Time Nick Message 22:21 caroline_crazycatlady good night! 22:03 ashimema Looking forward to seeing many of you next week in Marseille.. and will miss all those who can't attend. 22:03 davidnind[m] thanks! 22:03 ashimema thanks everyone who attended :) 22:03 huginn Log: http://meetings.koha-community.org/2019/development_irc_meeting_25_september_2019.2019-09-25-20.59.log.html 22:03 huginn Minutes (text): http://meetings.koha-community.org/2019/development_irc_meeting_25_september_2019.2019-09-25-20.59.txt 22:03 huginn Minutes: http://meetings.koha-community.org/2019/development_irc_meeting_25_september_2019.2019-09-25-20.59.html 22:03 huginn Meeting ended Wed Sep 25 22:03:14 2019 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) 22:03 ashimema #endmeeting 22:02 ashimema #info Next meeting: 09 October 2019, 14 UTC 22:02 thd +1 22:01 ashimema any objections/requests? 22:01 ashimema 14UTC for the earlier slow 22:00 ashimema 9th Oct would follow the usual pattern 21:58 ashimema #topic Set time of next meeting 21:58 ashimema #info Nothing to report 21:58 ashimema #topic Review of coding guidelines 21:58 ashimema so.. moving on 21:57 ashimema I'll try and sort that and put it in the next agenda.. 21:57 thd The link to the former proposal is missing, the only link is to the HAL documentation. 21:56 ashimema and without tcohen here to promote it I think we should postpone discussion untill the next meeting. 21:55 ashimema Not sure about anyone else, but I've not had a chance to get myself up to speed on this one.. 21:55 ashimema #topic API Proposal for related resources (HAL) 21:55 ashimema brill.. next topic 21:54 ashimema thanks :) 21:54 ashimema #info nkuitse is going to subit a new bug with the alternative of making the graphing availabilty syspref controlled. 21:54 nkuitse bye all -- i've got to eat dinner -- thanks for helping with this! 21:54 rangi no worries 21:54 nkuitse thanks again rangi 21:54 nkuitse i'm happy with this -- still going to patch something but i like the long-term solution a lot better 21:53 ashimema thanks nkuitse 21:53 nkuitse ok, yeah, i'll do that ashimema 21:53 ashimema that way you can see where the thinking came from 21:53 nkuitse hah! :-) 21:53 ashimema yeah, create a new bug and perhaps add the original into the 'see also' 21:53 caroline_crazycatlady [off] rangi says nothing the whole meeting and swoops in with simple yet effective solution from nowhere :) 21:52 nkuitse i don't want to muddy the waters, so to speak 21:52 ashimema that would be great 21:52 nkuitse do you think it would make sense to start a new ticket for it? 21:52 nkuitse i can take a look at the code and come up with a proposed patch 21:52 nkuitse :-) 21:51 nkuitse genius! 21:51 nkuitse charting should be on by default so as not to surprise people 21:51 rangi then those who can, can leave it on 21:51 nkuitse very clean, very easy 21:51 ashimema I like that appraoch rangi 21:51 rangi then for your use case, when you can't trust your users, you can disable charts 21:51 nkuitse ooh, i like that 21:51 nkuitse 's ok 21:51 rangi is to add a syspref, disable charts 21:51 cait sorr, got to leave 21:51 rangi what would make a lot more sense 21:50 nkuitse since there are other dedicated tools for visualizing data -- "do one thing well"! 21:50 nkuitse what i really wanted to do -- no offense to anyone who worked on the charting implementation -- was to remove charting altogether 21:50 nkuitse and Koha has no idea which is which 21:49 nkuitse but you can still try to chart the ones that don't make sense 21:49 ashimema totally 21:49 nkuitse just pointing out that charting is showing us that some reports make sense to chart and others don't 21:48 ashimema again.. I may be missing something.. it's been a while since I worked in this particular area 21:48 nkuitse i'm not suggesting munging in a GROUP BY 21:48 nkuitse but not knowing what you're doing can bring down the whole server 21:48 ashimema so how would we programatically be able to add in a GROUP BY ? 21:48 nkuitse yep 21:48 ashimema you can write pretty much any arbitrary read only sql in our reports system.. 