[00:03:40] My god, the oojs-ui docs are sexy as fuck [00:03:57] lol [00:04:00] rdwrer: Thanks. :-) [00:04:00] Like jsduck, but so much more than that [00:04:05] Everything is so great [00:13:48] OK, TrevorParscal, I have https://lh4.googleusercontent.com/q2uafYtSIdk_9Tg-RLDnOvAqUOHtd1_mLlh8TJk0rMiskCdnctpD4VBjOUWCWgJSISucamZD37mo6LFbYDZhvxEST8b92d8ZORn4Us1oRB3WTz-jWlpTRDhUZA [00:16:42] I'm using a ButtonSelectWidget for the HTML/wikitext switcher [00:17:17] I'm sure it will be easy to find the widget for the drop down [00:17:17] But is there a class I can use for the entire panel? [00:19:01] Ideally something with the little arrow bit already, else I guess I need to use the Window class or something [00:24:09] ok i'm running to class [00:24:46] cya later guys [01:07:42] I guess I will await an answer at home [01:14:46] (03PS1) 10Kaldari: Move some skin styling from ve.ce.Node.css to the skin styles [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112618 [01:27:23] (03CR) 10Jforrester: [C: 04-1] "Icon needs to be fixed." [oojs/ui] - 10https://gerrit.wikimedia.org/r/112603 (owner: 10Trevor Parscal) [01:32:56] (03PS26) 10Jforrester: Let users set #REDIRECT and __STATICREDIRECT__ status [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/95213 [02:01:25] (03PS10) 10Jforrester: Let users set __NOINDEX__ & __INDEX__ status [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/95724 [02:02:18] (03CR) 10Jforrester: "PS10 is a rebase. Still awaiting Trevor's new icon being available." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/95724 (owner: 10Jforrester) [02:02:25] (03PS13) 10Jforrester: Let users set __NEWEDITSECTION__ & __NONEWEDITSECTION__ status [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/95725 [02:02:41] (03PS12) 10Jforrester: [WIP] Create remaining meta items in DM [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/95726 [05:04:37] RoanKattouw, oh, haiii, how was your test?? [05:05:11] Hey [05:05:31] Went pretty well but I did get a trick question wrong [05:05:44] aww [05:05:50] I hate when that happens. [05:06:22] Did you know you were going to get it wrong as you were answering it, or was it one of those questions you were confident about until you found out you're wrong?? [05:06:23] Then when I got back to the office, I fixed a Parsoid outage [05:06:31] And now we're just starting the third outage of the day [05:06:37] I knew I was on shaky ground [05:06:49] ah, at least you had an inkling. [05:06:58] It's the surprising fail that getcha. [05:07:05] Yeah [05:08:34] wait, there's a parsoid outage? [05:09:05] Was [05:09:14] ~2 hours ago [05:09:22] A typo caused logrotate not to work [05:09:34] oops [05:09:34] And since the logs grow uniformly and the backend machines are mostly identical ... [05:09:45] oy [05:09:50] almost all backends ran out of disk space within 5-10 mins of each other [05:09:57] hah [05:10:10] Of the 24, 5 survived, and Sean brought back one before Erik pulled me in, so we had 6/24 backends alive [05:10:32] But LVS has a depool threshold where it won't let the pool shrink below a certain percentage [05:10:32] ouch [05:10:42] pool? [05:10:45] Presumably that threshold is >6/24, and so we were sending the occasional request to a broken machine [05:10:52] Pool of machines the load balancer balances over [05:11:00] Machines that stop responding are automatically depooled [05:11:17] oh, ok, I thought you were talking about logs, and it made no sense. Ahhh [05:11:31] But to prevent pathological behavior during overload-related outages, there is a minimum below which it'll stop depooling boxes [05:13:04] Yup, our depool threshold is 50% [05:14:15] interesting. [05:14:46] Hm. wait, how do you deal with a machine you can't reach? (might be a stupid question here, but..) [05:15:31] So, if a machine is unreachable or for any reason suspect, it's temporarily removed from the pool (depooled) [05:15:37] until such time as we trust it again, then it's repooled [05:16:45] Right. But then you had so many unreachable.. even if you fix the problem so no *more* logs are added in, you still need to access the machines to fix them up... [05:18:14] Also, I must point out, when you said machines stopped responding, I immediately thought of this: http://bash.org/?5273 [05:18:14] Yeah so we had 3/4 of the cluster broken with full disks [05:18:27] So the LB refused to depool any more after it had depooled half the pool [05:18:32] so half the requests were going to broken backends [05:18:56] Fortunately, 1) the Parsoid restart script clears the log file and 2) we have a utility for ssh-ing into lots of machines at ones [05:19:05] So modulo some bugs in the Parsoid setup and the restart scripts, that fixed it [05:19:08] ooh [05:19:21] Nice. [05:28:37] RoanKattouw, what dictates which image node (inline or block) VE builds from the parsoid output? [05:29:56] I have this bug, and it's really odd, and i'm not sure if this is an issue with how we get the output from Parsoid or if this is a problem on VE's interpretation of the node https://bugzilla.wikimedia.org/show_bug.cgi?id=61153 [05:35:35] mooeypoo: Usually whether it's a div or a span [05:35:47] The first question you should be asking is "what is the Parsoid HTML" [05:35:52] So this is a parsoiid issue? [05:36:06] as far as I saw, a span, but I need to double check [05:36:23] So, we're talking about the image wrongly being inline when the editor is loading, right? [05:36:34] Like, this image is present in the wikitext, it wasn't inserted by VE during this session [05:36:37] * mooeypoo helped with a persistent Parsoid bug today [05:36:44] That was fun. [05:36:50] What was that? [05:37:19] RoanKattouw, well, even if I write the wikitext manually it is the same effect, the image |frameless is inline if it doesn't have a position set (in other words, if position is 'default') [05:37:54] oh, there were errors in the console when pages were parsed [05:38:12] but it only happened in my supermassive complex-image-wikitext page, not in my "regular" wikitext page [05:39:02] turned out to be
 tags with surrounding text, which apparently resulted in this: https://gerrit.wikimedia.org/r/#/c/112608/ 
[05:39:04] 	 \o/
[05:39:04] 	 Hah, OK
[05:39:19] 	 lol, pre in p
[05:39:27] 	 Yeah I can see how that would be difficult to deal with
[05:39:40] 	 OK so if you can reproduce the bug by just writing wikitext and firing up VE, then you should look at the Parsoid HTML
[05:39:55] 	 If Parsoid produces a span, I expect an inline image, if it produces a figure, I expect a block image
[05:40:06] 	 s/span/span or /
[05:40:27] 	 yep
[05:40:36] 	 it produces a span
[05:40:50] 	 |frameless SHOULD be blocks, right?
