IRC log of #schooltool for Monday, 2009-03-16

*** replaceafill has joined #schooltool02:15
*** fsufitch has joined #schooltool05:01
*** fsufitch has left #schooltool05:56
*** replaceafill has quit IRC07:24
*** Aiste has joined #schooltool09:00
*** mgedmin has joined #schooltool10:01
*** Aiste has quit IRC11:23
*** mgedmin has quit IRC11:55
*** mgedmin has joined #schooltool11:58
*** alga has joined #SchoolTool12:04
*** mgedmin has quit IRC13:19
*** mgedmin has joined #schooltool13:19
*** balor has joined #schooltool13:49
*** mgedmin has quit IRC13:53
*** mgedmin has joined #schooltool13:56
*** th1a has joined #schooltool14:32
*** ignas has joined #schooltool14:55
*** fsufitch has joined #schooltool14:55
*** fsufitch has left #schooltool14:56
*** whaddon has joined #schooltool15:14
*** replaceafill has joined #schooltool15:16
Lumierehi th1a, replaceafill, ignas15:20
th1ahi Lumiere.15:20
replaceafillhi Lumiere15:20
*** whaddon_ has joined #schooltool15:21
th1aGood morning ignas, Lumiere, replaceafill, aelkner, yvl.15:31
replaceafillgood morning everybody15:31
th1aHow we doing aelkner?15:33
aelkneri finished the report card layout view last week15:33
aelknerand i started working on the report card pdf view itself15:33
th1ayvl around?15:33
ignasyeah around, will be here soon, was eating cake15:33
Lumieremmm cake?15:34
aelkneri have a question about testing pdfs15:34
lisppaste5th1a pasted "Sample import error" at
aelkneri noticed there wasn't any such tests for sla's narrative15:34
ignasCake! ;) (my birthday was 2 days ago)15:35
yvlhi everybody15:35
th1aignas: Does one test pdf generation?15:35
*** whaddon has quit IRC15:35
ignasth1a: i know, fields for persons do not match again15:35
* th1a checks his bank account for 100,000,000 isk.15:36
th1aYes, can we fix that?15:36
ignasth1a: just export once15:36
ignasto see how the columns should be15:36
ignasand shuffle the old columns around15:36
ignasmy friend did that when doing import recently, he managed to do it without me being anywhere close ;)15:37
* th1a has 27,546,357 isk to his name. :-(15:37
* ignas not rich either ;)15:37
ignasas for PDF's - we test that you get a "pdf" on the url15:37
ignasbut most of the tests are unit tests15:38
ignasthat test all the building blocks, as we can't really look into a pdf15:38
aelknerone could test the existence of the stories15:38
aelknerthat's what pdf files are built from15:39
aelknerhaven't you used reportlab?15:39
ignasit was looong time ago ;)15:39
*** whaddon_ is now known as whaddon15:39
aelkneryvl did this more recently15:39
yvlwell, testing PDF generation is a bit tricky15:39
yvltesting existence of stories is not enough15:40
yvlas it is same as testing existing of PDF15:40
yvlwe developed canvas stubbing method for testing purposes in one of POVs products15:41
yvlit's not perfect, but may be good enough15:41
yvlthough introducing it to schooltool would eat some development time15:42
yvl(and I think this should be done for schooltool, not just sla)15:42
th1aI'm in favor of not eating development time on it right now.15:43
algaGuys, I'm warning you15:43
algathis thing is a resource hog15:43
th1aWhich thing?15:43
algathe low level dumps take megs of disk space15:43
algathe canvas commands logging thing15:43
yvlok, so we can postpone PDF testing until we have more PDF reports15:44
algaif you have many branches on your disk, you end up having hundreds of megs taken by PDF tests15:44
th1aI'm in favor of quick and dirty testing of pdf generation at this point.15:44
aelkneri was thinking15:45
aelknersince pdf files are built from a list of stories15:45
ignasalga: we are not generating reports of such magnitude as you do ;)15:45
aelknerthe render step is the last that actually generates the pdf document15:46
aelknerbut if we had tests just for the stories themselves15:46
yvlaelkner, most of the objects in the story are some Reportlab classes15:46
aelkneri understand15:46
yvlwhich reslut in <Paragraph ...> <Table ...>15:46
aelknerbut we could dump those objects in a controlled fashion15:46
aelknerpresenting the attributes we find interesting15:46
aelkneryielding an output similar to queryHTML15:47
yvlgood idea15:47
yvlthough not very trivial to implement15:47
yvlok, maybe I'm wrong :)15:48
aelknerhow so?15:48
algaI convinced him it's not hard ;)15:49
