Time Nick Message 00:00 eythian hi tcohen 00:00 * tcohen accidentally deleted Ender's game copy 00:00 tcohen hmm 00:02 * tcohen thinks is not a big loss, though 00:02 rangi heh 00:03 * tcohen should grab a Xenoxide's copy to read this weekend 00:03 tcohen what are those places where people pick books named? 00:04 eythian Bookpickeries 00:05 tcohen ah, that 00:05 wahanui ah, that is, like, just to say that everything should be on one line 00:06 tcohen should check if there exists some system they use to xpose their catalog, online 00:07 tcohen rangi: i found eventbase to have a limited functionality 00:07 tcohen at least the free edition 00:07 tcohen will complete the schedule there 00:07 rangi ahhh the old freemium trick eh 00:08 tcohen it is uncomfortable to set, but its pretty neat once you see it on the mobile app 00:08 rangi is eventbrite any better? 00:08 rangi sweet 00:09 tcohen if evenbrite has a way to change the user's password, that's a win heh 00:09 rangi hehe 00:09 tcohen will try evenbrite tomorrow with bgkriegel 00:11 rangi ive never tried to set up either so i have no idea 03:15 eythian @wunder nzwn 03:15 huginn eythian: The current temperature in Wellington, New Zealand is 15.0°C (3:00 PM NZST on September 23, 2014). Conditions: Clear. Humidity: 63%. Dew Point: 8.0°C. Pressure: 29.86 in 1011 hPa (Falling). 04:22 rangi aleisha: 12281, 12959, 12893, 12775, 12944 04:52 rangi http://www.pressreader.com/bookmark/EGG26VFDWDV/ does this url work? 04:58 ibeardslee seems to 05:13 rangi eythian: does your elasticsearch code use teh SearchEngine syspref anywhere? 05:13 eythian yep 05:14 rangi thats the only bit of 12538 that overlaps i think 05:14 rangi ill sign it off, and note that 05:14 rangi its just a database update taht removes that variable 05:14 eythian I expect there will be quite a few conflicts. 05:15 eythian I'm planning to manually apply the patch and revert back the bits I want to keep. 05:15 rangi sounds good 05:21 cait aleisha? 05:22 cait i saw your comment on bug 12281 - i think maybe easier to download some records that are cataloged that way 05:22 huginn 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=12281 normal, P5 - low, ---, oleonard, Needs Signoff , RTL display for records using 880 broken in Bootstrap 05:22 cait [off] an example would be this: https://hfjs.bsz-bw.de/cgi-bin/koha/opac-detail.pl?biblionumber=30613 05:24 aleisha thanks cait i'll try that! 05:24 cait thx for taking a look :) 05:25 cait it's something not a lot of libaries use, but quite important to this one :) 05:31 aleisha so then what do i do after I've imported it into the catalog? 05:31 cait hm wait until it's indexed an dthen go to the detail page 05:32 cait it should look like in the catalog link, without the patch the script all runs to gether and is not on the other side 05:32 rangi i wonder how often the sandboxes index? 05:32 cait rangi: hm not sure 05:32 cait for the result list it needs to be reindexed to see it... else we could change an existing record 05:32 rangi you can get to the detail page without it being indexed though eh? 05:33 rangi ah right 05:33 cait the problem is on result too, but you'd see the effect on deail immediately i think 05:33 cait yo could use the link from th estaged import to get there 05:34 aleisha the link is clickable but doesn't take me anywhere 05:34 cait hm there should be 2 links i think, one to open the record as marc (from the title) and the biblionumber at the end of the table 05:35 cait the last should open in a new window i think 05:37 * magnus_away waves 05:39 aleisha cait is this what it's supposed to look like? http://pro.test14.biblibre.com/cgi-bin/koha/catalogue/detail.pl?biblionumber=9 05:39 magnuse rangi++ x100 for signing off bug 11401 05:39 cait i think it might index every 10 minutes if we are lucky 05:39 huginn 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=11401 enhancement, P5 - low, ---, magnus, Signed Off , Add support for Norwegian national library card 05:40 cait aleisha: sec, i have to look up the password 05:40 rangi magnuse: no worries 05:40 cait ah 05:41 cait you need to turn on the xslt display - i thougth that was on by default 05:41 rangi cait: only on new installs i think 05:41 cait inthe system preferences - look for xslt 05:42 cait rangi: makes sense 05:43 aleisha ah yes that's more like it 05:44 aleisha i can sign that one off now :) 05:45 cait cool :) 05:45 magnuse aleisha++ 05:45 cait thank you! 05:45 cait aleisha++ 05:58 magnuse at opl we are thinking about creating a standardized dev environment to make it as easy as possible to get up and running with development/testing patches. something like using vagrant to create a virtualbox with a working dev install, git bz, the qa tools etc. anyone else considered something like that? or interested in it? 05:59 eythian it seems like a good idea. 06:00 eythian That's about as much as I can contribute :) 06:03 ashimema good morning 06:05 ashimema magnuse.. what extra's are you envisioning getting from that vs the sandboxes? 