaps replaceafill should explain the issue further then.17:03
yvlwhat are the "wrong meetings"?17:04
yvlmeetings from another section of same teacher?17:04
replaceafillsometimes not even from the teacher17:05
replaceafilli remember i couldn't find a few in any of the teacher's section calendars17:05
yvlwell, the place I want to blame is getCurrentSectionTaught17:05
yvlbut that's just because I don't like that code ;)17:06
replaceafillbut that would send me to the wrong section, right?17:06
replaceafilli mean, url-wise17:06
replaceafilli was blaming more the adjacent_sections logic17:07
yvlaren't parts of journal retrieved with jquery?17:07
replaceafillpopups, yes17:07
replaceafillbut this happens even if i disable all js in it17:07
replaceafillso no validation, no popup code, etc17:08
replaceafillso, when you load the section journal view17:08
replaceafillcolumns have ids set to the __name__ of the meeting17:09
replaceafillwhat i realized is that when the bug happens17:09
replaceafillthose __name__s don't belong to the calendar of the current section17:09
replaceafillin jelkner's case, somehow, the first term spans from first day to last day IN THE SCHOOLYEAR17:09
replaceafilli meant the first term schedule17:10
yvlyes, I remember17:10
yvlweird configuration17:10
replaceafilli mean, the schedule of the first term section :)17:10
replaceafill(just to be clear) ;)17:10
replaceafillbut to add more weirdness, sometimes the __name__s don't come from that first term schedule anyway :(17:11
replaceafillso, that's why i blame journal.journal.adjacent_sections17:11
replaceafilli think it retrieves other sections (taken from instructors, students)17:12
replaceafillthen the journal view just uses dates in order to filter17:12
yvlthat is it's purpose17:12
yvlbtw, the way it's done is a bit outdated nowadays17:13
replaceafillyvl, that's pre-linked sections, correct?17:13
replaceafilli mean, that's what's trying to do?17:14
replaceafillso, if i understand this correctly, then in browser.journal.allMeetings17:15
yvlit tries to mash up journals of course taught by instructor to any of the students17:15
replaceafillevents are filtered by in term17:15
replaceafilland here:17:16
replaceafillmeetings = [by_uid[e.unique_id] for e in sorted(insecure_events)]17:16
replaceafillit's when sorting happens17:17
replaceafillbut sometimes that list has way more events than what's displayed in the journal view17:17
replaceafilli was getting 185 meetings for one term in jelkner's data17:18
replaceafill(which i don't trust ofc) :D17:18
yvland there were less meetings in all months of the term?17:19
yvldisplayed in the journal I mean17:20
replaceafillthe journal usually displays the scheduled periods right17:20
replaceafillthat's why you don't notice the bug easily17:20
replaceafillif you'd get like 80 columns you'd probably say "that's too many periods"17:21
replaceafillat some point i was tempted to break the subclassing of LyceumSectionJournalView17:21
replaceafillin the flourish view17:21
yvlit displays only the scheduled periods?17:21
replaceafillyvl, apparently17:22
replaceafillsame date17:22
replaceafillsame period name17:22
replaceafillso you get17:22
replaceafillbut if you inspect the id of the 03 cell, you'll see a different __name__17:22
replaceafillyvl, question17:23
replaceafillif i do sorted(calendar_events)17:24
replaceafillcould you point me to the calendar event methods that handle that?17:24
replaceafilli was looking for some __cmp__ there17:24
replaceafillbut couldn't get to it17:25
* yvl can't see that too17:26
replaceafillthat's one of my reasons to not trust the sorting line above17:26
replaceafillit's like we're relying on randomness there17:26
replaceafillyvl, if you have any advice on all of this, i could try anything :(17:27
replaceafillif you want to "see" it, i have jelkner's database17:27
replaceafillwhich you have to F5 F5 F5 F5 F5...17:28
replaceafilland at some point you'll see it :D17:28
yvlthere's not much I can advise without looking at it17:28
yvlif you could pass me a database and instructions where to look and what to do to reproduce :)17:29
yvland what to look for, ofc.17:29
yvlI could try debugging17:29
yvlit may be something simple17:29
replaceafillthe reproduce part is hard17:30
replaceafillok, i'll send you all of i got17:31
yvlwell at least you could tell me exactly what happened where instead of what was expected17:31
yvlsomething sometimes is wrong is not the best bug description ;)17:31
th1aThey are also having serious journal issues in Thailand, this exact bug and one where duplicate periods appear in the journal.  That's probably related, I'd think.17:31
yvlreplaceafill, could you also paste links to those bugs in email too?17:32
replaceafillone of my fears is: user recorded grades on wrong meetings17:35
replaceafillnow we display the right meetings17:35
replaceafillgrades are gone :(17:35
yvlthey are not gone17:35
yvlthey may be not visible though :)17:35
th1aIf you want to take a look at this yvl, that's fine, but don't put more than, say a day into it without telling me.17:35
replaceafilli mean, visually17:35
yvlok, th1a17:35
th1aThat is, I'm HAPPY for you to take a look at it now, just don't derail the release.17:36
th1aThis is a serious bug though.17:36
yvlfrankly, I was waiting for a similar bug for a while now17:36
yvlsomething does not feels right to me with journal's meeting code17:37
th1aAlso, I was telling replaceafill, it might be easier in the fix to dump the data into just a simple dictionary and then dump it back into a fixed journal.17:37
th1aRather than something more magical.17:37
th1aJust something to keep in mind.17:39
th1aOK, that should do it gentlemen.  Thanks!17:39
yvlthanks guys17:40
th1aSee you Wednesday.17:40
yvlsee you Wed17:40
replaceafillthanks everybody17:40
* th1a drops the bag of gravel.17:40
*** yvl has quit IRC17:40
*** th1a has quit IRC19:44
*** th1a has joined #schooltool19:48
replaceafillth1a, just sent yvl jelkner's database and how to "reproduce"20:46
th1aI see.20:47
replaceafillreproduce isn't the right word for this case :D20:47
*** replaceafill has quit IRC21:31
*** replaceafill has joined #schooltool21:58

Generated by 2.15.1 by Marius Gedminas - find it at!