15:01:10 #startmeeting Developer IRC Meeting, 2 February 2016 15:01:10 Meeting started Tue Feb 2 15:01:10 2016 UTC. The chair is cait. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:01:10 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:01:10 The meeting name has been set to 'developer_irc_meeting__2_february_2016' 15:01:18 #topic Introductions 15:01:18 #info wahanui, a bot that has become sentient 15:01:30 Please introduce yourself using #info like wahanui just did 15:01:39 #info Katrin Fischer, BSZ, Germany 15:01:46 #info Owen Leonard, Athens County Public Libraries 15:01:52 #info Jesse Weaver, ByWater Solutions, USA 15:02:02 #link https://wiki.koha-community.org/wiki/Development_IRC_meeting_2_February_2016 15:02:05 #info Joy Nelson, ByWater Solutions, USA 15:02:16 #info Julian Maurice, BibLibre, France 15:02:18 #info Jonathan Druart 15:02:25 #info Matthias Meusburger, Biblibre, France 15:02:32 #info Marcel de Rooy, Rijksmuseum, Netherlands 15:02:35 #info Andreas Hedström Mace, Stockholm University Library 15:02:50 #info Brendan Gallagher ByWater 15:03:03 #info Colin Campbell, PTFS-Europe 15:03:30 #info Nick Clemens, ByWater Solutions, USA 15:03:37 #info mirko tietgen, not really here 15:03:55 #info Tomas Cohen Arazi, Theke 15:04:28 #info Kyle M Hall, Bywater Solutions 15:04:29 #info Nate Curulla, BWS 15:04:30 hi mirko not really here :D 15:04:51 before we start with the topics on the wiki - any announcements? 15:05:02 RM notes? :) 15:05:23 Currently few notes 15:05:42 pushed the XSS so please test test that and let’s see if we can find any missing spots 15:06:00 the web-installer needs some touch up - but I think mtompset caught some of that 15:06:15 (there is one already, patched, passed qa) 15:06:23 #info XSS patches were pushed, please test and help to find any remaining problems 15:06:24 yes :) awesomeness 15:06:39 Still looking for more testers of Elastic Search 15:07:02 bag: which branch is recommended for testing? 15:07:07 catalyst or kc? 15:07:11 kc 15:07:12 there is a branch in kc 15:07:15 kc :) 15:07:20 #info Indranil Das Gupta, L2C2 Technologies 15:07:31 #info Elastic search needs testing as well - please use the branch on the main repository 15:07:40 instructions? 15:07:40 instructions are coming right now to the wiki near you or at http://i.imgur.com/oyZhY.jpg 15:07:49 * bag will catch up with tcohen sometime this week to get some more lessons - especially with jenkins and packaging 15:07:56 I think there is a page on the wiki with some notes at least 15:08:04 thanks cait 15:08:11 so that I start to hopefully clean up the tests and get master to be passing stable on jenkins 15:08:31 #link https://wiki.koha-community.org/wiki/Elasticsearch 15:08:34 the branch is named remotes/origin/new_12478_elasticsearch 15:08:40 yes ^^ 15:08:54 #info Elastic search branch: remotes/origin/new_12478_elasticsearch 15:09:01 that is it for me - I will keep pushing what I see from PQA 15:09:14 ok, let's move to our first big topic 15:09:23 if I miss something - just shoot me an @later and I’ll get on it 15:09:30 #topic Review of Coding guidelines 15:09:46 I have tried to group the various topics a bit - I hope people are ok with it :) 15:10:11 we have a lot to get through in this meeting, for some topics i think we can only start discussion and will need a draft to vote on next 15:10:46 I will go through them by sequence, but we have to watch the time a bit today with our long agenda 15:10:51 let's start with the first item 15:11:04 #info [marcelr] In relation to Plack: Should we add a PERL rule that prohibits defining lexical variables (my $var) in MODULES at the outermost block (file level), and also prohibits directly accessing file level lexicals in subroutines of SCRIPTS. 15:11:31 first part is less mandatory than the second part, i guess 15:11:34 I'd suggest that ideally all additions to the coding guidelines should include a good/bad example if possible 15:11:39 ooh.. meeting 15:11:50 #info Martin Renvoize, PTFS Europe 15:12:11 i am a bit out of my depth here - any comments, questions? 15:12:49 the first part needs some common sense 15:12:52 It would be more useful yo explain the rational behind this than to make it a rule it dosent apply to scripts which will never be used by plack 15:12:57 examples are good :) 15:12:58 I agree with the proposal, as it is also a clearer way to program 15:13:08 :-P 15:13:26 I totally agree 15:13:28 I almost agree with everything too 15:13:29 ok, so shoudl we note exceptions? 