[00:01:16] (03CR) 10Jforrester: Table of contents widget (031 comment) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/108945 (owner: 10Robmoen) [00:02:12] RoanKattouw: Hm.. "it can now get" comment nodes [00:02:26] You're saying the toDomElement of a dm comment data element will be a comment node? [00:02:33] Interesting [00:02:41] so that means even the root node doesn't have to be HTMLElement per se. [00:09:57] Krinkle: Comment meta item [00:10:03] Meta item != DM node [00:10:18] But yeah toDomElement in ve.dm.CommentMetaItem returns an HTMLCommentNode [00:10:27] *toDomElements [00:10:41] And CommentMetaItem subscribes to tagNames = [ '#comment' ]; [00:10:45] RoanKattouw: Ah, so there is indeed some dirt there we should clean up. [00:14:04] (03PS1) 10Jforrester: Enable VisualEditor by default in NS_USER as well as $wgContentNamespaces [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/118029 [00:14:13] RoanKattouw: Also, it should probably be nodeName in that case [00:14:19] but oh well [00:14:35] http://dom.spec.whatwg.org/#dom-node-nodename [00:14:47] Yeah internally it does use nodeName [00:14:51] But... yeah [00:14:52] OK [00:15:04] When I designed these API I didn't originally have this use in mind [00:15:23] As you can tell by the fact that tagNames = [ '#comment' ]; didn't work until that commit and comments were special-cased all over the place [00:15:30] *these APIs [00:15:49] So that's why the naming is all element-specific even though the internals are all at a node level [00:17:15] RoanKattouw, do you track dependencies between mw VE and core VE just in submodules or you also add Dependency line in commits? [00:18:59] We usually add some indication of dependency but we don't consistently use the Dependency line [00:19:00] Yet. [00:19:03] I want us to do that though [00:25:39] (03PS1) 10JGonera: Add createContext() to Surface [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/118032 [00:25:41] (03PS1) 10JGonera: Add MobileSurface [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/118033 [00:25:43] (03PS1) 10JGonera: [WIP] Split ve.ui.Context into DesktopContext and MobileContext [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/118034 [00:26:16] (03CR) 10jenkins-bot: [V: 04-1] [WIP] Split ve.ui.Context into DesktopContext and MobileContext [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/118034 (owner: 10JGonera) [00:26:47] (03CR) 10jenkins-bot: [V: 04-1] Add MobileSurface [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/118033 (owner: 10JGonera) [00:30:31] RoanKattouw, what do I need to do so that jsduck jenkins job doesn't complain? [00:31:19] Lemme see [00:31:45] jgonera: Edit .docs/categories.json [00:31:51] And add ve.ui.MobileSurface somewhere [00:31:52] also, is there a grunt task or something that runs jshint with your config? any precommit hook I can install? [00:31:54] ok [00:32:03] Yeah, running grunt should run it [00:32:27] As in, that should work already [00:32:51] yep, works, thanks [00:33:40] sorry for da grrrit spam [00:37:57] (03PS2) 10JGonera: [WIP] Split ve.ui.Context into DesktopContext and MobileContext [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/118034 [00:37:59] (03PS2) 10JGonera: Add MobileSurface [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/118033 [00:38:38] (03CR) 10jenkins-bot: [V: 04-1] [WIP] Split ve.ui.Context into DesktopContext and MobileContext [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/118034 (owner: 10JGonera) [00:41:51] (03PS3) 10JGonera: [WIP] Split ve.ui.Context into DesktopContext and MobileContext [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/118034 [00:42:47] (03CR) 10jenkins-bot: [V: 04-1] [WIP] Split ve.ui.Context into DesktopContext and MobileContext [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/118034 (owner: 10JGonera) [00:47:06] (03PS1) 10JGonera: Add createSurface() to Target [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/118042 [00:47:08] (03PS1) 10JGonera: Make ve.init.mw.MobileViewTarget use MobileSurface [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/118043 [00:47:10] (03PS1) 10JGonera: [WIP] Add DesktopContext and MobileContext [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/118044 [00:47:55] (03CR) 10jenkins-bot: [V: 04-1] Make ve.init.mw.MobileViewTarget use MobileSurface [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/118043 (owner: 10JGonera) [00:48:08] (03CR) 10jenkins-bot: [V: 04-1] [WIP] Add DesktopContext and MobileContext [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/118044 (owner: 10JGonera) [01:08:13] (03PS1) 10Catrope: [WIP] Red link support [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/118045 [01:09:17] (03PS2) 10Catrope: [WIP] Red link support [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/118045 [01:10:41] (03CR) 10jenkins-bot: [V: 04-1] [WIP] Red link support [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/118045 (owner: 10Catrope) [01:22:29] (03PS3) 10Jforrester: [WIP] Red link support [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/118045 (owner: 10Catrope) [01:23:51] (03CR) 10jenkins-bot: [V: 04-1] [WIP] Red link support [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/118045 (owner: 10Catrope) [01:31:00] Krinkle|detached: Can't work out what's happening with 118045 – grunt is passing locally. [09:53:57] (03CR) 10Siebrand: [C: 04-1] "Per PS1 L10n related comment." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/117343 (owner: 10Trevor Parscal) [15:13:19] * Elitre waves at her (probably sleepy) PM. [15:27:27] Hey Elitre. Not sleepy. :-) [15:28:46] yeah, I can see you writing a lot for 8. 30 am :) [15:31:46] Hey all, I'm an applicant for the FOSS Outreach Program for Women and I was told you might be able to help: [15:32:11] I'm interested in working on this raw project: https://bugzilla.wikimedia.org/show_bug.cgi?id=57804# [15:32:37] But as Erik mentions in the bug thread, the Visual Editor team might be working on something similar already [15:33:06] Is the WMF web service for citations he mentioned already under production? [15:50:41] mvolz: well, yes [15:50:59] let me grab a link for you [15:52:11] I guess that's good :) [15:52:23] hi James_F can you tell me anything about Save hanging in VE since yesterday? https://bugzilla.wikimedia.org/show_bug.cgi?id=62485 I'm seeing the same behavior when Mobile uses VE also [15:52:27] mvolz: that would be https://www.mediawiki.org/wiki/VisualEditor/Design/Reference_Dialog . [15:55:06] Elitre: thanks! Those mostly look like mock-ups- in terms of backend do you know of any code that's been pushed to repo? [15:58:26] mvolz, I know people are actively working on that, but f you wanna know more, I hope devs in this chan will be able to help you with that soon. It's almost 9am in SF so more people should show up here, and if they're not busy with meetings or something they'll be glad to help. [16:03:39] Elitre: thanks! [16:22:42] good morning VE peoples [16:23:29] Morning TrevorParscal. May I introduce you to mvolz who is interested in working with VE this summer via OPW. [16:23:46] mvolz, TrevorParscal is the lead developer of VE, if I recall correctly [16:25:05] sumanah: I think that's still true [16:25:13] hi mvolz [16:26:47] TrevorParscal: Jeez, I'd hope you would know [16:27:03] easy for you to say, you don't have Roan on your team [16:27:12] or Ed [16:27:15] or Timo [16:27:48] I guess I have a fun backwards experience [16:27:50] these guys are hard to keep up with [16:28:01] Being the most senior in the org and most junior in the industry on my team :P [16:28:15] Or...I think? [16:28:42] see, now you understand my lack of certainty [16:29:26] TrevorParscal: so you see her question in backscroll, right? [16:30:37] just read it [16:30:42] she still around? [16:30:48] mvolz: ping [16:30:56] TrevorParscal: Don't ask to answer :) [16:41:40] Back! [16:41:51] I was feeding my stick insects. [16:41:53] TrevorParscal: ^^ [16:44:26] Anyway, to sum up: I was interested in doing this raw project, https://www.mediawiki.org/wiki/Mentorship_programs/Possible_projects#Global.2C_better_URL_to_citation_conversion_functionality [16:45:01] But it seems that the VisualEditor team is already doing something similar, as Erik notes in this bug report: https://bugzilla.wikimedia.org/show_bug.cgi?id=57804 [16:45:45] And I was wondering if there was a potential project somewhere relating to that. I had drafted a very rough proposal: https://www.mediawiki.org/wiki/Improving_URL_citations_on_WikiMedia_(OPW_Project) [16:45:58] But I see it's potentially pretty outdated now :). [16:48:39] mvolz: it's on our roadmap to add lookup and auto-fill functionality to our transclusion dialog (the template editor) [16:48:56] I am the one primarily working on that dialog [16:49:44] TrevorParscal: Great! [16:49:51] no work has been done on ISBN or URL lookups [16:50:14] although, TrevorParscal, I think that's what's in the mockups? [16:50:50] so probably no code behind, but the design idea is there :) https://www.mediawiki.org/wiki/VisualEditor/Design/Reference_Dialog [16:53:24] I guess then my next question is: 1) Could URL lookup in the dialogue be separated out as an OPW project, and 2) Would you being willing to mentor or is there