06:05 ashimema anything beyond the ability to test a few extra patches that the sandboxes don't cope well with? 06:08 magnuse well, we'd use it for development too 06:09 magnuse the main point would be to be able to have a um, "pristine" environment to work in, quickly 06:09 ashimema in which case, I'm all for it.. 06:09 ashimema I do similar with a bash script at the moment.. 06:10 magnuse when i had trouble with my gitified setup the last couple of weeks it was caused by something i tested, it would have saved me a lot of time if could just have scrapped my current setup and got a new one with a couple clicks 06:10 magnuse ashimema: cool 06:10 ashimema once you have a debian installed with a root account, I run my script and it adds appropriate users, grabs koha master, installs it, sets up git-bz, qa tools, vim plugins etc.. 06:10 ashimema though I think at the moment it's set to do a gitified dev env.. 06:11 ashimema I'm tempted to swap back to a non-gitified 06:11 magnuse ashimema: sounds good 06:11 magnuse yeah, i cant quite decide between gitified and a tradtitional dev install either 06:11 ashimema I could sanitize my script and stick it on github if you like? 06:11 magnuse that would be awesome! 06:12 ashimema gitified is certainly quick and easy to get up and running.. 06:12 ashimema but it deffo has it's flaws still 06:12 magnuse yup 06:13 magnuse especially for testing things like zebra config, right? 06:13 ashimema yeah.. gitified is a bit of a mare for that sort of stuff 06:14 ashimema gets in the way rather than making things easier. 06:15 magnuse whereas with a trad dev install you can do make upgrade 06:15 ashimema yup.. 06:15 ashimema it's easy to get out of sync with a gitified install.. 06:16 ashimema I've often run into issues where my background scripts are running off one set of code at one version.. and the rest is running off the clone.. 06:16 ashimema and it leads to all sorts of wierdness 06:16 * magnuse can only imagine 06:24 ashimema https://github.com/mrenvoize/koha-dev 06:24 ashimema it's a hackish script as it wasn't ever really intended for anyone to see it ;) 06:24 ashimema but it works ;) and has been slowly evolving for a while now. 06:25 ashimema that is the gitified version.. seems the non-gitified version ahs dissapeared into history somewhere.. 06:25 ashimema I remember it taking allot more steps to work through with more human interactions.. 06:25 ashimema feel free to take it a buthcer it thoguh ;) 06:26 magnuse ashimema++ for sharing! 06:28 cait byeall 06:29 cait no work today, but will try to be here for the meeting :) 06:29 magnuse eythian: i seem to recall you saying it is a good idea to have a dedicated virtualbox for building packages, because if anything goes wrong, it can go really wrong? 06:29 ashimema I have a dedicated machine for building and hosting my packages.. 06:30 ashimema so far so good ;) 06:30 ashimema I think eythian goes a step further and has a dedicated place to build them.. then he uploads to a repo to server them 06:31 magnuse a dedicated place, like a separate building or something? 06:31 * magnuse imagines eythian's underground package building bunker 06:32 ashimema haha,. 06:36 * ashimema likes irc channels where the lead dev hangs out.. #mojo is great 06:39 reiveune hello 06:40 magnuse ooh some nice comment spam on http://koha-community.org/ 06:43 magnuse @later tell gmcharlt is it too late to fix bug 12883 "Holds tab on checkout screen never stops processing" in 3.16.4? 06:43 huginn magnuse: The operation succeeded. 06:50 alex_a bonjour 07:13 Joubu hello 07:15 magnuse bonjour Joubu 07:15 magnuse @wunder boo 07:15 huginn magnuse: The current temperature in Bodo, Norway is 3.0°C (8:50 AM CEST on September 23, 2014). Conditions: Clear. Humidity: 70%. Dew Point: -2.0°C. Windchill: -2.0°C. Pressure: 30.06 in 1018 hPa (Steady). 07:15 magnuse @wunder marseille 07:15 huginn magnuse: The current temperature in Realtor, CABRIES, France is 15.1°C (9:05 AM CEST on September 23, 2014). Conditions: Clear. Humidity: 77%. Dew Point: 11.0°C. Pressure: 29.89 in 1012 hPa (Rising). 07:15 magnuse hm... 07:15 ashimema @wunder stevenage, uk 07:15 huginn ashimema: The current temperature in Pin Green, Stevenage, United Kingdom is 10.9°C (8:10 AM BST on September 23, 2014). Conditions: Mostly Cloudy. Humidity: 95%. Dew Point: 10.0°C. Pressure: 30.15 in 1021 hPa (Steady). 07:21 ashimema bracing day there magnus 07:39 magnuse ashimema: yeah, i think winter might be approaching... there was fresh snow in the mountains yesterday 07:41 fridolin Hie all 08:21 Joubu too bad, nobody updated the dbic wiki page :-( 08:21 ashimema ? 