15:13:30 who likes global variables in modules? 15:13:39 or a scope where this applies? 15:13:43 There is no new things, we just need to update the wiki page 15:13:44 A good guideline would be to discourage globals where possible 15:14:13 cc: to clarify, by "scripts which will never be used by plack", do you mean things like cronjobs? 15:14:23 yes 15:14:59 would someone be willing to work out a draft? 15:15:06 i believe global variables can't be completely avoided in modules (for instance $VERSION, @EXPORT, ...), so why forbid it ? 15:15:23 obviously 15:15:36 jajm: those are read-only, though, correct? 15:15:42 jajm: but they are not storing state 15:15:55 * ashimema has gotta go collect kids.. sorry for me absence.. again! 15:15:56 #info Barton Chittenden, bws, Louisville, KY, USA 15:16:02 no they can be manipulated at runtime 15:16:05 * ashimema will read the minutes 15:16:15 cc: huh? 15:16:36 pianohacker, tcohen, ok, so forbid global variables that are storing state ? 15:16:43 VERSION and EXPORT are not readonly 15:17:35 cc: forbid touching VERSION and EXPORT could be another rule :-P 15:17:42 again, someone willing to make a draft? I am eager on an #action :) 15:17:49 jajm: I'd say so. Forbid state globals in modules/CGI scripts as it causes issues with plack, and discourage them in other places for style reasons? 15:18:13 lots of modules thouch EXPORT for good reasons 15:18:38 that sounds right pianohacker 15:18:43 i think that we all understand exceptions for EXPORT etc. 15:18:46 * cait volunteers pianohacker 15:19:01 Document what the issue is with plack lets try and minimise "magick" 15:19:11 pianohacker: can you wite a draft that we can review later? 15:19:25 cait: does the above work, or are you talking in the form of a coding guideline 15:19:26 cc: +1 15:19:31 cait: either way, sure 15:19:32 cait: i think we already had a draft 15:19:34 a coding guideline: :) 15:20:16 marcelr: hm? 15:20:17 I volunteer 15:20:30 just to summarize the discussion here in a short form 15:20:35 from the agenda 15:21:05 cc: https://perl.apache.org/docs/general/perl_reference/perl_reference.html#my____Scoped_Variable_in_Nested_Subroutines <- for context 15:21:28 marcelr: I thought adding the addtional remarks to it as discussed 15:21:41 ok 15:22:12 #action pianohacker will draw up a draft of a coding guideline regarding global variables in modules/CGI scripts, see https://perl.apache.org/docs/general/perl_reference/perl_reference.html#my____Scoped_Variable_in_Nested_Subroutines for context 15:22:31 thx 15:22:36 moving on to the next item for now 15:23:05 [marcelr] What about DBIx, Koha::Object versus old school MySQL statements in code ? [khall] I think the use of direct DBIx should be deprecated in favor of Koha::Object, and direct sql statements should be limited to queries that don't work well with the object model 15:23:34 there are some rules about use of SQL in the current coding guidelines 15:23:42 I think the above has been so strongly enforced on new patches that it's a de facto coding guideline 15:24:05 hm in part 15:24:16 the problem is that we have patches of different 'age' sitting in the queue 15:24:56 i think we can say DBIx before SQLfor sur, but not sure about enforcing Koha::Object for everything 15:25:05 #info Zeno Tajoli, CINECA, Italy 15:25:07 cait: I think we should add a grandfather clause for all new rules 15:25:13 khall: ? 15:25:16 to give them some leeway 15:25:23 I'm very unpersuaded by Koha::Objects in their current form 15:25:27 can you quickly explain a grandfather clause? :) 15:25:36 agreed, there's a lot of precedent for that 15:25:42 cait: meaning these new rules affect only patches submitted after the rules were voted in 15:25:57 for older patches, we can give them more latitude 15:26:02 grandfather clause: a case that violates the proposed rules, but will be allowed because it's already there. 15:26:03 it's not possible to force people using dbix::class, but we get more and more module using Koha::Objects in the Koha namespace 15:26:12 it will be more and more easier for devs to use it 15:26:19 we had this discussion before about allowing DBIx in all modules or not 15:26:22 especially because we will get ot of example 15:26:26 lot* 15:26:45 marcelr: Koha::Object(s) is what came out of that discussion 15:26:57 hm we only have this in the coding guidelines: PERL19: The use of C4::SQLHelper module is deprecated C4:SQLHelper has been superseded by DBIC (examples), please use this instead. 