[05:41:07] 	 I have no idea
[05:41:08] 	 the advanced wikitext is so complex, I'm having doubts here. 
[05:41:19] 	 What does the PHP parser do?
[05:41:44] 	 so, parsoid:
[05:41:48] 	 [[File:Westminstpalace.jpg|frameless]]
[05:41:48] 	 [[File:Westminstpalace.jpg|frameless|left]]
[05:41:48] 	 [[File:Westminstpalace.jpg|frameless|none]]
[05:41:53] 	 inline, block, block
[05:43:11] 	 ah, so does the php parser
[05:43:14] 	 is that weird or WHAT.
[05:43:22] 	 ... so... apparently this is "expected behavior" O.o
[05:44:05] 	 So, that's default floating (left for you), left floating, and not floating?
[05:44:17] 	 If so, they're all aligned the same, right?
[05:44:29] 	 (at least on a simple page)
[05:44:43] 	 That's a little crazy. having no position is basically equal to having "none" position, no??
[05:44:57] 	 they're all aligned the same, but the first is inline (valign middle, incidentally)
[05:45:01] 	 the second is float:left
[05:45:16] 	 and the third is div floatnone
[05:45:24] 	 but the first *should* be 'none', no?
[05:45:29] 	 this is conceptually weird.
[05:45:31] 	 I would think that having no position defaults to contentLang->isRTL() ? left : right
[05:45:42] 	 if so, then it should also be a block
[05:45:46] 	 But maybe that depends on the image type too and frameless is different somehow
[05:45:50] 	 instead it's an inline... that's just random.
[05:45:55] 	 I don't know, image syntax in MW is crazy
[05:46:13] 	 the bug is mostly annoying because we can't edit inline images right now
[05:46:20] 	 For whatever reason, it seems to be consistently agreed that frameless without an alignment means inline
[05:46:26] 	 Which I suppose is not entirely insane
[05:46:35] 	 so you get an image you can't edit. But that will change in a very short while, so I guess it's not quite a bug :\
[05:46:36] 	 If you omitted |frameless entirely, you'd get the same behavior
[05:46:50] 	 [[File:Foo.jpg]] is inline, [[File:Foo.jpg|right]] is presumably block
[05:46:55] 	 Not sure about |none though
[05:47:05] 	 how is it not insane? frameless without alignment should *either* be frameless|default or frameless|none
[05:47:14] 	 both of these render as block images.
[05:47:14] 	 Right
[05:47:21] 	 Yeah the fact that |none behaves differently is weird
[05:47:55] 	 Well, I guess it's a MW-wide thing :\ it's just a really counterintuitive wikitext result.
[05:48:19] 	 when we can edit inline images, the perceived bug that rummana posted will essentially go away, I guess.
[05:48:36] 	 Righ
[05:50:21] * mooeypoo  updated the bug report
[05:50:30] 	 images are weird.
[06:06:38] 	 btw RoanKattouw the plot thickens with the position fixup. Apparently, "default" is an image type that depends on wiki direction and does not appear in wikitext. That if statement is probably one of the more convoluded ones I produced.
