[01:22:39] James_F, do you happen to know how often events get inserted into the research db? [01:28:24] they get inserted into the master database at real time, but there may be a slight replication lag [01:28:47] re: Krenair [01:30:42] thanks ori, see -analytics, something else is broken in labs [01:30:50] oh you're not there [01:30:55] basically no events are even being logged [01:31:04] am looking into it now [01:32:00] (03PS1) 10Krinkle: mw.ViewPageTarget: Disable pointer events on dimmed elements while loading [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/201089 [01:38:52] (03PS4) 10Mooeypoo: [wip] Use DM HTML in previews [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/200762 (https://phabricator.wikimedia.org/T93042) [01:38:57] (03CR) 10jenkins-bot: [V: 04-1] [wip] Use DM HTML in previews [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/200762 (https://phabricator.wikimedia.org/T93042) (owner: 10Mooeypoo) [01:44:17] (03CR) 10Catrope: [C: 04-1] "Commit message says "while loading" but CSS is added to the ve-activatING selector?!" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/201089 (owner: 10Krinkle) [01:44:47] (03CR) 10Catrope: "Sorry: CSS was added to activatED, I would have expected it to be added to activatING" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/201089 (owner: 10Krinkle) [01:45:57] (03CR) 10Krinkle: "The same applies to the opacity. most of the elements are eventually permanently hidden by a stronger selector from ve-active. The rest re" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/201089 (owner: 10Krinkle) [01:46:24] (03PS2) 10Krinkle: mw.ViewPageTarget: Disable pointer events on dimmed elements [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/201089 [01:46:29] (03CR) 10jenkins-bot: [V: 04-1] mw.ViewPageTarget: Disable pointer events on dimmed elements [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/201089 (owner: 10Krinkle) [02:46:24] (03PS3) 10Alex Monk: Try to unbreak WikiEditor modules [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/199387 (https://phabricator.wikimedia.org/T93384) [02:47:33] (03CR) 10jenkins-bot: [V: 04-1] Try to unbreak WikiEditor modules [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/199387 (https://phabricator.wikimedia.org/T93384) (owner: 10Alex Monk) [02:52:01] (03PS4) 10Alex Monk: Try to unbreak WikiEditor modules [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/199387 (https://phabricator.wikimedia.org/T93384) [03:45:19] (03CR) 10Catrope: [C: 032] mw.ViewPageTarget: Disable pointer events on dimmed elements [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/201089 (owner: 10Krinkle) [03:47:49] (03Merged) 10jenkins-bot: mw.ViewPageTarget: Disable pointer events on dimmed elements [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/201089 (owner: 10Krinkle) [05:31:49] anybody around? we may have an issue with Citoid. [05:44:31] mooey|away: as per mvolz'instructions, I pinged you on https://phabricator.wikimedia.org/T94662 . [08:07:54] (03CR) 10Nemo bis: "User notified https://translatewiki.net/wiki/Thread:User_talk:Leli_Forte/Translations_in_wrong_language (I also rolled back on the wiki, t" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/201047 (owner: 10Jforrester) [08:20:09] (03PS1) 10Werdna: T93890: Allow intention flags for non-buttons [oojs/ui] - 10https://gerrit.wikimedia.org/r/201147 [08:22:24] (03CR) 10jenkins-bot: [V: 04-1] T93890: Allow intention flags for non-buttons [oojs/ui] - 10https://gerrit.wikimedia.org/r/201147 (owner: 10Werdna) [08:36:16] (03PS1) 10Nemo bis: Restore Leli Forte's mt translations [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/201149 [08:53:52] (03CR) 10Nikerabbit: [C: 031] "Easier this way than doing each one individually in a wiki." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/201149 (owner: 10Nemo bis) [09:47:54] (03CR) 10Raimond Spekking: [C: 032] Restore Leli Forte's mt translations [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/201149 (owner: 10Nemo bis) [09:53:15] (03Merged) 10jenkins-bot: Restore Leli Forte's mt translations [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/201149 (owner: 10Nemo bis) [10:22:07] (03CR) 10Bartosz Dziewoński: [C: 04-1] "MediaWikiTheme.php needs the same changes." [oojs/ui] - 10https://gerrit.wikimedia.org/r/201147 (owner: 10Werdna) [10:23:26] (03PS2) 10Werdna: T93890: Allow intention flags for non-buttons [oojs/ui] - 10https://gerrit.wikimedia.org/r/201147 [10:26:46] (03CR) 10Bartosz Dziewoński: [C: 031] "I want to try out if this doesn't change anything unexpected before I merge, but it looks sane." [oojs/ui] - 10https://gerrit.wikimedia.org/r/201147 (owner: 10Werdna) [10:27:07] (03CR) 10Werdna: "There are some test failures that I don't fully understand." [oojs/ui] - 10https://gerrit.wikimedia.org/r/201147 (owner: 10Werdna) [10:28:16] (03CR) 10Bartosz Dziewoński: "There aren't any anymore? They were because the JS and PHP versions were out of sync." [oojs/ui] - 10https://gerrit.wikimedia.org/r/201147 (owner: 10Werdna) [10:28:30] sup werdna [10:28:39] MatmaRex_mobile: oh thanks :p [10:28:44] I didn’t realise that’s what they were [10:29:40] it's not obvious when you have a hundred of tests failing with thousands of ines of output, heh [10:31:30] :D [10:32:33] (03CR) 10Bartosz Dziewoński: [C: 031] Only block for errors when saving [extensions/CodeEditor] - 10https://gerrit.wikimedia.org/r/200478 (https://phabricator.wikimedia.org/T70765) (owner: 10TheDJ) [10:36:50] werdna: ehi, do you have further thoughts about https://phabricator.wikimedia.org/T93595 ? I don't know how to move forward with that. [10:37:41] Elitre: I could help out, but I don’t know where Oliver is doing the analysis so all I can do is provide drive-by feedback [10:39:42] (03PS1) 10Bartosz Dziewoński: MenuLayout: Correct documentation [oojs/ui] - 10https://gerrit.wikimedia.org/r/201166 [10:45:29] werdna: can you get in touch with him about that? I'd really appreciate it. [10:47:09] (tnx!) [12:24:59] is [[Category:Something|{{PAGENAME}}]] correct wikitext? what is it for? [13:32:08] as per James' instructions, I will paste here some diffs from wikis which recently got VE as the main editor (phase 5). I can't say whether these are user manipulations or bugs. [13:33:53] https://az.wikipedia.org/w/index.php?title=%C6%8Fbu_B%C9%99kr_%C9%99l-Ba%C4%9Fdadi&curid=356301&diff=3233021&oldid=3233018 and https://az.wikipedia.org/w/index.php?title=D%C3%B6vl%C9%99t_T%C9%99bi%C9%99t_Yasaql%C4%B1%C4%9F%C4%B1&curid=114403&diff=3232972&oldid=2814654 - dunno how the empty headings were made. if I try to replicate, I get a nowiki tag instead. [13:37:03] I see this elsewhere as well (https://mk.wikipedia.org/w/index.php?title=%D0%95%D0%B2%D1%80%D0%BE%D0%BF%D1%81%D0%BA%D0%B8_%D0%A3%D0%BD%D0%B8%D0%B2%D0%B5%D1%80%D0%B7%D0%B8%D1%82%D0%B5%D1%82_(%D0%A1%D0%BA%D0%BE%D0%BF%D1%98%D0%B5)&curid=124138&diff=3328145&oldid=3274768 ). [13:52:57] here, I don't know what happened, and I couldn't find anything related on Phab: https://mn.wikipedia.org/w/index.php?title=%D0%A2%D0%BE%D0%BE%D1%86%D0%BE%D0%BD_%D0%B1%D0%BE%D0%B4%D0%BE%D1%85_%D1%88%D0%B8%D0%BD%D0%B3%D1%8D%D0%BD%D0%B8%D0%B9_%D0%B4%D0%B8%D0%BD%D0%B0%D0%BC%D0%B8%D0%BA&curid=39243&diff=424635&oldid=424634 [14:09:16] here, playing a bit at the bottom of the article means the category can end up on the same line of the final template: https://mdf.wikipedia.org/w/index.php?title=%D0%9D%D0%B8%D0%BA%D0%BE%D0%BD&diff=46505&oldid=45943 [14:10:20] in that same article, I don't seem able to see the categories so that I can change them. and that's all