[00:15:46] lzia, find me in the hammock if you're still around [00:16:03] I'm wfh today, Eloquence. :-( tomorrow morning? [00:16:13] or Hangout in a little bit? [00:16:26] sure [00:16:29] ping me whenever [00:18:27] okay. I will, Eloquence. [00:40:32] okay, how much do I have to pay one of you to do my taxes for me [00:56:11] Eloquence: how long will you be around? I'm stuck in something and would like to finish it before switching tasks if we can talk later. [00:56:24] a while :) [00:56:31] okay. :-) I'll ping then. thanks! [02:35:02] Eloquence: shall we talk in Hangout? [03:14:16] lzia, back now if you're still around [09:23:34] Analytics, Language-Engineering, MediaWiki-extensions-UniversalLanguageSelector, Mobile-Apps, and 5 others: there should be a comparison of clicks count on interlanguage links on different platforms - https://phabricator.wikimedia.org/T78351#1112414 (Amire80) [11:06:58] (PS1) QChris: Fix mediacounts description for status dumping script [analytics/refinery] - https://gerrit.wikimedia.org/r/196200 [11:07:00] (PS1) QChris: Add mediacounts top1000 files to status dumping script [analytics/refinery] - https://gerrit.wikimedia.org/r/196201 [13:41:48] Analytics-Cluster: Create HivePartitioner in Camus - https://phabricator.wikimedia.org/T92494#1112904 (Ottomata) NEW a:Ottomata [14:00:33] Analytics-Cluster: Create HivePartitioner in Camus - https://phabricator.wikimedia.org/T92494#1112973 (QChris) Not saying that such a partitioner is a bad idea, but just to call this out: * Bash's tab completion is a bit wonky around `=` *). This is already driving me insane around refined tables. * Hive's... [14:07:37] Analytics-Cluster: Create HivePartitioner in Camus - https://phabricator.wikimedia.org/T92494#1112986 (Ottomata) Agree, but if we use Hive's default,. then we can do MSCK REPAIR TABLE to automatically add partitions. This ticket is more of an open-source-work project for me. It would make Camus more useabl... [14:17:15] Analytics-Kanban: Implement adjustments to Visual Editor dashboards {lion} - https://phabricator.wikimedia.org/T91418#1113008 (Milimetric) [14:18:59] (CR) Ottomata: [C: 2 V: 2] Fix mediacounts description for status dumping script [analytics/refinery] - https://gerrit.wikimedia.org/r/196200 (owner: QChris) [14:19:03] hey qchris [14:19:09] Heya ottomata. [14:19:17] Analytics-EventLogging, Analytics-Kanban, operations: EventLogging query strings are truncated to 1014 bytes by ?(varnishncsa? or udp packet size?) - https://phabricator.wikimedia.org/T91347#1113013 (mforns) a:mforns>Ottomata [14:20:13] how hard would it be to make the raw partition status dump output [14:20:20] Analytics-EventLogging, Analytics-Kanban, operations: Eventlogging JS client should warn users when serialized event is more than "N" chars long and not sent the event - https://phabricator.wikimedia.org/T91918#1113018 (mforns) a:Nuria>mforns [14:20:26] show something useful about how bad things are [14:20:33] rather than just an X, if anything is bad? [14:21:22] You mean like %-different? [14:21:34] (From the sequence stats) [14:26:12] yeah, but percent-different is per host [14:26:30] Yes, but one can aggregate those things. [14:26:33] could we average it? that might hide outliers, but it would be better thna X [14:26:48] or maybe [14:26:59] if any host has more than i dunno 2 or 3% loss or duplicate? [14:27:35] It's just plain bash, so it should be easy to adapt. [14:27:53] I am not sure about what is alert-worthy. [14:28:26] Also ... since there is deduplication ... I am no sure how much duplicates matter on the raw partition. [14:28:32] true [14:28:47] i think i'd still ike to know though, i can tell if that number is positive [14:28:50] that i don't need to