aelkneri just figured we have the ability to control what we dump, like <para title='foo', etc.15:49
aelknerof we don't even need the '<' sign per se15:50
Lumiereaelkner / th1a: can you poke me on google talk when you get around to me (I'm working on a couple of things, and the audio alert from gt would help me15:50
th1aLumiere: OK.15:50
aelknerbut just a set of rules for dumpling reportlab objects15:50
aelknerand then one could test the entire pdf file before rendering it15:51
ignasyeah dumpling reportlab objects ;)15:51
aelknershrimp dumpling objects :)15:51
Lumierein soup with a side of fried rice?15:51
aelknerok, now i'm hungry15:52
Lumiereyour welcome15:52
th1adumplings fried in rendered pork fat?15:52
aelknerbut we'll leave out the rendering15:52
* Lumiere drools at work... darn it15:52
th1aI've rendered duck fat.15:52
th1aMmm... duck confit.15:52
aelkneri hope the duck didn't mind15:53
Lumiereit was dead15:53
th1aMy main concern is that whatever you do doesn't take long.15:53
th1aOr turn into an infinite regress of tiny problems.15:53
yvl1 dev day is ok?15:53
th1aSo is aelkner going to do this?15:54
yvlwell, maybe I should do it?15:55
aelknerid i do it, then do i decide the format of the output?15:55
yvlI mean, we have some of the code in POVs project already15:55
aelkneri'm fine with that as long as you can get something into schooltool along side of queryHTML15:56
th1aI'd rather have yvl do it.15:56
yvlon a related note - aelkner, which version of schooltool is SLA currently depending on?15:56
* yvl kind of lost track and broke his sandbox15:57
aelknertrunk as of a month of so ago15:57
yvlbut not the latest?15:57
aelkneri got sla to work with schoolyears15:57
aelkneri haven't tested it against the latest15:57
yvl(there's some fuzz about email demographics field)15:57
aelknerwhat's that?15:58
yvlwants email in IPersonAddForm15:58
* Lumiere is happy to see email as a demographics field...15:59
th1aI'm sure SLA will have to be rejiggered to work with demographics.15:59
aelkneramongst other things15:59
Lumiereit will really help us at some point if we want to use email for password reset15:59
aelknersla needs a serious evolution to properly use schooltool schoolyears16:00
th1aWell, that's a post 1.0 project.16:00
aelknerrather than having its own scholyears concept16:00
th1aThey're ok for the moment, right?16:00
yvlaelkner, I didn't find the right schooltool version wich sorks with SLA; gonna poke a roughly month old trunk16:01
ignasLumiere: it's not demographics anymore, email is part of contacts at the moment16:01
yvlthough it would help to know a more precise date/revision16:01
aelkneryvl, why the interest in sla at the moment?16:01
aelknerare we doing work for them that i don't know about?16:02
yvlI need to add PDF reports16:02
yvlletterhead stuff16:02
yvland as far as I understand, they're needed for SLA?16:02
yvlor am I horribly wrong? :)16:02
aelkneri wouldn't say you're wrong16:02
aelkneri'm just trying to understand what our arrangement with them is at the moment16:03
th1aNo, it is needed for SchoolTool in general.16:03
yvlok, so "horribly wrong"16:03
aelknerit's good to be horribly wrong16:03
yvlthat makes my life easier now, actually16:03
th1aIt is good to not have to figure out exactly how to get the sla branch running now.16:03
aelknerth1a: we're not worrying about sla until after april 16, right?16:04
th1aAs long as nothing breaks.16:04
aelkneri haven't heard a peep from them in months16:04
aelknerso i guess nothing's breaking16:04
th1aThat's good at this point.16:04
th1aThey know we're busy.16:04
th1aThey're busy.16:04
aelknerso i'm curious what yvl's going to come up with for reportlab story testing16:05
th1aOK, so yvl has to get the letterhead code into schooltool trunk.16:06
yvl+ section linking16:06
aelknerwhen you say letterhead, are you refrring to the logo thing that the narrative pdf has?16:06
th1aI guess.16:07
th1aLogo, name of school...16:07
th1aWhat you would expect to see on the top of an official document.16:07
aelknerthe thing that appears on every page16:07
aelkneras if it were a form16:07
yvlsomething along the lines of
th1aWell, top of the first page.16:07
th1aThat's the basic idea.16:08
Lumiereth1a: most places put letterhead at the top of evry page16:08
Lumiereor keep it to 1 page ;)16:08