someone else that could mentor? [16:54:27] Or is there a different project you'd be willing to mentor if that seems problematic? I notice you're listed as a mentor under the general VisualEditor project [16:55:10] Elitre: I would look at those mockups as a way to express ideas for what the UI might be, not a roadmap [16:55:48] mvolz: yes to both questions [16:56:04] Awesome! [16:56:32] James Forrester (James_F, currently James_F|Away) is the best person to talk to about working the details out [16:56:37] Ok. [16:56:58] he is our product manager and best judge of both roadmap and resourcing issues [16:57:11] I will /msg or e-mail him [16:58:05] the reason it's good for a OPW project is because it can be done in parallel to other work, it's unlikely to block anything or have complicated interactions (at least relative to other projects like real-time collaboration which will probably affect every line of code) [16:59:27] it's also something nobody on the team is very ... excited about .. personally? I don't know how to put that, I think we believe it is important but tend to be more interested in assuring it's done right than done quickly [17:00:10] so, it's gotten pushed back while we work on the infrastructure around the feature [17:01:02] Great, yeah all my wikipedia edits have C1 title missing errors because too lazy to input them manually (not exactly sure I should admit that!) [17:01:36] Ok I will try to poke around and see what you have in the repo so far and write a better proposal. [17:03:40] mvolz: cool, feel free to come to me with any questions about the transclusion dialog [17:03:53] it's been changing a lot lately, and there are a few more planned changes [17:04:38] (03PS1) 10Krinkle: Tag v1.0.8 [oojs/core] - 10https://gerrit.wikimedia.org/r/118095 [17:04:40] in general, I support using lookup services and have a plan to make them generic enough that we can use TemplateData to specifiy which field uses which service to populate values in which additional fields [17:05:55] but, more than that, the original design of reference dialog was basically the same as the reference dialog looks when you click the "use existing reference" when inserting a new reference, but the plan was to offer dynamic suggestions as you typed into the search box [17:06:23] this search-first interface didn't go over well because the suggestions weren't very good yet, so we moved away from that - at least for the time being [17:07:14] but if the suggestions were good, and as you typed a URL or ISBN number in that field it would offer you web and book citations respectively that were pre-populated, it would be far superior [17:08:41] the idea of populating a bunch of template fields by editing one of them is a little broken, because it has a lot of side effects, like overwriting existing values, adding parameter that weren't there previously, leaving parameter behind that may not be needed anymore, changing data without the changes being highlight visible and deliberate, ect. [17:09:41] mooeypoo: Can you link me to that bug related to a root install? [17:09:49] Anyway, something to think about - if you are working on the lookup service, the UI may be less of a concern, but just realize that the way it will be used would likely evolve quickly [17:10:04] and won't necessarily be what you see in mockups today [17:11:51] TrevorParscal: very cool [17:12:23] Re: autocomplete citations [17:13:14] And yes, good to keep in mind all the different ways the UI might be implemented. [17:13:55] (03PS5) 10Krinkle: ve.ce.Node: Emphasise slug transition with increasing outline offset [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/117896 [17:21:13] (03PS2) 10Krinkle: Make ve.dm.Converter assertions actually display message [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/117923 (owner: 10Catrope) [17:21:23] (03CR) 10Krinkle: Make ve.dm.Converter assertions actually display message (031 comment) [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/117923 (owner: 10Catrope) [17:21:28] (03CR) 10Krinkle: [C: 032] Make ve.dm.Converter assertions actually display message [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/117923 (owner: 10Catrope) [17:21:33] (03PS2) 10Krinkle: Make 'comment' its own metaitem type [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/117924 (owner: 10Catrope) [17:26:38] merci, monsieur James_F|Away :p [17:32:20] Krinkle, https://bugzilla.wikimedia.org/show_bug.cgi?id=62007 [17:33:13] I"m pretty sure the problem is with the API [17:33:59] but for the moment, I can't make my wiki work in root.. managed to set up apache document settings properly and changed LocalSettings but then I get routing errors when I try to load the wiki [18:08:33] mvolz: Hey if you're still around. Happy to talk. :-) [18:13:17] Great! I guess I don't have any further questions for the time being :). I'll probably get the proposal to you tomorrow sometime if that's okay (I'm on UTC -0 here). [18:16:16] (03CR) 10Jforrester: [C: 032] "Per Chris." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/117611 (owner: 10Cmcmahon) [18:16:32] mvolz: Sounds great. :-) [18:20:45] Krinkle: Wouldn't it make more sense to +2 from the bottom of the stack (or the top), rather than randomly in the middle? :-) [18:21:54] Krinkle: Also, did you want me to +2 the 1.0.8 tag for OOjs? [18:22:11] Yes [18:22:22] James_F: I +2 whatever is OK. [18:22:25] * James_F assumes that that's an out-of-order response. :-) [18:22:26] and ready to merge. [18:22:35] Krinkle: It clearly isn't "ready" to merge… ;-) [18:22:41] (03CR) 10Jforrester: [C: 032] Tag v1.0.8 [oojs/core] - 10https://gerrit.wikimedia.org/r/118095 (owner: 10Krinkle) [18:22:46] The other chnages I havent' reviewed the others yet, or there is something outstanding still [18:22:55] * James_F nods. [18:23:14] so later on you can re+2 if needed, or, if it still applies cleanly, gerrit will pick it up and merge it when the deendency is resolved. [18:23:34] just experessing that it is reviewed and approved, nothing wrong with that? I know it can't merge yet. [18:23:36] Yeah, though +2'ing something that's already outdated seems a little OTT. [18:23:48] It wasn't outdated when I reviewed it [18:23:56] and presumably it rebases cleanly [18:24:07] (03Merged) 10jenkins-bot: Tag v1.0.8 [oojs/core] - 10https://gerrit.wikimedia.org/r/118095 (owner: 10Krinkle) [18:24:32] Oh, sorry, yes, that was Roan, not you. [18:24:33] (03Merged) 10jenkins-bot: [Browser tests] Updates to transclusion tests should turn build green [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/117611 (owner: 10Cmcmahon) [18:24:45] Krinkle: Want me to do the MW-core and VE-core OOjs pull-throughs? [18:25:07] James_F: Creating the tag and distribution first [18:25:14] A comimit named "tag" doesn't magically create a tag [18:25:21] Krinkle: I know, but I can do that bit. [18:26:05] Can you? How will someone be able to review that commit with no place to verify the release against? It wouldn't be on npmjs.org [18:26:17] and different timestamp [18:26:59] Roan has still refused to allow me to give him release rights. He needs to sit down and let me share that key and have him set it up. [18:27:18] "Release" in your mind means GitHub still, right? [18:27:32] No [18:27:38] So… [18:27:40] Release is and never was on github [18:28:35] Release has always been on npm. The canonical copy to get and compare a distribution from. BUilding it locally is not the same as actually getting the release. Though in practice the only different will be the timestamp in the header, other differences can be caused by newer versions of grunt handling the concat slightly different etc. [18:29:19] Right. [18:29:25] And to save some time, I usually publish a plain copy on oojs.github.io/releases (which is just the result of cd /tmp, npm install oojs@v1.0.8 and copying node_modules/oojs/dist/oojs.js to that repo. [18:29:42] And we've now got you as a SPOF for "releases". [18:30:00] yes, I which is why I want Roan to get stuff set up [18:30:21] Built 1.0.8 and pushed to npm. [18:31:40] [13oojs.github.io] 15Krinkle pushed 1 new commit to 06master: 02https://github.com/oojs/oojs.github.io/commit/1b99d0fca1c9a6efbdcbb26d6a8e26d3a59c585f [18:31:41] 13oojs.github.io/06master 141b99d0f 15Timo Tijhof: Publish v1.0.8 release from npm [18:31:46] https://www.npmjs.org/package/oojs?purge [18:31:51] http://oojs.github.io/releases/oojs-1.0.8.js [18:32:43] James_F|Away: http://git.wikimedia.org/blob/oojs%2Fcore.git/v1.0.8/History.md [18:33:41] oojs/ui (is one release behind it seems), ve/ve, mw/core [18:36:30] [13oojs.github.io] 15Krinkle pushed 1 new commit to 06master: 02https://github.com/oojs/oojs.github.io/commit/86a65e19ea5e66376bd088cb2fc5c433ff22b54e [18:36:30] 13oojs.github.io/06master 1486a65e1 15Timo Tijhof: Regenerate docs for v1.0.8 [18:36:36] http://oojs.github.io/docs/ [18:42:31] (03PS1) 10Krinkle: readme: Fix broken Markdown list syntax [oojs/ui] - 10https://gerrit.wikimedia.org/r/118121 [18:43:08] (03PS2) 10Krinkle: readme: Fix broken Markdown list syntax [oojs/ui] - 10https://gerrit.wikimedia.org/r/118121 [18:46:36] (03PS1) 10Krinkle: Update oojs to v1.0.8 [oojs/ui] - 10https://gerrit.wikimedia.org/r/118123 [18:52:38] (03CR) 