21:48 nkuitse exactly -- i agree 21:47 ashimema I'm not sure how that would work nkuitse.. 21:47 nkuitse i don't know of a memnice kind of thing -- ionice might help, i suppose 21:47 davidnind[m] sorry, thinking out loud here :) 21:47 nkuitse rather than reading in a gazillion rows of raw data and then compiling that into the pretty pie chart the user's after 21:47 nkuitse the problem is, imho, that the graphing should be done on reports with a GROUP BY 21:47 davidnind[m] is there a way in Perl to limit what resources a script uses? something like nice that you can use on the command line (speaking from a point of ignorance about system admin + Perl) 21:46 ashimema at the same time I understand your worries 21:46 ashimema I don't see how a graph that didn't contain all rows would be useful to the end user 21:46 nkuitse and our users like reports in Koha, when we haven't told them to stop running them that is ;-) 21:46 cait we have had some libraries bring down their instance with reports pre-graphs 21:46 cait the sql reports are mighty, but they bare a bit of risk that's true 21:46 ashimema I think I'd prefer an appraoch that simply didn't return the all array untill requested 21:46 nkuitse i just head straight for mysql, i.e., from command line 21:45 nkuitse i'm woefully ignorant of fancy reporting tools 21:45 davidnind[m] could using something like metabase or Urungi in the meantime take the load of your server for reports (I haven't tried either of these, and not familiar with them except what I saw at KohaCon) 21:45 ashimema personally.. we just train all customers to take care when running reports 21:44 nkuitse right, DOS only -- though we've had some data loss (recovered from recent backup) 21:44 ashimema so not before 22 October in reality 21:44 nkuitse 888,000 bibs 21:44 ashimema it's not an imediate security issue so it won't get a special release. 21:44 cait yes, that's a lot 21:44 nkuitse we've got ~2.2 million items 21:44 nkuitse i don't know how many people have the requisite permisions to create reports, but i'm dreading a bad query that's missing a JOIN 21:43 nkuitse it's a pain, because bywater is managing our instance (we host it), but i don't want to wait much longer 21:42 nkuitse if it's that long then we'll patch it on our server and reverse it before we upgrade 21:42 cait next month 21:42 cait so it will be a little while 21:42 cait nkuitse: it was me saying on the bug that maintenance releases just happened 21:41 nkuitse as long as we don't paint ourselves into a corner in the process 21:41 nkuitse i'm fine with wording tweaks; i'm lobbying for a quick fix now and then improve on it later 21:40 cait hm yes - wonder if we shuld mention the pref, but too much text is not good either 21:40 nkuitse so it's pretty clear 21:39 nkuitse the wording is "Include first n rows of full report (ignore pagination)" -- that's the label for the checkbox 21:39 huginn Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=23624 normal, P5 - low, ---, paul, Passed QA , Count rows in report without (potentially) consuming all memory 21:39 nkuitse https://bugs.koha-community.org/bugzilla3/attachment.cgi?id=92994 21:39 ashimema oh good.. Tomas already did the followup I would have proposed in bug 23624 21:39 cait cool 21:39 nkuitse that's exactly what the latest rev of the patch does: 21:38 nkuitse yeah 21:38 cait i just want to avoid people not being aware of limits 21:38 cait a hint on the graph dialog maybe 21:37 nkuitse because that uses LIMIT and OFFSET 21:37 nkuitse but keep in mind that the limit would not affect browsing through report results 21:37 nkuitse yes, some recent tweaking in the work on the bug does just that 21:36 nkuitse oh, sorry, i misunderstood 21:36 nkuitse you mean a maximum that the user can set when running the report? 21:36 cait please note... only a maximum of x rows will be processed or something like that 21:36 cait as a hint? 21:36 cait could we include the pref setting int the gui? 21:36 nkuitse like $nrows * $ncols * 100 or something 21:36 nkuitse not that i know of, but we could certainly use some heuristics 21:35 thd Is there any efficient way to test quickly for the overall size of the returned data before actually returning it? 21:35 nkuitse but how do they know if a report will generate large results? -- they're not DB folks! 