IRC log of #schooltool for Thursday, 2010-09-23

*** menesis has quit IRC00:30
replaceafillaelkner, u around?00:44
*** replaceafill has quit IRC01:18
*** replaceafill has joined #schooltool01:38
*** replaceafill has quit IRC02:40
*** ignas_ has quit IRC04:08
*** alga has quit IRC04:09
*** replaceafill has joined #schooltool04:39
*** aks has joined #schooltool06:11
aksyvl: ping08:37
*** replaceafill has quit IRC09:32
*** menesis has joined #schooltool10:05
yvlpong, aks10:05
aksyvl: sorry was out for lunch10:27
aksyvl: see
aksit contains the repository containing the RPM spec files and sources, as well as the wiki contains the documentation10:28
yvlcool :)10:30
aksyvl: ok10:38
aksyvl: let me know if we require more documentation10:38
yvllooks good!10:41
yvlI'd personally prefer text from to be also included in the git repo10:41
yvlan, of course - scripts (however experimental) that download the sources10:42
aksyvl: yes, will do that in a while, cleaning up the mess in those scripts now10:45
yvlI'm not rushing you or anything, you did a great job :)10:45
aksyvl: Added the developer documentation to the repo as well10:56
yvlthanks :)10:57
aksyvl: one question - Do we have to build versus the trunk only or we do have stable releases?11:00
yvlsorry, didn't understand your question11:01
yvlwe have stable releases of 1.411:01
yvland once Maverick is out, we'll have stable releases of 1.511:02
yvl(this release was intended to be 1.6, so you can find eggs here at the moment: )11:03
yvlI hope this is what you were asking for :)11:03
aksyvl: yes11:04
aksyvl: the next thing I'll be packaging for schooltool would be version 1.5 for the latest build of Fedora (Fedora 13)11:04
aksyvl: what is the schedule for the release of next version (i.e 1.7)?11:05
yvlthe next release will be 1.611:13
yvland as we're aiming to get into Natty Universe repositories, it will match Ubuntu release schedule11:14
aksyvl: one problem, though in version 1.6, the files are labeled with 1.511:25
aksyvl: this shall create confusion11:25
yvlyes, it does11:26
yvlit's simply a wrong folder, and we didn't change it yet...11:26
yvlit is supposed to be 1.511:26
yvlapologies for the mess11:26
aksyvl: now i'm tweaking my build script to take argument of which version to build and hence fetches the source accordingly, it the version is "trunk" will do a bzr co/update, or if it is another version, pull the tar.gz file from the website11:27
aksyvl: and i'm forming the url as "" + schooltoolVersion, where the latter is the variable containing the version of the schooltool that is fetched by argparse11:28
*** alga has joined #schooltool11:28
aksyvl: so can we rename the directory to 1.5?11:29
yvlI'll ask menesis11:29
aksyvl: ok, let me know when it is renamed, and i'll re-test my build script against it, before pushing them to the repository11:29
yvlby the way, for trunk you might want to use
yvlthose are built by our buildbot11:31
yvlbuilbot also updates
yvl(you might want to use trunk eggs instead of bzr co, buildbot at least runs tests there)11:32
*** menesis has quit IRC12:12
aksyvl: ok12:37
aksyvl: can you manage to host the binary RPMS (in a repository structure)? It's 31M in size12:39
yvlI was about to write an email to Tom about that :)12:39
aksyvl: :)12:39
yvlah, and his nick is th1a, by the way :)12:39
aksyvl: oh yes, I remember now, I once had a talk with him12:39
*** menesis has joined #schooltool13:09
aksyvl, menesis: besides renaming the directory from 1.6 to 1.5, we also need to figure out how to find the most recent stable build inside the directory13:14
aksyvl, menesis: for example in, there is schooltool source from version 1.3.0 to 1.4.2, and my script needs a way to find which to download13:15
aksyvl, menesis: till now my script, asks for the version number (i.e. 1.3, 1.4, 1.5) which it appends to "" to find out the required directory, but while downloading the source for schooltool, it is confused13:17
menesisaks: ok I have renamed 1.6 to 1.513:37
aksmenesis: thanks, what about discovering the most recent build under the given version directory?13:37
menesiswe talked about how to name releases in the next development cycle, so it will be 1.613:38
yvlreleased in 2011 spring13:38
aksyvl: ok13:38
menesisbut the fact that 1.3.2 is in /1.4 is on purpose13:38
aksmenesis: so how would I discover that automatically using my script (python)?13:39
menesis1.3.x were development releases, 1.4.x stable13:39
aksmenesis: If we can maintain a text file containing the list of releases inside the version directory, I can parse the file, read out the contents, find the bigger one, and use it13:39