[06:06:57] 	 haha
[06:06:58] 	 image alignment, not type
[06:07:07] 	 Right
[06:07:30] 	 at first I added "Wiki's default" to the buttons but James didn't like that
[06:07:51] 	 so now I'm translating the default to the actual direction but ONLY if you chose another direction so it won't dirty diff :D
[06:12:12] 	 (03PS7) 10Mooeypoo: Revamp media edit dialog's position widget [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112041 
[08:45:40] 	 hi
[10:08:33] 	 (03PS2) 10Krinkle: core: Add a 'super' property to inheriting classes [oojs/core] - 10https://gerrit.wikimedia.org/r/112148 
[10:08:36] 	 (03CR) 10Krinkle: core: Add a 'super' property to inheriting classes (031 comment) [oojs/core] - 10https://gerrit.wikimedia.org/r/112148 (owner: 10Krinkle)
[10:16:10] 	 (03CR) 10Zfilipin: [Browser test] WIP Headless browser test(s) (032 comments) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/111890 (owner: 10Jhall)
[10:25:03] 	 (03CR) 10Zfilipin: [Browser test] WIP Headless browser test(s) (031 comment) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/111890 (owner: 10Jhall)
[10:28:57] 	 (03PS4) 10Krinkle: Fix more code that misreads API data with $wgDebugToolbar enabled [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112407 (owner: 10Alex Monk)
[10:29:40] 	 (03CR) 10Krinkle: Fix more code that misreads API data with $wgDebugToolbar enabled (031 comment) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112407 (owner: 10Alex Monk)
[10:38:54] 	 (03PS1) 10Hashar: [Browser test] @clean the Basic Edit feature [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112659 
[10:42:42] 	 (03CR) 10Hashar: "Triggered at http://integration.wikimedia.org/ci/job/visualeditor-doitall/61/console" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112659 (owner: 10Hashar)
[10:47:30] 	 (03CR) 10Krinkle: Make default OOJS UI styles optional (031 comment) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/109940 (owner: 10Jdlrobson)
[14:17:39] 	 (03PS1) 10Zfilipin: [browser test] Moved Ruby code that multiple repositories use to mediawiki_selenium Ruby gem [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112679 
[16:15:54] 	 (03CR) 10Jhall: [C: 032] [browser test] Moved Ruby code that multiple repositories use to mediawiki_selenium Ruby gem [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112679 (owner: 10Zfilipin)
[16:17:08] 	 (03Merged) 10jenkins-bot: [browser test] Moved Ruby code that multiple repositories use to mediawiki_selenium Ruby gem [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112679 (owner: 10Zfilipin)
[16:47:28] 	 (03PS1) 10Zfilipin: [browser test] Fix failing tests, Watir and page-object API changed [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112701 
[16:48:38] 	 (03CR) 10Jhall: [C: 032] [browser test] Fix failing tests, Watir and page-object API changed [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112701 (owner: 10Zfilipin)
[16:49:50] 	 (03Merged) 10jenkins-bot: [browser test] Fix failing tests, Watir and page-object API changed [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112701 (owner: 10Zfilipin)
[18:06:43] 	 (03CR) 10Jhall: [C: 032] [Browser test] @clean the Basic Edit feature [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112659 (owner: 10Hashar)
[18:22:17] 	 TrevorParscal, so what's wrong with resizeable things?
[18:27:10] 	 TrevorParscal: Could you look at https://bugzilla.wikimedia.org/show_bug.cgi?id=61224 quickly – I think Timo was fiddling with this recently and might have broken it.
[18:30:36] 	 hi kaldari
[18:31:20] 	 oh! I forgot to mention, TrevorParscal, I think I found a bug (or inconsistency?) in OO.ui.ButtonSelectWidget
[18:33:00] 	 TrevorParscal, visually, it seems the SelectWidget allows for only 1 selection at a time, but if you trigger a selection through .selectItem() it doesn't clear the previous selection, unlike "initializeSelection( item )" that clears the selection and selects that item. If this is intended (as multi-select?) then is there an inconsistency in the way the visual styles work? 
[18:33:53] 	 I had a bug in the media edit dialog with this, and ended up initializing selection each time the editor loads, because if I just kept selecting the relevant value, I ended up with an error about an endless recursion for "getItem()"
[18:40:44] 	 (03PS11) 10Jhall: [Browser test] Headless browser test(s) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/111890 
[18:40:54] 	 (03CR) 10jenkins-bot: [V: 04-1] [Browser test] Headless browser test(s) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/111890 (owner: 10Jhall)
[18:44:07] 	 (03PS12) 10Jhall: [Browser test] Headless browser test(s) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/111890 
[18:44:15] 	 (03CR) 10jenkins-bot: [V: 04-1] [Browser test] Headless browser test(s) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/111890 (owner: 10Jhall)
[18:51:25] 	 (03PS13) 10Jhall: [Browser test] Headless browser test(s) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/111890 
[18:54:52] 	 (03PS14) 10Jhall: [Browser test] Headless browser test(s) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/111890 
[19:03:32] 	 James_F: Lookin
[19:03:33] 	 Rehi TrevorParscal.
[19:03:46] 	 TrevorParscal: Thanks. See also mooeypoo's comments ^^^ re SelectWidget.
[19:04:24] 	 hm, ok, the problem with media search widget seems to be the API.
