[00:39:14] (03PS1) 10Jhall: [Browser test] Repair broken "header" browser test [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/114114 [00:55:43] (03CR) 10Jforrester: [C: 032] "Let's give this a go." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/114114 (owner: 10Jhall) [00:56:54] (03Merged) 10jenkins-bot: [Browser test] Repair broken "header" browser test [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/114114 (owner: 10Jhall) [12:32:42] (03CR) 10Esanders: [WIP] Centralize image handling into a stand-alone model (031 comment) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/113163 (owner: 10Trevor Parscal) [13:04:27] (03PS3) 10Krinkle: Refactor dialog and window sets [oojs/ui] - 10https://gerrit.wikimedia.org/r/112132 [13:04:42] (03CR) 10Krinkle: "Fleshed out commit message, no longer WIP" [oojs/ui] - 10https://gerrit.wikimedia.org/r/112132 (owner: 10Krinkle) [13:06:04] (03CR) 10Krinkle: "(This is a trivial rebase of patch set 9)." [oojs/ui] - 10https://gerrit.wikimedia.org/r/109006 (owner: 10Krinkle) [13:06:07] (03PS11) 10Krinkle: [WIP] Add demo for dialogs and layouts [oojs/ui] - 10https://gerrit.wikimedia.org/r/109006 [13:06:16] (03PS3) 10Trevor Parscal: [WIP] Convert styles to LESS format [oojs/ui] - 10https://gerrit.wikimedia.org/r/114093 [13:06:35] (03CR) 10jenkins-bot: [V: 04-1] [WIP] Convert styles to LESS format [oojs/ui] - 10https://gerrit.wikimedia.org/r/114093 (owner: 10Trevor Parscal) [13:34:55] (03CR) 10Krinkle: [C: 04-1] [WIP] Convert styles to LESS format (033 comments) [oojs/ui] - 10https://gerrit.wikimedia.org/r/114093 (owner: 10Trevor Parscal) [16:41:09] hey hey [16:41:24] any solution to VisualEditor/PrivateWiki/Parsoid ? [16:41:29] http://www.mediawiki.org/wiki/Thread:Extension_talk:VisualEditor/Your_wiki_requires_a_logged-in_account_to_access_the_API._Parsoid_will_not_work_for_this_wiki! [16:41:38] having that issue right now [16:54:57] I think I should file a bug about VE making it really hard to test a bug if you don't click Edit first. [17:08:30] (03PS5) 10Esanders: Allow editing of attributes of MW extensions [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/99655 [17:09:26] Elitre, unless the bug is with the "Edit" button... [17:09:57] Elitre, I don't think I understand the issue? [17:13:13] the issue is I really should click on Edit first thing if I want to test VE; I assumed I had done that earlier, and couldn't understand why I couldn't place the cursor where I wanted to :) [17:21:32] Grr, getting grunt to compile less for oojs-ui isn't as easy as I thought. We have our weird thing where individual files fragment the build target [17:21:40] right now just raster/vector though.. [17:21:50] need to duplicate it twice for less, and then again for the intro.css + prepend concat [17:29:09] do you collect pages where there are saving errors because VE tries to make changes it shouldn't? [17:29:35] Here's one: https://www.mediawiki.org/wiki/Manual:Pywikibot/Installation [17:30:06] ragesoss: We run a dirtydiff bot that tries to make changes to 100s of random pages and verifies the diff is empty [17:30:08] http://tools.wmflabs.org/visualeditor/dirtydiffs/ [17:30:15] That one hasn't yielded errors for a while now though [17:30:46] I tried to just fix a typo, but it wouldn't save, and there were a bunch of little changes throughout the page. [17:31:00] changing "bash" to just bash without quotes, inside code tags. [17:31:10] and disrupting some other tags. [17:33:53] ragesoss: Eurgh, looking at it that's a bad set of Parsoid regressions. [17:33:56] subbu: ^^^ [17:34:47] subbu: https://www.mediawiki.org/wiki/Manual:Pywikibot/Installation?veaction=edit – make a small edit (e.g. add a character to the first heading), then preview the diff. [17:35:01] ok, looking. [17:35:10] subbu: Note the old DSR "" -> "translate>" issue towards the button [17:35:13] Err. Bottom. [17:37:33] that does look bad. [17:38:58] subbu: I know we