aksmenesis: the problem now is that when you browse that directory, it lists the contents of the directory rather than displaying a page, and hence I'm not able to use urllib to read the page contents and retrieve the releases13:40
menesisthis is a page generated by apache13:50
menesisyou should be able to parse it as well13:50
menesisbut I don't think it is a good idea to package the highest version directory13:50
menesisyou cannot know if it is stable yet, maybe modules have been renamed, added or deleted13:51
menesisso the script won't work after we create 1.6 directory for development releases sometime next month13:52
menesisbut I know that this directory layout is not perfect13:52
menesisI have to update the download directory in my deb packages13:55
yvlFWIW, urllib.urlopen('')13:55
menesisbut for stable releases it works13:55
menesisaks: maybe it would be enough to add a symlink stable -> 1.4 ?14:02
aksyvl, menesis, i'll give a try14:03
aksmenesis: yes that would be great14:11
aksmenesis, yvl: using urllib, i was able to get the contents, but it also contains a lot of markup, I'll use regex and filter out, and try to find the most recent stable release in a given directory by this approach14:11
aksmenesis, yvl: leaving for home now, will see you guys tomorrow14:16
yvlsee you, aks  :)14:16
menesisu = urllib.urlopen('')14:16
menesisfrom lxml import etree14:17
menesish = etree.HTML(
yvlaks, alternatively you can open the ftp link14:17
aksyvl: open ftp link? how?14:17
yvlreadlines will give you something along the lines of "ls -al"14:17
yvl<yvl> FWIW, urllib.urlopen('')14:17
aksmenesis: yes the etree would be a good way to do as well14:18
aksyvl: i'll try out a few alternatives tomorrow, and let you guys know of it14:18
aksmenesis: ^^14:18
aksyvl, menesis: let for example in 1.5 directory all the schooltool tar.gz files are stable, aren't they14:19
menesiswell yes they are manually released14:20
menesisbut for example schooltool 1.5.1 broke schooltool.intervention until a newer version of it was made14:20
aksyvl, menesis: as my current RPMs are based on trunk checkout, I now want to base my RPMs on stable releases so as to help others use the tool14:20
menesiscurrectly the stable series is 1.414:21
aksyvl, menesis: so I want you to suggest me the most stable build that I should package14:21
menesisbut 1.5 is now in bugfix mode for release with Ubuntu 10.10 next month14:22
aksmenesis: the 1.4 contains a lot of tar.gz for schooltool, which one of them should i use14:22
menesisso you can use 1.514:22
yvlI really suggest to package 1.5 :)14:22
yvl(and package it again in a month)14:22
menesisthe latest tar.gz14:22
aksmenesis: you mean 1.5.1? I've already packaged it (but from the trunk)14:23
aksmenesis: if i get a green signal, i'll package 1.5.1 tomorrow, and update the spec file for schooltool at the repo, the dependecy should be the same in the 1.5.1dev and 1.5.114:24
aksyvl, menesis: so that means I'll have to re-package and test just schooltool and it would not be a big issue for me anyways ;)14:25
aksyvl, menesis: you'll find the spec for 1.5.1 tomorrow at the repo14:26
aksyvl: btw did you hear from tom about hosting the RPMs?14:26
yvlhe's not awake yet :)))14:27
aksyvl, menesis: after building against 1.5.1, I'll put the repo temporarily online somewhere from where you can download and place it to a permanent location (yes, if Tom agrees)14:27
aksyvl: ;)14:27
aksyvl, menesis: though I run some manual tests over my builds, I'd like others to help with QA. Keeping the repository online would help others use, test, and report about it14:28
aksyvl, menesis: I really need to leave now, bye14:29
yvland thanks ;)14:29
aksyvl: :)14:29
*** aks has quit IRC14:29
*** alga has quit IRC15:43
*** alga has joined #schooltool16:21
*** menesis has quit IRC16:24
*** yvl has quit IRC16:24
*** yvl has joined #schooltool16:24
*** menesis has joined #schooltool16:24
*** replaceafill has joined #schooltool16:56
*** replaceafill has joined #schooltool16:56
*** aks has joined #schooltool17:00
*** th1a has joined #schooltool17:04
aksth1a, hello17:10
th1ahi aks.17:24
th1aSorry I didn't notice you there.17:24
aksth1a: that's ok17:30
th1aSo you've got a working RPM?17:30
*** aks has quit IRC19:01
yvlth1a, I did a little gradebook refactoring Douglas (most likely) needed for Cambodia19:01
yvland sent an email to you all :)19:02
yvlcan you please poke Alan and Douglas some time about it?19:02
replaceafillyay! refactoring!!!19:03
th1aThis would be for Natty?19:03
yvlno, instant - Maverick :|19:03
yvlas Cambodia would use that19:04
replaceafillnice! adapters!19:04
yvlit's not that big, and doesn't touch UI whatsoever19:04
yvltests pass.... but - you know19:04
yvlchange is always has a risk of bugs19:04
replaceafillyvl, will test it today19:05