15:27:05 the lin goes here: https://wiki.koha-community.org/wiki/Examples_of_DBIC_in_Koha 15:27:06 i remember that this was not a decision 15:27:10 i'd leave it as a de-facto coding guideline, and let the RM have the last call on each situation 15:27:17 yup, we need something stronger 15:27:31 cait: this page has not been updated 15:27:50 Joubu: correct - so that's not ideal... but i suggest removing the PERL19 entirely 15:27:52 tcohen: that's the situation we should avoid. It's better to give devs the info they need to write it correctly the first time 15:28:01 what khall said 15:28:03 SQLHelper is no longer part of the codebase 15:28:21 cc: can you detail the problems you encountered ? 15:28:26 pianohacker: khall: I agree 15:28:40 PERL18 a PERL19 should be removed, yes 15:28:45 and* 15:28:49 agreed 15:29:04 +1 15:29:06 can we agree on removing those right away? 15:29:17 +1 15:29:20 one is C4::Dates deprecated, the other SQL::Helper 15:29:21 +1 15:29:26 +1 15:29:29 +1 15:29:30 +1 15:29:34 +1 15:29:35 +1 15:29:40 +1 15:29:47 +1 15:29:54 +1 15:29:59 +1 15:30:04 Performance, aldo negates some of the benefits of DBIIx::Class by obscuring the interface to the returned objects with an extra layer - it seems to add code rather than functionality 15:30:05 (add PERL11 to the list) 15:30:08 #agreed PERL18 and PERL19 to be removed, as the deprecated modules are no longer part of the codebase (C4::Dates and SQLHelper) 15:30:33 actually removing ws a later topic, i just jumped into as we were already discussing it 15:30:37 +1 15:30:44 let me read back a second for the dbix discussion 15:31:31 Joubu is right about PERL11 15:31:36 I am not sure if I remember correctly, but I think we didn't vote to enforce Koha::Object and agreed not to use DBIx in .pl files directly 15:31:45 And PERL12, by its own admission? 15:32:00 pianohacker: seems that way 15:32:03 cait: maybe add the DBIx rule with some clause that we need good arguments to bypass it or so 15:32:26 marcelr: sorry, not sure i understand - the rule about not using in .pl files? 15:32:41 the rule under discussion 15:32:53 cait: Perhaps to enforce usage where a Koha::Object already exists, to match what is current QA practice? 15:33:06 I am sorry, I got a bit lost 15:34:00 cait: for context: https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=14659#c4 15:34:01 04Bug 14659: enhancement, P5 - low, ---, jweaver, Needs Signoff , Allow patrons to enter card number and patron category on OPAC registration page 15:34:06 I think there are still some doubts about Koha::Object - see cc's comment, so giving some room to move would be good 15:34:53 so where a Koha::Object module exists - use that? 15:34:55 I'd like to understand cc's doubts. I don't think we've heard anything against from anyone else 15:34:58 not to pick on Joubu, just to point out what I mean 15:35:05 cait: yes 15:35:10 cait: yes, everybody has doubts about Koha::Object, but nobody never provided either a an alternative or valid arguments not to use it 15:35:19 The defacto recently has been to use Koha::Object(s) 15:35:44 I was under that thought too - the defacto 15:35:59 it was not decided yet 15:36:14 I've never seen a convincing rationale for using them 15:36:18 you can refer to "Koha and DBIC" from Septembre 2014 on Koha-devel 15:36:40 that discussion never reached a good conclusion iirc 15:36:41 then "Koha::Object" on July 2015 15:37:16 i would prefer a rule now with an escape for convincing arguments to qa/rm 15:37:31 marcelr++ 15:37:41 marcelr: basically every rule has that escape, out of necessity, but I agree :) 15:37:54 maybe we coudl summarize 15:37:57 pianohacker: nobody tends to provide the args 15:38:01 - use where an Object already exists 15:38:08 - use preferrably 15:38:15 ß 15:38:16 ? 15:38:26 that sounds reasonable and good to me! 15:38:37 i think every medium sized project for improving Koha (new module) should be discussed with the core team so (without interfering) we agree on different alternative approaches 15:38:39 again, someone willing to draft? 15:38:52 Joubu: are you referring to http://koha.1045719.n5.nabble.com/Koha-Object-td5848332.html ? 15:38:59 "core team" needs to be defined 15:39:06 arguments against an approach supporting the other will raise, win-win situation 15:39:11 what Joubu said 15:39:13 koha-devel 15:39:14 well, koha-devel is most likely the best koha list there is (Hey, Im a bot, im biased) and can be found at http://lists.koha-community.org/cgi-bin/mailman/listinfo/koha-devel 15:39:31 yep 15:39:44 Joubu: and http://koha.1045719.n5.nabble.com/Koha-and-DBIC-td5811156.html ? 