react [14:29:58] Since I am overloaded with work, [14:30:28] nuria, can you give me some details on the EL log size limit task? [14:30:28] I probably won't find time to do that the next weeks [14:30:30] yeah, ain't no thing qchris, i'm just wondering [14:30:49] But whoever wants to do it ... it should be easy and the relevant method would be: [14:31:24] dump_dataset_raw_webrequest_partition [14:31:44] of bin/refinery-dump-status-webrequest-partitions [14:34:25] Analytics-Cluster: Augment refinery-dump-status-webrequest-partitions script to show more useful status of webrequest raw partitions - https://phabricator.wikimedia.org/T92499#1113038 (Ottomata) NEW [14:34:50] Cool. [14:35:45] mforns: lemme look at ticket again ...... [14:35:58] nuria, sure: https://phabricator.wikimedia.org/T91918 [14:37:31] Analytics-EventLogging, Analytics-Kanban, operations: Eventlogging JS client should warn users when serialized event is more than "N" chars long and not sent the event - https://phabricator.wikimedia.org/T91918#1113048 (Nuria) [14:38:54] mforns: ok, just looked. If events are too large (over a configured limit N) they should not be sent and we should warn developers on the console. This limit shoudl also be enforced on the dev-server. [14:38:59] mforns: makes sense? [14:39:12] yes, what about the other clients? [14:39:58] nuria, should we start with this one, and then work on the others? [14:40:51] mforns: the mobile clients will be taken care of by mobile team, the server client does not use urls to post so limits are of different nature. [14:41:03] mforns: right now the iOS client is alredy changed [14:41:04] all right [14:41:09] oh, ok [14:41:09] *already [14:41:22] ok, and another question [14:42:01] the js client and devserver will share the same limit, you talked about having them both get the limit value from the same place? [14:43:42] also nuria, I am thinking... in the devserver if the log is greater than the limit, it means it was already truncated. [14:44:36] mforns: mmmm.. i do not see that, the devserver is just a receiver so it will receive any event [14:44:37] If the log makes it to the devserver, it should be valid... [14:44:59] mforns: no, not really, you might be logging to the devserver from the android app [14:45:17] mforns: makes sense? The devserver is just an endpoint [14:45:18] nuria, ok, but the devserver does not really have this limitation [14:45:46] mforns: right, teh devserver doesn't have nay limitations, it's just an aid when you are developing [14:45:57] nuria, aha [14:45:59] *any limitations [14:46:12] mforns: it only exists to make your development easier [14:46:54] nuria, sure, now I get it [14:48:00] mforns: very well, when you add limits to the client remember please to add unit tests [14:48:07] sure [14:48:17] mforns: Thank you ! [14:48:25] thank *you* :] [15:04:30] (PS1) Milimetric: [WIP] Add tabular layout with generic router [analytics/dashiki] - https://gerrit.wikimedia.org/r/196230 [15:17:02] I have destroyed the trie. It's a nice datastructure, but python's dicts are faster by an order of magnitude! [15:17:16] Oh well. [15:18:27] Analytics-Kanban: Implement adjustments to Visual Editor dashboards {lion} - https://phabricator.wikimedia.org/T91418#1113157 (kevinator) p:Triage>High [15:21:17] Analytics, Analytics-Kanban, Patch-For-Review: Code changes in VCL to add a Last-Acess cookie to track "top domain" uniques - https://phabricator.wikimedia.org/T92435#1113163 (kevinator) [15:21:18] Analytics-Cluster, Analytics-Kanban, Performance: Implement Last Visited cookie [34 pts] {bear} - https://phabricator.wikimedia.org/T88813#1113162 (kevinator) [15:22:58] Analytics, Analytics-Kanban, Patch-For-Review: Code changes in VCL to add a