08:21 ashimema what page Joubu? 08:22 Joubu ashimema: I sent an email yesterday on the ml 08:22 Joubu ashimema: http://wiki.koha-community.org/wiki/DBIC_discussion_overview 08:22 ashimema ah.. not read mail yet.. sorry 08:24 Joubu gmcharlt, rangi, khall: It would be awesome to see your points of view/notes added to the wiki page I created today (http://wiki.koha-community.org/wiki/DBIC_discussion_overview). I think we need a quick overview of our problems for the meeting (this afternoon) 11:57 oleonard Hi #koha 12:44 * magnuse gotta run 12:44 * ashimema needs some lunch 12:50 tcohen morning! 12:52 oleonard Hi tcohen 12:54 ashimema hello 12:54 tcohen i haven't heard from galne 12:54 tcohen galen, or kyle 12:54 tcohen once i get their feedback 12:55 tcohen we will coordinate releases 12:55 fridolin ok 12:56 ashimema does anyone know if the upgrade path between 3.14 and 3.16 has been fixed.. hat nasty db issues 12:56 ashimema i colleague came accross is again this morning in 3.16.3 12:56 fridolin ashimema: you mean pb with Bug 7372 ? 12:56 huginn 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=7372 enhancement, P5 - low, ---, jonathan.druart, Pushed to Stable , RoadTypes should be folded into Authorised Values 12:57 ashimema that was the one ;) 12:57 Joubu ashimema: I submitted a patch on it 12:57 ashimema Thanks Joubu 12:57 Joubu tcohen: you should have a look at this patch 13:16 oleonard There should be a check in the qa tools for "javascript:void(0)" 13:17 cait oleonard: can you tell a bit more about it? i can look out for it 13:18 oleonard '<a href="javascript:void(0)" onclick=...' It's usually used as a way to fake a "return false" on a link click 13:18 oleonard It's an obsolete way of doing things 13:18 cait Joubu: could you add a check? 13:41 indradg good evening #koha 13:51 oleonard So Bug 12755 didn't exist when I tested and signed off on Bug 10558, but now it does... 13:51 huginn 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=12755 major, P5 - low, ---, gmcharlt, NEW , MARC Preview doesn't always display in managed MARC record 13:52 huginn 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=10558 enhancement, P5 - low, ---, kyle, RESOLVED FIXED, Convert records table in manage-marc-import.pl to ajax DataTable 13:56 indradg is the mailman list manager having issues?? 13:57 indradg I seem to be missing a few mail lately 14:02 oleonard indradg: I regularly find Koha list messages marked as spam. Could that be it? 14:04 oleonard Ack, Bug 12755 is MY fault! 14:04 huginn 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=12755 major, P5 - low, ---, gmcharlt, NEW , MARC Preview doesn't always display in managed MARC record 14:04 * oleonard falls on his sword 14:04 cait noooooooo! 14:05 * cait jumps between oleonard and the sword 14:05 cait we can't lose you... or koha would turn out as ugly as.... 15:00 * cait waves 15:00 reiveune bye 15:03 cait hmmm 15:03 cait isn't it time? 15:03 tcohen it is 15:03 ashimema i beleive it is. 15:04 * tcohen is trying to find the meeting commands 15:04 ashimema hehe... 15:04 ashimema we've not used the proper vote syntax for a while.. 15:04 cait you are looking for #startmeeting topic 15:05 ashimema just a random observation.. I can never remember the syntax for it 15:05 cait ashimema: same here :) 15:05 tcohen #startmeeting Development IRC meeting, 23 September 2014: DBIC 15:06 huginn Meeting started Tue Sep 23 15:06:00 2014 UTC. The chair is tcohen. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:06 huginn Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:06 huginn The meeting name has been set to 'development_irc_meeting__23_september_2014__dbic' 15:06 gmcharlt good morning 15:06 tcohen #topic Introductions 15:06 wahanui #info wahanui, a bot that has become sentient 15:06 cait good morning gmcharlt 15:06 tcohen please introduce yourselves using #info name 15:06 ashimema #info Martin Renvoize, PTFS Europe 15:06 cait #info Katrin Fischer, BSZ, Germany 15:06 tcohen #info Tomas Cohen Arazi, Universidad Nacional de Cordoba 15:06 ColinC #info Colin Campbell PTFS Europe Ltd 15:07 ashimema do we need to ping ribasushi at some point ;) 15:07 gmcharlt #info Galen Charlton, Equinox 15:07 bgkriegel_ hi 15:07 bgkriegel_ #info Bernardo Gonzalez Kriegel 15:07 Joubu #info Jonathan Druart, BibLibre 15:08 tcohen ok 15:08 jcamins #info Jared Camins-Esakov, C & P Bibliography Services 15:09 tcohen khall? 