21:35 nkuitse for us, the problem is the outages -- our staff need to run reports but we've had to tell them to hold off on anything that might generate large results 21:34 ashimema 10,000 rows for a report that returns 50 fields is obviously not 21:34 nkuitse there's a very nice fix for 23624 -- very clean 21:34 ashimema 10,000 rows for a report that only returns two fields would be low memory.. 21:33 huginn Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=23624 normal, P5 - low, ---, paul, Passed QA , Count rows in report without (potentially) consuming all memory 21:33 nkuitse the query for the result set size was bug 23624 21:33 ashimema but it would be incredibly report dependant 21:33 thd Would it not help to introduce a slight delay to first query the size of the result set and if it would be over some large number of rows then prompt with an altered query form with some restriction set to avoid thrashing the system along with a warning. 21:33 nkuitse if you never have memory issues then you leave it unset, i.e., unlimited 21:33 nkuitse so, for example, if you feel pretty good about allowing 10,000 rows then you set it at 10,000 21:32 nkuitse a new sys pref that you can set to limit the number of rows 21:32 cait i tihnk it adds a pref 21:32 ashimema I've not had a chance to look at your patch yet so I may be misunderstanding your proposed solution 21:32 nkuitse the patches in the big report default to unlimited 21:32 ashimema to me a requiest to build a graph should be handled distinctly and with a warning of possible memory issues.. 21:32 cait i tihnk we need some message/tooltip in the gui 21:31 ashimema My gut feeling is that limiting the return by default is incorrect as it leads to meaningless output 21:31 nkuitse i'm not sure -- i haven't checked, or at least i don't think i have 21:31 davidnind[m] speaking from a point of ignorance about all things Perl :) 21:30 nkuitse the big one that first brought this to our attention was a report of all items at a library 21:30 davidnind[m] does that mean there is someting in the other report-related Perl scripts that stops them taking over the system resources? 21:30 nkuitse anything at all 21:30 ashimema sorry.. I didn't mean version.. I meant what reports allow graphs.. whether it's anything you build of just limited subset 21:30 cait nkuitse: there is the guided reports step by step thing - and sql - but when you run them it's the same yes 21:29 cait yes 21:29 cait 18.11 have them too 21:29 nkuitse no, we're running 18.11 21:29 cait um sorry no 21:29 cait 19.05 21:29 * ashimema can't remember off the top of his head where/when graphs become available 21:29 nkuitse aren't they the same thing? 21:29 cait i think they share templates 21:29 cait saved sql too? 21:28 nkuitse yes 21:28 ashimema so it's only for guided reports? 21:28 nkuitse that's the long and short of it 21:28 nkuitse (and only if they check the box that says "use all rows" or whatever when creating the chart!) 21:28 nkuitse but guided_reports.pl has no way of knowing whether the user will do that 21:27 nkuitse the full results aren't needed unless the user ends up making a chart 21:27 nkuitse so that the charting javascript can (if the user ends up charting) use all results 21:26 nkuitse then it passes \@allrows to the TT2 template 21:26 nkuitse } 21:26 nkuitse push @allrows, { cells => \@cells }; 21:26 nkuitse my @cells = map { +{ cell => $_ } } @$row; 21:26 nkuitse while (my $row = $sth2->fetchrow_arrayref()) { 21:26 nkuitse the problem is in guided_reports.pl 21:25 nkuitse one report recently led to (I think) 5+ GB above and beyond normal memory consumption 21:25 ashimema can you summarise the problem nkuitse 21:25 nkuitse the charting feature in guided_reports.pl consumes unbounded memory -- all data is loaded into memory, converted to JSON (so then 2 copies in mem) and only then sent back to browser 