th1aLumiere: Yeah, you're probably right.16:08
aelkneryeah, for instance, in a multi-student report card run16:08
th1aThat is one of many options we can add in coming years.16:08
Lumierereport cards should be 1 page16:08
aelknerper student16:08
aelkneryou type "of course", can't you?16:09
aelkneromg my bff just texted me16:09
th1aaelkner says "Get off my lawn!"16:10
aelknerno, it's goml16:10
th1aOK, so we need to keep aelkner and yvl from blocking each other.16:10
Lumiereaelkner: of course I can... doesn't mean I have to ;)16:11
yvlaelkner, what are you nearest plans?16:11
yvlfor the week16:11
aelkneri can write the pdf for now without a letterhead16:11
aelknerand i don't need to test the stories at first16:11
aelknerbut i'd like to be in on the discussion of how stories are dumped16:12
aelknersince we havn't done it yet, i'd like to vote on simplicity of output16:12
aelknerto make test-writing as clean as possible16:12
th1aLet's go around the rest of the folks and then come back to working that out.16:12
aelknerso that's my report16:12
yvldeal, aelkner16:13
ignasmerged most of the things to trunk16:13
ignasset up buildbots for trunk/release/release-116:13
Lumiereaelkner: just make sure to leave some space at the top for it ;)16:13
ignasplanning to branch our next release this week16:13
aelknerignas: what abotu schooltool.gradebook merges16:13
ignasand set up everything for it16:13
ignasaelkner: gradebook merges? i have merged douglas changes already16:13
ignasare there any others?16:14
Lumiereignas: including the section title change?16:14
ignaslooked at that now16:14
aelknermine from he sprint16:14
ignasand will have to fix it16:14
ignasaelkner: yeah, merged all DC work16:14
aelknerok, thanks16:14
LumiereI need an egg with that change that won't break cando semi-asap :)16:14
ignasLumiere: too much untranslatable inline html for my taste16:14
replaceafillignas, is it wrong?16:14
replaceafillignas, :O16:15
Lumierebefore dwelsh starts chasing me with a shotgun ;)16:15
ignasLumiere: I understand, it was not too high on *my* list of priorities ;)16:15
ignasreplaceafill: i16:15
LumiereI also apologize for the random email on it this weekend... he sent that before I could say I'll talk to ignas16:15
ignaswill copy the template to course/browser/remplates16:15
ignasLumiere: no problem, I am used to that ;)(16:15
ignasand make the title changes in the template16:16
ignasinstead of rendering html in python code16:16
replaceafillignas, Lumiere i guess we could use a schooltool.gradebook new egg, right?16:16
aelknerignas: is this the latest schooltool.gradebook trunk?16:16
Lumiereyes please16:16
aelkneri mean, is that what i should merge back to my branch?16:17
ignasaelkner: no16:17
Lumierereplaceafill: we need a gradebook egg and changes to cando to include it16:17
LumiereI guess16:17
Lumierebecause otherwise things are going to break16:17
Lumierereplaceafill: we're just going to need to be careful on pushing the release too fast ;)16:18
ignasaelkner: it's 0.2 release branch16:18
aelknerok, i found the regular schooltool.gradebook trunk16:18
ignasas soon as I get everything branched and buildboted16:18
aelknerit's not easy finding these branches in launchpad16:18
ignasI will get on with move to python2.516:18
ignasand try packaging things for jaunty16:19
ignasso we would know how many problems there are going to be16:19
th1aBetter now than later.16:20
aelknerignas: it looks like you didn't get all of the dc work in there16:20
ignasaelkner: in there being where?16:20
ignas here?16:20
aelkneri see replaceafill's external activities work16:20
aelkneryes, that's the trunk of schooltool.gradebook, right?16:21
aelkneri don't see my dc work16:21
Lumierelaunchpad isn't trunk of anything ;)16:21
Lumiereit's a mirror of trunk16:21
ignasaelkner: email me the branch16:21
ignasLumiere: yeah, mirror as of 4 hours ago16:21
aelkneri already did before, but ok16:21
ignasaelkner: that is missing16:21
ignassometimes emails get lost, i get a lot of them lately ;)16:22
* Lumiere knows that feeling... only so much you can do with a couple hundred emails in a weekend16:22
aelknerhow does one sort the branches so that one can find stuf16:22
*** jelkner has joined #schooltool16:22
Lumierehi jelkner16:22
aelknerlaunchpad just presents them all mixed up16:22
ignasyou can click on column headers16:23
