IRC log of #schooltool for Monday, 2013-04-15

*** menesis has joined #schooltool07:50
*** khildin has joined #schooltool11:01
*** ignas has joined #schooltool13:23
*** ignas has quit IRC14:52
*** chembian has joined #schooltool15:38
chembianhi15:39
*** chembian has quit IRC15:41
*** khildin_ has joined #schooltool15:58
*** khildin has quit IRC16:01
*** replaceafill has joined #schooltool16:30
th1ahi replaceafill, yvl, menesis.16:31
yvlgood morning16:31
replaceafillgood morning/afternoon16:31
th1areplaceafill:  For starters do you have any outstanding report issues?16:32
th1aThose hung reports?16:32
replaceafillnot anymore16:32
replaceafillah16:32
replaceafillyes16:32
replaceafillhere's what happened16:33
replaceafilli used the same database while i was learning the infrastructure16:33
replaceafillat the end i was able to convert all the cando pdfs16:33
replaceafillto remote ones16:33
replaceafilland then on friday i noticed that all my failed tasks16:34
replaceafillthat had tracebacks and everything16:34
replaceafillwere reset to "pending" in schooltool.tasks16:34
replaceafilland with no information about the traceback16:34
yvlyes, that happens at the moment16:35
replaceafilli haven't tried to "reproduce" (since i need tracebacks again)16:35
yvlthey are wiped out from redis in 24 hours or something like thet16:35
yvlthat16:35
replaceafillah right, i assumed redis was involved :)16:35
replaceafillbut the rest works great16:35
yvlthey should be sent as traceback "messages" to sysadmin16:35
replaceafilli even tried yvl's pending messages change16:36
replaceafill(the number of pending messages next to the login)16:36
yvl:)))16:36
replaceafilli also assume we're going to move the messages in person index into an accordion?16:37
th1aI wasn't thinking that.16:37
replaceafillyvl, do you want me to do that?16:37
th1aThat only shows up in your own person index, right?16:38
replaceafillyes16:38
th1aIn home.16:38
th1aYou don't really need to be looking in your own accordion that much.16:38
th1aYou're really doing there for the messages going forward.16:39
th1aRight?16:39
yvlyes?16:40
yvlyes :)16:40
replaceafill:D16:41
replaceafilland my last question for yvl today is16:41
replaceafillcan we start changing gradebook's pdf to the new style?16:41
th1aSo what's the conclusion with the accordion question?16:42
replaceafilli just noticed intervention has one too16:42
replaceafill(thanks to menesis bug)16:42
yvlit's a design decision, th1a, so I'll leave it up to you guys16:42
yvlgood either way, I guess16:42
yvl+0.25 for not in accordeon16:42
th1aI don't want it in an accordion at this point.16:42
yvlok :)16:43
th1areplaceafill and I should review the UI changes.16:43
replaceafillwill we have logic to allow the user "discard" the messages?16:43
replaceafillafter a while that table gets big16:43
th1aThat doesn't seem like a priority, just use our regular table mechanics, right?16:44
replaceafillah ok, batching16:44
yvlit has pagination16:44
replaceafillsure16:44
replaceafillyvl, i didn't check closely, but are the system messages translatable?16:46
replaceafillmaybe they don't need to be16:46
yvlsystem messages?16:46
replaceafillbut i remember seeing FAILURE and IN PROGRESS as regular strings16:46
yvlas in "Report X generated"?  Yes.16:46
yvloh, those16:46
yvlthose - no IIRC16:46
replaceafillah ok16:47
yvlthese are not messages actually16:47
yvlmore like status codes16:47
replaceafilltrue16:47
replaceafillyvl, what about my gradebook pdfs question?16:47
replaceafillsorry that i throwing lots of question around :)16:48
yvl:)16:48
yvlwell, yes and no16:48
yvllet's chat about that in 10 minutes16:48
replaceafillsure16:48
replaceafilljust to finish my report16:49
replaceafilli also checked some small bugs16:49
replaceafilltried to reproduce one in the report card16:49
replaceafillbut wasn't able to16:49
replaceafilli think i need to start with a fresh database16:49
replaceafillso i'll try that today16:49
th1aNo.16:50
replaceafillor not :)16:50
yvl:)))16:50
th1aWe need to work on reports or bugs that we can fix quickly.16:50