15:10 tcohen ok 15:10 tcohen This is an unusual meeting 15:10 tcohen to talk about a specific subject 15:10 tcohen #topic RM introduction 15:11 oleonard #info Owen Leonard, Athens County Public Libraries 15:11 tcohen The main subject is our adoption of DBIC, and how we do it 15:11 tcohen #link http://wiki.koha-community.org/wiki/DBIC_discussion_overview 15:12 tcohen the main discussion is related to how much stuff we expect DBIC to provide, and how to do it 15:12 tcohen there are mainly two discussed POV 15:13 tcohen on the one hand (a) benefit from DBIC magic, and put (some degree) of business logic into it 15:14 tcohen and on the other (b) relying on the ORM only for abstracting the DB and using it to implement our own API in which business logic should be put 15:14 tcohen this issue is related to point (1) on th elink 15:15 tcohen I've already expressed that my personal POV is that we should keep the current way of using DBIC 15:15 tcohen i.e. using it inside modules, that can be unit tested 15:16 fridolin +1 for UT 15:16 gmcharlt hmm 15:16 gmcharlt I think that point perhaps could be clarified - may I go through a somewhat lengthly example? 15:17 tcohen please 15:17 ashimema go for it 15:17 gmcharlt OK, so when we are talking about business logic, there are at least two ways in can manifest itself 15:17 fridolin point 1 : i think since there are aleardy modules nammed nearly like a db table (ie C4/Borrowers.pm), I'll vote for a) 15:18 gmcharlt umm, may I please have the floor 15:18 fridolin yep sorry 15:18 tcohen :) 15:18 gmcharlt so, one way business logic can manifest is as attributes of the underlying entities that we're dealing with 15:18 gmcharlt so take, for example, item records 15:19 gmcharlt one way of modeling an item record is to follow the DB schema closely 15:19 gmcharlt on the assumption that we've done a reasonably good job with our relational model 15:19 gmcharlt which, in many places, we have 15:19 gmcharlt in others, we haven't 15:19 gmcharlt but let's stick with items for now 15:20 gmcharlt one thing we could say is that, OK, a Koha::Schema::Result::Item is a decent base class for an item record 15:21 gmcharlt but that it doesn't /exactly/ give everything we need to ask of any item record 15:21 gmcharlt and the big example is ->effective_itemtype 15:22 gmcharlt there's a bit of business logic there, as of course the effective item type for circ polices depends on the item-level_itypes system preference 15:22 gmcharlt but it's somethat that can be well-encapsulated as (effectively) an attribute of K::S::R::Item 15:22 gmcharlt and there's a lot going for that approach 15:23 gmcharlt 1. it is definitely testable -- there's nothing stopping unit tests from being written on K::S::R::* objects 15:24 gmcharlt 2. it gives a *single* place to ask the question - "what is the effective item type of an item?" 15:24 gmcharlt so in that respect, I quite like it 15:24 gmcharlt next, I'm going to move on to an example where things are less clear-cut 15:24 gmcharlt namely, bibs 15:25 gmcharlt so, as everybody here knows, you need to look at both biblio and biblioitems to get all of the info regarding a bib entity 15:25 gmcharlt (and of course, that represents drift from Koha's original quasi-FRBR model, but that's in the past, water under the bridge, etc.) 15:26 gmcharlt but the upshot - this is an example where the current relational data model is flawed for our current purposes 15:26 gmcharlt as if you start from a K::S::R::Biblio object ... you have to go through some convulations to get at everything you might want to ask of a bib record object 15:27 gmcharlt now you /could/ do things like add K::S::R::Biblio->marc to grab the MARC record 15:27 gmcharlt and so on... but that's stretching things 15:27 gmcharlt and it's vulnerabile to cases where we would want to redo the data structure 15:27 gmcharlt (which, we probably want to do in the long run) 15:28 gmcharlt so there, I think the situation calls for (say) a Koha::Biblio object that can be *composed* of (say) a K::S::R::Biblio and a K::S::R::bibioitems object 15:29 gmcharlt but if we right such a class ... it would be useful to delegate as much to the underlying K::S::R attribute methods as possible 15:29 gmcharlt partcularly when we want to start traveling to related objects like items 15:30 gmcharlt so, now I'm going to move on to my third (and final) example 15:30 gmcharlt and that's business logic that pretty much doesn't naturally look like an attribute of an object 15:30 gmcharlt for example: the question of calculating the due date of a new loan 15:31 gmcharlt now you can envision procedural code 15:31 gmcharlt e.g., my $due_date = Koha::Loans::get_due_date($patron, $item) 15:31 gmcharlt or a more OO approach 15:31 gmcharlt my $loan = Koha::Loans->new($patron, $item) 15:32 gmcharlt where, in effect, the constructor of that object does things like calculate the due date 15:32 gmcharlt (and throw exceptions if the loan is against policy) 15:33 gmcharlt but however you do it, the upshot is that it wouldn't be a good idea to add lots of methods to Koha::S::R::Issue to try to do all of this 15:33 gmcharlt but rather, build a new class 15:33 gmcharlt (or build procedural code) 15:33 gmcharlt so what I've