Last-Acess cookie to track "top domain" uniques {bear} - https://phabricator.wikimedia.org/T92435#1113165 (kevinator) [15:23:56] Analytics-EventLogging, Analytics-Kanban: Change timestamp fields to reduce DB storage size [8 pts] - https://phabricator.wikimedia.org/T87660#1113170 (kevinator) p:High>Normal [15:24:10] Analytics-Kanban, Patch-For-Review: Failure Types by User Type - https://phabricator.wikimedia.org/T91123#1113172 (kevinator) p:Triage>Normal [15:27:00] Analytics, Analytics-Kanban: Dashboard Directory Design Feedback - https://phabricator.wikimedia.org/T92502#1113178 (Nuria) NEW [15:28:52] Analytics-Cluster, Analytics-Kanban: Read wiki dumps in Spark - https://phabricator.wikimedia.org/T92385#1113192 (kevinator) p:Low>High reprioritizing to high: prototyping revscoring [0] is a quarterly goal for the Research & Data team [1] [0] https://meta.wikimedia.org/wiki/Research:Revision_scori... [15:36:26] Analytics, Analytics-Kanban, Patch-For-Review: Code changes in VCL to add a Last-Access cookie to track "top domain" uniques {bear} - https://phabricator.wikimedia.org/T92435#1113201 (Nuria) [15:40:30] Analytics, Scrum-of-Scrums, Wikipedia-App-Android-App, Wikipedia-App-iOS-App, and 2 others: Avoid cache fragmenting URLs for Share a Fact shares - https://phabricator.wikimedia.org/T90606#1113222 (Fjalapeno) @dr0ptp4kt thanks! [15:40:49] Analytics, Scrum-of-Scrums, Wikipedia-App-Android-App, Wikipedia-App-iOS-App, and 2 others: Avoid cache fragmenting URLs for Share a Fact shares - https://phabricator.wikimedia.org/T90606#1113225 (Fjalapeno) a:Fjalapeno [15:44:15] Analytics-EventLogging, Analytics-Kanban, operations: Eventlogging JS client should warn users when serialized event is more than "N" chars long and not sent the event [8 pts] - https://phabricator.wikimedia.org/T91918#1113246 (kevinator) [15:45:31] Analytics-Cluster, Analytics-Kanban, Performance: Implement Last Access cookie [34 pts] {bear} - https://phabricator.wikimedia.org/T88813#1113252 (kevinator) [15:45:50] Analytics-Cluster, Analytics-Kanban, Performance: Implement Last-Access cookie [34 pts] {bear} - https://phabricator.wikimedia.org/T88813#1020908 (kevinator) [15:48:42] Analytics, Analytics-Kanban, Patch-For-Review: Code changes in VCL to add a Last-Access cookie to track "top domain" uniques [13 pts] {bear} - https://phabricator.wikimedia.org/T92435#1113279 (kevinator) [15:52:20] Analytics-Engineering, Analytics-EventLogging, Analytics-Kanban: Fix issue with database writes not keeping up with events - https://phabricator.wikimedia.org/T90029#1113290 (Nuria) https://gerrit.wikimedia.org/r/#/c/191231/ [16:00:31] Analytics-Cluster, Analytics-Kanban: Mobile product managers should have reports on session-related metrics from the Wikipedia Apps - https://phabricator.wikimedia.org/T86535#1113308 (Nuria) Daily seems that for a global user base wouldn't work so well, anything lower than weekly is not likely to be very... [16:03:00] gi11es: we have 1:1 now? [16:07:07] changing locations, back in a bit [16:14:28] Analytics-Cluster, Analytics-Kanban: Mobile product managers should have reports on session-related metrics from the Wikipedia Apps - https://phabricator.wikimedia.org/T86535#970748 (Nuria) @DarTar: Could you elaborate a bit on why do we need a geometric mean for this data, what does it represent? Thanks. [16:24:10] Analytics, Analytics-Kanban: Dashboard Directory Design Feedback - https://phabricator.wikimedia.org/T92502#1113428 (ggellerman) [16:34:44] Analytics-Engineering: Pageview dumps contain "zero" instead of project variants - https://phabricator.wikimedia.org/T92361#1113446 (Ironholds) Open>Invalid a:Ironholds [16:56:47] mornin' Eloquence. I collapsed last night at some point. Please ping