IRC log for #koha, 2022-11-25

← Previous day | Today | Next day → | Search | Index

All times shown according to UTC.

Time Nick Message
04:09 Oak joined #koha
05:45 Joubu @later tell tcohen What do we do for the "purge bg" failing test? Will you push the follow-up on 31969 or 32304?
05:45 huginn` Joubu: The operation succeeded.
06:10 mtj hmm, looks like my internet is busted today, rsyncing is taking forever :/
06:11 12 kB/s
06:20 Joubu hi mtj o/
06:20 Happy release day ;)
06:27 mtj hiya Joubu
06:27 Joubu @later tell tcohen see bug 32343 for the random failure from Patron.t!
06:27 huginn` Joubu: The operation succeeded.
06:28 Joubu bug 32343 # easy SO and QA
06:28 huginn` 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=32343 normal, P5 - low, ---, chris, Needs Signoff , Koha/Patron.t is failing randomly
06:30 mtj hmm, i might reboot my router, etc
06:33 upload from voda-wgtn to voda-chch is 100Mbps - but 3Mbps to spark-wgtn :/
06:34 ..so, a problem with outbound vodafone traffic
06:39 waiting 3 hours for an rsync that usually takes about 60 seconds 🍺
07:19 magnuse \o/
07:27 ashimema Morning
07:49 alex_ joined #koha
07:56 thibaud_g joined #koha
07:59 marcelr joined #koha
07:59 marcelr o/
08:00 alex_ Bonjour \o
08:01 lds joined #koha
08:01 reiveune joined #koha
08:01 reiveune hello
08:06 lukeg joined #koha
08:12 cait joined #koha
08:17 cait good morning #koha
08:21 marcelr hi cait
08:22 ashimema bug 32350
08:22 huginn` 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=32350 enhancement, P5 - low, ---, martin.renvoize, Needs Signoff , Should we die if TestBuilder is passed a column we're not expecting
08:22 ashimema would love your thoughts on the above Joubu
08:23 marcelr ashimema: He currently ignores the column in the value hash ?
08:23 My first thought would be: just ignore it
08:24 qa_team: bug 32304 please
08:24 huginn` 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=32304 normal, P5 - low, ---, m.de.rooy, Signed Off , Fix subtest search_limited and purge in BackgroundJobs.t
08:24 marcelr for Jenkins
08:24 ashimema so yes.. right now we just ignore such things..
08:25 the idea of this is to spot dev mistakes and thus make the test suite more reliably
08:25 see all the recent bugs Joubu's been fixing in tests because of misnamed fields passed to testbuilder
08:25 marcelr ok
08:25 ashimema I've QA'd at least three in the last few days I think.
08:26 marcelr you ought to think that the developers runs its own test once again
08:26 ashimema but yeah.. it may well not be worth the performance burden.. I don't think it's massive, but we lean on testbuilder a lot.. and there's lots of loops in there.. so this could end up being executed a lot.
08:26 well.. the issue is it only fails randomly..
08:26 marcelr you wont solve the random failures
08:27 ashimema you might get lucky in your runs and have testbuilder magically fill your value with something valid
08:27 marcelr true
08:27 but not depending on wrong column
08:27 ashimema huh
08:27 marcelr wrong column is just ignored
08:27 ashimema that's the point
08:27 ok..
08:28 see bug 32343
08:28 huginn` 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=32343 normal, P5 - low, ---, jonathan.druart+koha, Passed QA , Koha/Patron.t is failing randomly
08:28 ashimema that one was a misnamed column in the test..
08:28 marcelr ah ok
08:28 i get your point now :)
08:28 ashimema it passes half the time because by fluke testbuild spits out the right random bit for the field
08:29 marcelr well spotted
08:29 ashimema to make it pass 100% we just fixed the field name in the test
08:29 there's been a number of these recently basically
08:29 but yeah..
08:29 I'm still of two minds whether it's worth it 🙂
08:29 so really glad to have your feedback 🙂
08:29 marcelr i think jenkins wants your patch
08:29 these random fails are ugly
08:30 actually they come from lazy devs
08:30 * ashimema looks at your test fix now on bug 32304
08:30 huginn` 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=32304 normal, P5 - low, ---, m.de.rooy, Signed Off , Fix subtest search_limited and purge in BackgroundJobs.t
08:30 ashimema indeed
08:30 marcelr we just do not specify columns we need
08:30 or even type them wrong..
08:31 ashimema indeed 🙂
08:31 this would only spot typos and renames
08:31 but it's a start
08:32 marcelr i suppose that tomas wont push 32350 anymore ?
08:32 because it might influence all tests using TB
08:34 ashimema haha, yeah..
08:34 Joubu ashimema: yes, should die if a value is not correct
08:34 ashimema I reckon it's one for first thing next cycle.. followed by a walk through of fixing all the tests it may well highlight are somewhat broken 😜
08:34 I've not run it against the whole suite.. I really should
08:34 Joubu that will help us catch wrong things
08:35 ashimema do you agree it's a first thing next cycle job.. or should I stick it in this morning and watch the testsuite explode