sometimes had that issue before, but it got fixed (in November?) [17:39:51] i cannot tell offhand what is going on there ... will have to dig more. [17:41:12] ragesoss, is this the only page where you saw this? [17:41:36] subbu: yes. (I haven't gone looking for others.) [17:41:43] ok. [17:42:16] James_F, ragesoss will take a look in the afternoon after we have a checkin in a little while [17:42:26] subbu: Thanks. :-) [17:43:59] wondering if someone can diagnose the problem at https://fr.wikipedia.org/w/index.php?title=Brando_Eaton&curid=3990010&diff=101381845&oldid=95226942 [17:44:40] Elitre: Hmm. [17:44:59] * Elitre sees James scratching his head. [17:46:13] Elitre: If I make it with the link inspector with that external link, I get the slightly-wrong "[[:fr:Foo|Bar]]". [17:46:30] * Elitre too. [17:46:56] actually, I get https://fr.wikipedia.org/w/index.php?title=Utilisateur%3AElitre_%28WMF%29%2FProve2&diff=101404692&oldid=101404657 . [17:47:10] Same if I manually enter ":fr:Foo". [17:47:33] If I enter "Foo", I get "[[Foo|Bar]]" as expected too. [17:47:39] * Elitre tries with FF [17:50:43] (03PS1) 10Krinkle: [WIP] Implement less complilation and dist file creation [oojs/ui] - 10https://gerrit.wikimedia.org/r/114170 [17:50:52] oh, this is interesting: https://fr.wikipedia.org/w/index.php?title=Utilisateur%3AElitre_%28WMF%29%2FProve2&diff=101405500&oldid=101405469 now a word is missing? [17:51:13] (03CR) 10jenkins-bot: [V: 04-1] [WIP] Implement less complilation and dist file creation [oojs/ui] - 10https://gerrit.wikimedia.org/r/114170 (owner: 10Krinkle) [17:51:21] (03PS2) 10Krinkle: [WIP] build: Implement less compile and dist in Grunt [oojs/ui] - 10https://gerrit.wikimedia.org/r/114170 [17:51:57] Elitre: Firefox issue? [17:52:22] (03CR) 10jenkins-bot: [V: 04-1] [WIP] build: Implement less compile and dist in Grunt [oojs/ui] - 10https://gerrit.wikimedia.org/r/114170 (owner: 10Krinkle) [17:53:13] that was made with FF, yes. still different from what I wanted to get, but oh well. will file this in the meantime, under the scientific description "VisualEditor: now eating up my words". [17:54:13] :-( [18:01:48] :) actually James, it's not a problem with FF. Even under Chrome it will eat up the first word. The wikicode says Cabin is there, it doesn't show up in view mode though. [18:10:56] (03PS8) 10Alex Monk: VisualEditor: Mark hidden categories as different in the meta-data dialog so users understand they won't display [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/113464 [18:12:11] (03CR) 10jenkins-bot: [V: 04-1] VisualEditor: Mark hidden categories as different in the meta-data dialog so users understand they won't display [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/113464 (owner: 10Alex Monk) [18:13:17] '{boolean}' rather than '{bool}'? hmph [18:13:53] (03PS9) 10Alex Monk: VisualEditor: Mark hidden categories as different in the meta-data dialog so users understand they won't display [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/113464 [18:17:36] (now at https://bugzilla.wikimedia.org/show_bug.cgi?id=61550 .) [18:19:02] Krenair: Yep, {boolean} [18:19:24] php-doxygen seems to take int/bool but not in javascript. [18:32:38] Hm, that's strange. [18:32:44] I just got "Uncaught TypeError: Object # has no method 'updateLabel' " while trying to load VE [18:33:34] Must've been something I broke in my branch [18:36:26] Downloaded the same commit from gerrit and it works. :| what. [18:40:33] Nope there's different issues. Krinkle, got a moment? [18:41:31] Sure [18:41:48] `git checkout a498cb19f8c28cbaac86303832e44558e09a1238; git submodule update` -> this is what my patch is based on but stuff is really broken [18:43:01] Seems okay when I checkout master; submodule update;... Might just rebase [18:43:07] Krenair: master@{9 days ago}, okay, checked out a498cb19f8c28cbaac86303832e44558e09a1238 [18:43:45] Krenair: Getting the same error [18:43:48] Uncaught