[19:04:34] 	 well, first - edsanders, maybe I could answer your question about Resiable
[19:05:21] 	 So, what I am trying to do is unify all the different image data things we have scattered about, and make a single MWMediaModel
[19:05:24] 	 it won't return a result (at least not from local) for partial match. So, for image "Westminstpalace.jpg" it wont' return results for anything other than "Westminstpalace" :\
[19:06:35] 	 I'm seeing that parts of the image data are in the lin mod as attributes, parts are accessed through the API via a dm node, parts are processed in a CE node
[19:07:38] 	 what I need to solve my particular problem (hopefully without just copy/pasting a bunch of crap) is to be able to build an MWMediaModel which all the controls in the MWMediaEditDialog could listen to and control
[19:07:40] 	 well - sort of
[19:07:58] 	 that way we can have a mini MVC going on, like we do with MWTransclusionDialog
[19:08:51] 	 to be fair, with Transclusions, we don't use MWTransclusionModel anywhere else (yet) but theoretically we could if needed, because it already supports all the features of templates
[19:09:24] 	 Oh, TrevorParscal
[19:09:28] 	 same could go for media, we could have a unified media model with a common API for getting/changing info, and getting that info back out
[19:09:33] 	 rdwrer: ?
[19:09:34] 	 I was going to ask you about the best thing to use for my interface
[19:09:44] 	 https://lh4.googleusercontent.com/q2uafYtSIdk_9Tg-RLDnOvAqUOHtd1_mLlh8TJk0rMiskCdnctpD4VBjOUWCWgJSISucamZD37mo6LFbYDZhvxEST8b92d8ZORn4Us1oRB3WTz-jWlpTRDhUZA
[19:09:47] 	 edsanders: your thoughts?
[19:10:08] 	 I'm using a few different widgets already, but I'm not sure what's best to use for the actual pop-up with the embed/share business inside it
[19:10:24] 	 Frame and Layout look likely, but I'm not sure
[19:10:26] 	 well within the resizeable part currentDimensions is the only part which is image data
[19:10:29] 	 And there are no demos.
[19:11:10] 	 rdwrer: if you are using OOUI, use OO.ui.PopupButtonWidget
[19:11:22] 	 it will let you make a button, that then has a popup coming out of it
[19:11:22] 	 TrevorParscal: rdwrer means which layout.
[19:11:38] 	 Ooh.
[19:11:43] 	 Or maybe he doesn't?
[19:11:46] * James_F  shuts up. :_)
[19:12:02] 	 Where's *that* in the widgets demo
[19:12:32] 	 rdwrer: I think it's possible that PopupableElement doesn't yet support changing the gravity, only the alignment
[19:12:43] 	 something that needs to be done to support your case of popping up, above
[19:12:56] 	 (03CR) 10Hashar: "I have triggered PS14 on the experimental Parsoid/VE job but it fails :( http://integration.wikimedia.org/ci/job/visualeditor-doitall/62/" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/111890 (owner: 10Jhall)
[19:13:14] 	 Ah. Hm.
[19:13:24] 	 edsanders: yeah, but maxDimensions and originalDimensions are valuable pieces of data too, which would be nice if they were abstracted away
[19:13:47] 	 TrevorParscal: So maybe I should just do my own thing (which I already am) and use a Layout to contain the interface for now?
[19:13:49] 	 in any case, I realize the dm/ce approach is probably fine and doesn't need to use this new model
[19:13:50] 	 (03PS2) 10Hashar: [Browser test] @clean the Basic Edit feature [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112659 
[19:13:57] 	 I can manually show/hide it and style it easy enough
[19:14:39] 	 Hi guys, I left a wall of text at https://bugzilla.wikimedia.org/show_bug.cgi?id=52004#c3 (on how to go about showing a block notice to blocked users)
[19:14:51] 	 edsanders: but i wanted to hear your thoughts about this, because I felt like there was cleverness going on, but not in the direction I imagined - and you are a really smart person, so I'm not quick to assume you are wrong
[19:16:12] 	 rdwrer: or just add a config option to OO.ui.PopupWidget called gravity with sets a CSS class and affects positioning (I might be able to do that for you instead)
[19:16:19] 	 sure, I'd like to hear which specific bits of logic you want to move around and how you see it working, because I considered a number of alternative approaches
[19:16:41] 	 I'd appreciate the help...
[19:17:01] 	 I guess I'm not sure what the need is, though?
[19:17:22] 	 my main objective is to, at least within the MediaEdit dialog, have a single standalone model that can be modified and listened to, so that we can use MVC within the dialog
[19:20:30] 	 So resizeable and scalable need to be separate from mediaNode definitely
[19:20:53] 	 rdwrer: for now, just use your own thing, I have too many other things that are more urgent, but we should add support for placing a popup on the left/right/top of something, not just below
[19:21:34] 	 I also felt like there was something odd about MediaSizeWidget being so model like
[19:21:35] 	 James_F, eh. https://bugzilla.wikimedia.org/show_bug.cgi?id=61151
[19:21:51] 	 I get that you used a mixin to provide a common API, and that felt clever
[19:22:10] 	 but part of me felt like it was wrong to treat a widget like a model
[19:22:28] 	 but, then again, widgets often retain their own data for interaction purposes
[19:22:43] 	 so, it's not a hard-and-fast rule that they should be dataless, surely that have state
[19:22:57] 	 it just sort of made my idea crumble to the ground
[19:23:10] 	 my idea being, using pure MVC in the dialog
[19:23:10] 	 mooeypoo: Fun. :-(
[19:23:35] 	 I'm trying to figure out if this is "normal". Like, if commons behaves this way too
[19:23:59] 	 edsanders: maybe if there was a way to move the media info gathering bits out of the dm and ce nodes
[19:24:11] 	 I could build my model, and it could use those same utilities
[19:24:12] 	 'kay, I'm already on that path, so no problem
[19:24:22] 	 I can tear it out and replace it later, I guess :)
[19:24:35] 	 rdwrer: Or upstream it. :-)
[19:25:09] 	 but while they are stuck in there, I have to make my model pull from (at least initially) the nodes, rather than accept the attribute data from the lin-mod as a read-only starting point (like the Transclusion stuff does)
[19:25:34] 	 James_F: Well, "yes", but easier to build from the existing widget and then replace what I've hacked together, IMO
[19:25:52] 	 rdwrer: do what you gotta do
[19:26:26] 	 it's not a simple change to PopupWidget, we have to rethink the align option and consider how far we want to go with alignment features
[19:26:49] 	 TrevorParscal: 3D!