08:35 marcelr ashimema: you should do the check in build not in build_object
08:35 we are using build too still
08:35 and build_object calls bild
08:35 build
08:36 ashimema er.. I'll read back.. but I thought i put it inside he _buildColumnValues sub that's used by build..
08:36 it was a quick POC this morning
08:36 oop.. school run time.. back in half an hour
08:36 marcelr ah yes
08:37 i am reading wrong
08:37 forget it
08:39 Joubu we dont use Array::List ?
08:40 Joubu array::utils, sorry
08:41 marcelr i will add it on the patch
08:41 and test
08:42 Joubu marcelr: see last comment, there is a diff
08:43 ashimema Oh, do we already have Array::List as a dep.. I should have checked that better
08:43 Joubu marcelr: did you see my follow-up on bug 31969?
08:43 huginn` 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=31969 enhancement, P5 - low, ---, bjorn.nylen, Pushed to master , Options for cleanup_database.pl to remove finished jobs from the background job queue
08:43 marcelr no Array::Utils
08:44 ashimema: i am testing it now
08:44 Joubu Array::List does not make any sense :D
08:44 sorry for the confusion
08:44 marcelr lol
08:44 ashimema Sorry, I looked for Array:: Compare
08:44 Too many darn modules that do almost the same thing
08:45 Anyway, that short is that I agree to using a module if we've already got one were using for these types of compare .  It's much nicer code than my manual loop
08:45 marcelr no Joubu, didnt see that patch - i am fixing a wall that we are slamming
08:46 ashimema: its just one or two lines difference
08:46 ashimema It is, but this also has the advantage that it catches all failings in one hit
08:47 marcelr we need array_minus
08:47 not intersect
08:47 i will add it
08:48 and add a test :)
08:49 ashimema team++
08:49 Joubu I am expecting a lot of failures to be honest... :D
08:49 marcelr: ofc array_minus...
08:50 nope
08:51 yes, @values - @columns
08:51 ashimema Me too
08:53 Joubu and about the purge method, I strongly thing it's all wrong, and the follow-up should be pushed instead
08:53 and remove all the unecessary code
08:55 ashimema I was about to swing back around to that one
08:55 Joubu tuxayo: You send a lot of "Patch testing session", but no feedback about how it went. Do you get a lot of attendees? Which patches do you/they test? etc.
08:56 marcelr Joubu my second patch on 32304 should be obsolete, the first should be fine
08:56 Joubu mine makes yours useless as it remvoes the method..
08:57 marcelr sure
08:57 i fix two tests
08:57 i added another patch for jenkins this morning
08:58 Joubu talking about 'purge' or another thing?
09:03 marcelr same bug, another subtest search_limited
09:03 fails on JK
09:04 nice Error: value hash contains unrecognized columns: flag at /usr/share/koha/t/lib/TestBuilder.pm line 287.
09:07 TestBuilder uses FK hashes too in value => {} so we need a bit more logic
09:07 nice
09:15 TestBuilder.t itself contains a wrong column lol
09:15 ashimema: ^
09:15 ashimema LOL
09:17 marcelr but it survives the build of all objects..
09:18 which is a good sign
09:18 i needed something like this now too: my @passed_keys = grep { ref($original_value->{$_}) ne 'HASH' } keys %$original_value;
09:19 nested stuff for FKs
09:20 ashimema ah.. yes.. I forgot how nesting works on this
09:23 Joubu can you replace 'value' with 'values' where appropriate? :D
09:23 ashimema that would be lovely to do
09:24 it's bugged me for year.. but has never bugged me enough to actually do anything about.
09:24 marcelr and now the enhancements keep coming :)
09:24 ashimema tuxayo sessions have had mixed attendance.. some reasonable, some not so much.
09:25 we're trying to broadcast them a bit more from the ptfs-e side now (we've had some teething issues with a new comms channel for our customers)... and I intend to attend/send some ptfs-e devs to them to help tuxayo
09:30 JBoyer joined #koha
09:36 Joubu Who's using the pre-commit hook from the wiki?
09:37 and who added the check on "use"? (how can I 'git log' the changes?)
09:37 git blame even
09:38 ashimema: ?
09:38 is that you?
09:39 the `perl -wc` is also not correct I guess, you don't want to commit from ktd, or install the deps on the host
09:47 ok so basically the pre-commit hook on the wiki is wrong for more than 3 years, and nobody noticed
09:47 great :D
09:47 ashimema I found out that I'd lost my git hooks at some point
09:47 no idea when but when I went to update it with the vue stuff I found they were all missing
09:48 honestly, I can't remember what it did.. it's on my list to revisit
09:51 Joubu I am on it
09:58 ashimema: Can you review the last 3 revs please? https://wiki.koha-community.or[…]ks&action=history
10:10 ashimema yes.. but not for a couple of hours 😜
10:10 In a meeting now
10:22 marcelr ashimema: sorry for hijacking your report now
10:23 magnuse_ joined #koha
10:28 marcelr so yes, we start finding wrong testbuilder inits
10:29 ashimema not at all..
