[00:01:12] Reedy: Never. [00:02:57] Hello! If you attended the Summit remotely, would you mind filling out this form? We plan to use your answers to improve this experience for next year. (No need to sign in or anything - it's anonymous if you choose to put down a anonymous@email.com address.) https://docs.google.com/forms/d/e/1FAIpQLSeLavJQXW-Zo5AxNZHblbQI6hd0T4cFxr8KQYyZfD1Ly4DRMg/viewform [00:03:00] Thank you! [00:08:49] Actions to grow the diversity of our technical community - Livestream here: https://www.youtube.com/watch?v=ytTzofwJnsM, etherpad here: https://etherpad.wikimedia.org/p/devsummit17-diversity [00:09:01] just started in Hawthorne [00:13:39] What a strange session to stream. [00:20:50] SMalyshev: There is also the "sources" field that could be use to store the query [00:21:22] (and probably also an update rate (every X days...) [00:21:35] Tpt: are you in the room or is this being streamed? [00:21:46] milimetric: I'm in the room [00:21:49] !!- I just noticed at the bottom of .. https://en.wikipedia.org/wiki/Oakland%2C_California that there is a left - right- up -down link set (!) .. is that all manually created ? [00:21:51] haha :) ok, just making sure [00:25:51] Okay, I'm going home. Y'all have fun. [00:28:49] Bye [00:48:48] well, i guess i didn't make my point. if we decide the outcome is just to "thoughtfully consider technical debt" then we're just going to continue business as usual [00:49:05] unless you think that all of us have been failing to "thoughtfully consider technical debt" [00:50:57] some teams are definitely better than others at it [00:51:26] hmm seeing forest while in trees is always a challenge, no ? [00:51:28] so if you let each team continue to set their priorities, we will continue to do the same things we have been doing. [00:51:54] indeed [00:51:58] i believe that we all have been setting priorities with the best intentions, and with honest assessments about the priority of technical debt re other tasks. [00:52:36] we might be incorrect, of course! but we haven't been deliberately ignoring debt. we have been rationally setting other things as higher priority. [00:53:07] perhaps it is because we've been taking too short a view, or neglecting external factors, or our "rational" decisions aren't actually rational, etc. [00:53:32] but we need some external change if we are going to have a different result from continuing to set our priorities honestly as we have been doing. [00:54:42] my straw poll wasn't rhetorical; one way of affecting the result is by forcing (top-down) one of the tasks to be "technical debt". now, we might react by just relabelling something we were going to do anyway to satisfy the requirement [00:54:56] but at least that proposal would be some change. that opens the *possibility* of a different result. [00:55:38] if your change is just that we should "thoughtfully consider technical debt" -- the only way that will result in change is if you think that we have not been thoughtful about it previously. [00:58:30] hi all - another somewhat difficult topic... am I right in a dim recollection that a toolset for generating 2D of basic cartography map snapshots was never shared, and the author has declined to publish ? [00:59:46] .. this kind ? https://en.wikipedia.org/wiki/Oakland%2C_California#/media/File:Alameda_County_California_Incorporated_and_Unincorporated_areas_Oakland_Highlighted.svg [06:24:31] Does anyone know if there's a task for "Fixing non-latin fragment identifiers in MediaWiki URL" (was an unconference session today) - I can't seem to find any. [06:25:00] Ah, it's in the etherpad notes. Sorry :) Got it :) [06:41:07] hm I see this page (for the first time) https://www.wikidata.org/wiki/Wikidata:List_of_properties/Geographical_feature [06:41:50] but, the SPARQL examples of geo-queries .. includes a mysterious note about .. ".. this query written before 10/16" [06:42:11] .. I am not on a deep-dive of SPARQL, just a light pass.. but comments welcom [07:12:04] for those who were wondering: [07:12:08] [to the tune on "Country Roads", [07:12:09] John Denver version if you're doing Karaoke] [07:12:09] [verse] [07:12:09] i like parsing [07:12:10] wikitext yeah! [07:12:11] lots of templates [07:12:12] and recursive hash invoke [07:12:13] the code is old there [07:12:15] older than tstarling [07:12:16] younger than the jimbo [07:12:17] growing like a weed [07:12:18] [chorus] [07:12:20] wikitext [07:12:21] i love you so [07:12:22] all the ways [07:12:23] you keep me employed [07:12:24] hacking parsoid [07:12:25] in