ignasto sort them by time16:23
aelkneralphabettically would be ok16:24
aelknerso that all the schooltool-owners stuff os sorted together16:24
aelknerand all the ~aelkners and ~ignas too16:24
* jelkner is a bit hung over from celebrating the historic victory in el salvador last night!16:25
replaceafilljelkner, yay!16:25
jelkner(he also has to leave for work)16:25
jelkneri'll be back at 100% by the meeting tomorrow...16:25
jelkneri have to go now16:26
th1aCongratulations replaceafill.16:26
*** jelkner has quit IRC16:26
replaceafillth1a, :)16:26
aelknerignas: email sent16:27
ignasaelkner: thanks, will review it and merge it16:27
ignasas soon as I am done with section titles16:27
th1areplaceafill: Do you have an update?16:28
replaceafillyes just to say i met with welsh on friday16:28
replaceafillhe approved four of my fixes on the cando.message package16:28
replaceafillfound a new one, which i already fixed16:28
replaceafilli almost finish the last bug related to the cando.message package16:29
replaceafillhe told me his priorities for the remaining bugs, so i know where to go next16:29
Lumierereplaceafill: and... you wrote tests and fixed broken tests!16:29
replaceafillLumiere, :)16:30
replaceafillah and that too16:30
replaceafilli have no questions today16:30
th1athanks replaceafill.16:30
th1aLumiere: Anything else to add?16:30
th1aOK, I guess that ends the meeting proper.16:32
* th1a drops the bag of gravel.16:32
th1aaelkner & yvl: Can you continue sorting out the testing ideas?16:32
aelkneryvl: what do you think about my idea for a queryPDF?16:33
aelknerthat dumps the stories16:33
Lumiereth1a: I got what I needed from ignas ;)16:33
aelkneri mean there are only so many different types of reportlab objects that we'll use16:34
aelkneran we could just dump them with out own prefix and attribute output16:34
aelknerwith our16:35
yvllooking at queryXML16:35
aelknerkeep in mind that we don't need the '<' and '>' markup16:36
aelknerunless we wanted to come up with an xml version of pdf stories16:36
aelknerthat might be interesting to have16:37
yvldo we need XPath queries for output filtering?16:37
aelknerthat's an interesting idea16:38
yvlI was thinking something along the lines "extract the text from story objects"16:38
aelknerso if we convert pdf stories to xml16:38
aelknerthen we could use xpath to filter it16:38
aelkneram i on the same page as you?16:39
yvlI am not sure of benefits VS time16:39
yvl(time to implement)16:39
aelknerso the first step would be to convert the stories to xml16:39
yvlkeeping in mind that users want to write their own Reportlab flowables16:40
yvl(or we will some time in the future)16:40
aelknerwe don't need to worry about that until the future is now :)16:41
th1aBrittle tests in this case are probably not worth doing.16:41
aelknerbut just for argument sake, a custom flowable could have a custom tag16:41
aelknerand whoever writes the first one could put the custom part into place16:42
aelknerbut we don't one yet, so that part could be handled later16:43
yvlwait, I'm writing a small example of simple text output...16:44
yvl(another idea)16:44
ignaslisppaste5: url16:44
lisppaste5To use the lisppaste bot, visit and enter your paste.16:44
yvlthank you16:44
lisppaste5yvl pasted "Reportlab testing example" at
yvlthis is example from some of POVs tests16:46
yvl(not the ones alga and I talked about before)16:46
aelknerlooks useful16:46
aelknerkind of what i was suggesting prior to our xml discussion16:47
yvlthis could be stripped down even further...16:47
aelknerwithout it being xml though, we won't have the benefit of xpath16:48
*** whaddon has quit IRC16:49
aelknerbut i'm not sure we need to filter per se16:49
lisppaste5yvl annotated #77114 "simpler output" at
yvlwell it all depends on what we want to test:16:50
yvl1) just text / tables16:50
yvl2) text / tables + some layout (page breaks, etc.)16:50
aelkneri would include layout16:50
yvl3) all element layout16:50
yvl(and data)16:50
aelknerimages can be tested for existence without actually showing the image16:51
aelkneri would say that it should be possible to test for the existence of all elements16:52
aelknerbecause that is the closest to actually testing the rendered pdf16:52
yvlok, so we have two choices basically:16:54
yvldo an XML-style output and have benefit of xPath (and layout)16:54