15:39:45 the worse situation we've seen is starvation 15:39:52 cait: I think we can vote on what you just wrote 15:39:56 it's simple enough 15:39:57 no one answers, deve does what he wants 15:39:58 pianohacker: yep 15:40:00 it's the RM call 15:40:05 period 15:40:27 i'd like something a bit more polished in full sentences for the wiki:) 15:40:38 that someone will udnerstand outside of this meeting 15:40:39 cait: ok, I'll get something written right now 15:40:53 and with a clear escape clause - as the proof might be in the code... i don't know 15:40:58 thx 15:41:03 moving on for now 15:41:05 the grandfather clause 15:41:24 grandfather is implicit for more rules 15:41:31 probably 15:41:37 my suggestion woudl be: we should state the date and the meeting in the guidelines added clearly. Then we can say, code submitted before guidelines with a clear 'start from' date is still ok 15:42:05 and with "feel free to provide an alternative if you are not convinced with this one" clause? :) 15:42:12 or 'grandfather rules stated below applies' or whatever 15:42:41 cait: agreed 15:43:28 vote? 15:43:28 vote is probably going to the list regardless of what we decide 15:43:45 ... for exceptions to this rule we need convincing arguments for using or skipping DBIx/Koha::Object 15:43:57 formal vote or +1? 15:45:22 i'd like to stick with +1 actually, becuase it's a littler faster today 15:45:44 but again, i need someone to write it up - the grandfather rule 15:45:53 cait: will do 15:45:57 thx 15:46:26 moving on to next opic for now 15:46:32 [marcelr] New modules should be added if possible into the new Koha namespace (and not in C4). 15:46:36 while it sounds like we're pushing the K::O discussion to the mailing list, should we vote on the grandfather rule here and now? 15:46:41 i think this is pretty clear 15:46:44 sorry, just missed timing :) 15:46:58 pianohacker: will return to all the 'write up somethings' at the end 15:47:19 cait++ 15:47:35 cait: I've got those writeups, should we do a quick review / vote? 15:47:44 yes, where? 15:47:55 cait: I was just going to paste them here 15:47:58 PERL16 - Modules in the Koha namespace should be object oriented when possible, using Koha::Object(s) as a preferred base. 15:47:59 PERL16.1 - If an Koha::Object already exists, use it instead of other methods of table CRUD. 15:48:20 XXX - Patches submitted before the introduction of a new rule may pass qa even if they do not meet the current coding guidline requirements of the discretion of the QA team member. 15:48:37 also, we'll want to put the date on new rules from now on 15:48:42 but that's implied 15:49:18 cait: are there any patches old enough that we should do some wiki archeology to figure out the date-added of the existing rules? 15:49:33 pianohacker: i hope not, but probably will do so when the problem arises :) 15:49:42 seems reasonable 15:50:07 any comments? 15:50:57 don't go quiet... 15:51:06 going once 15:51:25 twice 15:51:33 no comments? 15:51:47 ok, voting on PERL16/16.1 please 15:51:50 +1 15:51:54 +1 15:51:55 +1 15:51:58 +1 15:51:59 +1 15:52:06 -1 15:52:06 +1 15:52:14 +1 15:52:17 +1 15:52:45 +1 15:53:06 +1 15:53:24 #agreed PERL16 - Modules in the Koha namespace should be object oriented when possible, using Koha::Object(s) as a preferred base. PERL16.1 - If an Koha::Object already exists, use it instead of other methods of table CRUD. 15:53:50 for the XXX rule - at the discretion of the qa team member... 15:53:55 +1 15:54:03 +1 15:54:04 +1 15:54:07 +1 15:54:12 do we want to keep it that way? so i can say 'no'? :) heh 15:54:37 +1 for XXX :) 15:54:56 +1 #QA team member*s* 15:55:14 yeah there are more 15:55:19 cait: my understanding of the wording (which I think should be slightly clarified) is that QA team members can choose to say _yes_, not no, on grandfathered patches 15:55:41 khall: is that what you had in mind? 15:55:52 yes sounds about right khall 15:55:56 pianohacker is correct 15:56:04 :) 15:56:13 pianohacker: ah right - can say 'yes' 15:56:23 The 's' is important I think 15:56:36 qa team members? 