me if you'd like to chat. I'll be available in 30 min. [17:19:16] Analytics-Tech-community-metrics, ECT-March-2015: KPI pages in korma need horizontal margins - https://phabricator.wikimedia.org/T88670#1113664 (Dicortazar) All of the KPI pages were updated. Please, have a look at them. We're still dealing with the horizontal scrollbar. [17:19:28] Analytics-Cluster, Analytics-Kanban: Mobile PMs has reports on session-related metrics from Wikipedia Apps - https://phabricator.wikimedia.org/T86535#1113665 (kevinator) [17:30:44] Analytics, Scrum-of-Scrums, Wikipedia-App-Android-App, Wikipedia-App-iOS-App, and 3 others: Avoid cache fragmenting URLs for Share a Fact shares - https://phabricator.wikimedia.org/T90606#1113713 (dr0ptp4kt) The iOS client side code is good. Merged that in. [17:31:32] Analytics, Scrum-of-Scrums, Wikipedia-App-Android-App, Wikipedia-App-iOS-App, and 3 others: Avoid cache fragmenting URLs for Share a Fact shares - https://phabricator.wikimedia.org/T90606#1113721 (dr0ptp4kt) a:Fjalapeno>dr0ptp4kt [17:31:40] Analytics, Scrum-of-Scrums, Wikipedia-App-Android-App, Wikipedia-App-iOS-App, and 3 others: Avoid cache fragmenting URLs for Share a Fact shares - https://phabricator.wikimedia.org/T90606#1063059 (dr0ptp4kt) a:dr0ptp4kt>None [17:48:56] ottomata: wow, https://gerrit.wikimedia.org/r/#/c/196073/ is super nice [17:49:06] ottomata: since the kafka writer is busted at the moment anyway, should we just delete that code for now? [17:52:37] Analytics-Cluster, Analytics-Kanban: Mobile PMs has reports on session-related metrics from Wikipedia Apps - https://phabricator.wikimedia.org/T86535#1113814 (kevinator) I spoke to Ironholds on IRC: **Run the report weekly on a 30 day window** 30 days looking backward is what the previous ad hoc reports... [17:52:49] ori: , i'm going to fix it! [17:52:54] now that i packaged it, i started working on it already [17:53:00] should I do it in this patch or a separate one [17:53:07] ori, EL code is so slick! [17:53:10] i hadn't hacked on it before [17:53:16] took me a bit to understand how the handlers worked [17:53:19] but that is sliiiiiick [17:53:31] :)))) [17:53:35] separate patch is fine [17:53:39] k [17:54:19] another q for you ori, should i be doing something with unicode stuff? all your other streams seem to give an iterator where events are converted or somethign? [17:54:23] didn't understand what was going on there [17:54:42] message.value is already decoded, no? [17:54:47] i guess? [17:54:51] it yields a dict, not a string, iirc [17:55:00] i think it just Does the Right Thing™ [17:55:26] ok, hm. k. also, shoudl I try to make the kafka_reader have a raw=False arguemnt, where it knows to return a parsed json object if true? [17:56:22] I think it already returns a parsed json object, no? [17:56:27] * ori looks at statsv code [17:56:49] ah no, message.value is a string [17:58:01] I think you have it right [17:58:10] I am getting senile [17:58:16] and conflating things [17:58:28] keep it as-is [17:58:37] well, i mean, your other methods can use iter_json if raw=False [17:58:51] and return a json object by passing the message to json.loads [17:59:08] i can't use that method because the iter_unicode -> iter_file thing doesn't work with the kafka message iterator [17:59:14] it isn't a file object [17:59:23] but, it might be cool if you could do [17:59:28] kafka_reader(..., raw=False) [17:59:34] and that would get you a stream of parsed json objects [17:59:37] rather than a stream of json strings [17:59:39] ottomata: do you have things to talk about in the meeting now? [17:59:40] well, you don't have to rely on iter_unicode() / iter_json() [17:59:45] yes! [17:59:50] leila: we are gonna talk about spark! [17:59:52] you