for now. [14:16:33] lucky subbu lost all of my bug reports above, so he doesn't need to deal with them :p [14:17:31] good morning .. [14:18:43] subbu, :) [14:39:20] James_F|Away: RoanKattouw_away: ^^^ if you're interested in possible bugs from Phase 5 wikis. [15:45:05] (03PS1) 10Esanders: [BREAKING CHANGE] Split setupSlugs in to setup(Block|Inline)Slugs [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/201205 (https://phabricator.wikimedia.org/T91245) [15:45:49] (03PS1) 10Esanders: Rename setupSlugs to setupBlockSlugs [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/201206 [15:49:49] Morning Krinkle. [15:57:24] https://commons.wikimedia.org/w/index.php?title=Category:Wikimedia_servers_by_location&diff=155484686&oldid=155484639 [15:57:33] I wonder how we're generating those interwikis. I certainly didn't insert those as interwikis. [15:57:39] And I don't like :en: and I don't like labsconsole: [15:57:50] :/ [15:58:38] maybe it's parsoid magic [15:59:04] Krenair: Was that a VE edit? Why's it not tagged as such? [15:59:15] The edit before that was a VE edit [15:59:24] Yes. [15:59:25] the one on the left [15:59:39] the one on the right is me going back in the source editor and changing some stuff [15:59:39] https://commons.wikimedia.org/w/index.php?title=Category:Wikimedia_servers_by_location&diff=prev&oldid=155484639 looks fine to me. [15:59:55] labsconsole: and :en: are the 'ranking' interwikis for those. [16:00:19] while we should be using wikitech: and w: :/ [16:00:33] it's a matter of changing the MW config afaik [16:00:48] (03Abandoned) 10Esanders: Rename setupSlugs to setupBlockSlugs [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/201206 (owner: 10Esanders) [16:00:59] only one of the interwikis can be primary per wiki.. [16:01:29] Yeah. [16:01:31] :en: actually looks okay to me [16:01:34] (03CR) 10Esanders: "Not likely to be a real breaking change - but changes some function names around." [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/201205 (https://phabricator.wikimedia.org/T91245) (owner: 10Esanders) [16:01:35] Agreed. [16:01:40] Wikipedia is the ranking wiki. [16:01:52] It might be awkward, but it's what MediaWiki is based on. [16:02:33] It shouldn't even work on commons. [16:02:44] I'm guessing it does for historical reasons [16:03:51] * James_F shrugs. [16:04:15] We certainly shouldn't be encouraging it [16:07:55] OK MatmaRex, back on the horse...I want to take the PanelLayout that I created on the backend, detach it from the DOM, and put it into a Dialog. [16:08:46] For some reason, content in the config doesn't do anything...well, at least not with an HtmlSnippet, but maybe I can just use the jQuery object [16:09:51] Nope. Nope I cannot. [16:10:09] marktraceur: can you pastebin the code somewhere? [16:10:15] Yeah, sec [16:12:04] I saw a fatal error pop up last night from OOUI code being called from Extension:Collection [16:12:10] I suppose I should talk to cscott about it [16:12:24] https://phabricator.wikimedia.org/P467 [16:12:26] MatmaRex: ^^ [16:13:03] MatmaRex: This is where 'message' doesn't work either, but it's the same either way [16:14:02] At least the dialog gets opened. [16:15:10] hmm. weird. [16:15:18] James_F: On to my 5th year at WMF now :) [16:15:25] April 1st 2011 [16:15:30] Krinkle: Such a fool, you are. [16:15:46] MatmaRex: I wouldn't be surprised if the answer were that it doesn't like jQuery elements. But then what *does* it like? [16:17:25] no, it should take jQuery elements just fine [16:17:32] Oh hrm [16:17:36] It's there but hidden. [16:17:38] I don't know why. [16:17:53] Maybe because some herp styled it funny [16:18:04] Excuse me while I finish this and then walk away in shame [16:18:53] James_F, I was wondering what you thought about https://phabricator.wikimedia.org/T94084#1169234 [16:19:26] should we alter core to make the edit source button do it? [16:19:27] :D [16:19:33] Krenair: Maybe. [16:21:29] MatmaRex: OK, what about dialog width [16:22:28] Hm, maybe setSize [16:22:38] marktraceur: MyDialog.static.size = 'large'; [16:23:35] MatmaRex: Oh, I need a subclass for that? [16:24:16] marktraceur: probably. i think setting a size yourself will interact funnily with things [16:24:18] but maybe it won't [16:24:26] I think it just doesn't work at all. [16:25:01] Oh, actually, it worked OK. [16:25:12] marktraceur: if you do it at the wrong time, WindowManager might overwrite your change [16:25:16] Hm, the z-index is a bit funky... [16:25:23] or something [16:25:30] MatmaRex: I doubt it, there's only one interaction with this form [16:25:42] i know something was funky about this, but can't exactly remember what right now [16:26:06] MatmaRex: The personal bar has a higher z-index than the dialogs for some reason... :P [16:26:25] marktraceur: try resizing the browser window and make sure that the OOUI dialog doesn't respond in weird ways [16:26:25] That's going to look pretty when this rolls out. [16:26:36] if it doesn't, it's probably good [16:26:50] Naw, it looks pretty great actually [16:27:49] okay. yay [16:28:34] A few more tweaks, and making it actually do something, and I'll be off to the races [16:30:10] MatmaRex: Damn, the labels for accept and reject are static, too. *fine* I'll subclass. [16:30:29] accept and reject? [16:30:33] the buttons? [16:30:36] Yeah [16:30:46] nnnnno [16:30:47] shouldn't be [16:31:30] Oh. Maybe not digging deep enough yet. [16:32:04] getActionWidgets() I guess [16:32:07] marktraceur: you can pass 'actions' to openWindow, like 'title' and 'message' [16:32:28] Huh, where is that documented [16:32:29] the docs for this are, ughhh [16:32:32] Hahaha [16:32:34] on OO.ui.MessageDialog#getSetupProcess [16:32:36] apparently [16:33:05] in the same format as OO.ui.MessageDialog.static.actions [16:33:17] Right. [16:33:27] docs are lacking, but we have examples for this in the demo [16:37:30] MatmaRex: Primary constructive ActionOptionWidgets don't get the full body colour, just a little highlight colour on hover? [16:39:31] marktraceur: in MessageDialogs? probably not. you know, these are Apex's styles, since no one ever bothers to design anything for the existing interfaces around here. [16:41:55] Just confirming [16:42:05] MatmaRex: Unless you want me to shoehorn in a change. [16:43:23] marktraceur: to the design? if you feel like it, i don't really care for the button styles. you might want to poke TrevorP|Away [16:43:52] mooeypoo: hello, you mentioned in this https://gerrit.wikimedia.org/r/#/c/200945/ extension.json file, but i didn't changed anything in the json file [16:44:31] rits, yeah, it's just a way to add files just like you added it to the VisualEditor.php [16:44:51] when you add a file you need to add it to VisualEditor.php -- and also to extension.json [16:45:13] mooeypoo: oh, ok i didnt knew about that i will add it thank you :) [16:45:22] no problem :) [16:45:31] it's easy to miss, it's relatively new [16:53:27] (03CR) 10Divec: [C: 031] "Very much nicer." [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/201205 (https://phabricator.wikimedia.org/T91245) (owner: 10Esanders) [16:55:22] Krenair: Looked in EL beta labs and all looks good, let me know if you were able to test [16:56:07] MatmaRex: And...last thing, I think. What are the events to listen to for accept/reject? [16:56:26] Or is it just one "close" and I should check some other value for which button got pressed [16:57:10] nuria, max(timestamp) is 20150401165401 [16:57:11] marktraceur: it's passed as "closing data" [16:57:21] so sounds like its working again [16:57:22] thanks nuria [16:57:23] one sec [16:57:32] Krenair: ok. [16:59:47] marktraceur: second-to-last example on https://www.mediawiki.org/wiki/OOjs_UI/Windows/Message_Dialogs [16:59:47] Eugh [17:01:19] OK I think I've got it from here, thanks [17:14:33] (03PS3) 10Rits: Assigning width and height to alien extension nodes [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/200945 [17:15:03] (03CR) 10jenkins-bot: [V: 04-1] Assigning width and height to alien extension nodes [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/200945 (owner: 10Rits) [17:15:10] \o/ success, thanks a lot MatmaRex [17:15:42] yay [17:18:15] rits, does it work now in your testing? with the css file [17:18:18] mooeypoo: https://gerrit.wikimedia.org/r/200945 , -1 this time also [17:18:49] mooeypoo: i didnt tested in the inspector just a moment [17:20:59] mooeypoo: it didn't worked, nothing in the inspector also [17:21:25] rits, do you see any of the classes you wrote in the CSS file applied? [17:21:41] when you inspect the node in the browser, do you see the new css rules ? [17:22:19] mooeypoo: no i searched for it the classes are not showing up [17:22:49] rits, then something's happening with your css [17:22:56] either the definitions aren't right or the file's not there [17:23:16] mooeypoo: oh ok actually i used the same css as of the template one [17:23:31] rits, yes, but if it's not being applied, it wont' work :) [17:23:56] (03CR) 10Legoktm: "See my comment on Iabc322c7b92c768c9a6e7d272a5816e32743a501 for an example." [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/200305 (https://phabricator.wikimedia.org/T93616) (owner: 10MarkAHershberger) [17:24:11] mooeypoo: that could be the reason actually git stashing sometimes delete the file i will just double check it [17:24:15] make sure it's applied. If not, then there is probably an issue with either the class names or a missing class definition or something like that [17:24:42] rits, the file is there in gerrit, but if the definitions in it aren't applied, it might be something in the code [17:25:40] rits, take a look at the css file and see if there might be classes that are missing from the node itself [17:25:54] it might be something where a class wasn't applied or some definition is missing or something like that [17:26:00] mooeypoo: ok i will check it again, no actually yesterday stashing removed my css file and i got to know it today [17:26:21] stashing? you mean git stash? [17:26:42] mooeypoo: yes i meant that git stash [17:27:04] rits, if you git stash, you can git stash apply [17:27:19] it will apply the stashing..? i'm a little confused, if you run into git problems let us know, we can help [17:28:02] mooeypoo: i used git stash pop but the git issues doesnt resolved so easily i changed it all over again [17:28:33] okay I'm not sure I understand what git issues, but if you run into more specific ones, we can help as they happen [17:28:52] rits, in any case, the current problem should be about some definition in the existing files in your gerrit patchset [17:29:10] if you have mismatching stuff in your local version, just reset everything and pull the gerrit patchset [17:29:26] mooeypoo: ok then i should work over the css file [17:29:44] no its the same now as the gerrit patchset [17:29:47] yeah it should be either in the css or in the code, but something with the styles is probably wrong or missing [17:29:51] ok [17:30:20] mooeypoo: ok i will work over that now, thank you [17:30:25] sure [17:38:47] (03PS5) 10Mooeypoo: Create reference preview widget [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/200762 (https://phabricator.wikimedia.org/T93042) [17:41:54] (03PS1) 10Legoktm: Update extension.json, get rid of PHP entry point [extensions/Citoid] - 10https://gerrit.wikimedia.org/r/201235 [17:52:14] (03CR) 10MarkAHershberger: "The only relevant comment I see there is" [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/200305 (https://phabricator.wikimedia.org/T93616) (owner: 10MarkAHershberger) [17:56:12] (03CR) 10Esanders: "Could reference preview widget be made generic? There doesn't seem to be much in there that is specific to references." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/200762 (https://phabricator.wikimedia.org/T93042) (owner: 10Mooeypoo) [18:06:14] (03CR) 10Mooeypoo: [C: 04-1] Assigning width and height to alien extension nodes (032 comments) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/200945 (owner: 10Rits) [18:06:55] (03CR) 10Mooeypoo: "Good point. I didn't see current use of this outside references for the moment, but there probably is - if not for now, for the future for" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/200762 (https://phabricator.wikimedia.org/T93042) (owner: 10Mooeypoo) [18:07:23] edsanders, what shall I call this new widget? Just MWPreviewWidget or specifically MWGeneratedContentPreviewWidget ? [18:07:51] well, I guess it's not *just* for GCNs, though that's its main purpose [18:09:11] edsanders, mooeypoo: The fact that it uses dm.ReferenceModel and ce.InternalItemNode is specific to references, at lesat [18:09:52] RoanKattouw, yeah but I can probably generalize it if I find another example use, which is what I'm trying to do [18:11:39] RoanKattouw, though I'm having problems finding those. General templates, maybe? [18:13:28] I don't know that there are any right now [18:13:38] I think Ed wants to use this for captions [18:14:05] hm [18:14:08] so sub documents? [18:14:08] :D [18:15:07] Yeah :D [18:15:44] I coudl call it 'MWDocumentPreviewWidget" but I am going to have to do soemthing creative with config option to allow for various models and nodes [18:15:50] I have an idea but it's a little convoluded [18:16:11] For subdocuments that's what it should be [18:16:16] But we don't have subdocuments right now [18:16:33] So let's deal with the reality we have rather than the reality we wish we had [18:16:50] * mooeypoo nods [18:17:03] we can generalize it later, I guess. I thought there might be another example use right now [18:18:06] James_F the ref name bug is now https://phabricator.wikimedia.org/T94712 [18:18:13] RoanKattouw, okay, not even generalize the name? [18:19:05] Hmm [18:19:16] Maybe it should be called MWPreviewWidget and take an InternalItem? [18:19:22] And then I can make it take a Document later? [18:19:46] RoanKattouw, I need the document to create the getDomFromModel though [18:19:54] var preview = ve.dm.converter.getDomFromModel( this.model.getDocument(), true ) [18:20:23] or do you mean take both model and internal item? [18:22:25] A dm.InternalItemNode has a .getDocument() method [18:22:57] James_F, I also think you should read the comments here: https://en.wikipedia.org/wiki/Wikipedia:VisualEditor/Feedback#Copying_and_editing_links [18:23:01] hm [18:23:16] RoanKattouw, so I'll create the internal nodes outside the widget? [18:23:17] I can do that [18:23:44] just means it's a step back in the abstraction, but if it will mean better use cases, we can go with it [18:24:10] No, no [18:24:12] I'm saying [18:24:18] Take a *DM* InternalItemNod [18:24:24] And *create* a *CE* InternalItemNode [18:24:33] oh [18:24:35] mh [18:24:49] From a ReferenceModel I think there's a getter for the dm.InternalItemNode? [18:24:56] this.model.getDocument().getInternalList().getItemNode( this.model.getListIndex() ) <-- dm node [18:25:06] i think? [18:25:09] should be. [18:25:18] this.model instanceof ve.dm.MWReferenceModel ? [18:25:25] yep [18:25:35] according to the docs, that's right [18:25:39] See ve.dm.MWReferenceModel.prototype.findInternalItem [18:25:49] ve.ce.InternalItemNode takes