21:25 ashimema This one was only braught to my attention about an hour ago.. not had a great deal of time to look into it.. 21:24 nkuitse we've had several outages and service degradations recently due to this 21:24 nkuitse that's my bug 21:24 huginn Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=23626 major, P2, ---, paul, Signed Off , Add a system preference to limit the number of rows of data used in a chart when viewing report results 21:24 ashimema #info Bug 23626 21:24 thd I am trying to keep additional undefined values out of wiki migration where I can. Undefined values might not be avoidable for some columns. 21:23 ashimema #topic What's the best way to limit the amount of memory consumed when running a guided report? 21:23 ashimema lets jump on 21:23 ashimema nope.. not yet covered davidnind[m] 21:22 ashimema well.. it sounds like there's interest in such guidance so I'll try to come up with some words and we can continue discussion next meeting. 21:22 davidnind[m] is this covered in the devlopment guide? looks like it should be if it isn't already 21:21 ashimema thd++ 21:20 thd Undefined values provoke bugs where the code is not expecting undefined. I personally find seeing undefined to be ugly, difficult to read in some context, and seldom necessary. 21:20 ashimema yup... 21:20 cait empty could still be counted? 21:20 ashimema I think when the return is expected to be a set rather than a singular then we should always return a set (even if empty).. but for the cases where a singular is expected then undef checks are more sane 21:20 cait ashimema: if we have checks for 'number of objects returned' could that be a problem? 21:19 cait I tihnk in this case there is not really a difference between empty and undef 21:18 thd Undefined may have a logical status which may be distinct from an empty set value, however, it seems to be too often used without any logical distinction between undefined and empty set. 21:18 cait d 21:18 cait bugs are ba 21:18 ashimema it's mostly a matter of style.. but I've certinly seen a number of cases lately where the defined check was missed and as such we ended up with bugs 21:18 cait it sounds like the nicer way, but not sure i have enough knowledge to be helpful here :) 21:18 ashimema I mostly wanted to get some opinions on it.. before really suggesting a coding guideline 21:17 ashimema or... we could always return an object, object set and as such do away with the need for the defined check 21:16 ashimema $thing = defined($result->link_method) ? $result->link_method->related_method : undef; 21:15 cait if you get an empty object - no checks required? 21:15 cait hm 21:15 ashimema we could either fall to QAers to make sure any calls to such methods are always of a ternary form 21:14 huginn Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=23272 enhancement, P5 - low, ---, tomascohen, RESOLVED FIXED, Koha::AuthorisedValue should use Koha::Object::Limit::Library 21:14 ashimema So.. whilst QAing bug 23272 I noticed that we're starting to develop a pattern bad practice. 21:14 cait is there a best practice there we could follow? 21:13 ashimema #topic Should we allow Koha::Object methods that normally return Koha::Object or Koha::Objects type object to ever return 'undef'? 21:12 ashimema #topic General development discussion (trends, ideas, ...) 21:12 cait yep 21:12 ashimema moving on?.. 21:12 ashimema my bad 21:12 ashimema oops 21:11 cait not 26 in the qa queue... but i would like the to be right now 21:11 ashimema #info QA queue: 26 Major/Critical, 77 Total 21:10 cait i hope to have a little session next week with the attending parts of QA team to discuss a few of them together 21:10 cait 77 in the QA queue 21:10 * ashimema has a headache coming on so is rushing through a bit ;) 21:10 cait most are new/nso, so work for everyone there 21:10 ashimema anything else cait 21:10 cait 26 majors/criticals that need attention, in different stages 21:10 ashimema #info cait and ashimema will work on some triage next week at Marseille hackfest. 21:09 ashimema #info It's been a quiet summer, but some bigger bugs have recently hit the QA queue and need attention. 