could implement raw=False in kafka_reader [17:59:54] :D [17:59:57] I'll be there, ottomata [18:00:01] ja, think i might ori :) [18:00:05] ok meeting time! [18:00:09] ttyl [18:50:43] test [18:54:08] i didn't hear anything [19:05:47] ori: got your CR, will correct and try to deploy EL master Monday as i have other things to do today and I am not ahuge fan of deploying on Fridays [19:06:17] ori: EL master-> EL "master branch" that is [19:15:06] nuria, I wonder if we should check the query string length or the full url length... [19:15:25] maybe the log is being truncated counting also the url host [19:16:03] and after that split [19:16:56] nuria, if this was so, and we set a limit at 1014 bytes... if EL host changes, then the limit would not be valid any more [19:17:12] I'll try to test that [19:27:53] Analytics, Engineering-Community, ECT-March-2015: Analytics Team Offsite - Before Wikimania - https://phabricator.wikimedia.org/T90602#1114312 (Rfarrand) Sent request to first choice hotel, will email Analytics team for travel dates today. @leila: don't remove the Mexico City room hold yet! Good to hav... [19:27:58] mforns: EL host? [19:28:02] yes [19:28:30] I'm having a look at that [19:28:35] mforns: we should just check the event url, for say "n" [19:29:00] the whole url or just the query string, that's my question [19:29:37] I'm trying to discover if the truncation is done on the whole url and then split to get the query string [19:29:47] mforns: sorry, we should just check the whole event url, for say "n chars" or "n bytes" , that N will be configured [19:29:47] or the truncation happens after the split [19:31:08] but imagine the actual problem being with the query string only; imagine that when the qs is larger than 1014b it is truncated. [19:31:18] mforns: regardless, if we truncate at a given url length we are covered, if you look at varnishncsa logging it is logging according to a format, truncation would happen at the time the logging of chars goes over the limit. [19:32:18] mforns: At least i think that is the way it works [19:32:39] mmm, the query string is only a part of the varnishncsa format [19:32:40] mforns: check the format: [19:32:57] there are other parts posterior to the truncation that pass and get not truncated [19:33:24] mforns: right, the format also includes the UA, [19:33:28] I think the truncation happens before the log gets formatted [19:33:30] mforns: and timestamp [19:34:10] mforns: likely, when those are coherced into the logging strings, [19:34:24] my point is, if we set a limit = 1014 + host_length, and after that EL host changes.. [19:34:38] I don't know how to explain that [19:35:33] mforns: We are not solving the problem precisely, at all. We are just giving the developer an aid: "if your event is larger than this N" [19:35:37] if the actual problem is truncating query string and we set a limit to the whole url, we can have problems if the EL host changes [19:35:40] "it will for sure not arrive" [19:35:47] ok, makes sense [19:35:55] whole url then [19:36:17] mforns: But it could be that the event is on the limit and gets rruncated, sure, that can happen too. [19:36:38] *truncated [19:36:45] so maybe I'll leave some margin [19:37:01] mforns: k [19:37:12] nuria, all right, thanks! [19:45:26] (CR) Ottomata: [C: 2 V: 2] Add mediacounts top1000 files to status dumping script [analytics/refinery] - https://gerrit.wikimedia.org/r/196201 (owner: QChris) [19:50:48] ori: , yt [19:50:49] ? [19:58:23] Analytics, Engineering-Community, ECT-March-2015: Analytics Team Offsite - Before Wikimania - https://phabricator.wikimedia.org/T90602#1114501 (leila) Thanks, @Rfarrand. You can find the list of attendees here: http://etherpad.wikimedia.org/p/Analytics_Wikimania [20:19:56] ottomata: i am now [20:20:17] Umm, what was I going to ask you UmmmMMmm [20:20:57] totally forgot, something about this kafka eventlogging stuff [20:21:06] this isn't what I was going to ask you, but! here's a different q [20:21:15] i've got the raw vs json kafka reader working [20:21:17] pretty awesome [20:21:17] OH! [20:21:18] i remmeber. [20:21:32] what's the difference between a forwarder and a consumer...it seems the consumer could do the same thing [20:21:46] if it was given a udp input and a 0mq output [20:22:04] the forwarder is just a special cased consumer? [20:23:49] ori ^ [20:24:12] oh, yeah, i was going to bring that up [20:24:22] so, the consumer *could* just consume the udp log [20:24:47] but zeromq gives you the ability to have multiple subscribers [20:25:35] so as a first order of business, whenever iterating on something that processes a udp stream, i transform it to a zeromq stream [20:25:36] but, i mean, the consumer forwards too, no? [20:25:46] but kafka provides all of that too [20:25:49] udp -> 0mq with consumer? [20:26:01] or kafka -> kafka [20:26:06] aye, i just got that working too :) [20:26:18] but, ja, for sure, kafka could take the place of 0mq here, but i'm trying to start as small as possible [20:26:22] and just replace the udp part [20:26:38] but, i'm just wondering: why -forwarer when -consumer is almost the same? [20:26:44] is it just a special cased consumer? [20:26:55] e.g. [20:26:55] why not [20:27:12] eventlogging-consumer udp://localhost:1234 tcp://localhost:8422 [20:27:14] or whatever? [20:27:29] (or kafka instead of udp) [20:28:44] because the consumer transforms the stream [20:29:05] it does? [20:29:06] and i wanted to be able to iterate with production data pre-transformations [20:29:12] ohh [20:29:13] oh i guess it does json stuff [20:29:14] fuck [20:29:16] i'm senile [20:29:24] i think the forwarder just combines streams [20:29:26] haha, i can't remember stuff I wrote years ago anyway [20:29:42] mm, the forwarder does the same thing as consumer, but just wihtout calling eventlogging.drive() [20:29:45] it just consolidates the server-side and client-side stuff [20:29:49] it does [20:29:50] for line in f: [20:29:50] sock_out.send_string(line) [20:30:01] it does? [20:30:07] i should look over all of this before saying anything [20:30:08] are there multiple forwarders? [20:30:10] haha [20:30:10] i could be getting it wrong [20:30:11] sock_in = udp_socket('0.0.0.0', args.port) [20:30:39] so https://github.com/wikimedia/mediawiki-extensions-EventLogging/blob/master/server/bin/eventlogging-forwarder#L47 [20:30:40] yeah there are multiple forwarders on vanadium [20:30:41] forwarder 8421 start/running 7963 [20:30:41] forwarder 8422 start/running 7966 [20:30:45] ah ok [20:30:55] one for server-side, the other for client-side [20:31:02] but still, i'm just trying to understand the reasoning behind having a separate concept of forwarder, when coinsumer does the same thing [20:31:06] you could have two consumers running [20:31:28] each with their output being a 0mq tcp port? [20:31:58] forwarder does seem more direct than consumer, it doesn't use the eventlogging.drive method which ends up calling a bunch of stuff [20:32:13] mforns: you around? [20:32:21] but actually, drive() doesn't really do that mcuh either [20:32:23] milimetric, yes [20:32:30] it just gets a reader and write from the input and output urls passed to consumer [20:32:43] and sends events from reader to writer [20:32:50] milimetric, what's up? [20:32:55] forwarder does that too, but just in the eventloggin-forwarder script itself [20:32:58] mforns: did you see that flake8 shows some style problems on your change? [20:33:01] I was just about to