15:56:40 Joubu: member is the member at hand 15:56:51 the one doing the qa 15:56:59 it's adifference indeed 15:57:10 I'd ask for another QA POV in this case, so I won't be alone 15:57:25 that's probably a good thing to put in the guideline 15:57:29 What we did for the recovery pwd patchset for instance 15:57:40 it's an existing habit, should be written down :) 15:57:47 it did not fit the requirements, *we* agreed to let it pass 15:57:48 anyway 15:57:56 hm what about 15:58:08 but is not a coding rule 15:58:20 Grandfather clause: XXX - Patches submitted before the introduction of a new rule may pass QA even if they do not meet the current coding guideline requirements in agreement with the QA team. 15:58:25 ok, let's forget that, not important 15:58:49 whole team is not needed i guess 15:59:12 #agreed Grandfather clause: XXX - Patches submitted before the introduction of a new rule may pass QA even if they do not meet the current coding guideline requirements after consulting with the QA team members. ? 15:59:17 imo, more than 1 is needed 15:59:23 i am nto sure how to phrase that there shoudl be a window for discussion 15:59:25 at discretion of QA team member (and possibly another QAer) 15:59:36 if noone chimes in... that woudl be ok 15:59:37 some things are trivial 16:00:12 I like that wording, cait. 16:00:31 I would also like to bring up bug 8753 for a quick discussion and vote. In that but I wrote two scripts, one for display, and the other for CRUD 16:00:32 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=8753 enhancement, P1 - high, ---, charles.farmer, Pushed to Master , Add forgot password link to OPAC 16:00:43 consulting would practically mean asking for opinions - if response is low... so it's up to those who work it out 16:00:52 the general consensus is that we shouldn't do this, so I think this should have a rule as well 16:01:09 khall: do you mean bug 14610? 16:01:10 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=14610 enhancement, P5 - low, ---, kyle.m.hall, Signed Off , Add ability to place article requests in Koha 16:01:11 one thing after the other please 16:01:20 which is the preferred phrasing now? so i can add it to the minutes 16:01:21 cait: yes, you are correct ; ) 16:01:50 quick please 16:01:51 yes cait is correct there - phew 16:01:52 khall: you cannot catch every exception in a rule (or you don't want to) 16:02:30 marcelr: I was failed qa for something not in the qa guidlines, I don't want that to happen to future developers or forgetful me ; ) 16:02:31 i think what rangi said in argentina always applies... rules are good, but rules can be broken (I choose to read broken as questioned :)) 16:02:52 i thin we need to find somethin gin between 16:03:00 right, it's better to allow a rule to get broken than to fail someone who hasn't broken a rule 16:03:17 maybe less rules but more recommendations 16:03:18 anyway consulting, dicrestion or agreement? 16:03:19 good point 16:03:19 I know! The blade went right through that child! 16:03:26 i think it should be possible to fail although there is no rule 16:03:35 but you should explain of course 16:03:58 the rules are just tools :) 16:04:13 transparency is also key 16:04:30 yes, and a failure without a rule should generate discussion and a new rule 16:04:44 so can we finish please with the XXX rule? 16:04:48 so we can move on? 16:04:59 I thought the XXX rule was finished 16:05:00 cait: I'm ok with any of consulting, dicrestion or agreement. 16:05:09 there was discussion about the wording 16:05:10 I would propose PERLXX: CGI scripts that handle CRUD operatiosn ( Create, Rad, Update, Delete ) should all be handled in the same script 16:05:24 ok, let's table my recent discussion and finish XXX 16:05:31 thx 16:05:38 just tell me which, i will log and we can move on 16:06:11 the idea behind XXX is that even if a patch fails the current qa guidlines, a QA'er can still give it a pass if the rules it fails were made after the patch's submission 16:06:20 yes 16:06:39 should we revote, or are we just discussing the specific wording? 16:06:40 i think the question was if one qa'er or after asking other qa'ers for opinions 16:06:47 i think only the wording 16:06:50 if noone insists 16:07:03 well, we can always elevate it to the head of QA ; ) 16:07:10 heh 16:07:22 only one rule 16:07:23 head of QA is mean i have heard... better not 16:07:40 i will log as suggested if there is no strong opinion 16:07:50 #agreed Grandfather clause: XXX - Patches submitted before the introduction of a new rule may pass QA even if they do not meet the current coding guideline requirements of the discretion of the QA team member. 