21:09 cait let me take a quick look 21:09 cait it's been a bit of a quiet summer 21:09 ashimema #topic Updates from the QA team 21:08 ashimema #info The latest releases have all gone out, well done RMaints and Team. 21:08 cait maybe we should do some triage work next week 21:08 cait I think some bigger things just hit the QA queue, but I am bit behind 21:08 ashimema #topic Updates from the Release Maintainers 21:08 ashimema #info As above, we're now moving into the polishing phase of the cycle.. there's still time for a few more minor enhancements but generally I would like the team to concentrate on polishing what already pushed and bugfixes. 21:06 ashimema #topic Update from the Release manager (19.11) 21:06 ashimema rmaints 21:06 ashimema tcohen around? 21:06 ashimema yeay :) 21:06 nkuitse #info Paul Hoffman, Fenway Library Organization, USA 21:05 cait #info Katrin Fischer, BSZ, Germany 21:05 ashimema anyone have anything else? 21:05 ashimema #info Marseille Hackfest is upon us... 30 Sept - 4th Oct 21:05 ashimema #info The 'adventurous pushing' window is now closed.. though I will contemplate a few more during Marseile hackfest. Lets move into the polishing phase. 21:03 ashimema #topic Announcements 21:01 thd #info Thomas Dukleth, Agogme, New York City - no more jury duty for a few years again 21:00 davidnind[m] #info David Nind, NZ 21:00 * ashimema is aiming for a short one tonight 21:00 ashimema #info Martin Renvoize, PTFS Europe 21:00 ashimema #topic Introductions 20:59 davidnind[m] Will be back on my desktop soon, on mobile phone at the moment... 20:59 huginn The meeting name has been set to 'development_irc_meeting_25_september_2019' 20:59 huginn Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 20:59 huginn Meeting started Wed Sep 25 20:59:50 2019 UTC. The chair is ashimema. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:59 ashimema #startmeeting Development IRC meeting 25 September 2019 20:58 thd Yes, I survived jury duty ;) 20:57 ashimema anyone around for the dev meeting? 20:47 huginn caroline_crazycatlady: The operation succeeded. 20:47 caroline_crazycatlady @later tell cait nvm found the picture frame... now trying to find out what the URL is for my image... 20:40 huginn caroline_crazycatlady: The operation succeeded. 20:40 caroline_crazycatlady @later tell cait can you walk me through adding an image in a news item? I can't find the picture frame thing you were talking about 20:33 giuseppep_ Hello, after upgrade to 19.05 I'm having this issue when booking from OPAC, it return me Template process failed: undef error - The method Koha::Hold->is_cancelable is not covered by tests! . What is it ? 20:30 * ashimema is feeling rough tonight and really just wants to go to bed.. 20:29 wahanui i heard Next meeting was https://wiki.koha-community.org/wiki/Next_IRC_meetings 20:29 ashimema Next meeting 20:29 wahanui Meeting is over :) 20:29 ashimema Meeting? 20:17 davidnind[m] thanks! 20:08 caroline_crazycatlady I'll let you know what I think of it 20:08 caroline_crazycatlady the book has been ordered! :) 19:53 davidnind[m] Would be interested to know if it is useful or not - it looks like it would be 19:52 davidnind[m] Having never done a migration, the table of contents seems to cover some of things people talk about on the lists or IRC 19:51 caroline_crazycatlady I think I will ask my boss to buy it for me 19:51 caroline_crazycatlady davidnind[m]: interesting! 19:20 davidnind[m] Has anyone every used it, or could recommend it either way? 19:19 davidnind[m] This looks like an interesting book - Migrating Library Data: A Practical Manual https://www.alastore.ala.org/content/migrating-library-data-practical-manual 17:26 huginn Bug 23675: enhancement, P5 - low, ---, koha-bugs, NEW , UseACQFrameworkForBiblioRecords does not work with diacritics (MARC21) 17:26 huginn caroline_catlady: The operation succeeded. 