[19:26:51] 	 ;-)
[19:27:04] 	 because there's more than just N, NE, E, SE, S, SW, W, NW
[19:27:05] 	 TrevorParscal, I'm not really familiar with how the TransclusionModel stuff works
[19:27:09] 	 and even that is a ton of options
[19:27:58] 	 edsanders: it's pretty simple, it takes the parsoid attribute from the linmod and builds out a transclusion object with parts, template parts have parameters
[19:28:05] 	 and it abstracts the template data stuff
[19:28:12] 	 which pulls from the API to get more stuff
[19:28:24] 	 meta info
[19:28:37] 	 so, it's quite similar to the media model idea
[19:29:10] 	 it would start with a plain object (attributes of am image node, for instance) and then grab a couple things from the API
[19:30:33] 	 then it would have all the business logic about media
[19:30:55] 	 methods of changing media, events are emitted when things change
[19:33:46] 	 edsanders: this is like the 2nd to most boring convo you've ever had, huh?
[19:33:48] 	 lol :)
[19:34:21] 	 ha, what's the first?
[19:34:35] 	 edsanders: Any conversation with me? ;-)
[19:35:08] 	 so we need to have the api requests somewhere where both the ce.resizeable and mediaSizeWidget can trigger them
[19:35:12] 	 and access the results
[19:36:26] 	 I think the reason we have logic in ce is because we don't have dm.resizeable
[19:37:04] 	 James_F, it seems to be the same thing in Commons, but it would be best to consult with someone who may know more about how search works? Do you know who I can talk to?
[19:37:22] 	 the alternative to a mixin was a Scalable object as a property, that gets passed to the mediaWidget when you initialise it
[19:37:37] 	 mooeypoo: I've asked Deskana about it.
[19:37:46] 	 James_F, oh, awesome
[19:37:48] 	 thanks!
[19:37:49] 	 because at the moment we have to copy the scalable properties from the node to the widget
[19:38:20] 	 edsanders: yeah, that's starting to get to the problem, any time I see a lot of properties being copied by specific names I get worried
[19:38:35] 	 ve.Scalable#setPropertiesFromScalable is the nasty bit
[19:39:05] 	 of course if you have scalable as a property we have to use events/listeners in place of methods
[19:41:04] 	 TrevorParscal / edsanders  I may have missed something, but why shouldn't Scallable be a mixing? Won't that solve everything, the quetion of 'original' (rather than current) size is only relevant to resizable stuff anyways, no? Isn't it most logical to have it in a scalable and use it as mixin when relevant?
[19:42:42] 	 I thought the entire point of mixins was to be able to have shared functionality between "types" of similar objects... seems to me to be the exact point in this case, no? What am I missing?
[19:43:11] 	 as long as we have 2 nodes and 1 widget all containing some or all of the information about scalability/resizability we are probably doing something strange
[19:43:51] 	 i think the ce/dm have need of certain and limited things
[19:43:59] 	 The widget shouldn't really contain stuff, it should present it, and the nodes can have no functionality related to size and get all that functionality from the mixin.
[19:44:04] * mooeypoo  might be missing something
[19:44:27] 	 and the model I want to create is a superset of that, so as ed suggested, perhaps some things should be pulled out of dm/ce
[19:44:28] 	 mooeypoo, the widget at the resizeabl
[19:44:31] 	 e
[19:44:40] 	 are more linked than that even
[19:44:43] 	 I'm thinking maybe MWMediaModel could have some static methods that ce/dm use
[19:44:45] 	 they point to the same data
[19:45:13] 	 it could have a static API cache, a static method to get dimension and type info from the API
[19:45:14] 	 edsanders, but you have one place that actually messes/changes/updates the data, no?
[19:45:25] 	 mooeypoo, every time we create the mediawidget we have to copy over the scalable properties from the node, which feels a bit hacky
[19:45:38] 	 then they could all share that, and it would live in a logical central place
[19:46:05] 	 hm.
[19:46:19] 	 i'm probably over-simplifying a bit
[19:46:42] 	 but, I will just stumble when I stumble, I think this is the general direction I'm going
[19:46:50] 	 What about multiple invocations of the same image (e.g. flag icons' repeated use)? Or is that a complexity too far in the name of efficiency?
[19:49:23] 	 James_F: we would still cache API results...
[19:51:24] 	 TrevorParscal: I was thinking about reducing objects as much as possible, but yes.