replaceafillth1a, kk16:50
th1aI don't want to spend too much time reproducing right now unless we've got all the lower hanging fruit.16:51
replaceafillsure, i understand16:51
th1aMaybe ask a follow-up question?16:51
th1aCould this be a unicode thing?16:51
replaceafilli don't think so16:52
replaceafilli'm talking about this btw16:52
replaceafillhttps://bugs.launchpad.net/schooltool.gradebook/+bug/100817616:52
replaceafillit's like the layout not having a title16:53
th1aOK.16:53
replaceafillok, that's it from me16:53
th1aI'm just saying, there are probably equally important things that don't require reproducing.16:53
th1aThanks replaceafill.16:53
th1ayvl?16:53
yvlok16:54
yvladded the tiny indicator for new messages16:54
yvlnear login name16:54
yvlpushed a different report class lookup16:55
yvlthen found a weird bug16:55
yvland was trying to reproduce it, but couldn't16:56
yvlone report failed to generate because of DB conflict16:56
yvlit could be because of work-in-progress code16:56
yvlor it could repeat itself under heavy usage16:57
yvlso we can either try to do some load testing16:57
yvlor table this for now16:57
yvlthis brings us to gradebook reports16:58
th1aWell, we probably are at gut check time here...16:58
th1aOK, go ahead yvl.16:59
yvlok16:59
th1aWe'll get to the gut check in a minute.16:59
yvlso, gradebook report workflow is really really really convoluted16:59
yvlbut they will probably work as-is16:59
yvlprobably16:59
yvlso17:00
yvlreplaceafill did implement GradebookReportTask or such17:00
yvlsimilar thing is needed for gradebook17:00
replaceafillyvl, can you explain briefly the "convoluted" part?17:00
yvlthe gradebook report utilizes gradebook overview, right?17:01
replaceafillright17:01
yvlthat in turn uses gradebook base17:01
replaceafillwhich in turn uses section finder :D17:02
yvlyes17:02
yvlthere are pathways that will not work17:02
replaceafillah17:02
yvlbut I don't think that they are used17:03
replaceafillfor instance?17:03
replaceafillpreference stuff?17:03
yvlanything that leads to person annotations17:03
yvlwith remote reports you can schedule, then wonder off somewhere else and change stuff17:04
yvlso we'll just need to hope it works :)17:05
replaceafillyvl, do you think we'll see if there are problems really quick?17:06
replaceafillor maybe we can code "new" reports17:06
replaceafillthat don't depend on the same base class17:06
replaceafill(would take more time obviously)17:07
yvlwell17:07
yvlwe could test manually17:07
yvladd some large time.sleep for report task17:08
yvland then fiddle with preferences17:08
yvlnavigate to other sections17:08
yvlreuqest other reports17:08
yvland see if it breaks down17:08
replaceafillah ok17:08
replaceafillwhy does the preference (or annotation) would fail exactly yvl?17:09
replaceafilli mean17:09
replaceafillwhy specifically that part?17:09
yvlit's stored in database17:09
yvlif report looks something up when it is generated17:09
* replaceafill is just trying to see if his reports would have this issue :)17:10
yvlrather than when it was requested...17:10
yvlsay you unhide some column, click to generate report, then hide some column again17:11
yvlif the report is actually generated before hiding, you're lucky17:11
yvlfrankly, same goes for all data17:12
replaceafillbut in that case you'll end up with unexpected output right? not a "failure"?17:12
yvlstudent grades, anything17:12
yvlyes17:12
replaceafillah ok, i understand now17:12
yvlunless it looks up getCurrentSectionTaught or something similar17:12
yvlthen it will be really unexpected ;)17:12
replaceafill:(17:12
yvlif!17:12
yvlprobably not17:12
yvlbut not sure17:13
yvlneed testing ;)17:13
replaceafillyes17:13
th1aI wouldn't be super concerned at this point about changes between the request and printing at this point.17:13
th1aSince at worst it is an easy issue for the user to figure out.17:13
yvlok17:13
yvloh, random thing:17:13
yvlreplaceafill, if you ever need to pass extra parameters from request17:14
yvlyou can override RequestRemoteReportDialog.schedule17:14