done so far is provided three examples of the sort of business logic we have to deal with 15:33 cait thx gmcharlt - that#s really helpful 15:33 gmcharlt next, I'm going to talk about testing 15:34 gmcharlt first, I submit that it doesn't matter for the prupose of being able to test things whether code exists as a K::S::R method or as a K::Object method to be reachable for testing 15:34 gmcharlt either way, it can be tested 15:35 gmcharlt IOW, I'm sayng that how we decide the DBIC usage questions need not be inlfuenced all that much by testing concerns 15:35 gmcharlt as we have both db-dependent and non-db-dependent ways of testing 15:35 * tcohen agrees 15:35 gmcharlt so as far as my current thinking on the overall matter is concerned 15:36 gmcharlt 1. I dislike adding layers of classes needlessly -- in particular, if a K::S::R object represents an entity adequately, I see no reason not to use it as a base class or a very thinly-encapsulated entity 15:37 gmcharlt e.g., K::S::R::Item->effective_itemtype is a good example 15:38 gmcharlt 2. Not *everything* can be done that way -- e.g., I would call for a Koha::Biblio class to be written that (among other things) contains K::S::R::Biblio and K::S::R::Biblioitems objects, and makes some effort to hide the details of those tables 15:38 gmcharlt as we KNOW that those tables are eventually going to significantly change 15:39 gmcharlt so the upshot: on question #1, I fall firmly in the middle 15:39 gmcharlt now we may want to provide some syntactic sugar to the reduce the typing to instantiant a K::S::R object 15:40 gmcharlt but for areas where our data model is good -- I think a lot can be done by adding attribute methods to the underlying K::S::R classes 15:41 gmcharlt and (to answer in part a question that cait had asked me) borrowers is an example where the database schema mostly matches how we view the objects 15:41 gmcharlt so I think that for now K::S::R::Borrowers would be a good base class 15:42 gmcharlt now that's not to say that we couldn't have a Koha::Borrowers, but for now it shoudl delegate/pass along as much as psosible to K::S::R::Borrowers 15:42 cait hm so we would put something there that pulls information from ohter tables as well? 15:42 cait sorry i am still trying to wrap my mind around it :) 15:42 gmcharlt well, K::S::R via DBIC does that for us for free 15:43 tcohen welcome ribasushi 15:43 cait so it would get the extended patron attributes etc. 15:43 gmcharlt at least in the cases where the underlying tables have FK relationsips 15:43 cait ok thx :) 15:43 gmcharlt as far as our quesinon #6 is concerned -- I'm not a fan of a Koha::Object and/or Koha::Object::Set as a universal base class 15:43 gmcharlt too much weight 15:44 gmcharlt if anything should be a universal base class, I'd go for Class::Accessor 15:44 gmcharlt and thus ends my monologue for now 15:44 gmcharlt but to summarize: I'm suggesting a hybrid approach, one of whose big advantages is that it will allow us to continue to switch thigns over to using DBIC incrementally 15:45 * gmcharlt goes under his desk to look for his hands, which have fallen off 15:45 tcohen other opinions please? 15:45 * cait helps searching... hard to put a hand back on with no hands... 15:47 Joubu gmcharlt: how do implement set/list? 15:47 ashimema I think gmcharlt described it pretty succinctly really. 15:47 Joubu you* 15:48 gmcharlt Joubu: one moment before I answer you, want to check something 15:50 Joubu rangi: around? if I remember correctly you was not fond of extending K::S::R into "Koha::Object" 15:50 gmcharlt Joubu: so for now, my view is to use DBIC::ResultSet where possible 15:51 tcohen gmcharlt: even on front end stuff like the .pl scripts¿ 15:51 cait Joubu: 3 am in nz i think :( 15:51 ColinC Yes it makes for clear robust code 15:51 gmcharlt provide constructures where needed so that you can (say) take a Koha::S::R::Biblio and initiatialize a Koha::Biblio object from it 15:51 gmcharlt and uses plain old arrays and hashes for sets ;) 15:52 * jcamins agrees with gmcharlt on everything that gmcharlt said. 15:52 gmcharlt and as far as the scripts are concerned... yeah, I personally have no problem fetching data directly using DBIC in them 15:52 Joubu ok, my fieling is the same, except I would create a Koha::Biblios or Koha::Biblio::Set 15:53 gmcharlt but the key thing is this: the scripts should be focused on GETTING the objects, but they wouldn't be doing much if any business logic 15:53 gmcharlt and where possible, would be passing the K::S::R and K::* objects directly to the template so that the TT code can take care of fetching attributes for display 15:53 Joubu gmcharlt: do you have a poc to show us? A class with search+crud? 15:54 gmcharlt Joubu: I disagree in the general case -- i.e., I wouldn't want to see a Koha::Vendor::Set or the like 15:54 gmcharlt however, catalog searches are a special case 15:55 tcohen patron searches? 