16:07:51 ok 16:07:52 next 16:08:06 Kyle suggested a rule about the handling of CRUD operations 16:08:16 PERLXX: CGI scripts that handle CRUD operatiosn ( Create, Rad, Update, Delete ) should all be handled in the same script when possible. 16:08:21 triggered by bug 14610 16:08:23 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=14610 enhancement, P5 - low, ---, kyle.m.hall, Signed Off , Add ability to place article requests in Koha 16:08:24 s/Rad/Read 16:08:37 any questions, comments? 16:08:56 that looks sensible to me. 16:08:57 I have failed it, so I am ok with the proposition 16:09:05 I also commented, so ok as well 16:09:08 sounds reasonable 16:09:11 I think it's a good rule for future patches; a majority though not all of Koha follows this pattern 16:09:19 hah 16:09:24 I have rewritten the admin scripts to follow this pattern 16:09:37 using the same $op value would be better 16:09:56 we can include a script to point to as the best example 16:10:11 add_form, add_validate, list, delete_confirm, delete_confirmed 16:10:31 admin/cities is the smaller one I guess 16:10:32 afk 16:10:39 hm modify? 16:10:46 add_form 16:11:23 There is no need to have a different op, if an id is passed, you are modifying, otherwise you are adding 16:11:36 Joubu: I don't think we need to be *quite* that specific in the rule, but yes, that's what we are looking for. 16:12:13 ok, no general disagreement 16:12:25 i think? 16:12:34 no disagreement here 16:12:35 khall: if we don't specify, we will get different wording for the same action 16:12:46 but not important 16:13:01 i think adding an example would be good 16:13:08 as suggested 16:13:13 pointing to an existing file 16:13:20 so you can add admin/cities.pl 16:13:29 Joubu: I understand. We can work out a basic action list later, but I think the example should suffice for now 16:13:33 can we vote please? 16:13:34 if everybody had a look at it already 16:13:40 no, though I think we should revisit Kyle's general concern about failing QA for rules not in the coding guidelines after we vote on this :) 16:13:47 khall: yep, agreed 16:13:50 We might want to show good examples - an example is more useful than a rule for someone learning the codebase 16:13:55 vote. 16:13:58 cc: totally agree 16:14:11 cc: also agreed 16:14:24 can we have +1 or -1 please? 16:14:29 +1 16:14:32 +1 16:14:33 +1 16:14:34 +1 16:14:35 +1 16:14:36 cc: and it can help address Joubu's concern about consistency in the details without having to put them in stone 16:14:38 +1 16:14:39 +1 16:15:14 +1 16:15:26 +1 16:15:47 #agreed PERLXX: CGI scripts that handle CRUD operatiosn ( Create, Rad, Update, Delete ) should all be handled in the same script when possible. 16:16:06 #info PERLXX shoudl link to admin/cities.pl for a good example 16:16:23 ok, any loose ends right now or can we move to next? 16:16:34 next. 16:16:36 cait: what I mentioned above (:13t 16:16:46 ok, go for it 16:16:47 It's *my* good example, I don't know if everybody agrees with that 16:17:09 Joubu: I think everyone is happy with it 16:17:12 Joubu: noone disagreed so far... we will see if the topic reappears on next agenda ;) 16:17:28 Joubu++ 16:18:07 I'm not pushing for anything dramatic, but I think we should codify that, if a QA team member has concerns about coding style (as opposed to intended functionality not working correctly), they should be either in the coding guidelines or the patch should be used to introduce a new coding guideline in a meeting / koha/devel 16:18:09 I think failing without a coding guileline should actually be an 'in discussion' - it indicates a disagreement about something and it should lead to a discussion with more people / dev meeting 16:18:43 i see this is a defacto actually 16:18:58 cait: I think that's somewhat the practice, but we should connect such things more strongly to the coding guidelines 16:19:05 pianohacker: if a dev wants to use a new pattern, the easier is to ask on koha-devel or here on #koha 16:19:05 I think you both agree. It'd like it to be dejure 16:19:13 as it is, the guidelines are incomplete and somewhat unloved 16:19:18 khall: now i need the dictionary 16:19:35 cait: just means it should be a codified rule 16:19:53 de facto - rule in practice 16:20:05 cait/Joubu: the main reason I'd like this (and khall as well, to my understanding), is that it would be much better to have up-to-date coding guidelines that reflect as much as possible of our current practices 16:20:06 de jure - rule codified in law ( i.e. coding guielines ) 16:20:12 this is one of the thing sthat