Time  Nick      Message
00:05 rambutan  what sporting event is being discussed? 9-11 would seem to not be soccer, not basketball
00:07 tcohen    a tennis tie-break
04:05 bgkriegel bye
04:25 mtompset  Have a great day, #koha.
08:30 cait      good morning #koha
08:56 cait      bgkriegel++
09:01 cait      signed off 46!
09:02 * cait    feels a little like that guy with the hill and the stone at times
09:02 cait      but i had asked for it...
10:40 cait      native speaker around?
13:32 rambutan  It's morning.
13:33 cait      goodmornign rambutan :)
13:35 cait      jcamins helped me out already - jcamins++
13:56 rambutan  I thhink you and jcamins must be on #koha 24/7
13:57 cait      I am not going to comment on this :)
13:58 rambutan  I should go check on my pancakes, brb
14:29 rambutan  jenkins?
14:29 wahanui   somebody said jenkins was http://jenkins.koha-community.org/
14:32 rambutan  image of jenkins reminds me of Jeeves (askjeeves)
14:38 rambutan  How does the distributed build work in Jenkins?
14:39 cait      rambutan: I think i am the wrong one to ask this :)
14:39 cait      as far as i unddrstand it... there are different servers doing different builds and reporting back?
14:40 rambutan  yea, but what in Jenkins actually sets this up? I'm reading the docs, but haven't found anything yet about
14:40 rambutan  distribuiting the build over difference CPUs
14:41 cait      i think one build is done by one server
14:41 cait      another does the test suite iwth mariadb and so on
14:42 rambutan  OK, so it's not really a distributed task manager then.
14:42 cait      not sure
14:42 rambutan  More like Boinc then
14:44 cait      to me it looks like both are very different software
14:45 rambutan  yea, but conceptually the functionality must be analogous
14:45 rambutan  this is interesting - trigger builds by email
14:45 rambutan  jenkins-foo: "|/bin/wget -o /dev/null http://YOURHOST/jenkins/job/PROJECTNAME/build"
14:48 rambutan  wonder if we're using mavin
14:48 rambutan  Maven
14:48 rambutan  Ah Ha! Eureka! https://wiki.jenkins-ci.org/display/JENKINS/Distributed+builds
14:49 rambutan  Wonder who hosts Jenkins Master?
14:49 cait      rangi does
14:50 rambutan  ah
14:50 rambutan  appears quite simple to add/manage slaves
14:52 rambutan  of course, jenkins is all java, there must be just a little twitch factor there :)
14:55 rambutan  OK, so we're not compiling Perl, so what does a "build" actually do?
14:56 jcamins   rambutan: it runs the complete installation process and all the unit tests.
14:56 rambutan  ic ic
14:56 jcamins   And a sub-task generates a package.
14:56 rambutan  so that means it's running a shell script
14:57 jcamins   continuous integration?
14:57 rambutan  OK, this is easy as cake
14:57 jcamins   continuous integration is http://wiki.koha-community.org/wiki/Continuous_Integration
14:57 jcamins   That has the build script.
14:59 rambutan  Wish I could have attended that (Chris') presentation
15:40 rambutan  @later tell rangi do VM's work OK as Jenkins slaves, or is bare metal preferred?
15:40 huginn    rambutan: The operation succeeded.
15:55 rambutan  @later tell rangi How many Jenkins slaves are we currently running?
15:55 huginn    rambutan: The operation succeeded.
17:21 cait      bgkriegel++ :)
17:22 bgkriegel now for what? :)
17:23 bgkriegel thx :)
17:24 bgkriegel you have been very busy! cait++
17:24 cait      not busy enough :)
17:24 cait      still 13 enh/new features remaining
18:26 cait1     t
18:26 bgkriegel :)
18:26 bgkriegel done 10854
18:28 cait      bgkriegel++ :)
18:49 bgkriegel cait: 10969. Russian file has a TRUNCATE, but can't be truncated because foreign key constraint, so 1) remove truncate or 2) replace with delete 3) lift constraint for that sentence
18:56 cait      bgkriegel: hm
18:56 cait      you are right - i just kept it how it was
18:56 cait      maybe comment it?