15:55 gmcharlt i.e., I could certainly see a Koha::CatalogSearchResultSet class 15:55 Joubu "passing the K::S::R and K::* objects directly to the template" => to me, we should not manipulate both. Only Koha::* 15:55 Joubu manipulate in pl 15:55 ColinC No need for a Set object you just have a set of object much as ResultSet itself is a set of objs 15:56 gmcharlt Joubu: I think our main difference of opinion is that I think K::S::R objects should be used when they suffice, and that Koha::* classes should be created when they are necessary, but that we can do both and be happy 15:57 Joubu gmcharlt: I just want consistency, I would like to create all objects using the same way 15:57 gmcharlt from the POV of the templates, either way it's going to look like [% object.attribute %] anyway 15:57 Joubu my $patron = Koha::Patron->new($patron_info)->insert 15:58 gmcharlt Joubu: I want consistent *interfaces* for accessors -- I care much less about the base class 15:58 ColinC From the callers vire both should look the same 15:58 Joubu and @patrons = Koha::PatronSet->search(name => 'smith') 15:58 Joubu s/PatronSet/WhatYouWant 15:59 gmcharlt Joubu: I don't want to (or want anybody) to have to write 100+ sets of boilerplate objects to do that when K::S::R can handle that for most objects 15:59 gmcharlt HOWEVER 16:00 tcohen I don't like the hybrid approach: when does it suffice to use K::S::R? 16:00 gmcharlt there might be the possibility of some AUTOLOAD magic 16:00 gmcharlt tcohen: items, for now 16:00 gmcharlt but to continue my thought 16:01 gmcharlt there might be the possibility of some AUTOLOAD magic or the like so that we can chop out the "Schema::Result" part of the names to use the DBIC classes directly 16:01 gmcharlt and where needed, e.g., Koha::Biblio, override when something that's hand-constructed 16:02 Joubu (11518 (effective_itemtype) still NSO, for 6months, and we didn't get enough thoughts) 16:03 cait i think we becasue we needed ot have this discussion - hopefully the bugs stuck now can then start moving forward 16:04 Joubu I really would like to see how you (everybody) would like to write code in pl file. It's clear to me, but I didn't get ideas/thoughts from others. Don't you have a POC/try somewhere to show? 16:04 Joubu I think it's the first question to consider 16:05 Joubu When we answered, we should try to have a consistent, simple, easy to use and flexible API 16:07 Joubu hum, ping? 16:07 wahanui Using deft allegory, the authors have provided an insightful and intuitive explanation of one of Unix's most venerable networking utilities. http://www.amazon.com/Story-about-Reading-Railroad-Books/dp/0448421658 16:08 tcohen i'm here 16:08 ColinC been using dbic and its nice to have a consitent reliable approach where we need classes that span tables like biblio the api should be as similar as possible 16:08 cait i think the dual approach makes sense tome, but it might be hard to judge at times which route is better 16:09 tcohen i think in the long run we will have a better model, which makes DBIC classes represent our objects better 16:09 tcohen BUT 16:09 tcohen we will still need to put business logic on top of that 16:09 tcohen and in that case, extending Koha::Schema::ResultSet::* should be the way to go 16:10 tcohen we could think Koha::Biblio makes sense "only" because of a bad underlying model 16:11 tcohen if we can rely on DBIC for everything 16:12 gmcharlt right - my inclination is to have the RDBMS model match the desired class structure where possible 16:12 gmcharlt but knowing that it it won't be always possible 16:13 gmcharlt as far as Joubu's question is concerned - well, current DBIC use in scripts like admin/categorie.pl isn't actually horrid, but I would grant that it's certainly verbose 16:13 gmcharlt the trick would be improving that without burying the DBIC classes udner too many layers 16:14 tcohen by *too many* you mean one, right? 16:14 gmcharlt yep, if we can avoid it :) 16:15 Joubu it's for the lisiblity, maintenability 16:15 Joubu s/lisibility/readability 16:16 Joubu Does nobody want to write some code? 16:17 Joubu just some lines here? 16:18 tcohen any more comments/opinions? 