yvl* change always has ...19:05
th1aAre there additional code changes aelkner and replaceafill need to do?19:05
yvlaelkner - no19:05
yvlhopefully - I haven't checked it against his code19:05
yvlreplaceafill would use the adapters to get the behaviour he wants19:06
yvlI didn't figure out an easier way to get to that19:06
yvlit isn't merged to trunk, and won't be unless our developers (and you) give the green light19:07
yvlreplaceafill, I forgot to mention one thing in the email19:09
replaceafillyvl, the way i figured was what we discussed: subclassing :(19:09
yvlyes, for now19:10
yvlbut of the linked activity19:10
yvlas opposed of the whole gradebook19:10
yvland several views19:10
th1aI'm just emerging from a nasty head cold, btw.19:11
replaceafillth1a, same here :(19:11
replaceafillnot that nasty though19:11
yvlwell - that's autumn for you.19:12
replaceafillyvl, i have a question19:12
yvlreplaceafill, you can get around without subclassing there if you want, but you'll have to implement the ILinkedColumnActivity interface19:13
yvlshoot :)19:13
replaceafillyesterday i noticed the average logic returns integers19:13
replaceafilland it loses precision19:13
replaceafilli mean the current logic19:14
replaceafillwill your branch replace that logic?19:14
replaceafilli see getWorksheetAverageScore for instance19:14
replaceafillah it uses gradebook.getWorksheetTotalAverage(worksheet, evaluatee)19:14
yvlyou can try replacing the results if you need19:15
replaceafillso, since the bug is still there heres my other question19:15
yvl(that's one of the fun things with adapters :) )19:15
replaceafillthat method returns total, average19:15
replaceafilltotal being a Decimal19:15
replaceafilland average an int19:15
replaceafilli was thinking if i should return float, float19:15
replaceafilland let the caller the responsability19:16
replaceafillof formatting and stuff19:16
replaceafillrounding, etc19:16
th1aYes, it is better to retain precision.19:16
* yvl agrees19:17
yvlI don't know how the views would react to such change19:17
replaceafillyvl, right19:17
replaceafillwill have to check19:17
yvland if anything, you can do your own average calculation19:18
yvlwith your own precision19:18
yvlthough I'd rather see it fixed in schooltool.gradebook19:18
replaceafilli was thinking of fixing gradebook core19:18
yvlsome time soonish :)19:18
yvlwell, one thing at a time, yes? ;)19:18
replaceafill1. change my view logic to use your approach19:19
replaceafill2. fix the rounding bug in gradebook trunk :P19:19
replaceafillthis is the one btw
* yvl started thinking today, that the score system part got a little bit abandoned for a while now19:21
yvlit needs some love19:21
yvllike - a score system that shows percentage with 'n' decimal point accuracy19:21
th1aUI needs love more.  ;-)19:21
yvland making views and other code use those score systems instead of magical ints19:22
yvlth1a, you are right, as usual :)19:22
yvlok guys19:22
yvlget better!19:23
th1aGetting there...19:23
yvlreplaceafill, don't hesitate to blame if something is not working right19:23
replaceafillthanks yvl, will flood your email then :P19:23
yvlthank you ;)19:23
replaceafillwith questions/requests/complains/etc19:23
replaceafillthanks yvl19:23
yvlsee you all tomorrow :)19:24
*** aks has joined #schooltool19:37
aksth1a: sorry got disconnected at that time19:38
aksth1a: yes, I've got a working set of RPMs19:38
th1aNo problem.19:38
th1aExcellent news.19:38
th1aWe should be able to host them once we get everything sorted.19:38
*** menesis has quit IRC19:39
th1aLet us know what else you need -- particularly reports.19:39
aksth1a: what sort of reports?19:40
aksth1a: the current RPMs is compiled against trunk version, tomorrow I'll be packaging agains version 1.5.119:40
th1aWhatever kind of reports the schools need to generate for parents, data analysis, government, etc.19:40
aksth1a: hmm... we'll be launching out a pilot test in two of the schools in october, will inform you about the progress and/or additional feature requirements19:41
th1aYes, please.  Keep us informed as much as possible.19:41
aksth1a: depending upon the outcome of the pilot test, we'll be going out with a overall implementation at all our OLPC deployment sites/schools19:41
th1aWe're a little different than most open source projects in that we have paid staff!19:41
th1aHow many OLPC schools do you have now?19:42
aksth1a: oh ok19:42
aksth1a: we're currently doing with 26 schools19:42
th1aOK... I have to go pick up my daughter at school.  Thanks for the good news aks!19:43
aksth1a: :)19:44
*** aks has quit IRC20:28
*** menesis has joined #schooltool21:04

Generated by 2.15.1 by Marius Gedminas - find it at!