TypeError: Object # has no method 'updateLabel' [18:43:56] from ve.ui.Toolbar.onSurfaceAddCommand [18:44:01] from ve.init.mw.ViewPageTarget.setUpToolbar [18:44:02] Did Trevor's label refactor miss obe? [18:44:07] one, even. [18:44:56] * eca0a08 - (Stand back again) Even moar label manias! (label refactoring) + VE core update (7 days ago) [18:45:01] that commit is more recent [18:45:09] so that, if anything, fixed that [18:45:19] or maybe it broke it even worse, and then something else fixed it between then and master [18:45:24] either way it is fixed in latest master [18:45:26] Okay, just seems weird that I didn't encounter this error last week.. [18:45:31] I'll just rebase my patch. Thanks. [18:45:41] James_F: Is there a branch pointer based on master from before 9 days ago? [18:46:05] Krinkle: Yes – the wmf branches? [18:46:38] (03PS8) 10Alex Monk: Show preview of edit summary in review screen [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112490 [18:46:54] Krinkle: https://git.wikimedia.org/log/mediawiki%2Fextensions%2FVisualEditor.git/refs%2Fheads%2Fwmf%2F1.23wmf12 is the newest one older than 9 days. [18:47:00] (03CR) 10Alex Monk: "(just did a rebase)" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112490 (owner: 10Alex Monk) [18:47:33] James_F: Is 12 still deployed though? [18:47:37] 13/14 right now [18:47:44] Krinkle: No. [18:47:47] So the rebase didn't fix anything until I updated the submodules [18:48:11] Okay, so it's in the submodule updates. [18:48:30] Probably something Trevor's change in ve broke, and then fixed in commit immediately after [18:48:34] (instead of in the same commit) [18:48:45] (03PS1) 10Jhall: [Browser test] Alternate solution for "headings" browser test [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/114182 [18:49:02] Still seems weird that I didn't run into this issue last week... must've been working with an outdated lib/ve or something I guess. [18:49:48] Krinkle: ISTR there were a few follow-ups RoanKattouw_away did. [18:50:11] So the only danger is in a branch cut (13 or 14) that was cut before it was fixed [18:50:21] unlikely though as we'd have heard reports from prod by now [18:50:36] Yeah. [18:54:12] (03PS10) 10Jforrester: Mark hidden categories as different in the meta-data dialog [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/113464 (owner: 10Alex Monk) [18:54:52] fixing capital letters after colons on Bugzilla, so James is happy. [18:55:35] Elitre: :-) [18:56:08] (03PS6) 10Jforrester: Prevent clicks on top-most shield for centered image nodes [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/113258 (owner: 10Trevor Parscal) [18:56:34] (03CR) 10Jforrester: "PS6 is a rebase to remove out-dated lib/ve dependency that'd been superseded." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/113258 (owner: 10Trevor Parscal) [18:56:40] (03CR) 10Cmcmahon: [C: 032] [Browser test] Alternate solution for "headings" browser test [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/114182 (owner: 10Jhall) [18:57:10] (03PS2) 10Jforrester: Split VE-MW's ve.ce.Node.css file into per-node files [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/113487 [18:57:15] also, slightly edited https://bugzilla.wikimedia.org/show_bug.cgi?id=61335 . I don't think Erik really meant that, but can revert in case. :) [18:58:29] (03Merged) 10jenkins-bot: [Browser test] Alternate solution for "headings" browser test [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/114182 (owner: 10Jhall) [18:58:56] Elitre: :-) [18:59:56] doesn't the idea of infinite universes with infinite en.wikis give you a headache? [19:00:11] Germans, everywhere... [19:00:37] Elitre: We already have two… [19:01:17] I'm not a sysop to test that on simple, anyway :) [19:04:03] I also found https://bugzilla.wikimedia.org/show_bug.cgi?id=55951 , which is probably a duplicate. [19:04:52] RoanKattouw: https://gerrit.wikimedia.org/r/#/c/113487/ and