yvland update task.request_params dict17:14
yvlwhen report is executed, data from there will be used as normal request17:15
yvlyou know, data like task.request_params['FILTER'] = 'something'17:15
yvltask.request_params['widget.0.stuff'] = '42'17:15
replaceafillah17:16
replaceafilli'll need that for the completion report pdf (if it stays as it is)17:16
yvlok17:17
* yvl done17:17
th1aOK, so, this isn't feeling like we'll be releasing it to universe on Thursday.17:18
th1aOTOH, we pretty much have to release it to our PPA on the 25th.17:19
th1a"Have to" in the sense that if we put it off beyond that, I fear it will NEVER be finished and possibly none of the other things we're planning on will ever happen.17:19
th1aSo I think we're looking at our usual scenario release-wise, in other words.17:20
yvltrue17:20
th1aSo, Ubuntu release is just everything except the celery stuff, PPA release includes celery.17:21
yvlhm17:22
yvlmaybe replaceafill could remake gradebook reports without remoting?17:22
th1aI'd think the Ubuntu release we already have ready, more or less?17:22
th1aWhat do you mean, yvl?17:23
yvl(to be released to Ubuntu)17:23
yvlremake - as in update to new style17:23
yvlby guidelines17:23
yvlnot sure how much work is that, actually :|17:23
yvlapologies for interruption, please continue17:24
th1amenesis:  This making sense to you in general?17:25
th1aOr am I ruining your life?17:25
menesisI have reached the same conclusion in the last few days17:25
th1aOK.17:26
menesisthat no way are we landing such big unfinished features days before the final freeze17:27
th1aYes.17:27
menesisthe 2.4 release17:27
menesis without the celery/offline reports/messages stuff17:27
th1aYes.17:28
menesisis ready17:28
th1aRight.  OK.17:28
menesisI have been working on that today since the morning17:28
menesisready = is in good shape17:28
th1aShould replaceafill try to jam in some bugfixes in the next couple days or is that more trouble than it is worth?17:28
menesisI have some finished cleanup work to merge17:29
menesisI will release schooltool 2.4b1 soon17:29
menesisothers already have a b1 release17:29
menesisthen if you can, try to squeeze in some bugfixes17:30
menesisI will do one more round of final 2.4.0 releases on April 18th (final freeze date)17:31
menesisthen there is one week for critical fixes only17:31
th1aWill the b1 actually be in the universe repos?17:31
menesisyes, 2.4b1 is in the universe17:31
menesisexcept schooltool core that I am finalising today17:32
menesisthe last release in universe is 2.1 for oneiric17:32
menesis2.3 is only in the oneiric ppa17:33
menesisand then celery stuff lands as the first feature of 2.517:34
th1aOK.  We should probably just release 2.5 beta on april 25 to the dev ppa, or something like that.17:34
menesisyes, that can be done as planned17:35
menesisin the dev ppa17:36
menesisI will also create a 'latest' ppa for 2.417:37
menesisalso, 2.4 will be the last "flourish" release17:37
th1aOK.17:38
menesisI plan to merge flourish and celery, and push that to "trunk"17:38
menesisbecause the new schooltool will break all older branches17:39
menesiscurrently there is a "trunk" on launchpad17:39
menesisit will be renamed to 1.7 (or 0.x)17:39
menesisand marked Obsolete17:39
menesison May 9, the last of ubuntu releases containing schooltool 1.x17:40
menesiswill reach end-of-life17:40
menesisSay goodbye to Zonki and the red old skin!17:41
replaceafill:'(17:41
replaceafill:D17:41
th1aOK, good.17:41
th1aI'd done that in my mind a long time ago.  ;-)17:41
replaceafill:D17:41
menesisme, too, but I have to do that on launchpad, ftp, buildbot and elsewhere, too :)17:42
yvl:)17:43
menesisI have written about some of that to the book already17:43
menesisabout supported schooltool releases17:44
th1aOK.17:44
th1aAnything else, menesis?17:46
menesismoved buildbot to the new server17:47
menesismarius wanted to shut down schooltool.pov.lt buildbot17:47
menesisso I redirected buildbot, coverage to the new server17:48
th1aThe DNS change is progressing.17:48