10:29 it's great to see
10:30 marcelr should we fix them on the same report ?
10:30 or another follow-up report
10:31 finding few more now in db_dependent..
10:32 so definitely 23.05
10:49 Joubu marcelr: IMO it is better to fix on the same bug
10:51 magnuse joined #koha
10:51 ashimema cait around?
10:52 marcelr Joubu we have a quite a bunch; I prefer to separate
10:52 * ashimema is sure that used to give me a mass of cait1, cait2 etc
10:53 ashimema @later tell cait can you switch the status to 'RESOLVED FIXED' on bug 32186 for me.. I'm testing a dashboard change.
10:53 huginn` ashimema: The operation succeeded.
10:53 marcelr bug 32351 is a silly example of the TestBuilder errors
10:53 huginn` 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=32351 normal, P5 - low, ---, m.de.rooy, Needs Signoff , Fix TestBuilder calls in Koha/Holds.t
10:53 Joubu marcelr: ok, but they will have to be pushed at the same time.. no much benefits
10:53 marcelr right
10:53 but it allows team effort too
10:53 Joubu ha, 1 bug per failure?!
10:53 marcelr i do not feel obliged to fix them all
10:54 i didnt create them too ;)
10:54 32351 is just an example; we can combine obviously
10:55 Joubu that's also a good example of.... why is the test passing if there is such obvious typo? :D
10:56 marcelr yeah it doesnt make sense at all
10:56 great dev
10:56 ashimema scary
10:56 * ashimema hopes that wasn't him
10:56 marcelr i think you arent
10:56 lds_ joined #koha
10:57 ashimema my mistake rate has picked up ☹️
10:57 hopefully it'll come down again
10:57 marcelr hmm i added a name Pass in a failing test
10:57 or something like that
10:58 Joubu `biblio_id` int(11) NOT NULL DEFAULT 0 COMMENT 'id for the bibliographic record the group belongs to',
10:58 default 0?
10:58 marcelr in a test that passes for a failure
10:58 man
10:58 whats that
10:59 Joubu marcelr: did you run the whole test suite?
11:00 marcelr yes only mentioned a few exceptions like selenium
11:00 Joubu what's the approximative number of errors? 10, 100?
11:00 ok, didn't see you comment
11:00 your*
11:00 marcelr 10 and some tests that we need to check
11:00 Joubu I was expecting much more
11:01 marcelr some fails due to my env probably
11:01 t is not relevant here
11:01 maybe we should fix them all on 251
11:01 Joubu why is t/db_dependent/yaml.t db dependent?..
11:01 marcelr 351
11:02 probably because some one pushed it
11:02 Joubu and the same one wrote it
11:07 marcelr: I would remove the wrong param actually, instead of fixing its name
11:07 if it's wrong it's not needed
11:08 marcelr ok
11:08 Joubu with special attention given to occurrences like item_group where it can hide something else
11:10 marcelr done
11:11 This more or less could apply to all cases ? Since they probably passed..
11:12 And now die on wrong name
11:12 It feels a bit awkward
11:13 htg now; please test 32350 and submit new fixes on 32351 :)
11:28 Joubu @later tell tcohen 32352 - jenkins failing on xt/check_makefile.t. Just ignore it for 22.11 I'd say.
11:28 huginn` Joubu: The operation succeeded.
11:30 ashimema OK.. I have an hour before my next meeting
11:51 Joubu ashimema: can we simplify 32350?
11:51 first and last patches are useless
12:03 ashimema yes.. we should squash to clean them up
12:08 Joubu tcohen: I think we need 32353 for 22.11, see last comments from 31447 for context
12:20 ashimema: something weird happening on 32353
12:20 tcohen hola #koha
12:21 ashimema I was just looking there
12:22 weird
12:23 Joubu ashimema: ok found it
12:23 the order of the if, elsif
12:24 ashimema ah... interesting
12:24 Joubu but it is still failing...
12:24 ashimema dang
12:24 Joubu have to go, I don't know if I have much time this afternoon :-/
12:28 * ashimema will pick it up after his next meeting
12:28 ashimema thanks for all the work today Joubu
12:40 tcohen Joubu++
12:41 thank you, Joubu
12:50 alex_ joined #koha
15:37 ashimema marcelr or cait still around?
15:57 where did we get with bug 32351 marcelr
15:57 huginn` 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=32351 normal, P5 - low, ---, koha-bugs, ASSIGNED , Fix all TestBuilder calls failing due to wrong column names (see 32350)
15:57 ashimema was it ready for SO?
16:55 tuxayo Joubu: «feedback about how it went» I do it in meetings (except the two last ones that I missed) Poor attendance in general, 2 people usually, 2 or 3 patches out of the SO queue in average I think. It's not looking great to continue at this pace.
16:58 ashimema ptfs-e will bring more along to the next ones
16:58 right.. clocking off time
16:58 have a great weekend #koha
16:59 * ashimema looks forward to the next release and another great cycle..
16:59 tuxayo see ya :)
17:03 reiveune left #koha
18:10 tcohen rmaints
18:10 tuxayo: around?
18:10 need your help :-D
18:36 have a great weekend y'all
21:22 ashimema tcohen++
22:00 emlam joined #koha

← Previous day | Today | Next day → | Search | Index

koha1