[19:53:29] 	 it's important to identify that sometimes copying is good - you don't want the widget to be changing values in the linear model's attributes object
[19:53:42] 	 but at some point you want to copy back, in the form of a transction
[19:53:56] 	 I think right now though, we are copying more than needed
[19:54:15] 	 but that's really not the issue, the issue is how scattered the media code is
[19:57:40] 	 (03PS2) 10Trevor Parscal: Create new "settings" icon, move "settings" icon to "advanced" [oojs/ui] - 10https://gerrit.wikimedia.org/r/112603 
[19:58:26] 	 (03CR) 10Jforrester: [C: 032] Create new "settings" icon, move "settings" icon to "advanced" [oojs/ui] - 10https://gerrit.wikimedia.org/r/112603 (owner: 10Trevor Parscal)
[19:58:50] 	 (03Merged) 10jenkins-bot: Create new "settings" icon, move "settings" icon to "advanced" [oojs/ui] - 10https://gerrit.wikimedia.org/r/112603 (owner: 10Trevor Parscal)
[19:59:02] 	 TrevorParscal, I'm in favour of you moving media widget code around, but I do think we need to keep a separate abstraction for resizeable things, where we can unify the logic for min/maxDimensions and aspect ratio lock.
[20:01:59] 	 (03PS1) 10Trevor Parscal: Add icons demo [oojs/ui] - 10https://gerrit.wikimedia.org/r/112789 
[20:02:39] 	 (03CR) 10Jforrester: Add icons demo (031 comment) [oojs/ui] - 10https://gerrit.wikimedia.org/r/112789 (owner: 10Trevor Parscal)
[20:02:54] 	 i'm fine with keeping Resizable, I will hit you up if I start thinking about changing it in a significant way, but as I said before, I think it's not the problem really
[20:03:08] 	 (03PS1) 10Trevor Parscal: Moved PNG icons to their own variant [oojs/ui] - 10https://gerrit.wikimedia.org/r/112790 
[20:04:31] 	 (03PS2) 10Jforrester: Moved PNG icons to their own variant [oojs/ui] - 10https://gerrit.wikimedia.org/r/112790 (owner: 10Trevor Parscal)
[20:08:21] 	 Hmm... Wondering what the best way to mark a category as hidden for https://bugzilla.wikimedia.org/show_bug.cgi?id=50763 is...
[20:09:19] 	 Krenair: Yeah, not very helpful request from me. :-) Maybe hidden?
[20:09:23] 	 Err.
[20:09:34] 	 That was meant to be:
[20:09:45] 	 "Maybe show hidden categories with an icon of some sort next to them?"
[20:09:55] 	 Perhaps some slight but noticeable transparency and hovertext 'Hidden category'? Or can icon of some sort?
[20:10:01] 	 an*
[20:10:08] * James_F  nodes.
[20:10:12] 	 :p
[20:10:17] * James_F  sighs.
[20:10:41] 	 Also, inherited categories from templates on the page aren't in this list (because they aren't set on the page), but do appear in the list of categories in read mode (obviously). Maybe a read-only list underneath it?
[20:12:36] 	 (03PS2) 10Trevor Parscal: Update OOjs UI to v0.1.0-pre (1fcaf6a9a0) [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/111675 
[20:12:56] 	 Yep, sounds useful.
[20:13:08] 	 Do you think we already have some 'hidden category' type icon available? Or perhaps a designer at WMF can come up with something?
[20:13:21] 	 (03CR) 10jenkins-bot: [V: 04-1] Update OOjs UI to v0.1.0-pre (1fcaf6a9a0) [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/111675 (owner: 10Trevor Parscal)
[20:13:54] 	 Krenair: Don't think we currently have one; happy to ask Design if they have any ideas (or Trevor, in his Copious Free Time™ :-)).
[20:14:02] 	 Krenair: Maybe put on the back-burner, I guess. :-(
[20:15:07] 	 James_F, I wonder if we really need to show the category as hidden *on the list*... A small popup appears (with the delete icon etc.) - perhaps we could just make a note there?
[20:15:31] 	 (03PS2) 10Trevor Parscal: (Stand back again) Even moar label manias! (label refactoring) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/111682 
[20:15:32] 	 Krenair: Possibly. Maybe italicise them?
[20:15:39] 	 (03CR) 10jenkins-bot: [V: 04-1] (Stand back again) Even moar label manias! (label refactoring) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/111682 (owner: 10Trevor Parscal)
[20:15:45] 	 (03PS3) 10Trevor Parscal: Add tooltips to the drop-down menus explaining what they do [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/108855 (owner: 10Jforrester)
[20:15:54] 	 (03CR) 10jenkins-bot: [V: 04-1] Add tooltips to the drop-down menus explaining what they do [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/108855 (owner: 10Jforrester)
[20:16:30] 	 TrevorParscal: Thanks. :-)
[20:16:38] 	 James_F: https://gerrit.wikimedia.org/r/#/c/108855/ has been rebased and fixed up - jenkins seems pissed but it works great
[20:16:52] 	 James_F, sounds good, I'll try adding that
[20:16:55] 	 (03PS15) 10Jhall: [Browser test] Headless browser test(s) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/111890 
[20:17:10] 	 TrevorParscal: Cross-repo builds don't pass Jenkins if it's not merged.
[20:17:25] 	 yeah, it's on crack
[20:17:32] 	 TrevorParscal: If https://gerrit.wikimedia.org/r/#/c/111672/ is OK to merge (it looks sane to me) we can proceed…
[20:18:06] 	 MERGE IT!