https://gerrit.wikimedia.org/r/#/c/112618/ [19:05:05] RoanKattouw: (Will need manual rebase either way - pick whichever you want to go first.) [19:05:40] RoanKattouw: https://gerrit.wikimedia.org/r/#/c/110434/ has a +1 from Timo which you asked for before +2ing. [19:06:26] RoanKattouw: Your commit of https://gerrit.wikimedia.org/r/#/c/113473 LGTM but I'd rather Timo or someone who's used to "our style" decides to merge. [19:06:35] RoanKattouw: Hm.. something just truck me, don't accesskey- message allow blanking? [19:07:01] We may have to do this conditionally, especially the selector query where we embed that key could get tricky [19:07:07] RoanKattouw: Finally, https://gerrit.wikimedia.org/r/#/c/112490/ and https://gerrit.wikimedia.org/r/#/c/113464/ [19:07:14] Not sure if that's the case though [19:08:09] Yep, per Linker.php, accesskey can be empty or '-', which will disable it for that button and free the key for usage elsewhere [19:08:44] Krinkle: Hmm. We were told by LE to not allow per-language or per-wiki over-riding of short-cut keys. [19:08:59] Krinkle: I'm confused as to why they said that if accesskeys can be… [19:10:08] (03CR) 10Krinkle: "Not sure if it applies here, but I think we have to make this conditional on whether the message is not disabled." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/110434 (owner: 10Alex Monk) [19:10:22] (03CR) 10Krinkle: Use accesskey=s on the save buttons [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/110434 (owner: 10Alex Monk) [19:10:36] James_F: This has been in core for a long time, I don't know how that relates to LE [19:10:59] wikis do whatever wikis want to do, this isn't i18n related specifically, can be, but by no means has to be. [19:11:17] changing it to a different letter, yes, but that should only be done by translatewiki. [19:11:34] but disabling it is not i18n related, but a local wiki choice. including non-wmf. [19:11:53] Krenair: Did you sort out your VeUiHistoryTool breakage? You may have to upgrade/downgrade MediaWiki core because OOUI lives there now [19:11:54] (03PS9) 10Alex Monk: Show preview of edit summary in review screen [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112490 [19:12:13] RoanKattouw, yeah, just rebased the patch in the end [19:12:53] Oh that might be related to the issue I was having. I'm working on a newer version of core than I was last week. [19:14:00] Forresterian spelling: http://i.imgur.com/O7o7Ii0.jpg [19:14:25] haha [19:14:29] edsanders: :-) [19:15:28] It needs a few ligatures [19:16:15] (03CR) 10Catrope: [C: 04-1] "Surely the shared CSS should be loaded *before* the skin-specific CSS, not after?" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112618 (owner: 10Kaldari) [19:17:36] edsanders: I'm amazed. Is this is common for post game riot prevention or something? [19:17:52] rmoen: Selling of alcohol is locally regulated. [19:18:06] rmoen: So the local council can impose pretty random rules. [19:18:11] rmoen, they don't do it for every game, but more so recently [19:18:31] rmoen: Especially near foöt-båll grounds. [19:18:33] ;-) [19:18:46] we rarely have any sort of trouble here [19:19:05] The only US equivalent I can think of is Caltrain banning open containers after 9pm on game nights [19:19:07] Anyone see Terry? [19:19:12] also the pub on the other side of the road is open [19:19:24] Which is more remarkable because they allow open containers at all, they're usually banned in public places [19:19:25] rmoen: No, sorry. [19:19:36] RoanKattouw: Imagine if they stopped selling alcohol prior to a giants game? [19:19:48] Yeah that wouldn't fly here [19:19:51] People would lose their shit [19:20:27] Well Terry's 20 minutes late to the 1:1 think its save to call it ? [19:20:36] * safe * [19:20:47] Also you can buy beer in the stadium [19:20:59] and from the Tesco half a mile away [19:21:04] rmoen: I dunno. [19:21:13] edsanders: oh right, so they stop selling