the dark [07:12:27] wikitext [07:12:28] i love you so. [07:12:30] [verse] [07:12:30] all my patches [07:12:32] gather round me [07:12:33] wish that i could get [07:12:35] just one more review now [07:12:35] php is rusty [07:12:36] the manual's close by [07:12:37] this here regexp's making [07:12:39] teardrops in my eye [07:12:40] [chorus] [07:12:41] [bridge] [07:12:42] i hear a voice [07:12:44] in the morning hour it's calling me [07:12:45] jenkins reminding me that +2's far away [07:12:46] and driving down the road i start remembering [07:12:47] that i should have run phpcs [07:12:48] yesterday [07:12:49] [chorus, 2x] [07:12:51] [outro] [07:12:52] wikitext [07:12:53] you leave me vexed [07:12:54] wikitext [07:12:56] i wonder what's next [07:13:28] * Isarra backs away slowly. [07:14:01] wow - John Denver in your ear, what a fate [07:14:14] .. could go on for a long time (!) [08:00:54] cscott_away: https://meta.wikimedia.org/wiki/I_Will_Revise - https://meta.wikimedia.org/wiki/Category:Humor_songs_and_poems [08:00:57] would make a good addition :) [09:24:56] I like the Olivia Newton-John version [14:48:26] Interesting essay on open source that was shared on the open source listserv I'm on: https://www.adamhyde.net/new-intro/ [15:32:53] melodykramer: sounds like he agrees with https://www.gnu.org/philosophy/open-source-misses-the-point ;) [16:14:37] Nemo_bis: It does! I think he may be looking for feedback - not sure of the best way to convey that! [17:07:39] huh. I probably should have been in here two days ago [17:08:04] anyways: milimetric I am running late but it's because I am Getting Things Done relating to getting paid [17:08:26] it will probably be another hour or so before I make it to the venue [17:11:54] Hmm, I know of something that would have been good to have as a remote participator... being able to participate in sessions about STUFF THAT I WORK ON. >.> [17:17:51] _joe_: I have a plan for fixing the video scalers - https://phabricator.wikimedia.org/T155098 [17:17:57] i'll try and knock out a patch [17:17:57] I know you can't accomodate everyone with what's streamed and what isn't, and if the technology just plain fails to work, that's a separate problem, but now that I look back, it's a little worrying how very few relevant ones I got to even see. >.> [17:20:35] Not that the others weren't interesting and all. [17:21:05] brion: Yay! [17:22:00] Isarra: Could you write that in the feedback form? We had limited equipment this time around, but knowing which sessions might be useful would help us figure that out for next year. [17:22:24] Or the next event, which will likely be before next year. [17:22:25] basically i'll have the jobs pull 'next available' with customized priority instead of fixing the list at job enqueueing time; then the faster jobs can be prioritized over slower high-res jobs [17:24:22] i can then make that even fancier in future [17:24:26] CKoerner "name the day" list -- great ! [17:24:28] and it shouldn't require retooling the entire queue :D [17:33:13] TimStarling: Do you know if the ArchCom Planning Meeting at 1pm will happen, and if does, if you need the room in the SF office for it? [17:36:02] cscott: Should we get together to talk annotations now or later? I'm in the middle of reading up on the spec. [17:37:17] guillom: I don't think we'll be at the office, we have rooms here at the venue [17:37:31] thanks, RoanKattouw. [17:38:30] Could the owner of a red bicycle with green bags on the back move their bike please? It's parked parallel along the bike rack and blocking about 2/3 of all the bike parking [17:44:18] <_joe_> brion: I'll take a look :) [17:58:56] Krinkle: are you around at the summit? wanted to poke you on https://gerrit.wikimedia.org/r/#/c/325949/7 [18:09:13] mforns: bo! [18:09:23] hey :] [18:10:20] For impressions: T152650 [18:10:20] T152650: Spike: Impressions abnormally low for Ireland - https://phabricator.wikimedia.org/T152650 [18:11:42] Notes from another task: https://etherpad.wikimedia.org/p/T152122_notes [18:12:16] Check out line 212 for queries creating temporary tables [18:13:53] iPython notebook https://gist.github.com/AndrewGreen/3a512d0cfe01c5cf8db6ae217d7b66c7 [18:27:45] https://collab.wikimedia.org/wiki/Fundraising/2015/Impression_Notes#Correlating_pageviews_and_impressions [18:34:14] jupyter +1 [18:36:56] melodykramer: I already did the feedback form. Should I just do it again? >.