[20:18:18] 	 Have someone look it over quickly, but it's legit man
[20:18:26] 	 "Someone"?
[20:18:31] 	 almost buy the team lunch if you find a problem good
[20:18:33] 	 almost
[20:18:41] 	 The only devs working today are you, edsanders and Moriel.
[20:18:43] 	 I don't make bets like that, but if I did, I would
[20:18:52] 	 OK, you owe me lunch then. :-)
[20:18:53] 	 can't Ed take a peek? I don't know...
[20:19:24] 	 (03CR) 10Jhall: "Good catch Hashar! We had too many patch sets all merging around the same time. Should be fixed now." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/111890 (owner: 10Jhall)
[20:19:28] 	 TrevorParscal: Maybe if the commit title didn't read like it was a WIP? :-)
[20:20:22] 	 TrevorParscal: Did you work out what was wrong with https://bugzilla.wikimedia.org/show_bug.cgi?id=61224 ?
[20:20:37] 	 not yet
[20:20:42] 	 gotta eat some cow
[20:20:50] * James_F  tsks.
[20:21:28] 	 (03PS3) 10Jforrester: Refactor labels and titles to be consistent [oojs/ui] - 10https://gerrit.wikimedia.org/r/111672 (owner: 10Trevor Parscal)
[20:28:46] 	 greetings MW family
[20:29:17] 	 James_F, what am I reviewing?
[20:29:30] 	 edsanders: https://gerrit.wikimedia.org/r/#/c/111672/
[20:29:35] 	 rkitect: Heya.
[20:30:38] 	 I have recently installed VE on my 1.22.2 MW setup, but when I click edit, nothing happens. I've tested Parsoid succesfully. No errors are being thrown. I was hoping someone could point me in the direction of some "common causes" 
[20:31:17] 	 the URL updates to show "action=edit" at the end, but no editing is available
[20:31:54] 	 rkitect: ?veaction=edit or ?action=edit 
[20:32:04] 	 ?veaction=edit sorry
[20:33:03] 	 ok, I actually just found a new symptom. If I click "edit" from the read tab, nothing happens, but if I click "edit" from the "edit source" tab, I get ?veaction=edit
[20:33:38] 	 rkitect: Hmm. That's odd.
[20:33:45] 	 rkitect: Do you have any JS console errors?
[20:33:55] 	 url: http://http://gibsonhh.info:8081/KSQWikiCore/
[20:34:07] 	 well check, didn't think aboutt hose errors
[20:34:56] 	 rkitect: You're getting "Uncaught Error: Unknown dependency: jquery.collapsibleTabs" on every page, which is probably breaking VE.
[20:35:06] 	 oh nice
[20:35:16] 	 I'll check into that
[20:42:50] 	 James_F, friendly reminder re forms :p
[20:43:00] 	 should VE include an edit link for all skins or only vector?
[20:43:00] 	 mooeypoo: Argh, yes.
[20:43:37] 	 Thanks! :)
[20:46:24] 	 rkitect: Vector, Monobook and Apex; if you've got a third-party skin you'll need to add support for that skin.
[21:21:40] 	 (03PS1) 10Jhall: [Browser test] Update list of prod wikis to be verified. [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112809 
[21:57:40] 	 ryasmeen, are you around? 
[21:57:52] 	 I'm going over the bugs you posted, I can't replicate most of them so far in master 
[21:57:59] 	 hey mooeypoo, yup i am
[21:58:05] 	 :S
[21:58:07] 	 okay
[21:58:12] 	 let me recheck again
[21:58:18] 	 weird, right? 
[21:58:20] 	 i see your comment in one of them
[21:58:29] 	 (03PS1) 10Trevor Parscal: Fix margins after they were broken in 9d988924 [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112814 
[21:59:11] 	 just posted another
[21:59:18] 	 (03PS2) 10Trevor Parscal: Fix margins after they were broken in 9d988924 [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112814 
[21:59:39] 	 ryasmeen, there's also a pending commit that fixes a couple of known bugs -- https://gerrit.wikimedia.org/r/#/c/112041/
[21:59:58] 	 okay, when is it going to be merged?
[22:00:12] 	 ryasmeen, most notable, the things that are related to "none". But I still can't replicate the two bugs I posted comments on. I'm not sure if I may be missing something?
[22:00:32] 	 oh, uh, whenever RoanKattouw_away and James_F have time to review it, hopefully :)
[22:00:43] 	 okay :)
[22:01:23] 	 James_F: fixed that bug, btw
[22:10:40] 	 ryasmeen, another one I couldn't replicate https://bugzilla.wikimedia.org/show_bug.cgi?id=61059
[22:11:33] 	 oh... wait.. I.. think I see what you mean here in that bug, but I am not sure, and if it's what I think it is, it's probably intended behavior. Did you start with 600x800 and then delete '6' or *replace* 6 ?
[22:13:41] 	 ryasmeen, if you deleted the 6, it made the original width value 0, which adjusted the height to 0 too. But then you can simply write 500 as you meant to which adjusts both values. The values are updated each time you insert/change/adjust the input on either the width or height. Does this explain the bug? I am not sure I understood it correctly.