I am nt sure it needs to be a rule... it seems more like a right to me... people disagree, it should lead to discussion 16:20:25 it's impossible to list all we can do or not do :) 16:20:42 i'd just like to say that i thik we have followed that 16:20:52 but if you need to introduce a design change, ask other devs 16:20:55 Joubu: you are correct, but we should do out best to help out those devs that haven't been around for years 16:21:00 so that code can be written to match those practices from the beginning 16:21:03 and what khall said 16:21:05 and disagreeing with QA is fine (well... maybe not that TOO often ;) ) 16:21:20 Joubu: I agree not everything can be codified, but as much as is reasonable _should_ be :) 16:21:24 i am just not sure how that would read as a rule 16:22:03 pianohacker: yes, that's what we are doing for 90min :) 16:22:17 yeah... running out of time soon :) 16:22:20 :D 16:22:21 ha 16:22:36 aaand on that note... next topic? 16:22:39 i thnk maybe this can be both sides 16:23:00 cait: I'll send a possible wording to the dev list after the meeting :) 16:23:06 if you introduce a design change - aks others... if you don't agree with a qa ruling... bring it to a dev meeting 16:23:09 thanks pianohacker! 16:23:18 QA court! 16:23:26 yeah... i'd like to avoid that a bit heh 16:23:33 jk ; ) 16:23:43 I definitively agree to have a list of "best pratices" to fill when QAing and vote them on the next dev meeting 16:23:48 i mean it goes both ways - devs shoudl talk about what they do.. as should qa 16:24:09 Joubu: can you give a quick example? 16:24:17 pianohacker: ok i wil action 16:24:20 no, I don't have any in mind 16:24:36 the problem is often we get either radio silence or a split down the middle when we discuss things 16:24:37 #action pianohacker to suggest a possible guideline for handling disagreements 16:24:44 but I will try to think about that when failing qa 16:24:57 ok 16:25:10 we i think the namespace rule is already passed? 16:25:17 yes 16:25:26 yeah, very very defacto, should be dejure 16:25:27 how about the update to JS4? 16:25:39 Sorry, but what is JS4? 16:25:45 oleonard_: https://wiki.koha-community.org/wiki/Coding_Guidelines#JS4:_Avoid_joining_multiple_language_strings_with_other_variables 16:25:52 khall: because of the new %s support? 16:25:58 right 16:26:01 sorry, help me 16:26:08 did we vote on 'new modules shoudl be in Koha::'? 16:26:26 oh, I mis-understood 16:26:28 It has been voted years ago, isn't it?? 16:26:39 not in the coding guidelines as of today 16:26:41 an oversight 16:26:48 ok, quick vote on that? 16:26:51 let's get that outta the way 16:27:02 +1 to add 'new modules should be in Koha::' 16:27:09 +1 16:27:11 PERL XXX: New modules should be added to the Koha namespace as C4 is deprecated. ? 16:27:13 +1 16:27:16 +1 16:27:21 wording can be refined of course 16:27:30 I agree with cait's proposed wording 16:27:31 +1 16:27:40 also agree 16:27:53 +1 16:28:05 +1 # yes please... 16:28:19 #agreed PERL XXX: New modules should be added to the Koha namespace as C4 is deprecated. 16:28:21 ok 16:28:24 +1 16:28:27 new subroutine as well actually... 16:28:30 bag, you are too slow :) 16:28:43 lol 16:28:54 Joubu: would you have tomove the wholemodule in this case? or just make a new one for the routine? 16:28:57 Joubu: I think that's a bit too tricky to codify 16:29:14 cait: I'd like to table my sub-rule on that ( single argument stuff ) for another meeting so we can get though more important stuff this meeting ; ) 16:29:25 we already voted on removing perl 18 / perl 19 16:29:32 i'd like to finish removals if ossible 16:29:37 then we will probably have to stop 16:29:42 (Ginny was distracting me) ;) 16:29:44 what's left for removals? 16:29:57 perl11 and perl12? 16:29:58 [kfischer] 'HTML5: Deprecation of the 'prog' and 'CCSR' OPAC themes.' Templats are long gone. 16:30:05 [kfischer] 'PERL11: No CVS - Development has moved from CVS to git. Therefore the use of CVS keywords $Id$ and $Revision$ should be discontinued.' Is this rule still needed? 16:30:08 Joubu: things like adding new circ functionality as Koha:: modules without just calling back into a million C4 modules would be a pretty high bar 16:30:10 [kfischer] 'PERL12: VERSION [Deprecated as of start of 3.12 release cycle] Each module should have a $VERSION variable to preserve progressive levels of compatibility with dependent code.' A bit confusing - should remaining VERSION variables be removed? 