> [18:37:20] But which sessions is going to vary by who's eve nthere and not, too.. [18:38:03] ...I keep realising things I didn't think of after. I'm hopeless. [18:39:15] AndyRussG|devsum: what are you using to make the notebook? :) [18:40:22] madhuvishy: that was just ipython running on stat1002, then copied the notebook contents into a Github gist. Automatically displays! :) [18:41:50] ah - IPy 5 has some nice keybd tweaks for the CLI oriented [18:42:14] * darkblue_b sees SPARK in the future :-) [18:51:48] Why is preview before the edit for even a thing, let alone the default? >.> [18:52:37] if I don't remember incorrectly, because it reloads the page and you would have to scroll down to find it? [18:53:13] You have to REALLY scroll down to do anything with it. [18:54:48] And then you also need to warn people it is a preview, which easily gets lost on a colourful page. >.> [18:57:31] cscott: Catching up on backscroll, I just read your comments about tech debt. The main things I hope will change, compared to the past, are: 1) Encouraging engineers to advocate more for the prioritization of reducing tech debt AND for the ability to clean up code as part of feature development to avoid piling up more tech debt [18:58:26] 2) Raising awareness among product managers that execs value tech debt reduction, so they feel less pressure to prioritize only new stuff and to get the new stuff done so fast that more tech debt is incurred [18:59:00] which is probably compatible with the shorthand about being more conscious, but I can see where that shorthand in isolation isn't really helpful [19:00:06] i hope you will continue participating in the conversation as it moves forward. We need lots of diverse voices. [19:01:50] quick heads up: Dependency Injection session now, at table 16 in Cypress. [19:02:15] Oh so, I guess there was a meeting about third parties in another room [19:02:23] indeed [19:02:40] * bawolff didn't even realize there was meetings in other rooms [19:03:06] Isarra: If you didn't write that the first time around, it'd be good to capture again! [19:04:19] melodykramer: Can you put the link in the topic or something? [19:04:49] pinging TimStarling, addshor, ejegg|afk, tgr_, volans, RoanKattouw, subbu, leszek_wmde: DI session now at table 16 in Cypress [19:04:55] ...if anyone shows up :) [19:05:32] Done, good idea Isarra [19:05:42] btw, is this camera in Chapel Hill room recording? cc melodykramer rfarrand [19:05:44] Thanks. [19:06:02] or was it just left over from the previous days [19:07:07] DanielK_WMDE: thanks for the ping, unfortunately I'm busy at the moment, I might join later, sorry [19:07:27] I think it's left over greg-g - the only streaming session today is the recap: https://www.mediawiki.org/wiki/Wikimedia_Developer_Summit/2017/Program#Wednesday.2C_January_11th [19:07:37] greg-g: Please wave to the camera. greg-g There is no camera. :P [19:07:53] greg-g: Don't turn around. [19:09:14] whoever tried to pair your chromebook to my mac, i turned off bluetooth ;) [19:10:30] greg-g no [19:10:35] left over [19:12:07] melodykramer: rfarrand :) thanks! [19:12:54] i'll be at the venue later to stream the recap. will bring a longer cable so that we can get a better camera angle [19:23:23] Why are the little teeny x things to remove tags on phabricator SO DAMN SMALL?! [19:28:35] _joe_: ok i _think_ i did this right for an initial stab at config for queue split, lemme know if needs adjusting. :) https://gerrit.wikimedia.org/r/#/c/331668/ & https://gerrit.wikimedia.org/r/#/c/331669/ [19:29:12] plus there's MW changeset to set the priorization which sets the job command name [19:31:22] * brion wanders back to coffeeland [19:37:05] lzia: you about? [19:39:24] DanielK_WMDE: Sorry I have conflicting things [19:43:52] is anyone working on the production fatal errors that pine just reported to wikitech-l? [19:44:54] where are the releng folks? [19:46:01] cscott: The spec requires IDs for each annotation, which must be an IRI...I'm not sure if I should fake that for now (like https://annotations.wikimedia.org/annotations/1 even if that doesn't exist) or if I should try to write an API module for individual annotations and use that (like https://commons.wikimedia.org/w/api.php?action=annotation&id=1) [19:46:09] Your input welcome. [19:46:19] TimStarling: It was probably fixed before his email hit it [19:46:38] right, thanks Reedy [19:49:57] brion: https://phabricator.wikimedia.org/T155114 [20:04:44] marktraceur: I think the ID would be a combination of the revision ID and some index w/in the collection of annotations. I'd fake it for now, we can substitute an appropriate URL to the MCR API