review [20:33:24] milimetric, yes you're right [20:34:15] not only that but it will also crash limn-mobile-data flake8 [20:34:29] milimetric, ok I'll have a look at that [20:34:38] cool [20:35:22] milimetric, thx, just ping me if you need anything [20:36:30] ori: I updated the change [20:36:32] https://gerrit.wikimedia.org/r/#/c/196073/ [20:36:40] kafka_writer works (sorry, it was so simple I did it in this patch) [20:36:42] will look in a bit [20:36:46] kafka_reader now supports json parsing [20:36:49] \o/ [20:36:55] its pretty awesome [20:36:55] give me a few mins :) [20:36:59] no worries [20:37:17] i'm heading out fairly soon, but we can talk later if i'm not around [20:52:34] wikimedia/mediawiki-extensions-EventLogging#365 (master - 50aa502 : Translation updater bot): The build was broken. [20:52:34] Change view : https://github.com/wikimedia/mediawiki-extensions-EventLogging/compare/a2bf69e169ab...50aa502762b1 [20:52:34] Build details : http://travis-ci.org/wikimedia/mediawiki-extensions-EventLogging/builds/54153886 [22:26:18] Analytics-Kanban: Missing dialect-specific directories in the pageviews definition - https://phabricator.wikimedia.org/T92020#1115074 (kevinator) [22:30:39] Analytics-Kanban: Missing dialect-specific directories in the pageviews definition - https://phabricator.wikimedia.org/T92020#1115089 (Ironholds) A more useful thing to do than update the documentation, might be actually updating the definition. [22:41:52] Analytics-Kanban: Missing dialect-specific directories in the pageviews definition - https://phabricator.wikimedia.org/T92020#1115158 (kevinator) [22:42:45] tnegrin: I think I declined that one a couple of days ago? it was conflicting with another meeting [22:43:23] the monthly wmf/wikidata meeting [22:43:29] Analytics-Kanban: Update Pageview UDF with dialect-specific directories - https://phabricator.wikimedia.org/T92020#1115163 (kevinator) [22:44:27] Analytics-Kanban: Update Pageview UDF with dialect-specific directories - https://phabricator.wikimedia.org/T92020#1115166 (Ironholds) Not "per definition"; the changes are not in the definition. The phabricator ticket gets registered as a def update, thence patched and logged. [22:44:42] Analytics, Language-Engineering, MediaWiki-extensions-UniversalLanguageSelector, Mobile-Apps, and 5 others: there should be a comparison of clicks count on interlanguage links on different platforms - https://phabricator.wikimedia.org/T78351#1115167 (Halfak) [22:46:15] Analytics, Language-Engineering, MediaWiki-extensions-UniversalLanguageSelector, Mobile-Apps, and 5 others: there should be a comparison of clicks count on interlanguage links on different platforms - https://phabricator.wikimedia.org/T78351#1115170 (DarTar) @Amire80, we reviewed this card. I was a... [23:10:39] milimetric, yt? [23:10:48] yes mforns [23:11:12] hi! I remember once you commented about some EL QUnit tests that were failing? [23:11:18] was it you? [23:12:33] uh [23:12:46] yes, right, I remember [23:13:06] mforns: it was my mistake I think... something to do with not having the right branch checked out... [23:13:10] or maybe something weirder [23:13:16] milimetric, ok I'm having the same problem [23:13:28] shoot, I don't remember what it was [23:13:43] the tests that fail are the ones that use mw.track() [23:14:10] maybe you need a different branch of mediawiki checked out in vagrant? [23:14:13] the error log says something about centralauth... [23:14:30] I think I added this to the wiki, but if I go look there I'll never find it [23:14:47] ok, thanks for the hints! [23:49:31] omg best css framework evah! :) (semantic-ui) [23:53:01] (PS1) Milimetric: [WIP] Do NOT Merge [analytics/dashiki] - https://gerrit.wikimedia.org/r/196489