17:26 caroline_catlady @later tell cait can you test? https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=23675 17:00 koha-jenkins Project Koha_18.11_U18 build #174: SUCCESS in 49 min: https://jenkins.koha-community.org/job/Koha_18.11_U18/174/ 17:00 cait ok, got to go :) cya al later 16:58 cait maybe :) 16:58 * oleonard guesses German ghosts say "üüüüüüüüüüüüüööööööööööööööääääääää" 16:58 caroline_catlady ok 16:57 cait but if you file it, i might be able to later 16:57 cait can't test right now, just have a few minutes until bus time 16:57 cait üöä would be some typical German ones 16:57 caroline_catlady ok. I'm filinf a bug, but I wanted someone to confirm I wasn't hallucinating :) 16:57 cait because we want to - but updates haven't happened yet 16:56 cait caroline_catlady: not yet, not good to hear 16:55 caroline_catlady oh well, thank you tcohen :) 16:55 wahanui it has been said that tcohen is the man you want to thank 16:55 caroline_catlady tcohen? 16:54 caroline_catlady or magnuse maybe? I think norwegian has some non standard characters 16:49 caroline_catlady *some 16:48 caroline_catlady I'm having problems with diacritics and I wondered if it was the same with come german characters 16:47 caroline_catlady cait, do any of your libraries use UseACQFrameworkForBiblioRecords ? 16:11 koha-jenkins Project Koha_18.11_U18 build #173: SUCCESS in 50 min: https://jenkins.koha-community.org/job/Koha_18.11_U18/173/ 15:51 koha-jenkins Project Koha_18.11_D8 build #180: SUCCESS in 26 min: https://jenkins.koha-community.org/job/Koha_18.11_D8/180/ 15:51 koha-jenkins Project Koha_18.11_D9 build #181: SUCCESS in 26 min: https://jenkins.koha-community.org/job/Koha_18.11_D9/181/ 15:27 reiveune bye 15:04 koha-jenkins Project Koha_18.11_D8 build #179: SUCCESS in 38 min: https://jenkins.koha-community.org/job/Koha_18.11_D8/179/ 15:01 koha-jenkins Project Koha_18.11_D9 build #180: SUCCESS in 32 min: https://jenkins.koha-community.org/job/Koha_18.11_D9/180/ 14:54 koha-jenkins Project Koha_18.11_U18 build #172: SUCCESS in 27 min: https://jenkins.koha-community.org/job/Koha_18.11_U18/172/ 14:29 caroline_catlady hi wizzyrea! 14:29 wahanui privet, wizzyrea 14:29 wizzyrea hi 13:51 wahanui OK, oleonard. 13:51 oleonard wahanui: everyone is headed to the hackfest in Marseille 13:50 wahanui oleonard: I forgot everyone 13:50 oleonard wahanui: forget everyone 13:50 wahanui everyone is headed to kohacon that's where everyone is 13:50 oleonard wahanui: everyone? 13:25 wahanui somebody said everyone was headed to kohacon that's where everyone is 13:25 caroline_crazycatlady hey everyone! 11:48 calire :D 11:47 oleonard Still waking maybe 11:47 oleonard Backing? 11:46 * oleonard is backing, keeping his seat warm 11:44 calire come back! 11:44 oleonard-away Oh no, I'm not away! 11:44 calire hi oleonard-away 11:38 oleonard-away Hi #koha 09:25 koha-jenkins Project Koha_Master_D8 build #424: STILL UNSTABLE in 42 min: https://jenkins.koha-community.org/job/Koha_Master_D8/424/ 09:18 koha-jenkins Project Koha_Master_U18 build #362: FIXED in 37 min: https://jenkins.koha-community.org/job/Koha_Master_U18/362/ 09:18 wahanui Congratulations! 09:18 koha-jenkins Yippee, build fixed! 09:10 koha-jenkins Project Koha_Master_D9 build #880: FIXED in 31 min: https://jenkins.koha-community.org/job/Koha_Master_D9/880/ 09:10 wahanui Congratulations! 09:10 koha-jenkins Yippee, build fixed! 08:39 huginn News from kohagit: Bug 23517: (follow-up) AddReserve expects a priority parameter <http://git.koha-community.org/gitweb/?p=koha.git;a=commitdiff;h=06bd256019f519124676719db976d4f6e8258131> 07:01 magnuse \o/ 07:00 liliputech_asu salut #koha! 06:48 alex_a Bonjour 06:34 reiveune hello 06:19 huginn magnuse: The operation succeeded. 06:19 magnuse @later tell ashimema already making plans for the next hackfest in marseille? 06:16 * magnuse waves 00:38 koha-jenkins Project Koha_18.11_D8 build #178: SUCCESS in 38 min: https://jenkins.koha-community.org/job/Koha_18.11_D8/178/ 00:37 koha-jenkins Project Koha_18.11_U18 build #171: SUCCESS in 35 min: https://jenkins.koha-community.org/job/Koha_18.11_U18/171/ 00:25 koha-jenkins Project Koha_18.11_D9 build #179: FIXED in 26 min: https://jenkins.koha-community.org/job/Koha_18.11_D9/179/ 00:25 wahanui Congratulations! 00:25 koha-jenkins Yippee, build fixed!