10Jforrester: "https://git.wikimedia.org/blob/oojs%2Fui.git/3027b13f749bc465149edc4f085b751225569708/README.md doesn't look good — is this just gitblit b" [oojs/ui] - 10https://gerrit.wikimedia.org/r/118121 (owner: 10Krinkle) [18:54:23] (03CR) 10Krinkle: "It worked when I previewed it on GitHub (click "Edit" when logged-in on GitHub.com for any .md file anywhere, and paste the contents and h" [oojs/ui] - 10https://gerrit.wikimedia.org/r/118121 (owner: 10Krinkle) [19:06:35] (03PS1) 10Krinkle: Update oojs to v1.0.8 [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/118125 [19:17:49] (03CR) 10Jforrester: [C: 032] "Happy to ignore gitblit. :-)" [oojs/ui] - 10https://gerrit.wikimedia.org/r/118121 (owner: 10Krinkle) [19:18:12] (03Merged) 10jenkins-bot: readme: Fix broken Markdown list syntax [oojs/ui] - 10https://gerrit.wikimedia.org/r/118121 (owner: 10Krinkle) [19:22:07] Krinkle: Want me to merge https://gerrit.wikimedia.org/r/#/c/118123/ https://gerrit.wikimedia.org/r/#/c/118125/ https://gerrit.wikimedia.org/r/#/c/118126/ ? [19:22:16] Yep [19:23:37] You should also do an OOjs UI build for MW-core that uses it. [19:23:49] Given that we're about to make OOjs UI's primary implementation give it. [19:24:00] (No breaking changes, but it's the point of the thing.) [19:24:07] (03CR) 10Jforrester: [C: 032] Update oojs to v1.0.8 [oojs/ui] - 10https://gerrit.wikimedia.org/r/118123 (owner: 10Krinkle) [19:24:14] (03CR) 10Jforrester: [C: 032] Update oojs to v1.0.8 [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/118125 (owner: 10Krinkle) [19:24:36] (03PS1) 10Krinkle: Drop file "OO.ui" file prefix [oojs/ui] - 10https://gerrit.wikimedia.org/r/118127 [19:25:06] (03Merged) 10jenkins-bot: Update oojs to v1.0.8 [oojs/ui] - 10https://gerrit.wikimedia.org/r/118123 (owner: 10Krinkle) [19:25:12] (03Merged) 10jenkins-bot: Update oojs to v1.0.8 [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/118125 (owner: 10Krinkle) [19:25:18] (03CR) 10Krinkle: "Used a shell script like these per directory:" [oojs/ui] - 10https://gerrit.wikimedia.org/r/118127 (owner: 10Krinkle) [19:28:03] Krinkle: 118127 causes grunt to fail for me on recess. [19:28:26] yeah, still draft [19:28:50] Not according to gerrit. :-) [19:28:52] The fact that the failure is a success on jenkins worries me [19:28:54] https://integration.wikimedia.org/ci/job/oojs-ui-npm/305/console [19:28:56] no it does fail there [19:29:14] Also, how did V+2 work? [19:29:17] Yeah. [19:29:35] Is it being run with --force or something? [19:29:46] Some module somewhere (npm, grunt, grunt-recess, recess, whatever that uses..) is not returning the proper bash exit code [19:30:06] it's not jenkins, npm or grunt. Those I know to do it right. [19:30:20] But why does it fail locally and not on the server? [19:32:00] (03PS2) 10Krinkle: Drop file "OO.ui" file prefix [oojs/ui] - 10https://gerrit.wikimedia.org/r/118127 [19:32:07] James_F: It fails on both [19:32:10] look at the url I gave you [19:32:14] :) [19:33:01] the "result: SUCESS" is Jenkins exitcode.toString(), that's not grunt, you'd never see that locally even when it passes. [19:33:04] https://integration.wikimedia.org/ci/job/oojs-ui-npm/305/console [19:33:14] Krinkle: "But why does it fail locally and continue on the server?", happy? [19:33:15] same error I got locally. Recess returning help instead of actual results [19:33:35] both locally and on the server it aborts as soon as it starts recess and quits. [19:33:44] grunt-contrib-clean doesn't get run locally for me. [19:33:56] no difference. And on both it does not continue and exits immediately. And on both it exists with status code 1 [19:34:10] at least for me it did. [19:34:13] So… the link is a lie? [19:34:23] James_F: What command did you run locally? [19:34:26] clean is always before build [19:34:30] Krinkle: "grunt". [19:34:43] "19:25:22 GENERAL HELP: http://git.io/recess " was the last line I got. [19:34:57] yes, me too, jenkins too [19:35:18] grunt -> default -> test -> git-build + clean + recess .. [19:35:54] James_F: Is there no " Running "clean:dist" (clean) task" line in the output when you run grunt locally on oojs/ui ? [19:35:54] So… the rest of https://integration.wikimedia.org/ci/job/oojs-ui-npm/305/console after that line is just my hallucination? [19:36:22] James_F: The rest after that line is the Jenkins UI (not the command line) showing in human english what the exit code is (zero or non-zero ? SUCCESS : FAILURE) [19:36:56] and I bet if you run it locally and woudl echo $?, it'll be success, not error. [19:37:31] Krinkle: https://gist.github.com/anonymous/0769c08ec6139c04dc14 [19:37:34] Krinkle: I get no "19:25:22 grunt-contrib-clean@0.5.0 node_modules/grunt-contrib-clean " [19:37:44] Krinkle: I get no "grunt-contrib-jshint@0.8.0 node_modules/grunt-contrib-jshint" [19:37:47] Krinkle: I get no … [19:37:54] .. [19:37:57] what is something completely different [19:38:00] https://gist.github.com/anonymous/0769c08ec6139c04dc14#file-gistfile1-txt-L4 [19:38:02] line 4 [19:38:05] Running "clean:dist" (clean) task [19:38:07] it's right there [19:38:15] Yes. [19:38:18] the line you pasted with the @0.5.0 is from installing dependencies [19:38:35] you already have that installed, isn't not going to redownload it every time you run the tests [19:39:05] But it aborts before it returns locally. [19:39:24] * James_F gives up, given that you won't answer the question. [19:39:44] No, you're wrong james and when I'm not called for dinner I'll explain it to you if you're interested. [19:40:13] Sure. [19:40:23] James_F: so hey, now a bad time ? [19:40:27] When you're back, the question I'm asking is: "Why is the same script not returning the same result on the server as on my client?" [19:41:23] James_F: I believe there are additional rules to the generation of the TOC... And now i'm sad [19:42:18] rmoen: Sure. [19:42:28] rmoen: Oh. :-( [19:42:39] Got a quick minute for a hangout ? [19:43:02] James_F: This is a bit confusing to explain on irc [19:43:20] Sure. [19:43:46] James_F: Excellent, ring me when you are ready. [19:54:26] (03CR) 10Jforrester: [C: 032] "Built locally and confirmed no changes in dist files. Thanks!" [oojs/ui] - 10https://gerrit.wikimedia.org/r/118127 (owner: 10Krinkle) [19:54:58] (03Merged) 10jenkins-bot: Drop file "OO.ui" file prefix [oojs/ui] - 10https://gerrit.wikimedia.org/r/118127 (owner: 10Krinkle) [19:55:26] Yay. [20:08:34] What do you guys think of https://bugzilla.wikimedia.org/show_bug.cgi?id=52414 ? [20:09:24] we could have it resolve redirects when showing the suggestions, but that might make it seem broken to users who don't realise it followed a redirect to get to something that clearly doesn't match their search [20:11:35] Yeah. [20:12:07] If you add [[Category:Foo]] to an article in WT, and Foo redirects to Bar, do you get [[Category:Bar]] at the bottom of the page? What about [[Category:Foo]]? [20:28:29] hi TrevorParscal, would you have some time to have a look at https://gerrit.wikimedia.org/r/#/c/118034 ? it's still WIP because I have more questions [20:28:46] James_F, you get [[Category:Foo]] at the bottom of the page, in italics [20:30:51] jgonera: lookin [20:32:34] Krenair: Nothing about [[Category:Bar]]? [20:32:34] Not that I see James_F [20:32:34] jgonera: why did you choose to retain hide but not show? [20:32:34] in the base class [20:32:35] Krenair: Hmm. In that case, maybe we should offer rather than just silently substitute? [20:32:38] TrevorParscal, no particular reason, I just removed everything and looked at the JS console to see what was missing. it's probably more logical to leave show as well [20:32:54] James_F, possibly... wonder what sort of UI would be best for that [20:32:56] looking some more [20:33:05] but I'd make it noop anyway and since it seemed it wasn't used, I just didn't put it in there [20:34:39] Krenair: Maybe offer it in the drop-down directly underneath its source? [20:34:39] Krenair: So [ Matching categories: Foo, Bar (-> Bang), Baz, Boom (-> Bong) ] [20:34:52] With Bar and Boom italicised, and Bang and Bong indented a bit? [20:34:57] * James_F isn't sure [20:35:39] jgonera: you should examine callers and figure out when hide/show would be called and what the intention of the caller is when they run hide/show, and that should give you an idea as to whether to retain both or remove both [20:36:19] TrevorParscal, hide() was definitely called by something, but I don't remember what. I'll double check both of them [20:36:32] jgonera: also, this sort of makes me thing we may want to have DesktopSurface and make Surface abstract (same as context) [20:36:45] James_F, hm, trying to imagine that one. Do you mean having the redirect target as a submenu that appears to the right (or left on rtl? ugh) when you hover over the redirect suggestion? [20:36:54] my main doubt now however is how to implement the drawer that slides down from the top of the screen [20:37:00] TrevorParscal, yeah, I thought about that too [20:37:02] Krenair: I was thinking inline. [20:37:17] Krenair: This is where I need an online sketching tool. :-) [20:37:51] I think the VE motto, if we have one, which we don't, but if we did, would be something like, "when in doubt, refactor" [20:38:02] TrevorParscal, this is definitely