ve.dm.InternalItemNode [18:25:53] and that is what I feed it now [18:26:15] Which word for word does what you said except that for some strange reason it takes a surface [18:26:26] RoanKattouw, @return {ve.dm.InternalItemNode|null} <-- #findInternalItem [18:26:31] Yeah [18:26:55] RoanKattouw, in all cases of use right now I have the referenceModel anyways [18:27:10] I can just use the same logic *outside* the widget to create the dm internal node and feed it. [18:27:45] Yeah [18:27:55] just trying to verify that i'm not losing any functionality. I don't think so, but I'll test it out [18:28:01] No I think you're fine [18:40:22] (03CR) 10Catrope: [C: 04-1] "I'm a bit confused. It seems like a lot of these changes don't change the behavior, there's only one of them that does?" (033 comments) [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/201205 (https://phabricator.wikimedia.org/T91245) (owner: 10Esanders) [18:41:02] ve-ce-documentNode is bouncing to { top: 0 } ... now I need to set a watchpoint on CSS changes [18:41:23] * hexmode hates that he is learning more about IE11 [18:42:49] hexmode: When does that happen? [18:43:07] When inserting a bullet [18:43:12] RoanKattouw: ^^ [18:43:18] hah [18:44:26] is my pain funny to you? :P [18:44:34] thats ok, no pain, no gain [18:44:42] I was surprised [18:44:43] I'm doing a lot of growing [18:44:59] I see scrolling issues on copy-paste in IE11 but I didn't realize it did something weird when inserting a bullet [18:45:15] Let me verify on enwiki [18:48:25] RoanKattouw: verified on enwiki. I can point you to my phab task if you're interested. [18:48:53] RoanKattouw: but even more helpful would be telling me how to trap whenever the screen jumps like that. [18:48:59] I have no idea [18:49:14] (03PS6) 10Mooeypoo: Create a preview widget [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/200762 (https://phabricator.wikimedia.org/T93042) [18:49:32] IIRC, I was stepping through the code and the BUG DIDN'T HAPPEN [18:49:42] grmmbl [18:49:52] Scrolling can sometimes be due to focus changes, so you could listen for those [18:50:29] $(document).on('focusin focusout', function ( e ) { console.log( e.type, e.target ); } ); [18:50:29] yeah, looking at fixes for similar bugs, that seemed to be what was fixed [18:50:39] :) [18:56:01] (03CR) 10Esanders: "I'm trying to change as few thing as possible. There are some redundant checks that can be removed later, but at the moment it's clear wha" [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/201205 (https://phabricator.wikimedia.org/T91245) (owner: 10Esanders) [18:57:26] (03CR) 10Catrope: [C: 032] [BREAKING CHANGE] Split setupSlugs in to setup(Block|Inline)Slugs [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/201205 (https://phabricator.wikimedia.org/T91245) (owner: 10Esanders) [18:58:21] edsanders|away: That patch fixes the bug (setupSlugs called twice), right? [19:05:35] (03Merged) 10jenkins-bot: [BREAKING CHANGE] Split setupSlugs in to setup(Block|Inline)Slugs [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/201205 (https://phabricator.wikimedia.org/T91245) (owner: 10Esanders) [19:08:31] James_F: i just filed two more blockers to https://phabricator.wikimedia.org/T92551 :/ [19:16:11] MatmaRex: Yay. [19:18:33] not exactly yay. [19:18:33] ;) [19:18:33] Not exactly not sarcastic. ;-) [19:18:33] * MatmaRex grumbles [19:37:09] uh, MatmaRex, quick OOUI PHP question [19:37:14] I'm trying to debug something in Collection [19:37:21] using eval.php [19:37:26] Caught exception OOUI\Exception: OOUI\Theme::singleton was called with no singleton theme set. [19:37:44] hm [19:37:54] Krenair: missing call to OutputPage::setupOOUI()? [19:38:03] enableOOUI* [19:38:14] which does: [19:38:14] OOUI\Theme::setSingleton( new OOUI\MediaWikiTheme() ); [19:38:14] OOUI\Element::setDefaultDir( $this->getLanguage()->getDir() ); [19:38:14] $this->addModuleStyles( 'oojs-ui.styles' ); [19:38:48] that did it, thanks [19:38:52] Krenair: i think cscott started writing his patch before we added that, heh [19:39:25] this is not actually using Collection's code [19:39:30] just trying something very similar in eval.php [19:39:42] I took a tiny snippet and probably missed such a call [19:47:30] RoanKattouw: ty for the snippet. Focus isn't changing, according to that. Which shows, I think, why the other fixes didn't work for this. [19:47:33] (sorry for the delay... life.) [19:48:08] anyone have other ideas about what causes screen jumping? [19:48:17] * hexmode goes to look for his bug so he can update it [19:57:54] James_F|Away: have you looked at mediawiki-l recently? VE's REL1_24 is somewhat broken. are we maintaining it (backporting stuff)? [20:05:33] Krinkle: whenever you feel like thinking about RL cache invalidation… https://gerrit.wikimedia.org/r/#/c/201267/ [20:06:53] MatmaRex: OK. Will check out tomorrow morning [20:07:05] thanks [20:15:10] RoanKattouw, thinking about https://phabricator.wikimedia.org/T49779#486219 again [20:15:56] so...on the edit source page, make the (VE) edit button open the same sort of OOUI dialog we show for switching from VE to edit source [20:17:40] if they choose to keep changes, load VE into the current screen, VE detects that there's a textbox open, gets the metadata for the page, and send the current wikitext off for conversion? [20:27:27] where's wikibugs? :( [20:28:45] forgetting to join channels again [20:28:46] legoktm [20:35:59] MatmaRex: having issues [20:37:05] (03CR) 10Legoktm: [C: 04-2] "And this will probably break i18n :/" [extensions/Citoid] - 10https://gerrit.wikimedia.org/r/201235 (owner: 10Legoktm) [20:53:42] (03PS4) 10Mooeypoo: Assigning width and height to alien extension nodes [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/200945 (owner: 10Rits) [20:55:33] mvolz: around? ankita-ks is wondering what's special about "lookup" in Citoid. [20:55:45] how would it be different from just "search" and "find". [20:56:41] (03CR) 10jenkins-bot: [V: 04-1] Assigning width and height to alien extension nodes [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/200945 (owner: 10Rits) [20:57:20] aharoni, you mean the button? [20:57:59] aharoni, it's "lookup" because we aren't really searching or finding a specific reference that already exists, we're looking up the source (the site) and building it [20:58:25] mooeypoo: https://translatewiki.net/wiki/Special:Translate?filter=&action=translate&group=ext-citoid [20:58:33] MatmaRex: I think it should be back now [20:58:43] mooeypoo: As you can see, I translated it to Hebrew with the same word as "search" ;) [20:58:48] would you suggest something different? [20:59:00] ankita-ks asks the same question for Hindi. [20:59:19] hm, yeah I'm trying to think [20:59:25] it's not quite 'search' [20:59:34] but i'm not sure i have an alternative word. Hmmm [21:00:21] aharoni, google translate the almighty (ha) says "lookup" is בדיקה [21:00:35] which isn't quite right, but it's not quite search either [21:01:09] aharoni, how would you translate 'retrieve' ? [21:01:42] mooeypoo: retrieve = אחזור, that's an easy one [21:01:49] לשלוף? [21:01:50] oh [21:03:22] mooeypoo: The Academy of the Hebrew Languages has a terms dictionary, which is sometimes very useful and sometimes very ridiculous :) [21:04:19] aharoni, lookup is more like cross referencing [21:04:27] it actually has "look up", and it's חיפוש (the same word as "search"), although it appears under the topic of Geodesy and not Information Technology. [21:04:30] RoanKattouw is making a good point about dictionary comparison [21:04:38] 10VisualEditor, 7Browser-Support-Internet-Explorer: Screen jumps in Internet Explorer 11 when clicking "Bullet list" or "Numbered list" in toolbar - https://phabricator.wikimedia.org/T93929#1172025 (10MarkAHershberger) Using a similar function to the one @Catrope gave me with way too many events, I narrowed it... [21:04:43] thanks legoktm [21:05:16] aharoni, "search" implies an uncertainty about where something is, while lookup is going to the place where you know it is to retrieve information [21:05:25] ... that makes sense? I have NO idea what alternative word for this in Hebrew though [21:05:42] what do you do to a word in the dictionary? You look it up, you don't search for it... how do we say that in hebrew? [21:05:50] That's what I said anyway [21:05:50] ... do we... say that in Hebrew...? [21:06:06] mooeypoo: according to a thesaurus, the synonyms of חיפוש are איתור, בדיקה, ביקוש, חיטוט, נבירה, סריקה, פשפוש [21:06:08] Like, if the user inputs an ISBN or DOI, getting the title for that isn't a "search" but it is a "lookup" [21:06:19] ביקוש! [21:06:24] yes, that sounds better to me [21:06:30] I think [21:06:43] Isn't this that you're asking for something you're know is there [21:06:51] Oh really. Curious. You're a native speaker, I'm not :) [21:06:52] or איתור [21:06:54] Krenair: So, AFAIK we lack proper Parsoid support to do real wikitext-to-VE switching [21:07:01] איתור ! [21:07:04] that's the one. I think! [21:07:12] aharoni, well that doesn't mean anything. You studied linguistics [21:07:14] I just use it [21:07:19] RoanKattouw, so... we need to make dependencies? [21:07:20] :D [21:07:23] RoanKattouw, what are we missing from parsoid? [21:07:32] Krenair: We were saying earlier that bidirectional switching is something we'd really like to do in Q4 (I see it's nominated as a Q4 bug) but it needs Parsoid/RB support [21:07:47] Well, you know how selser restores old wikitext for unmodified HTML? [21:08:05] Not how. [21:08:12] OK but you know that it does [21:08:13] I know it does that. [21:08:23] aharoni, I would say איתור is better. Sounds closer to something like 'pinpoint' or 'spot' [21:09:06] mooeypoo: yeah, that's the one I'd pick if not חיפוש. Nice discussion :) [21:09:07] Roughly speaking, it does this by knowing that a certain HTML node started as a certain substring of the wikitext (I believe that this is what dsr or one of the Xsr things is: [12,20] meaning "this HTML node was origWikitext.slice( 12, 20 )) [21:09:17] aharoni, yes! :) thanks for that [21:09:26] But so when serializing, it needs to refer back to the original wikitext [21:09:35] Which is fine when the original wikitext is a revision that's in the database [21:09:54] But when it's "random crap the user had in their editor when they chose to switch", it's not necessarily stored anywhere [21:10:49] This problem is probably solvable, and the Parsoid and RESTbase teams probably have better ideas for solving it than I do -- I'm just saying that selser won't obviously work without some additional effort [21:11:01] ankita-ks: So, for Hebrew mooeypoo and I found another word that is not "search", but more like "locate". If you can think of something like that for Hindi, it would be good. If not... not a big deal, just use the same word as "search". [21:12:48] RoanKattouw_away: Parsoid now has the ability to pass in the wikitext as well [21:13:01] we don't expose that in RB yet, but we can [21:15:55] heya rits [21:16:21] okay. let me think. [21:16:28] mooeypoo: hello sorry for getting offline, just didn't when i got to sleep its 3AM here actually :) [21:16:37] rits, aw! go sleep! [21:16:39] :) [21:16:58] mooeypoo: no i had now :) [21:17:29] mooeypoo: i tried the changes but the css classes is not showing up [21:17:34] btw does "איתור" translate to "locate"? [21:17:38] aharoni ^ [21:18:09] RoanKattouw_away, ... but that's exactly what we want [21:18:29] RoanKattouw_away, we want all the diffs etc. to be against the revision in the DB [21:18:44] ankita-ks: yeah, something like that. do you have a Hindi thesaurus? (Dictionary of synonyms.) [21:19:20] I am looking online. I will get something like that from our libarary tomorrow. that will be a more reliable version. [21:19:42] RoanKattouw_away, we're not just going to send them straight back to the wikitext editor when they're done in VE, are we? [21:19:51] mooeypoo: i think there is something in the class file, css classes should first show up , changing the css file would not fix it right [21:21:11] rits, okay I went over your code [21:21:19] first off, when you run it you should see that there are errors in the console [21:21:45] rits, do you see any errors in the console now when you run this? [21:22:27] mooeypoo: no i didnt got any [21:22:48] mooeypoo: just checked it again [21:23:54] rits, that's really weird. I'm getting an error when I run it [21:24:12] rits, can you update your master, pull, then review the change again from gerrit? [21:24:15] Krenair: Suppose you start with wikitext, make some changes then switch to VE, then make more changed [21:24:16] *changes [21:24:18] then save [21:24:18] I want to make sure we're both looking at the same code [21:24:19] mooeypoo: i mean i just checked it i didn't got any errors [21:24:27] Parsoid will parse from your modified version [21:24:29] mooeypoo: ok i will just do it [21:24:47] But it somehow needs to preserve enough information so as to selser cleanly against the unmodified version [21:25:17] RoanKattouw, the unmodified version is the version that we started with before you touched it in the wikitext editor [21:25:29] not what we entered into VE from the wikitext editor [21:25:30] yes [21:25:44] The latter I called the modified version [21:25:55] rits, you're not getting the error "Uncaught TypeError: Cannot read property 'detach' of undefined" in your console? [21:26:43] mooeypoo: no [21:26:56] RoanKattouw, Can Parsoid not take some wikitext and the revision ID it was based on, and generate the HTML to put in VE? [21:27:01] rits, what browser are you using, and what code are you pulling? Are we looking at the same code? [21:27:39] mooeypoo: google chrome, yes i am on the same code as i submitted the patchset to gerrit [21:27:58] Krenair: Probably [21:28:06] rits, you have a couple of patchsets -- let's make sure we're looking at the same one. Which url or patchset number? [21:28:11] Krenair: That's a good suggestion actually [21:28:17] Krenair: But such functionality does not exist right now [21:28:27] what's missing from it? [21:28:42] mooeypoo: this one https://gerrit.wikimedia.org/r/#/c/200945/ [21:28:43] rits, are you looking at the chrome console itself? ctrl-shift-j ? or some plugin? [21:29:07] Right now it's always the case that the wikitext it generates HTML from and the wikitext it uses as its selser base are the same [21:29:11] mooeypoo: no directly the chrome console, no plugins [21:29:13] That might need to change, or something else might need to happen [21:29:24] Anyway, we should discuss this with the Parsoid people [21:29:33] We have our weekly VE/Parsoid dependencies meeting tomorrow after the standup on IRC [21:29:38] rits this is really weird. I'm getting an error, and that error actually makes sense to be an error [21:29:55] rits, is VE loading for you at all when you are with this patch pulled? [21:30:00] or is it stuck during load? [21:30:32] mooeypoo: yes its loading i am also surprised by this thing [21:30:41] RoanKattouw, I don't really understand the problems involved here... It seems to me like we already have everything we need [21:30:47] no errors at all its working fine [21:31:26] Parsoid has an API to take