replaceafillmenesis, please try to keep the irc logs alive :)17:50
replaceafilli usually depend on them when i have no connection17:50
menesisalso completed merging some my branches, will land to schooltool soon17:50
menesisand release 2.4b117:50
menesisthat's all I think17:51
th1aOK, thanks menesis.17:52
th1aLet's talk about real and abstract deadlines.17:53
th1aAs always, these Ubuntu release deadlines are a bit abstract.17:53
th1aBut at least they're a bit harder than ones I pull completely out of the air, which slipped a lot more.17:53
th1aThere are two things to keep in mind now.17:53
th1ayvl is still literally stuck on his first post-it for February, so our plans for the year, which seemed fairly conservative on the whole, are in danger of completely falling apart.17:54
th1aThere is certainly a real "end of the year, end of job" deadline looming for yvl.17:55
yvltrue17:55
th1aAnd his tasks.17:55
th1aSo we're really falling behind there.17:55
th1aOTOH, we didn't actually have a full year of tasks.17:55
th1aBut we can't go on like this forever.17:55
yvlof course17:56
th1aBeyond that, there is the risk that this beta won't be tested.17:56
th1aFortuanately, there is a real task for that.17:56
th1aCanDo sites still want to be able to run a dump of all section reports on pdf.17:57
th1aWhich last year we hacked in by running a script that requested them.17:57
th1aI'd like to do it sanely with celery.17:57
replaceafillth1a, in a single pdf?17:57
th1aProbably not.17:57
replaceafilltar.gz?17:57
th1aI don't have a strong opinion.17:58
replaceafillah ok17:58
th1aWhatever makes sense.17:58
replaceafill(just curious)17:58
th1aJust, with a celery task or tasks.17:58
th1aSo basically, that's a heavy, real test, to have it ready for all the CanDo sites in June.17:59
yvlgood18:00
th1aAlso, we should be adding new celery reports to the ppa release pretty regularly.18:00
th1aOK.  I guess that's enough for now.18:02
th1areplaceafill:  We should go over which seem like the best low hanging fruit.18:02
replaceafillth1a, kk18:02
th1aHave a good week, guys.  See you Wednesday.18:03
th1aThanks menesis.18:03
* th1a drops the bag of gravel.18:03
yvlsee you Wed18:03
replaceafillthanks everybody18:03
replaceafillth1a, should i focus on core packages for bugs?18:04
replaceafillleaving cando and virginia out?18:04
replaceafillfor now18:04
replaceafillif i understand correctly virginia and cando are in our dev ppa, right?18:05
th1aCanDo is core.18:05
replaceafillah ok18:05
th1aMy recent batch of CanDo bugs should have a number of easy ones, unless you already picked those out.18:08
th1aI guess the search thing is a hairball.18:08
replaceafillyes18:08
replaceafilli started on that and stopped18:08
replaceafillhttps://bugs.launchpad.net/schooltool.cando/+bug/110371018:08
replaceafilllow18:08
th1aI'm still mad at aelkner about that one.18:09
th1aI told him how to do it, and he did it the most convoluted way possible instead.18:09
replaceafill:D18:09
th1aYou may just have to fix it for real.18:09
th1aDo you know the way it is supposed to be?18:09
replaceafillyou mean remove the assumption that the last two levels are skillset and skills?18:10
th1aHere was what I wanted.18:10
th1aFor a given document, store alternative names for skillset and skill.18:10
th1aLike, as attributes.18:11
th1aThat's it.18:11
th1aWhen I even try to understand what he did, I lose my mind.18:11
replaceafill:D18:11
th1aIt was so convoluted I couldn't even explain why it was a bad idea.18:11
replaceafillbut what if you have no layers?18:12
replaceafilli thought cando should still be usable18:12
replaceafilljust with skillsets and skills18:12
th1aDocument is a layer.18:12
replaceafillyes, but i mean no layers attached to skillsets18:12
th1aSo, what's the problem?18:12
th1aYou have to search for layers, skillsets and skills.18:13
replaceafillsearch right now depends on layers18:13
th1aThree things.18:13
th1aWell, clearly that's the way aelkner thinks.18:13
th1aIs it really so hard to aggregate three searches?18:13