not final and I'm open to any suggestions. I just want to have something working more or less how we envisioned it on mobile, and then tweak and refactor it [20:38:10] so, don't hold back if you see something that could be reoranized [20:38:19] jgonera: understood [20:38:26] seems sensible to me so far [20:38:39] cool, glad to hear that [20:38:43] so, about the drawer [20:38:51] yes [20:39:06] Roan said it could either be a new widget in OOUI or I could animate a PanelLayour in MobileCOntext [20:39:15] PanelLayout* [20:39:29] James_F, that's what I thought, then I realised that then depending on where you clicked in the box, you'd get a different result. At the moment the whole box is one link [20:39:29] a PanelLayout is probably not what you want [20:39:50] Krenair: I meant in the suggestions box, perhaps? [20:39:57] it's very simple, but by default has CSS properties to make it position absolute, top 0, left 0, right 0, bottom 0 [20:40:08] James_F, um yes, that's what I'm talking about [20:40:09] or something like that [20:40:27] Krenair: Each result is its own click link, isn't it? [20:40:32] anyway, you would probably just create a DrawerLayout [20:40:51] TrevorParscal, I see. would that be a subclass of PanelLayout? [20:40:53] or, perhaps a new class of thing [20:41:05] like how Dialog and Inspector are their own thing [20:41:27] but, I guess you are just putting inspectors in it [20:41:38] ok, I've decided, you should do none of the suggested things [20:41:58] James_F, yes, that's what I meant [20:42:00] 'box' in "then I realised that then depending on where you clicked in the box, you'd get a different result. At the moment the whole box is one link" referred to each suggestion result [20:42:03] ... an online sketching tool would be really useful for trying to explain this, yeah [20:42:14] * James_F grins. [20:42:24] just add styles that will apply to the MobileContext's DOM element, add additional DOM elements as needed, and write styles for ve-ui-mobileContext [20:42:56] TrevorParscal, ok, so no new widgets in OOUI for now, right? [20:43:07] Even if we do eventually start using drawers, which I'm not saying we will, we would do it differently anyway [20:43:23] and if we ever get to the point where we want something similar on desktop, then we'll think about abstracting it, right? [20:43:29] I see [20:43:30] yeah, just stylize MobileContext like we did with the existing context [20:43:56] I think the existing context might use PopupWidget or something, but you have a very simple case, I wouldn't bother generalizing it until we have a second use for it [20:43:58] so, the existing context actually does use a widget, it uses a PopupWidget [20:44:04] yeah [20:44:10] but I guess I can achieve what we need on mobile with just MobileCOntext styling [20:44:23] originally it did not, we generalized it when people were like "we want one of those popup things over here too" [20:44:30] I see [20:44:44] ok, I'll do that now and see how it goes then [20:44:58] thanks TrevorParscal [20:47:09] James_F: So, Jenkins jobs using npm, like all jenkins jobs, just execute a few simple bash commands (npm install && npm test in this case), and aborts if one of them fails, and uses the exit code to determine the final status (reflected with the appended "Finished: SUCCESS/FAILURE"). 'npm install', if run on a directory where you alreayd have the correct versions of all dependencies just echos [20:47:09] an empty string and returns code 0 (ok), then npm test == grunt test == grunt [default=test], runs a few tasks, and like Jenkins, grunt executes a few subtasks and immediately aborts if one of them fails. The thing you mentioned about "clean" not running is probably something you misinterpreted. That task is unrelated to this and works fine, so it moves on to the 'recess' task which (as I've [20:47:09] filed upstream in the mean time) doesn't exit with nonzero if the files aren't found, instead the 'recess' fail helpfully returns 0 and displays the man page (the files aren't found in Patch Set 1 because I didn't update modules.json yet). Both for you and me locally and on Jenkins the grunt task aborted after the recess failure, but because it returned 0 (ok) inside recess (for the man page), [20:47:10] that is what cascades up and shows SUCCESS everywhere. [20:47:10] Krinkle: OK. But clearly I missing something still. :-) [20:47:10] Krinkle: When run locally, my log aborted after the recess task. [20:47:13] Krinkle: When run on the server, it did not. [20:47:31] James_F: Fixed that bug :) Seems to be working on a skipped level now :) [20:47:33] Krinkle: Is