18:56 cait      it seems quite dangerous
18:56 bgkriegel If they want truncate, then i go for 3)
18:56 cait      ok
18:56 bgkriegel just disable, truncate and enable
18:57 cait      ok
19:10 rangi     morning
19:10 rangi     2 new devs in september
19:10 cait      morning
19:11 rangi     i should do a patch for that
19:12 rangi     @later tell rambutan VM work fine, we have 2 already, i still need to configure bgkriegel's one to do a bit more, galens one is running master against mariadb, i can show you how at kohacon
19:12 huginn    rangi: The operation succeeded.
20:09 mtj       hey rangi, have you got round to looking at travis.cli.org ?
20:09 rangi     yep, it's good as far as proprietary SaaS goes
20:10 mtj       s/looking/look/
20:12 jcamins   rangi: it looks like their code is on Github.
20:12 mtj       hmm, i actually thought it travis was FOSS :/
20:12 rangi     oh cool
20:12 jcamins   MIT license.
20:12 wahanui   well, mit license is compatible, but upgrading to GPLv3+ is a good idea anyway.
20:12 rangi     i might be confusing with coveralls
20:13 jcamins   rangi: https://github.com/travis-ci/travis
20:13 rangi     mtj: it seems fine, doesnt do anything jenkins doesnt
20:13 rangi     its integration with github is nice, if you like github
20:13 rangi     but there is a plugin to do the same for jenkins
20:15 mtj       i started experimenting with travis coz it was quick, free and i couldnt spare a jenkins box
20:17 mtj       i think jenkins does a bunch more stuff than travis
20:18 mtj       i liked getting a clean VM for every build in travis….  tho i'm sure jenkins could be set up to do the same thing
20:19 rangi     yep
20:19 jcamins   mtj: easy peasy. Worst case scenario, you use Vagrant.
20:19 jcamins   jenkins can run a bash script that starts Vagrant.
20:19 rangi     https://wiki.jenkins-ci.org/display/JENKINS/Vagrant+Plugin
20:20 rangi     even easier
20:20 mtj       touche
20:21 jcamins   Cool.
20:22 mtj       is it easy to get different OSs booting in vagrant?
20:22 rangi     havent tried, but i cant imagine it is
20:22 mtj       ie: as a vagrant 'box'
20:23 mtj       looks like it might be easy peasy
20:24 jcamins   mtj: I only use it with Ubuntu.
20:24 jcamins   But it's dead simple with Ubuntu.
20:25 mtj       woah! -> http://www.vagrantbox.es
20:25 jcamins   mtj: https://github.com/biblionarrator/biblionarrator/blob/master/Vagrantfile and https://github.com/biblionarrator/biblionarrator/blob/master/bootstrap-deb.sh
20:26 mtj       awesome, i've been looking for something like vagrant for a while…
20:26 jcamins   That script doesn't actually work because Cassandra's packaging is broken.
20:27 rangi     vagrant + chef = awesome
20:27 jcamins   It did for a while, though.
20:27 rangi     mtj: you should look at docker.io too its pretty neat
20:27 mtj       i failed badly trying to get ubuntu VMs booting on a debian XEN system, a few weeks ago
20:27 rangi     different use cases i think
20:27 rangi     but still neat
20:28 jcamins   Yeah, very different. But both very cool.
20:29 mtj       rangi, yep thanks - docker.io was next on my TODO list
20:29 mtj       ( i saw it mentioned on #koha recently - and re-remembered it )
20:30 rangi     mtj: travis + coveralls = https://coveralls.io/r/ranginui/ncip-server-chris
20:30 rangi     which is kinda fun
20:31 rangi     Devel::Cover::Report::Coveralls
20:31 rangi     cover -test -report coveralls
20:34 mtj       very cool :)
20:34 mtj       i had some success recently, testing Koha install-docs on travis…   https://travis-ci.org/KohaAloha/koha-install-docs
20:35 rangi     cool
20:35 mtj       was nice to 'prove' that some instructions actually worked...
20:35 rangi     i really need to get jenkins to mark a build failed
20:35 rangi     if coverage drops
20:37 cait      :)
20:37 cait      +1
20:37 * rangi   grabs coffee brb
20:37 mtj       hmm, perhaps the 'cover' test could be added to the qa-tool.pl even...
20:38 mtj       like the perl-critic test
20:38 mtj       (might be a really slow test tho)
20:39 jcamins   That would be an incredibly slow test.
20:39 mtj       yeah - i thought it might :p
20:39 jcamins   You'd have to run the entire test suite twice.
20:40 cait      another slave?
20:40 cait      i mean jenkins could do it
20:40 jcamins   cait: no, I meant with koha-qa.