arbitrary wikitext and turn it into VE. So we take the user's modified wikitext and send it through that, then let them edit the HTML further in VE [21:31:32] Turn it into HTML* even [21:31:51] rits, are you looking at a page that has an MWAlienExtensionNode already in it? [21:31:58] some graph data or something ? [21:32:04] krenair has arrived ;) [21:32:21] mooeypoo: yes its having a graph [21:32:22] Then VE uses the normal process to turn the resulting HTML back into wikitext, based on the original wikitext before they started changing anything in wikitext [21:32:27] gwicke, .. what? [21:32:34] Krenair: in VE land [21:32:40] rits, did you ever work with the Chrome console before? just veritfying [21:32:43] heh [21:32:53] Krenair: But then when converting that HTML back to wikitext, selser needs access to the original wikitext, which has already vanished in a puff of ephemerality at that point [21:33:12] we can know which revision ID they started with in the wikitext editor [21:33:13] mooeypoo: yes i did for my blog project [21:33:22] rits, we just verified, and I'm not the only one that has the console error output -- if you're not seeing it, we need to see if you're on the same patchset [21:33:26] that's all we need, isn't it? [21:33:37] can you go to your terminal and hit 'git status' and tell me what the output is? [21:33:38] Krenair: the problem is that the last parse would use modified wikitext [21:33:38] Yes, but the wikitext we converted to HTML and the wikitext of the revid they started with could be completely different [21:33:47] so the dsr is different [21:33:51] i am familiar to it i dont why its not giving a error [21:33:54] they're supposed to be completely different [21:34:16] rits, let's try to make sure you're in the right patchset in git? [21:34:17] the HTML you put into VE and the HTML you end up with when saving can be completely different [21:34:18] that's ok [21:34:19] which in turn means that Parsoid can't use the original wikitext for anything unless it can establish some DOM equality of the original HTML & the modified one [21:34:38] alternatively, it can use the modified wikitext for selser [21:34:59] but then we'll need to keep around intermediate HTML and data-parsoid as well [21:35:22] which isn't impossible either, just a complication [21:35:22] mooeypoo: http://pastebin.com/UZq2yVpq this was the output to git status [21:36:07] mooeypoo: RoanKattouw mvolz - I updated the qqq for lookup, FWIW: https://translatewiki.net/wiki/MediaWiki:Citoid-citeFromIDDialog-lookup-button/qqq [21:36:31] mooeypoo: i dont know why its showing changes to lib/ve i didnt did any actually [21:36:56] Krenair: one potential issue with the intemediate html / data-parsoid solution is that dirty diffs can accumulate from each switch, but I think at this point Parsoid might be good enough for that to not matter very much [21:37:00] rits, it's not, it's a submodule, so it's pointing to the wrong thing. hang on [21:37:03] let's do this: [21:37:07] gwicke: Yeah storing the intermediate wikitext / HTML / data-parsoid was my first idea; I defer to you guys as to what the best solution is [21:37:25] the other issue is performance, especially if you want to do full-page editing on large pages [21:37:29] rits, okay, let's do this -- 'git checkout master' 'git pull' 'git submodule update' -- when you're done, do 'git review -d 200945' and reload your page [21:37:31] mooeypoo: ok [21:37:37] switching from edited wt to html can take some time [21:37:42] ok [21:37:45] gwicke: But regardless of how it's solved; we'd really like Parsoid support for wikitext->VE switching to happen soon since it's a Q4 task for VE [21:38:13] So you think we might get dirty diffs out of allowing wikitext editing in the process? [21:38:18] (I guess today is the first day of Q4 so "soon" doesn't necessarily mean "ZOMG we need it next week", but you should know that it's on our roadmap for this quarter) [21:38:35] Krenair: I think he's saying that dirty diffs could /accumulate/ with each switch [21:38:41] Krenair: there is a bigger chance of that with many switches / edits, yes [21:38:53] might no longer matter in practice though [21:39:47] I think from a usability pov wt2html performance for large pages would be the biggest problem [21:40:03] Right [21:40:20] rits, did it work? [21:40:23] section editing could help with that [21:40:36] I suppose in the longer term, optimizations could be made by referring to the original wikitext (from the DB) and the HTML that's already in RB for that? [21:40:40] (Kind of a reverse selser?) [21:40:44] Or yeah something section-based [21:40:45] or some section edit like optimization in parsoid [21:40:53] mooeypoo: no i thought to just clear the cache might be beacuse of that ve is loading [21:40:53] which we have been talking about forever now.. [21:41:01] Yes [21:41:03] perfectly now also [21:41:07] That's another thing that's probably gonna be a Q4 thing [21:41:11] Section-based stuff [21:41:27] rits, did you do the above steps in git ? [21:41:31] yeah, that's our top prio [21:41:43] along with other API tweaks [21:41:45] mooeypoo: yes i followed each one [21:41:48] Excellent [21:42:06] rits, can you show me what your page has with the wikitext ? [21:42:45] rits, also, did you install the Graph extension, and is it working for you in "view" mode ? [21:42:50] ok i will send you it in pastebin one second [21:42:56] great [21:43:28] mooeypoo: yes its working i can see the graph in the view mode and i have installed the graph extension [21:43:38] ok awesome, just checking [21:43:50] trying to think of anything that could happen :) [21:44:22] mooeypoo: yes its to wierd why is not showing a error :) [21:46:00] mooeypoo: this is the wikitext http://pastebin.com/uVGjhUjr [21:46:04] 10VisualEditor, 6Editing: In Firefox: after inserting a link in VE, the cursor stays inside the link as you continue to type - https://phabricator.wikimedia.org/T94765#1172201 (10DannyH) 3NEW [21:47:04] 10VisualEditor, 6Editing: In Firefox: after inserting a link in VE, the cursor stays inside the link as you continue to type - https://phabricator.wikimedia.org/T94765#1172210 (10DannyH) [21:47:50] mooeypoo: these is the git log i followed, if might something wrong in this http://pastebin.com/msydt3Gg [21:48:08] (03PS1) 10Esanders: Re-apply selection when ContentBranchNode changes [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/201357 (https://phabricator.wikimedia.org/T94696) [21:48:46] (03CR) 10Esanders: "This fix feels a little sledgehammery, but I'm not sure there's a lighter way that's still reliable." [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/201357 (https://phabricator.wikimedia.org/T94696) (owner: 10Esanders) [21:49:37] (03CR) 10Catrope: [C: 032] Re-apply selection when ContentBranchNode changes [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/201357 (https://phabricator.wikimedia.org/T94696) (owner: 10Esanders) [21:52:06] rits, excellent, we're both on the same code now. When you load VE now, it's still not throwing any errors? [21:52:22] RoanKattouw: https://github.com/jquery/jquery/issues/2145 [21:52:40] WTF [21:52:48] https://github.com/jquery/jquery/issues/2145#issuecomment-85094548 [21:52:51] specifically that one [21:52:51] mooeypoo: no [21:53:03] Krinkle: In case I hadn't already told you: iOS Safari is the worst browser in existence, and I say that as somebody who worked on IE support professionally [21:53:23] rits, okay, let's do another thing. Open the console on Sources [21:53:24] mooeypoo: amm could be my browser problem? [21:53:31] (03Merged) 