th1aIs this really just one of those places where Zope still makes the simplest thing impossible?18:14
replaceafilli think the skills search in projects does that18:14
replaceafill:D18:14
th1aAnyhow, we'll have to come back to that in a few weeks.18:14
replaceafilli think it's just a matter of defining where are we going to use this search18:15
replaceafillbecause18:15
replaceafillwhat stopped me was that for assigning *skillsets* to *courses*18:15
replaceafill(if we decide to keep the same behavior)18:15
replaceafillyou need to look only layers (nodes), and course18:16
replaceafills/course/skillset18:16
replaceafillbut ok, we'll come back to it18:16
replaceafillmenesis, thanks for merging my cando branch!18:33
replaceafilli was about to do it when i noticed "Merged" in the status :D18:33
replaceafillth1a, zyt?19:11
th1ahi replaceafill19:11
replaceafillhey th1a quick question19:12
replaceafilli'm on https://bugs.launchpad.net/schooltool.cando/+bug/110370819:12
replaceafilland noticed19:12
replaceafillskill sets -> Label, no hint19:12
replaceafillskills -> Short label, no hint19:12
replaceafillnodes -> Label, hint19:12
replaceafillhint for nodes: "Limit to 7 characters or less."19:13
th1aYeah, it is all fucked up.19:13
th1aFirst, for the labels, is it always true they are limited to 7 characters?19:13
replaceafillno19:14
replaceafilljust for nodes and skillsets19:14
th1aIt is most important for skills!19:15
th1amumble grumble.19:15
th1aActually, that's probably the only one that matters.19:16
replaceafillchecking the code..19:16
replaceafilldata['label'] = u'%02d' % (len(self.context) + 1)19:16
replaceafillthe label for skills is a number based on the container19:16
replaceafillwhen there is no skill label19:16
replaceafillif there is, i don't see any other constraint19:16
* replaceafill goes check manually19:16
replaceafillyep, " this is a huge label for this skill " is a valid label19:17
replaceafilllol19:18
replaceafilli'll better save how it's done19:18
replaceafillth1a, how should it be?19:18
th1aThey all should be called label.19:19
replaceafilljust limit for skills19:19
replaceafillok19:19
th1aThey all should have a limit of 7 characters.19:19
th1aThey all should have the same hint "Limited to 7 characters or less."19:19
replaceafillkk19:19
replaceafillth1a, so nodes and skillsets could have no label, right?19:24
replaceafillskill must19:24
th1aYes.19:25
replaceafillthe label change resulted in a much higher fruit than i expected21:14
replaceafilli added a test just to check21:14
replaceafillmoving on...21:14
*** ignas has joined #schooltool21:16
replaceafill:D21:25
replaceafilli wonder what's the diff between description and full description ;)21:25
*** ignas has quit IRC21:26
*** ignas has joined #schooltool21:26
th1areplaceafill:  We should fix that.21:32
replaceafillth1a, on it21:32
th1akk21:32
th1aI figured there was an underlying problem there.21:32
replaceafillit's an easy fix, we just need to be consistent i guess21:33
th1aJust description.21:33
th1aDescription.21:33
replaceafillupdating the tests is the b..ch21:33
th1aYes.21:33
replaceafillth1a, the plan is keep doing small fixes tomorrow too, correct?21:34
th1aSure.21:34
th1aIf we have them.21:34
replaceafillnext: https://bugs.launchpad.net/schooltool.cando/+bug/112347121:35
th1agood one21:38
*** th1a has quit IRC21:40
*** th1a has joined #schooltool21:42
* replaceafill goes to get lunch, bb in ~1h21:50
*** Tranquil has joined #schooltool22:15
Tranquilhello22:15
*** Tranquil is now known as Guest2246122:15
Guest22461How do I go about installing School Tool on a Ubuntu installation?22:16
replaceafillGuest22461, http://book.schooltool.org/install-2_0.html22:17
replaceafillbasically "apt-get install schooltool"22:17
Guest22461thank you so much22:18
replaceafillno problem Guest22461 :)22:18
Guest22461My name failed to register... lol22:18
replaceafill:D22:19
Guest22461Love unix but Ubuntu (Linux) is so much nicer22:19
*** Guest22461 has quit IRC22:23
replaceafillth1a, zyt?23:45

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