[00:16:58] leila: the survey is live https://en.wikipedia.org/wiki/Book?quicksurvey=true [00:17:19] * leila claps, bmansurov. :-) checking. [00:19:42] bmansurov: I submitted a sample response and it's working. We will wait for more real responses to come in and we will start the QA later tonight. Thank you, bmansurov. [00:19:52] cool [03:38:02] hi dr0ptp4kt. are you around? [15:10:29] dbrant: hey! question about the auto-completion a/b test. https://phabricator.wikimedia.org/T125984 [15:10:53] dbrant: i can't find any documentation of it being a generator [15:11:08] dbrant: only action=cirrus-suggest [15:14:29] dbrant: should i start coding up an a/b test of our existing prefixsearch vs action=cirrus-suggest? [15:15:40] mdholloway: hmm, i could have sworn that it's just a new option that they built into the existing prefixsearch generator... [15:15:58] mdholloway: better to ask someone in the search team. ebernhardson or dcausse, perhaps [15:16:59] dbrant: from the emails it looks like that's the plan but it doesn't look like it's been done yet. but will follow-up with the search team. [16:04:40] dbrant: http://en-suggesty.wmflabs.org/suggest.html is using action=cirrus-suggest for autocomplete [16:13:29] niedzielski: looks like our brainwaves were in sync yesterday. I'll rebase and resubmit [16:13:29] dbrant: your patch (https://gerrit.wikimedia.org/r/#/c/271181) looks good to me but needs a rebase due to my recent changes. mind if i rebase? [16:13:39] ^ [16:13:46] :) [16:17:01] mdholloway: i'd still check with the discovery crew. the task for integrating it into a generator is resolved... [17:16:01] mdholloway: is https://gerrit.wikimedia.org/r/#/c/269876/ ready for merge? [17:16:36] niedzielski: yes. [17:16:47] 👍 [17:21:33] mdholloway: is there a test i should run for https://gerrit.wikimedia.org/r/#/c/270971/ ? [17:24:28] yes, assuming you're testing against a test config, you can set the whitelisted langs to, say, just 'en', then clink an external link away from another language wiki [17:24:55] niedzielski: i'll pm you an example [17:29:11] dr0ptp4kt, jdlrobson, nzr, bmansurov: something's come up that'll require me to shift the rest of my hours today to work quite a bit later. Shouldn't cause any disruptions besides missing standup, for which I'll send an email tonight [17:29:36] I'll still be reachable by phone over IRC for questions [17:36:27] jhobs: thx for the heads up, take good care [17:37:33] sorry if that sounded ominous btw, I just have to deal with a leaking problem coming from the apartment upstairs [17:38:07] and the maintenance crew can be a hassle to lock down times for [17:40:06] jhobs: oh joy [18:04:50] bd808: time conflict another meeting [18:44:00] bmansurov: taking a look at your patches now [18:44:16] cool [18:45:47] bmansurov: shouldn't beta always show the new language overlay? [18:46:09] jdlrobson: it should and it should be configurable [18:46:18] jdlrobson: it's done in the config patch [18:46:49] bmansurov: yeh i see. I was just curious if "A" should be 1 in beta. [18:47:00] yes [18:47:37] (would also save you a SWAT deploy) [18:50:00] bmansurov: quick question on https://gerrit.wikimedia.org/r/#/c/270882/ [18:51:11] jdlrobson: ok, i'll keep the device language unchanged, but add language to cover the new code [18:54:03] sweet [18:54:32] jdlrobson: updated it [18:56:46] bmansurov: so now i am looking at https://gerrit.wikimedia.org/r/271263 [18:57:00] I think you should have languageOverlayBeta and languageOverlayStable [18:57:28] it makes the contents of wgExperiments inconsistent [18:57:45] alternatively just have one for stable and enable unconditionally on beta [18:57:49] that's what Gather did [18:58:01] jdlrobson: should it be consistent? Can experiments have different configurations? [18:58:49] I'd rather not hardcode beta, but I can create 2 different experiment [18:59:17] I still think that experiments should be independent of each other and we should be able to configure them however we want [18:59:46] the different level of nesting could cause broken configurations [19:00:24] if I know an experiment languageOverlay exists I imagine it would be pretty easy to accidentally set it to have a name, buckets and enabled feed and then break things [19:00:53] (given there is no interface in js/php to it) [19:01:43] jdlrobson: ok, i'll create two [19:02:22] sounds good [19:13:03] jdlrobson: done [19:55:57] dbrant: I think I can't make grooming today. Looks like it's a Needs Triage kind of day, anyway. You guys good on yer own? [19:56:25] mbinder: sure, no worries [20:06:02] dr0ptp4kt: I'm sorry I think I have forgotten where to look to learn how the sampling is done for QuickSurveys [20:06:13] dr0ptp4kt: do we sample requests? users based on a token? [20:22:52] bmansurov: jenkins no longer complaining - https://gerrit.wikimedia.org/r/#/c/270429/ [21:00:45] lzia: it starts from https://github.com/wikimedia/mediawiki-extensions-QuickSurveys/blob/84699d73c5ee635d9dfd59dfd2b15408281f0f3f/resources/ext.quicksurveys.lib/lib.js#L137 [21:00:53] lzia: look for getBucketForSurvey in there [21:01:03] lzia: that in turn uses the file https://github.com/wikimedia/mediawiki/blob/f89552db15222de6760be8ccd1e69e3ba22a312a/resources/src/mediawiki/mediawiki.experiments.js [21:01:53] lzia: and that uses a combination of hashing and integer / binary arithmetic operators [22:55:10] bearND: oh, i see you just submitted a patch for the summary extract test failure [22:55:15] : [22:55:16] :) [22:55:33] bearND: i submitted a patch to change the article, but i don't mind either way [22:56:48] bearND: as for the other one, i think you just need to upgrade domino. that failure was a big issue while you were gone but it's resolved now (or should be...) [22:59:25] mdholloway: lol, what a coincindence. :) [22:59:35] mdholloway: i like your patch better [23:00:00] bearND: background on the text-decoration issue: https://phabricator.wikimedia.org/T124446 [23:04:03] mdholloway: thank you. That fixed it. I've closed the task as invalid. [23:04:14] bearND: cool. i was worried for a second! [23:04:17] i mean the one i submitted today