20:40 cait      and you could test coverage only on the files changes
20:40 cait      changed
20:40 cait      not so slow then maybe
20:41 jcamins   cait: the problem is you have to run every single test to find out which ones include coverage of the changed files.
20:45 cait      hm yeah you are right
20:46 rangi     it shouldnt fail qa
20:46 rangi     nessecarily
20:46 mtj       jcamins, thanks for the vagrant examples, they are just what i need :)
20:46 rangi     but jenkins can make the build fail and highlight coverage dropped
20:48 * jcamins contemplates tracings.
20:48 * cait    runs the qa script on random bugs because she is really too tired to continue testing
21:26 eythian   hi
21:26 cait      hi eythian
21:26 cait      how are you?
21:26 cait      :)
21:27 eythian   Daylight saving changed, so slightly jetlagged :)
21:27 cait      oh did it?
21:27 cait      what time is it for you now?
21:27 cait      11 hours now? that always confuses me
21:30 wizzyrea  10:30 here
21:30 eythian   yeah, that
21:30 wahanui   well, yeah, that is cool - a way to enhance the OPAC without bloating Koha...
21:31 cait      eythian: i was wondering if you could look at 2 bugs for me?
21:31 cait      bug 10733 is all packages
21:31 huginn    04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=10733 enhancement, P5 - low, ---, tomascohen, Signed Off , Memcached on package installs
21:31 cait      and i wonder if we need additional changes for bug 10729
21:31 huginn    04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=10729 enhancement, P5 - low, ---, colin.campbell, Signed Off , Support phrases-icu.xml in install and upgrade
21:37 cait      eythian: are you hiding now? :)
21:37 eythian   just distracted :)
21:37 eythian   I'll try to have a look today.
21:37 cait      that would be awesome, thx
21:44 jcamins   So it turns out there are some disadvantages to a discovery layer like Biblionarrator.
21:44 cait      oh?
21:44 jcamins   When you do a search you get all the really relevant results that have been cataloged *horribly*.
21:44 jcamins   And all you can say is "oh my god, I hope I didn't write that."
21:45 jcamins   While trying to ignore the sneaking suspicion that it may have been one of your first records.
21:45 cait      ah
21:45 cait      that happens with people using koha too
21:45 cait      when they first try the publication date search :)
21:45 cait      and other searches
21:45 jcamins   lol
21:46 cait      9190... that doesn't look quite right
21:46 jcamins   True, but the really horribly cataloged records don't show up in searches, usually.
21:47 jcamins   This record...
21:47 * jcamins sighs.
22:03 cait      good night :)
22:11 jcamins   @marc 952
22:11 wahanui   The 952 field is used by Koha to store item data in MARC21 and is described at http://wiki.koha-community.org/wiki/Holdings_data_fields_%289xx%29
22:11 huginn    jcamins: unknown tag 952
22:41 jcamins   Good morning, cait.
22:41 cait      *hides*
23:41 jcamins   rangi: hey, I don't suppose you know of anyone who has collected statistics on which indexes are actually used on the OPAC?
23:42 rangi     not me
23:44 jcamins   I computed it for one library, and it matched what I expected: 78.61% keyword, 10.09% author, 4.83% title, and 3.8% subject.
23:45 jcamins   The other 9 indexes used on the OPAC make up the remaining 2.67% of searches.
23:47 rangi     sounds what id expect
23:48 jcamins   Yeah, but I'd be interested in sanity checking those results, to confirm that it is as consistent as I think.
23:49 wizzyrea  i'd guess it has a lot to do with what kind of library, and what kind of search they set as default
23:49 jcamins   Anti-confirmation bias, don't you know.
23:49 wizzyrea  but yea, that seems like what I'd expect
23:49 rangi     ill see if i can pull some stats
23:50 jcamins   wizzyrea: this was a specialized research library.
23:50 wizzyrea  default set to keyword?
23:50 rangi     that makes me think
23:50 jcamins   Yep.
23:56 jcamins   A study at Nanyang Technical University has keyword 68.9%, title 16.5%, author 8.2%, and subject 6.4%.
23:57 rangi     we should totally make a way to export those stats, anonymised
23:57 rangi     so that libraries can click and export and then we can collect them
23:59 jcamins   Yeah, that'd be useful.
23:59 jcamins   Well, interesting.
23:59 wahanui   interesting is sometimes good and sometimes bad