to fetch from the slot. [20:04:57] Cool. [20:05:21] Coming soon to a wiki near you, thousands of annotations with the URL https://annotations.wikimedia.org/annotation/1 [20:07:13] SMalyshev: You around to talk about adding a service line for federated queries from our alternative query service? [20:08:04] marktraceur: but i'd expect that would be like https://en.wikipedia.org/api/rest_v1/page/html/Main_Page/749836961 except referencing a specific slot [20:08:24] *nod* [20:08:37] i think you might still need to assign sequential numbers to each individual annotation on the image? [20:09:56] hi J-Mo [20:10:36] hey matanya! [20:10:46] harej: https://github.com/tarrow/librarybase-pwb [20:10:48] cheers [20:11:46] tarrow: remote today, but we can talk here [20:13:11] SMalyshev: ok, cool. So where should I I take a look? Does it require actually poking a .java file or is it some kind of config line? [20:13:30] tarrow: I'm writing the answer to your email :) [20:13:47] SMalyshev: Awesome! [20:16:08] cscott: That could work. At least the response would be internally conistent then. [20:19:57] matanya: hey are you using the 'async' option when doing chunked uploading on the video2commons tool? looks like that will do chunk reassembly in a background job, which in theory shouldn't time out [20:26:52] brion: we are [20:26:56] In the past, we had timeouts with async [20:27:00] and it's still failing? dang [20:27:06] when you had a large number of chunks [20:27:08] yes, on very long videos [20:27:09] try increasing chunk size [20:27:17] ah that might be fixed by the recent improvements to timeouts on job queue too [20:27:25] ah, cool [20:27:33] * bawolff working off memory of what the state was years ago, so may be wrong [20:27:53] increasing chunk size so that there is a smaller number of chunks used to help things at least [20:28:18] ooh lunch! [20:28:26] it is all a matter of length, at some lengths nothing really helps anymore [20:43:59] no more pink! thanks RoanKattouw ! [21:09:44] Could we maybe turn off the microphones in the main room? its picking up the room conversation and amplifying background noise [21:11:10] Snrk. [21:33:22] legoktm: fix you damn phan php7.0 -> php patch! [21:37:46] taking notes for the archcom session https://etherpad.wikimedia.org/p/devsummit17-role-of-archcom [21:39:33] RoanKattouw: Krinkle will you be joining this one ^ [21:40:00] I will. Where is it? [21:40:09] Oak [21:40:12] upstairs [21:46:10] greg-g: are mortals allowed to join? [21:46:57] legoktm: if you bring a chair or stand :) [21:47:10] also, where are these sessions being organized/planned [21:47:30] https://www.mediawiki.org/wiki/Wikimedia_Developer_Summit/2017/Get_Stuff_Done_Day has nothing [21:48:06] because I mostly missed the mwstake one earlier today :/ [21:48:37] legoktm: the wall on the corner of the room seems to have a schedule thing [21:48:42] what he said ^ [21:48:47] ok [21:48:50] does anyone here like to knit a lot? [21:48:50] thanks [21:49:18] (and yeah, I missed the first pat of the mwstake one too) [21:49:58] Ward's wife likes to knit and will be at the social event tonight and would be happy to have some suggestions about where to go around the city. [22:14:05] I might be wrong, but does Fluffernutter knit? [22:17:59] Reedy: fix phan? [22:24:00] thedj: https://gerrit.wikimedia.org/r/#/c/331660/ [22:30:04] Who wants to sign my PGP key :D [22:37:08] ostriches: https://phabricator.wikimedia.org/T155130 [22:37:51] ty [22:38:18] legoktm: what user do i need? [22:38:31] jenkins-workspace [22:38:35] ty [22:38:36] I think [22:38:41] ls -l and check [22:38:49] ahh, jenkins-deploy [22:41:57] does anyone want to volunteer to run part of the closing session where people will be presenting their work? basically we can meet and discuss, you will introduce and then moderation by making sure people are sticking to the time limit or following up with clarifying info / comment [22:46:34] found someone :) [22:53:50] legoktm: might have fixed it.... [22:53:57] making a patch now... [22:54:13] yaaaay [22:56:16] If anyone knows where Daniel Kinzler is, I have a question for him. [23:02:23] anomie: He's next to greg and maxsem [23:02:40] Reedy: Which room and table? [23:02:46] big room [23:02:51] and he's just moving... [23:37:31] addshore: https://integration.wikimedia.org/ci/job/mwext-php70-phan-jessie/99/console passed [23:38:27] legoktm: YAY! merge the CI thingy? :D [23:38:31] woo!!!! [23:38:58] addshore: now move it out of experimental for those extensions? [23:39:04] legoktm: will do!