[22:19:15] 	 (03CR) 10Cmcmahon: [C: 032] [Browser test] Update list of prod wikis to be verified. [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112809 (owner: 10Jhall)
[22:19:47] 	 yes that is what the bug is about
[22:20:39] 	 (03Merged) 10jenkins-bot: [Browser test] Update list of prod wikis to be verified. [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112809 (owner: 10Jhall)
[22:21:06] 	 okay so I was thinking since I am changing only one dimension of the image , so it should not change the other one .But since it adjusts itself depending on the new value , I guess thats alright
[22:24:11] 	 also mooeypoo: I just reproduced https://bugzilla.wikimedia.org/show_bug.cgi?id=61152 on  betalabs
[22:24:22] 	 (03PS3) 10Hashar: [Browser test] @clean the Basic Edit feature [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112659 
[22:24:30] 	 ryasmeen, it is adjusting itself according to aspect ratio 
[22:24:33] 	 James_F: thanks for your help. I've fixed the missing dependency error, only getting a deprecation warning from vector now. Still not seeing any action from clicking "edit" from the read tab
[22:24:44] 	 ... what the...
[22:24:54] 	 ryasmeen, beta labs is master?
[22:25:06] 	 (03CR) 10Hashar: "The 'Review your changes' button not being find can be related to https://gerrit.wikimedia.org/r/#/c/112701/1 . So I rebased the patch onc" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112659 (owner: 10Hashar)
[22:26:22] 	 mooeypoo, its showing version 052b039
[22:26:38] * mooeypoo  pokes James_F 
[22:28:00] 	 you around? Can it be that a bug is in beta labs but not on master, considering last merged change related to media edit dialog was merged on 2/7 ?
[22:28:02] 	 this is so weird.
[22:31:25] 	 (03CR) 10Hashar: "That PS being run on http://integration.wikimedia.org/ci/job/Visualeditor-doitall/63/" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112659 (owner: 10Hashar)
[22:32:49] 	 Silly question, but is oojs-ui looking towards being separate modules so we don't need to load...everything
[22:38:55] 	 (03CR) 10Hashar: "Rebasing fixed one scenario which was blocked due to the iframe:" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112659 (owner: 10Hashar)
[22:39:01] 	 rdwrer: Possibly; how large is it nowadays?
[22:39:45] 	 I think spagewmf was saying something on the order of tens of kb, maybe over 100, but I'm not sure
[22:41:24] 	 (03PS3) 10Jhall: [Browser test] WIP New test for VE Page Settings menu. [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/110659 
[22:41:33] 	 (03CR) 10jenkins-bot: [V: 04-1] [Browser test] WIP New test for VE Page Settings menu. [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/110659 (owner: 10Jhall)
[22:42:15] 	 rdwrer: Looking, the current (uncompressed) built OOjs UI JS is 170 KiB, and the CSS is another 41 KiB.
[22:42:18] 	 mooeypoo, can you try reproducing them on betalabs?
[22:43:04] 	 Sounds right
[22:43:07] 	 rdwrer: This compares to 870 KiB for jquery.ui.
[22:43:14] 	 Heh
[22:43:22] 	 I think I stopped using jquery.ui too
[22:43:26] 	 I certainly meant to
[22:43:48] 	 rdwrer: But that's not an entirely fair comparison, as it does some things OOjs UI doesn't yet (as well as not doing a bunch of things OOjs does).
[22:43:51] 	 I use dialog, but I'm replacing it
[22:44:43] * James_F  nods.
[22:44:50] 	 Use OOjs UI's!
[22:46:18] 	 I mean, basically I am
[22:46:26] 	 I'm sort of replacing it with a redesign though
[22:46:45] 	 Also I may have been overzealous in using ButtonSelectWidget for my navigation...aw.
[22:50:18] 	 ryasmeen, I just tried one of the bugs in betalabs and can't replicate it there either (the 'none' alignment one, answered in the bug report)
[22:50:44] 	 James_F: how away is ed?
[22:50:48] 	 but ryasmeen issues with "none" are likely going to be fixed with the commit I mentioned there. There were a couple of undelying ones. Still, I don't know what I'm doing wrong but I can't get it to replicate O.o
[22:51:00] 	 ed ed and away
[22:51:09] 	 lol
[22:51:22] 	 TrevorParscal: "Idle: 2:21:58"
[22:51:34] 	 What was the original position of the image that you tried shifting?
[22:51:37] 	 probably very away
[22:51:39] 	 TrevorParscal: So I'd say "very". He's not been connected to IRC for that time.
[22:51:41] 	 TrevorParscal: IM him?
[22:51:50] 	 i'll try
[22:52:22] 	 mooeypoo, I tried on an image which was left aligned , so selecting none moves it to right
[22:52:30] 	 ryasmeen, on insert, the image was aligned to the right/
[22:52:53] 	 right is the default alignment of LTR wikis -- so you tried on an existing image?
[22:53:25] 	 ryasmeen, ok, I just picked a random image in that page that was align=left, changed it to 'none' and it's still left (just without floats)
[22:54:28] 	 ryasmeen, can I change your test page ? 
[22:55:17] 	 okay sure mooeypoo
[22:55:35] 	 Hm.
[22:55:54] 	 SelectWidget's code is magical enough that I can't tell if it's a