16:20 gmcharlt Joubu: I'm not sure you get to make demands on anybody's time but your own -- but I think I have some ideas on how to implemnt some concision 16:20 tcohen ok, i think we have exhausted the discussion for now. I'll try to sumarize on the wiki Joubu set for this topic 16:20 gmcharlt in the meantime, I've got more utf8 test cases to write today 16:20 tcohen :-D 16:21 tcohen gmcharlt: do u have some sample test we can take a look at? 16:21 Joubu gmcharlt: I just would like to know how continue to develop 16:21 gmcharlt utf8 test cases first, then we'll see 16:21 Joubu gmcharlt: I had some tries, but they seem to be bad. 16:22 tcohen Joubu: you're talking about using DBIC? 16:22 Joubu gmcharlt: I don't ask a lot. Just how instanciate an object, save it. And possibly a search 16:23 gmcharlt Joubu: I know; I'll get to it, but I'm not making promises as to when 16:23 Joubu I need to unfreeze this situation 16:23 Joubu s/I/We 16:23 gmcharlt I don't disagree 16:23 gmcharlt I think the discussion today has been helpful 16:23 gmcharlt and should get even more help when the NZers weigh in 16:24 gmcharlt IOW, I think we're making progress, as slow as it may seem 16:24 gmcharlt Joubu++ # keeping at it 16:25 tcohen ok, i think the meeting concluded 16:25 Joubu There are tries on 12830, 12896, 8007, 10363 16:25 tcohen some more thoughts will be expressed later 16:26 Joubu so all approches is bad, and I will try something else 16:26 tcohen and we will work on this in person, during the hackfest 16:26 tcohen I encourage people that disagrees with a specific dev approach, to make it explicit 16:26 Joubu or 1 is good, and it would be good to know :) 16:26 tcohen how it should be done 16:26 cait i think the hackfest will be a chance and we can post the code for others to look at 16:26 Joubu http://www.gliffy.com/go/publish/6160409 16:27 tcohen because otherwise it discourages participation 16:27 gmcharlt Joubu: on a completely different topic, I just now learned of a code2bib French mailing list... is it like code4lib? 16:27 Joubu for information, maybe not a lot of people is aware of that. 16:27 Joubu gmcharlt: what the name of the ml? 16:27 Joubu is* 16:28 gmcharlt Joubu: code2bib 16:28 tcohen ok, thanks every1 16:28 tcohen #endmeeting 16:28 huginn Meeting ended Tue Sep 23 16:28:25 2014 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) 16:28 huginn Minutes: http://meetings.koha-community.org/2014/development_irc_meeting__23_september_2014__dbic.2014-09-23-15.06.html 16:28 huginn Minutes (text): http://meetings.koha-community.org/2014/development_irc_meeting__23_september_2014__dbic.2014-09-23-15.06.txt 16:28 huginn Log: http://meetings.koha-community.org/2014/development_irc_meeting__23_september_2014__dbic.2014-09-23-15.06.log.html 16:28 gmcharlt Joubu: https://groupes.renater.fr/sympa/info/code2bib - I think 16:28 ribasushi folks, hi 16:28 Joubu gmcharlt: I don't know, Paul follows it, not me 16:28 ribasushi I am really sorry I couldn't make it, delayed train :( 16:28 ribasushi I am here now if of any use 16:29 Joubu ribasushi: you can have a look at the minutes if you want :) 16:29 ribasushi link? 16:29 wahanui somebody said link was broken. Also, we don't know what tools you're using to create the marc, but there is probably an option in that tool. 16:29 ashimema we haven't done an end meeting yet 16:29 oleonard Yeah-huh. 16:29 gmcharlt ribasushi: http://meetings.koha-community.org/2014/development_irc_meeting__23_september_2014__dbic.2014-09-23-15.06.log.html 16:29 tcohen http://meetings.koha-community.org/2014/development_irc_meeting__23_september_2014__dbic.2014-09-23-15.06.txt 16:30 Joubu ashimema: yes, just before ribasushi said hi :) 16:30 ashimema oh yeah.. 16:30 ashimema missed it 16:30 cait Joubu: what's the pic made with? it looks interesting 16:30 cait oh 16:30 cait with gliffy 16:30 cait ok 16:31 tcohen ribasushi: we have splitted opinions on how show the ORM be used on the project 16:31 Joubu cait: I made it to show how apply a patch, for testers. There are too many deps, and it's hard to see what apply when you are on a bug report. 16:31 tcohen i'm sorry, i have to leave now 16:31 Joubu cait: it's non pushed patches 16:32 Joubu and it's just the plumbing... 16:32 cait i know, big project 16:32 cait I saw paola tackled some of it:) 16:32 cait i hope after 26th i can do more again 16:33 cait i have the exam this and a presentation next week... and then the trip, finding time is too had right now 16:34 fridolin thanks for this meeting 16:35 ribasushi tcohen: I will read backlog and write a summary for Joubu 16:35 ribasushi once again sorry for not making it, pesky deutchebanhn 16:36 cait ribasushi: oh, you are in Germany? :) 16:38 Joubu thanks ribasushi! 16:40 ribasushi yes I am 16:40 ribasushi sorry need to attend to something else first, be with you in 30 mins folks 16:42 cait ribasushi: me too :) thx for helping us out 16:53 Joubu gmcharlt: sorry, I forgot your question about code2lib. So yes it is a "french local group of code4lib" (from the link you sent me) 17:18 gmcharlt Joubu: thanks 17:46 oleonard Patron attribute type limitation by category seems to be broken 17:48 cait oleonard: oh :( where are you looking? 17:48 cait and did it save correctly? i think sometimes it does not if you create a patron category from scratch 17:51 oleonard I created a new patron attribute type and chose one patron category to limit it to. When I go to add a patron with that category the patron attribute type doesn't show up. 17:52 oleonard Only if I remove the category restriction does the patron attribute entry field appear. 