this because the server is set up differently, or because it's running something else, or is it just a mystery of the Cosmos? [20:47:37] rmoen: Ooh, yay. [20:47:58] rmoen: Glad that my as-evil-as-possible test page found this, then. :-) [20:47:59] James_F: I'm sorry but you are indeed missing something still. Both in your gist paste and on the jenkins job, it most certainly did abort after the recess task. [20:48:06] https://gist.github.com/anonymous/0769c08ec6139c04dc14 https://integration.wikimedia.org/ci/job/oojs-ui-npm/305/console [20:48:22] scroll all the way down [20:48:26] James_F, so, what do we do re: redirect category suggestions? [20:48:37] James_F: it's definitely helped. I'm still on track to get this done today I hope, spent a couple hours tracking that down [20:48:38] don't confuse "grunt build" as part of "npm install" with the "grunt test" after that [20:48:42] Krinkle: Oh, I see, the first abort didn't abort the npm install process, just the grunt test? [20:48:54] s/abort/come at the end of/ [20:49:37] James_F: the unit tests are only run as part of grunt-test inside npm-test. The build is done in both grunt-build and grunt-test. Normally you would' run this twice, but because npm, we do install and test, and build happens to be run by both. [20:49:43] if npm-install fails, we don't run test [20:49:50] but because it doesn't return a proper exit code.. [20:50:04] the same reason Jenkins marks the job success, is also the reason npm test happens still after teh failed npm install grunt build [20:50:26] because it didn't actually throw an error [20:50:31] not for you locally either [20:50:33] it just stops [20:51:33] Krinkle: Right. Thank you for the explanation. [20:52:22] Krenair: I've spent 10 minutes trying to get Chrome to let me edit the HTML of a menuWidget without success, so I give up trying to show a change. [20:52:41] heh, okay. [20:53:12] Krenair: My suggestion: For results that are redirects, make them italic, and insert underneath them (styled differently – maybe the redirect arrow and indented a bit?) the redirect target as an option to select instead. [20:53:43] Krenair: But it's just an idea, not necessarily "what you must do". We're open to you using your judgement on what you think works. :-) [20:53:54] okay, I think I will do that actually. [20:54:38] One minor problem: There's no SVG version of the redirect icon (and we don't just use a normal Unicode glyph). [20:55:01] All the VE interface is CSS, SVG or text; no raster graphics. We'd like to avoid adding them just for this. :-) [20:55:18] Okay. [20:55:28] Perhaps just text for now then? [20:56:20] ↳ (8627) is the Unicode glyph closest to the image used in MW. [20:56:24] So we could just use that? [20:56:27] ok [21:00:42] huh, the redirect icon has no svg version? [21:00:50] someone file a bug and i'll make one :D [21:03:08] MatmaRex: Can I file a bug about you filing a bug? :-) [21:03:26] MatmaRex: We should just use the Unicode glyph, then fall back to SVG, then fall back to PNG. [21:03:31] MatmaRex: It's stupid. :-) [21:03:35] heh [21:04:40] TrevorParscal, .ve-ui-surface-overlay has height of 0, so setting height: 100% on .ve-ui-mobileContext won't do anything. should I compute the height in JS? how do you usually deal with such things? [21:05:13] the contents of the overlay are absolutely positioned and it's overflow is visible [21:05:39] are you trying to make a child element 100% width and height? [21:05:58] yes [21:06:13] use position fixed, like dialogs do [21:07:49] even pretty old mobile devices support that [21:08:10] and, it's a superset of those that support conenteditable anyway [21:11:46] James_F, this is frustrating, I don't appear to be able to insert a linebreak into a MenuItemWidget label [21:21:31] Krinkle: so, I was hoping to pick your brain about something [21:21:36] take a look at this: https://gist.github.com/trevorparscal/9495342 [21:22:27] I already have it working so that if I have data like that I can generate the cite tools when VE loads [21:23:00] the question is, where and how do I make these stored and curated on wiki and then delivered to the client when VE loads [21:23:07] Krinkle: ideas? [21:25:09] So far: http://cl.ly/image/0i183J042W2t [21:26:01] icons need another pass, and when you click on them we don't actually do the right thing (yet) but I'm generating those tools dynamically based on a mock data structure (like the one in the gist) [21:26:06] Krinkle: are you alive? [21:26:12] oh, interesting [21:26:31] TrevorParscal, yeah, I just forgot about this, it should work fine, thanks [21:26:42] jgonera: cool [21:26:51]