16:30:25 html5 - is easy i think 16:30:27 I'm up for a vote to remove all of the above, any objections? 16:30:30 the code is no longer there .) 16:30:39 pianohacker: yes indeed there are exceptions :) 16:30:45 for the last one, just a question: shoudl re remove existing VERSION variables? 16:31:06 they appeared earlier today in discussion as not being readonly - but i am not sure what they are used for actually 16:31:13 easy enough to do 16:31:17 We never used them 16:31:34 I volunteer to remove them if we so wish 16:31:39 would that break any plugin that put the version number in the use statement? 16:32:01 pianohacker: not plugins I'm aware of do that. 16:32:12 plugins can specify a *Koha* version though 16:32:15 then I'd say kill 'em 16:32:16 I think the koha plugins check version differently? 16:32:22 we say run perlcritic at one point but it complains if VERSION is missing 16:32:28 oh 16:32:32 cc: is that configurable? 16:32:55 hm haven't seen that in my qa script - possible we already did? 16:33:09 there is certainly a perlcritic rule to remove it 16:33:10 almost certainly -- (if you want to spend time configuring it) 16:34:02 Joubu: are you comfortable volunteering to make perlcritic happy? 16:34:03 hm right now i wonder where i got my perlcritic file from... that it uses 16:34:14 let's postpone that one and vote on the other 2 for now? 16:34:17 yar 16:34:27 sounds good 16:34:29 please vote on removal of html5 and perl12 16:34:32 cait: well, I think we're in agreement about removing the rule 16:34:33 argh 16:34:33 Actually I don't understand, we have added some new module without the $VERSION var defined, and the tests passed 16:34:36 perl 11 16:34:39 +1 16:34:44 html5 and perl11 16:34:48 but +1 16:34:52 +1 16:34:53 +1 16:34:59 +1 16:35:00 Joubu: yeah i think we need to investigate :) 16:35:04 +1 16:35:05 +1 16:35:08 +1 16:35:11 +1 16:35:37 +1 16:35:50 #agreed to remove PERL11: No CVS - Development has moved from CVS to git. Therefore the use of CVS keywords $Id$ and $Revision$ should be discontinued. 16:36:02 #agreed to remove HTML5: Deprecation of the 'prog' and 'CCSR' OPAC themes.' Templats are long gone 16:36:21 ok 16:36:29 *templates 16:36:31 who is going to amek the agreed to changes? 16:36:40 barton: toolate :) 16:36:55 I can try to - but woudl encourage people to check the wiki changes 16:37:02 i can send an email to the list 16:37:18 once done 16:37:41 ok? 16:37:43 cait: let me know when you've done it and I can review it 16:37:46 np. 16:37:53 ok 16:38:08 #action cait to apply agreed to changes on the wiki 16:38:21 that way... if i messed up the minutes i will have to sort it out :) 16:38:30 bag: next meeting? 16:38:31 it has been said that next meeting is December 5th 18 UTC 16:38:42 forget next meeting 16:38:43 cait: I forgot next meeting 16:39:06 looking 16:39:27 should we go for sooner than later, as there's plenty we couldn't get to? 16:39:35 2nd of March ? 16:39:37 would be nice if we could do bi-monthly dev meetings ( 2 a month ). Is that too often ? 16:39:48 I think sooner would be better 18:18:49 @quote random 18:18:49 gmcharlt: Quote #240: " we will have no hazing of RM's" (added by gmcharlt at 10:22 PM, April 05, 2013) 18:23:05 bug 12321 18:23:06 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=12321 normal, P5 - low, ---, gmcharlt, NEW , indicators not editable after linking authority 18:23:09 cool 18:23:12 gmcharlt++ 18:42:16 gmcharlt: any idea why it died? 18:42:38 pianohacker: Linode in Atlanta sufferred a routing snafu 18:43:20 ah kk 18:46:17 My error logs these days are all "CGI::param called in list context from package..." Makes it hard to see the real issues. 18:47:41 do we even have a bug for that yet? 18:49:58 oh you got huginn breathing again 18:50:05 gmcharlt: anyhtng we could do about the meeting logs? 18:52:05 pianohacker: Bug 14121 similar? 18:52:06 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=14121 minor, P5 - low, ---, mtompset, RESOLVED FIXED, Silence warnings t/db_dependent/Auth_with_cas.t 18:56:21 oleonard: yeah, that's the same error 19:00:40 bye 19:06:06 I imagine there exists a project where the newer bug always gets marked as the duplicate even if it has patches on it. 19:08:16 maybe it's a language thing? 19:08:36 i haven't read all of today's bug mail yet - guess i will find it sooner or later 20:45:54 hmm 20:45:58 #endmeeting 20:46:29 hm, he forgot :( 20:47:10 hmm, could you try it, cait? since you started the meeting? 20:47:18 sure 20:47:20 #endmeeting