17:54 cait ah, i was thiking patron categories, got confused 17:54 cait sorry 17:54 cait i somehow thought about patron categories and branch limists 18:20 oleonard Oh hey someone commenting on the K-C web site also checks website posts there in the early hours in the dawn as they love to learn more and more 18:21 * oleonard would be happy to help delete comment spam on K-C web site if someone gave him the proper permissions 18:22 * oleonard wonders if wizzyrea is the only one who can do that 18:22 tcohen whic comment? 18:23 oleonard http://koha-community.org/koha-3-16-3-released/comment-page-1/#comment-54924 18:23 cait oleonard: hm don't know but i think she turned off the comments on most sites now 18:23 cait there was some dicussion about that some days ago i think 18:23 cait hm maybe only for new posts? 18:27 tcohen which comment? 18:27 wahanui which comment are you refuting? 18:28 oleonard tcohen: http://koha-community.org/koha-3-16-3-released/comment-page-1/#comment-54924 18:29 tcohen i was kidding, i've just flaged it as spam 18:48 * oleonard submits a patch for his patron attribute type bug 18:52 huginn New commit(s) kohagit: Bug 12162 - Add class="branchcode" to body tag to make OPAC CSS-styleable per branch <http://git.koha-community.org/gitweb/?p=koha.git;a=commitdiff;h=6246f2c70049baf86bc942c2473793c83057a64f> / Bug 11672: (followup) warnings tested <http://git.koha-community.org/gitweb/?p=koha.git;a=commitdiff;h=b2839a7c9c8554097bd36de0ad439967e2035c5c> / Bug 11672: Untranslatable dropdown on Guided Reports and dictionary <http://git.koha-c 18:58 tcohen bye #koha 18:58 oleonard Bye tcohen 18:58 * oleonard is off as well 19:54 rangi @later tell oleonard comments are off for new posts, old ones will switch off after 30 days or so ill kill that spam in the meantime, comments on the site are not really useful 19:54 huginn rangi: The operation succeeded. 19:56 rangi and fwiw im against coupling objects tightly with the db, because the db wont always be there in its current state, and 2 we might want to cache objects a lot more than we do now, im mostly in the middle but lean more to Koha::Item than Koha::Schema::Resultset::Item .. because im not sure i want to add a caching layer to DBIx::Class level code .. however working code wins, probably time to finish the discussion and write code 20:55 cait gmcharlt++ 22:01 oleonard Hi again #koha 22:03 rangi heya oleonard 22:03 wahanui somebody said oleonard was still here, if you just wish hard enough. or Koha's master UI designer 22:03 cait hi oleonard 22:03 cait hm no tcohen yet 22:04 oleonard Is there another dev meeting? 22:04 rangi not sure, i dont think i have anything worth adding 22:04 cait hmthere should be one 22:05 cait i am actuallystill a bit confused by who proposes what 22:05 oleonard I thought the first one seemed productive, but it's all over my head. 22:05 cait yeah mine too 22:05 cait rangi: maybe we could have a quick... comments on first meeting one and you repeat what you siad for the logs? :) 22:06 rangi actually naw 22:06 cait heh kind of expected that answer 22:06 cait someone else wants to add something? 22:06 rangi im kinda sick of dev meetings 22:08 cait i think there was supposed to be as ummary written 22:08 cait people could comment on that 22:09 cait i think the caching might be a good point 22:09 cait to have the code in a Koha: module? 22:09 rangi like gmcharlt said, probably in some modules 22:11 rangi ill wait and see the summary, as long as its guidelines, not rules, i can probably work with anything 23:13 dcook eythian: Remember how I mentioned that reserves vs old_reserves issue I was having? 23:13 eythian yup 23:13 dcook Turns out that MySQL's "OPTIMIZE TABLE" command resets your AUTO_INCREMENT 23:13 eythian ah 23:14 eythian so that's something you never want to do 23:14 dcook Yep 23:14 dcook Looks like it does a max(id)+1 for the reset 23:15 eythian oh, I see 23:15 eythian that's not so terrible 23:16 dcook Yeah, it only causes occasional problems 23:17 dcook Like..if you have 50 reserves, and move the 50th to old_reserves 23:17 dcook reserves gets reset and then you start producing duplicate IDs 23:17 dcook So you can't cancel/fill holds for those duplicates 23:17 dcook And it just seems like random chaos 23:17 dcook Kudos to the boss for finding that one 23:17 dcook Or rather the root cause 23:18 dcook Looks like it's a bug in MySQL that was supposed to be fixed in 5.5? 23:18 dcook But we're using MariaDB now 23:18 dcook http://bugs.mysql.com/bug.php?id=18274 23:19 dcook 5.5.33 :S 23:20 * dcook shrugs 23:20 eythian hmm, that's certainly worth being aware of 23:21 dcook Totes