10jenkins-bot: Re-apply selection when ContentBranchNode changes [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/201357 (https://phabricator.wikimedia.org/T94696) (owner: 10Esanders) [21:53:35] ok [21:53:37] not really? shouldn't..? I dunno :\ [21:53:49] This is the second time in less than a year that jQuery has to do a minor release for a serious regression in iOS Safari that apple isn't fixing or backporting. [21:54:02] Both times the bug was fixed in WebKit master in a few days, but no backport or release [21:54:08] rits, okay on the right hand side you should have the box for watch expressions and call stack [21:54:08] "The problem likely only occurs after the code has been executed a few thousand times." [21:54:15] mooeypoo: i did [21:54:22] Krinkle: What was the first time? [21:54:23] yes [21:54:25] above it are the icons to move forward/pause etc, right? [21:54:29] RoanKattouw: Much much worse. [21:54:35] RoanKattouw: Basic css selectors not working [21:54:38] the last button on that row should be a stop-sing with "pause" in it [21:54:45] http://blog.jquery.com/2014/12/18/jquery-1-11-2-and-2-1-3-released-safari-fail-safe-edition/ [21:54:46] rits, "Pause on exceptions" [21:55:09] yes they are [21:56:18] mooeypoo: i paused it [21:56:23] rits, click on tha button and refresh VE [21:56:35] it's not really pause, it's just causing the code to stop on exceptions [21:56:38] mooeypoo: ok [21:56:44] refresh VE now [21:56:52] tell me if the code stops [21:57:48] mooeypoo: no it didnt the ve loaded again [21:57:56] rits, I don't know what to say. [21:58:09] rits, I'm loading your page with your wikitext [21:58:10] mooeypoo: me too , i am sorry [21:58:25] the code should fail, I see exactly where it should fail, too ... I have no idea why it's not failing for you [21:58:31] rits, let's try Firefox? [21:58:43] ok firefox just a second [21:59:14] 10VisualEditor, 10VisualEditor-ContentEditable, 3Editing Department 2014/15 Q4 blockers, 5WMF-deploy-2015-04-08_(1.25wmf25): View/model out of sync when undoing paragraph creation - https://phabricator.wikimedia.org/T94696#1172270 (10Jdforrester-WMF) p:5Triage>3High a:5dchan>3Esanders [21:59:20] 10VisualEditor, 10VisualEditor-ContentEditable, 3Editing Department 2014/15 Q4 blockers, 5WMF-deploy-2015-04-08_(1.25wmf25): View/model out of sync when undoing paragraph creation - https://phabricator.wikimedia.org/T94696#1172273 (10Jdforrester-WMF) 5Open>3Resolved [21:59:46] Krinkle: any clue to tracing down this anonymous function? https://phabricator.wikimedia.org/T93929 [22:00:20] Krinkle: I've set a watch on the scroll event but that doesn't show me who set it... [22:00:32] mooeypoo: no errors there [22:00:44] i will do it with the source mode [22:00:50] just a moment [22:01:07] rits, are you sure you're loading the page from the same place we're pulling the code? [22:01:56] mooeypoo: yes , from localhost and i updated the changes to that only [22:02:09] ok rits go to the console in VE when it's loaded [22:02:10] mooeypoo: same in firefox [22:02:36] mooeypoo: there [22:02:55] rits, type ve.ce.MWAlienExtensionNode.prototype.onSetup.toString(); [22:03:47] then copy the result of what you see in the console and put it in some pastebin [22:03:51] mooeypoo: got this Uncaught TypeError: Cannot read property 'prototype' of undefined [22:04:07] ... whaaat [22:04:24] mooeypoo: this was the output [22:04:34] amm yes [22:04:49] rits: Was this while VisualEditor was already loaded? [22:05:20] Or did you run it before loading VE? [22:05:22] RoanKattouw: yes it is loaded [22:05:27] WTF [22:05:35] hexmode: I'd defer to RoanKattouw, but I think we've got either multiple or one central issue where we invoke scroll for some reason. [22:05:48] It happens when focussing, when pasting, when typing sometimes. [22:06:01] rits, can yo utake a screenshot of the browser with the console ? [22:06:06] And now the find/replace also seems to be doing something smart with the scroll offset [22:06:21] which accounts to growing offsets under certain window/document size ratios [22:06:28] amounts* [22:06:30] causes [22:06:30] mooeypoo: yes where should i send it [22:06:50] rits, you can email it to me moriel@wikimedia.org [22:06:56] Krinkle: yeah, my grep for scroll found several things. Right now the debugger has me in some jqeury bit and I'm trying to figure out who sets that up. [22:07:01] rits: If you type ve.version.id in the console, what do you get back? [22:07:39] RoanKattouw: de5275e this [22:07:39] hexmode: To track down anonymous functions like that I recommend using debug=true [22:07:48] hexmode: Or Chrome, but since you're tracking down an IE bug that's not an option [22:07:51] mooeypoo: just a second [22:07:56] rits: Hmm that is the right verison [22:08:09] * hexmode loves debug=true for VE :) [22:08:15] rits: What about typeof ve.ce.MWAlienExtensionNode [22:08:25] hexmode: I mean sure, it'll take approximately a year to load :P [22:08:42] Clippy!! https://review.openstack.org/#/q/status:open+project:openstack-infra/nodepool,n,z [22:08:45] RoanKattouw: undefined this [22:08:54] wtf [22:08:59] HOW [22:09:04] * RoanKattouw is very confusd [22:09:09] RoanKattouw: IE does clean up the code so at least I'm not left staring at onelonglineofgibberish [22:09:46] but I'll try debug=true now [22:09:49] ARGH [22:09:51] OF COURSE [22:09:58] It's a freaking experimental module [22:10:07] Krinkle, that's awesome! We should've had him for April fools.... [22:10:15] * RoanKattouw throws a few rocks and a kitchen sink in the general direction of the person that came up with that [22:10:19] Oh wait, that was me [22:10:40] rits: In LocalSettings.php, add $wgDefaultUserOptions['visualeditor-enable-experimental'] = 1; [22:10:51] Otherwise MWAlienExtensionNode doesn't get loaded [22:10:54] Hope you ducked your rocks, RoanKattouw [22:10:57] This is entirely my fault and I feel bad :( [22:11:03] oh wow [22:11:22] RoanKattouw: ok [22:11:43] haha [22:12:12] rits, phew! okay, *now* it should work :) thanks RoanKattouw ! [22:12:31] or rather, now it should break. It's kind of funny but we're trying really really hard to break rits' wiki. [22:12:35] mooeypoo, still about 12 hours of april fools day somewhere in the world left [22:12:52] Krenair, don't look at me, I've been pushing for a Clippy in VE for april fools for 2 years [22:13:13] someone James_F is a party pooper. [22:13:41] It's currently 12:13 HAST [22:13:50] mid day on 1st april [22:13:52] whoa, what the... there's a bright blue progress bar now? [22:13:59] yes greg-g [22:14:02] mooeypoo: yay got one now Uncaught TypeError: Cannot read property 'detach' of undefined [22:14:09] rits, yay! [22:14:15] rits, okay, now we can talk about why that happens :D [22:14:17] RoanKattouw: thanks [22:14:33] Krenair: interesting [22:14:37] mooeypoo: yes :) :D [22:14:54] rits, this is actually not your fault at all; the problem is that render is running *before* 'onSetup' -- this means that this.$icon is not set up yet when we call fro it to be detached() so the code breaks [22:15:24] mooeypoo: ok [22:16:28] mooeypoo: so how should it attached to it then, but the same is followed in the template case [22:16:43] rits, the reason it works in the MWTransclusionNode case is that there, we load the mixins in a different order. Here, though, MWAlienExtensionNode has its mixins in its own parent, so tweaking the order and putting OO.ui.IconElement in between will be a problem [22:16:59] rits, so, instead, add a condition before you detach [22:17:12] first check if this.