yvlor do simple text output and have very human readable output (like table representation)16:55
ignasor both16:55
yvlor both16:55
ignasif you have a standard xml output16:55
ignasfor tables16:55
yvlgot it16:55
ignaswriting an xpath table formatter is easy16:55
ignasso you can xpath the table and render the table you need16:55
yvlso, going for XML ouptut... any objections? :)16:56
aelknernone here16:56
Lumiereimo test layout of text separately from text generation16:56
ignasLumiere: text generation will be done in unit tests most of the time anyway, we need xml to test how all the parts come together16:57
ignasand whether the text that was generated in some method that we unit tested16:57
ignasis actually present in the PDF16:57
*** replaceafill has quit IRC17:47
Lumierethat works18:06
*** fsufitch1 has joined #schooltool18:10
*** fsufitch1 has left #schooltool18:10
*** fsufitch has joined #schooltool18:24
*** fsufitch has left #schooltool18:24
*** balor has quit IRC18:44
*** mgedmin has quit IRC19:31
ignasLumiere: are you there?20:17
* Lumiere notes that he doesn't get audio alerts from IRC anymore... so I may not respond instantly to initial messages ;)20:19
ignasno problem20:19
ignasyou want the title fix in schooltool trunk20:20
ignasor schooltool 2008.1020:20
Lumiereyes please :)20:20
Lumiereoh... 8.10 I guess20:20
ignasgot it20:21
Lumieresec let me make sure20:21
Lumiereis that 8.10 or trunk?20:21
*** elarson_ has joined #schooltool20:22
ignasnot sure if cando is compatible with the newest merge though20:23
Lumiereignas: that's why I need to test20:23
ignasargh, why are you using trunk anyway20:23
Lumiereignas: good question20:24
ignas2008.10 had everything that trunk had up until the last merge20:24
ignasthat added custom demographics20:24
Lumierewell... maybe we should move to 2008.1020:24
*** jcrowley has joined #schooltool20:24
Lumiereit isn't that hard20:24
Lumiereit's just changing the egg20:24
Lumiere(I assume)20:24
Lumierewe really should sit down and try to work out workflow of eggs etc20:25
Lumierefor both sides20:25
Lumiere(especially as cando gets closer to svn2bzr and auto-building dpkg20:25
Lumieregiven that we're going to want to either build our own system like ST or use ST's system)20:26
Lumiereand we should try and get close to the same workflow to do that20:26
ignasif only it was that easy when it comes to versions of eggs20:26
ignasif you hardcode it in - you will have trouble developing (most probably)20:26
ignasso it is advised to set it in buildout.cfg20:26
ignasbut buildout.cfg can't handle "wildcard" versions20:27
ignasand wants precise versions20:27
Lumierecan we do >=?20:27
ignasin buildout - no20:27
Lumierethis sounds like a problem with buildout imo20:27
ignasyou can in one place, but it will be ignored most probably ;)20:27
Lumieremaybe it would be worth making alias'd eggs20:27
ignasbuildout will blame setuptools, or will suggest you rewrite the configuration file parser ;)20:28
Lumierefor schooltool-2008.10-latest20:28
ignasnope, you don't want the same version number point at 2 different things20:28
Lumiereand schooltool-1.0-latest20:28
ignaswon't work20:28
ignasbecause setuptools will not know if it should upgrade20:28
ignasargh, somehow I am managing to mess up my zope component cache, and I don't like it20:29
*** dwelsh has joined #schooltool20:32
dwelshth1a:  you there?20:32
*** elarson_ has quit IRC20:33
Lumiereignas: so... I guess we should figure out what the step is20:46
ignasLumiere: latest 2008.10 is schooltool-2008.10.9dev-r2473.tar.gz20:47
ignasit would be nice if you could try out different strings in buildout.cfg20:47
ignasand see if you can filter out and depend on20:47
ignasschooltoo-2008.10 , schooltool-2008.10.920:48
ignasor if that fails - at least on the precise version of schooltool egg20:48
ignaslater we can start thinking of something like automatically updated versions.cfg for schooltool releases or something like that20:48
ignasbut now switching to 2008.10 and then switching to the 2008.10 with the fix is what's important20:49
Lumierelet me try 2008.10.920:49
ignasschooltool = 2008.10.920:50
Lumiere*runs buildout -n20:50
*** ignas has quit IRC21:08
*** jcrowley has quit IRC21:08
*** alga has quit IRC21:34

Generated by 2.15.1 by Marius Gedminas - find it at!