$icon isn't null, and only if it's not null, detach it. [22:17:44] mooeypoo: okay got it [22:17:50] rits, also, look at the other 2 comments I wrote on patch set 3 [22:18:11] rits, now it should work :) [22:18:30] mooeypoo: yes i did for the config [22:19:17] great. It should work after those fixes. [22:19:51] mooeypoo: and for the inheritdoc i will fix that, thank you so much :) its great working with you thank you :) [22:24:02] 10VisualEditor: Newly made links are red until page is saved - https://phabricator.wikimedia.org/T75741#1172352 (10Whatamidoing-WMF) I don't know, but I'm seeing red links in Citoid. It's in the initial display, not when you go back to look at it later. I suspect that this is a separate bug, though. (Steps to... [22:24:30] mooeypoo: yay yay , it worked thank you so much :) [22:24:43] RoanKattouw, were you doing some work on subdocuments at some point? [22:25:22] ah it was https://gerrit.wikimedia.org/r/#/c/179940/, ok [22:25:23] (03CR) 10Esanders: "Is that a no then?" [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/201357 (https://phabricator.wikimedia.org/T94696) (owner: 10Esanders) [22:25:30] very very wip, haha [22:26:16] (03CR) 10Catrope: "I think what you did is pretty reasonable. I suppose David might have better ideas." [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/201357 (https://phabricator.wikimedia.org/T94696) (owner: 10Esanders) [22:31:17] rits, one more thing (without driving you crazy) -- can you also add a short comment above the 'if not null' condition? Just explain that render is triggered before onSetup so we have to make sure we only detach when we actually have an icon first [22:32:09] mooeypoo: yes i will add it [22:38:54] (03Abandoned) 10Mooeypoo: Hide slugs in context menus [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/200086 (https://phabricator.wikimedia.org/T93042) (owner: 10Mooeypoo) [22:48:45] James_F, RoanKattouw: so what's the plan re loading HTML directly from RB on group 0 wikis? [22:49:02] I think we want to do that in prod soon? [22:49:17] But the code to support that has been on group0 wikis for like... an hour? [22:49:22] James_F knows the plan [22:49:24] gwicke: Not yet. [22:49:29] :-) [22:49:35] Labs first, then group0 tomorrow. [22:49:41] is it enabled in labs? [22:49:42] If all goes well, enwiki this time next week. [22:49:43] No. [22:49:48] Ahm, yes? [22:49:51] no? [22:50:01] Wait. [22:50:02] (03PS5) 10Rits: Assigning width and height to alien extension nodes [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/200945 (https://phabricator.wikimedia.org/T92868) [22:50:02] Yes. [22:50:04] Never mind. [22:50:13] It works fine in Labs. [22:50:15] Err. [22:50:18] Beta Cluster. [22:50:18] ;) [22:50:24] That's what I meant [22:50:31] sooo.. group 0? [22:50:37] (03CR) 10jenkins-bot: [V: 04-1] Assigning width and height to alien extension nodes [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/200945 (https://phabricator.wikimedia.org/T92868) (owner: 10Rits) [22:50:47] Yeah, let's do that now for SWAT. [22:50:49] 10VisualEditor, 10Parsoid, 3Editing Department 2014/15 Q4 blockers: mw:WikiLink with fully resolved url doesn't serialize to the right piped wikilink syntax - https://phabricator.wikimedia.org/T94723#1172426 (10ssastry) p:5Triage>3Normal [22:51:00] * gwicke is keen to give it a try on prod hw [22:51:06] Yeah. [22:54:00] (03PS6) 10Rits: Assigning width and height to alien extension nodes [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/200945 (https://phabricator.wikimedia.org/T92868) [22:59:18] mooeypoo: please review the patchset https://gerrit.wikimedia.org/r/#/c/200945/ ,thank you for your time :) [22:59:36] rits, awesome, will do [22:59:42] rits, thanks! [23:00:11] mooeypoo: okay, :) , my pleasure [23:01:31] (03PS4) 10Mooeypoo: [breaking change] Make default property interfaceText [extensions/TemplateData] - 10https://gerrit.wikimedia.org/r/195451 (https://phabricator.wikimedia.org/T54966) [23:02:23] (03PS5) 10Mooeypoo: [BREAKING CHANGE] Make default property interfaceText [extensions/TemplateData] - 10https://gerrit.wikimedia.org/r/195451 (https://phabricator.wikimedia.org/T54966) [23:04:54] (03CR) 10Mooeypoo: [C: 04-1] "Small style comments" (033 comments) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/200945 (https://phabricator.wikimedia.org/T92868) (owner: 10Rits) [23:05:36] Can someone take a look at https://gerrit.wikimedia.org/r/#/c/200801/ ("Add pluggable talk page poster and use it for mediawiki.feedback")? [23:10:03] 10VisualEditor, 10WMF-Product-Strategy, 3Editing Department 2014/15 Q4 blockers: Perform a weekly review of edits made with VisualEditor - https://phabricator.wikimedia.org/T94767#1172445 (10Jdforrester-WMF) [23:23:24] MatmaRex, didn't we fix the popups in RTL? we had a bug where they popped up outside the window when at the edge? [23:24:10] MatmaRex, we seem to have a regression in the popoup inside the tempalte dialog, the one that shows the description of params -- it goes outside the dialog (hidden) in RTL [23:24:30] mooeypoo: we did fix something like that at some point, affecting toolbar popups [23:24:54] MatmaRex, ahh, it was only toolbar ones?I thought we fixed popup widget in general [23:26:17] mooeypoo: yes, it was popupwidget, and in general. [23:26:24] hm [23:26:25] toolbar was just where that was visible [23:26:33] trying to figure out what happened and how to fix it [23:26:34] The popup also anchors in the wrong place [23:26:56] yeah i was going to say that 00 in rtl it anchors slightly to the left, it seems on the edge of the icon [23:33:13] TrevorP|Away: https://support.office.microsoft.com/en-gb/article/Office-language-interface-pack-LIP-downloads-d63007c2-e8ae-41fd-8bfb-fce2857010e1?ui=en-US&rs=en-GB&ad=GB [23:35:24] TrevorP|Away: http://ckeditor.com/support/faq [23:38:35] 10VisualEditor, 10VisualEditor-MediaWiki-Media, 7Verified-in-Phase0, 7Verified-in-Phase2, 5WMF-deploy-2015-03-25_(1.25wmf23): Setting a image to reflow after inserting prevents saving, offset out of bounds - https://phabricator.wikimedia.org/T93204#1172499 (10Etonkovidova) [23:38:50] (03PS1) 10Bartosz Dziewoński: [WIP] Properly support LTR/RTL icon versions in colorize-svg.js [oojs/ui] - 10https://gerrit.wikimedia.org/r/201379 (https://phabricator.wikimedia.org/T94560) [23:40:37] blergh. [23:41:42] 10VisualEditor, 10VisualEditor-DataModel, 7Verified-in-Phase2, 5WMF-deploy-2015-03-25_(1.25wmf23): VisualEditor: Whitespace-only headings should be collapsed to (blank) paragraphs - https://phabricator.wikimedia.org/T51452#1172508 (10Etonkovidova) [23:42:56] 10VisualEditor, 10VisualEditor-MediaWiki, 3Editing Department 2014/15 Q4 blockers, 7user-notice: VisualEditor: Provide a way for the user to switch between VisualEditor and wikitext source editor modes without saving - https://phabricator.wikimedia.org/T49779#1172511 (10Krenair) We should talk to #Parsoid... [23:44:30] 10VisualEditor, 10VisualEditor-MediaWiki, 3Editing Department 2014/15 Q4 blockers, 7user-notice: VisualEditor: Provide a way for the user to switch between VisualEditor and wikitext source editor modes without saving - https://phabricator.wikimedia.org/T49779#1172521 (10Jdforrester-WMF) >>! In T49779#11725... [23:51:29] 10VisualEditor, 7Verified-in-Phase0, 7Verified-in-Phase2, 5WMF-deploy-2015-03-25_(1.25wmf23): Comment body in context showing a scrollbar - https://phabricator.wikimedia.org/T92450#1172549 (10Etonkovidova) [23:56:15] Are you guys okay with https://gerrit.wikimedia.org/r/201388 ? It only affects MW.org and test wikis. it's so the VE API doesn't bail on Flow pages (at least in NS_TALK) until we figure out a better solution. ^ James_F