it around town but in the stadium its fine [19:21:56] well the beer companies that sponsor the league wouldn't be happy otherwise [19:22:30] also due to the hooliganism in the 80s, it's illegal to drink alcohol *in signt* of a football pitch [19:22:38] so you have to neck your beer before returning to your seat [19:28:53] eh. Does anyone have any idea what 'thumbwidth' in the imageinfo API reply *actually* means? Is that the thumbnail size that was uploaded to wherever we're getting the response from? There's no data about it in api.php and in the imageinfo reference in the docs ( https://www.mediawiki.org/wiki/API:Properties#imageinfo_.2F_ii ) [19:29:07] Hmm [19:29:08] anyone ever run into issues with it just hanging on api.php call? [19:29:20] i click on edit the blue bar on the left spins and spins and spins and the call never completes [19:29:23] mooeypoo: I think that might only apply if you set iiurlwidth / iiurlheight ? [19:29:31] I've been going in circles trying to figure out where to get the user's default preference and the wiki default preference regarding thumbnails. Without wgUserOptions it seems to be impossible :\ [19:29:33] tmkt: That happens sometimes, usually due to a JS error [19:29:40] tmkt: Are there any errors in your JavaScript console? [19:29:49] i took the URL out of firebug and just ran the url by itself [19:29:50] RoanKattouw, so, that' the image's *actual* dimensions, not default wiki ones. Right? [19:29:51] and it hangs [19:29:58] mooeypoo: Yeah thumb preferences are nasty. Also, surely they're Parsoid's responsibility? [19:30:00] the image in its thumbnail version. [19:30:12] RoanKattouw, not when you insert new images :\ [19:30:18] This url /api.php?format=json&action=visualeditor&paction=parse&page=SSL [19:30:22] mooeypoo: I think it's a property that only applies when you ask imageinfo for a thumbnail [19:30:30] Right, new images :S [19:30:33] i've even waited for it to timeout without any luck [19:30:42] tmkt: You're sure it's timing out on that URL? [19:30:45] just keeps saying in the status bar waiting for.. [19:30:55] doesn't time out..just sits there waiting for response from the server [19:30:59] Oh, hold on [19:31:03] I remember this bug [19:31:05] also, RoanKattouw, we need to figure out when to add mw-default-size and when to take it off. With it, parsoid removes the |120x120 (or whatever) bounding box for 'default', and only displays these without the class. So, I need to figure out when it's on and when it's off. [19:31:10] tmkt: Private wiki, right? [19:31:16] Connecting...spinning and spinning.. [19:31:16] yup [19:31:21] And without checking what the default dimensions are, it's turning out to be very tricky [19:31:37] set the cookie option on parsoid [19:31:45] tmkt: I know how to fix this, give me a minute to look it up [19:32:07] you are my best friend [19:32:22] tmkt: Try setting $wgSessionsInObjectCache = true; [19:32:27] mooeypoo, the good news is that CSS takes precedence over width & height attributes [19:32:29] (Obviously this should be in the docs) [19:32:30] RoanKattouw, we are currently building the new MWBlockImageNode with width/height that are set from imageinfo's thumbwidth/thumbheight. It seems that's not ideal at all... I thought maybe to cache those, and only take off "default" if they're changed, but the problem is that these in themselves are *not* the default values for the wiki. [19:32:37] RoanKattouw: in parsoid settings? [19:32:49] or wiki settings [19:32:49] tmkt: No, MediaWiki [19:33:00] ok [19:33:19] The problem is that with the default session handler, MW will deadlock if you have an MW request waiting for a Parsoid request waiting for another MW request [19:33:22] gwicke, yeah... two problems though: 1. Telling VE's CE what size the image should be when we just insert it, and 2. Adding or removing the 'mw-default-size' class so Parsoid can either make it default or not. [19:33:39] mooeypoo: thumbwidth/thumbheight should only be set if you asked for them AFAIK [19:33:47] mooeypoo, as I said, on way is to use CSS [19:34:06] for 1) [19:34:21] Hm. [19:34:28] thank you so much [19:34:33] 2) is indeed trickier, but I guess you can remove the default size class when resizing happens [19:34:52] ideally we'd specify most thumb sizes relative to the normal thumb size [19:35:08] gwicke, yeah, which means I need to cache the original size that was inserted. I'm trying to figure out if I'm caching the right thing now. [19:36:12] (03PS7) 10Alex Monk: Use accesskey=s on the save buttons [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/110434 [19:36:14] what a productive day [19:36:16] finally [19:39:36] James_F, so, to be absolutely clear on https://bugzilla.wikimedia.org/show_bug.cgi?id=51260 , we're going with ? [19:49:24] James_F, so for https://gerrit.wikimedia.org/r/#/c/32700/ we could create 'termsofusepage' (like 'privacypage') then hope each wikimedia wiki sets it properly. [20:00:13] Krenair: Eww. [20:00:23] Krenair: Isn't there a WikimediaMessages hook we can steal? [20:01:15] James_F, don't think so... [20:01:56] although my copy is out of date [20:02:28] Krenair: Hmm. [20:02:40] Latest version has FlowTermsOfUse, heh [20:03:29] James_F, did you mean to change it to point to the privacy policy instead? [20:03:43] RoanKattouw_away: https://gerrit.wikimedia.org/r/#/c/114100/ is sane? [20:03:45] I thought legal vetted the version saying Terms of Use... did that change or? [20:03:57] Krenair: Yes. Legal did, then the TOS concept was removed from MW. [20:04:09] Krenair: Or something. [20:04:23] Krenair: TOS is a WikimediaMessages thing now only, I think? [20:04:44] seems like it [20:06:14] James_F: ve.ui.MWTitleInputWidget [20:07:10] (03PS1) 10Jforrester: Make the #REDIRECT field an MWTitleInput rather than just plain [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/114245 [20:07:54] (03PS2) 10Catrope: Make the #REDIRECT field an MWTitleInputWidget rather than just plain [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/114245 (owner: 10Jforrester) [20:12:24] James_F, I'm having problems with this image size thing. I can work without sizes (default) which fixes the wikitext issue, but leaves 'size widget' empty. To fill it, even if I add a 'default size' button i need to *get* the default wiki size, which I just can't find in the API and the $wgUserOptions seems to be unavailable. [20:13:18] James_F, the best so far I could come up with is mostly for new images or for ones that had no size set in wikitext: They will get the 'default' tag which means no wikitext size will be added on save -- unless you change their size actively. That means, though, that the size widget for these images start out *empty* (and hence "invalid" and marked red) [20:13:37] not perfect, but without getting the default wiki (and user...) width/height values I'm not sure what else I can do. [20:15:38] James_F, also, while that fixes the wikitext problem (which means the images will be default-size in php-parser, it doesn't quite solve the CE problem. Right now, images without explicit size appear as their thumbnail-file dimensions. That's not the default, though, so it might be that in CE the image will be bigger than in the page view after save. [20:25:56] * Elitre wishing a nice day, and leaving [20:38:07] James_F, still there? [20:59:10] omg, what is up with the font on this page? http://www.zayo.com/abovenet Someone thought it was a good idea to add text-shadow.. [20:59:11] (03PS5) 10Jhall: [Browser test] WIP New test for VE Page Settings menu. [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/110659 [21:06:30] (03PS6) 10Esanders: Add MW-specific keyboard shortcut help. [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/113118 [21:17:31] (03PS7) 10Esanders: Add MW-specific keyboard shortcut help. [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/113118 [21:18:12] Krinkle, I've seen it in quite a lot of sites. Doesn't bother me much when it's that subtle, to be honest. [21:19:13] it's quite prominent on my display [21:19:52] Really ? interesting. It's barely visible on mine. I only noticed after I took it off to see [21:20:59] I thought it was a PNG instead of text, retina/hidpi display [21:21:03] http://i.imgur.com/Pc8LPxm.png [21:30:40] (03PS7) 10Catrope: Move some skin styling from ve.ce.Node.css to a shared skin css file [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112618 (owner: 10Kaldari) [21:31:01] (03CR) 10Catrope: [C: 032] Move some skin styling from ve.ce.Node.css to a shared skin css file [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112618 (owner: 10Kaldari) [21:31:55] RoanKattouw: Thanks. Sorry I hadn't been able to fix that myself, stuck in meetings all day [21:32:34] (03PS4) 10Catrope: Converting ViewPageTarget CSS files to LESS [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/114004 (owner: 10Kaldari) [21:32:39] kaldari: No worries [21:33:06] (03Merged) 10jenkins-bot: Move some skin styling from ve.ce.Node.css to a shared skin css file [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112618 (owner: 10Kaldari) [21:40:27] (03CR) 10Catrope: [C: 032] Use accesskey=s on the save buttons [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/110434 (owner: 10Alex Monk) [21:42:15] (03Merged) 10jenkins-bot: Use accesskey=s on the save buttons [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/110434 (owner: 10Alex Monk) [21:48:28] (03PS1) 10Catrope: Disable save button (and Alt+S) on conflict [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/114343 [21:49:11] (03CR) 10Catrope: "Alt+S-in-conflict-pane issue addressed in https://gerrit.wikimedia.org/r/#/c/114343/" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/110434 (owner: 10Alex Monk) [21:58:36] (03PS9) 10Mooeypoo: Revamp media edit dialog's position widget [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112041 [22:00:50] (03CR) 10Alex Monk: [C: 032] Disable save button (and Alt+S) on conflict [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/114343 (owner: 10Catrope) [22:02:02] (03Merged) 10jenkins-bot: Disable save button (and Alt+S) on conflict [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/114343 (owner: 10Catrope) [22:11:26] meh,no more "CheckboxWidget" ? [22:11:45] TrevorP|Away, whyyyy no more CheckboxWidget? It had a conveniently attached label... [22:15:30] mooeypoo: FieldLayout stuff replaces that [22:15:53] oh [22:16:50] (03PS10) 10Catrope: Show preview of edit summary in review screen [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112490 (owner: 10Alex Monk) [22:18:56] Sigh. Only in PHP do you have 'gzdecode' and 'gzuncompress' which do different things. (Same with 'gzencode' and 'gzcompress') [22:19:26] sigh [22:21:42] (03PS1) 10Jforrester: Revert "Set href attribute of newly inserted image to image title" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/114352 [22:21:52] (03CR) 10jenkins-bot: [V: 04-1] Revert "Set href attribute of newly inserted image to image title" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/114352 (owner: 10Jforrester) [22:24:45] (03PS2) 10Jforrester: Revert "Set href attribute of newly inserted image to image title" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/114352 [22:25:13] (03CR) 10Jforrester: "PS2 is a rebase." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/114352 (owner: 10Jforrester) [22:26:47] (03PS10) 10Mooeypoo: Revamp media edit dialog's position widget [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112041 [22:27:29] (03PS3) 10Jforrester: Revert "Set href attribute of newly inserted image to image title" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/114352 [22:28:54] James_F, okay so MWHttpRequest + Accept-Encoding: gzip [22:29:05] Depending on the engine used, this header should be sent already [22:29:18] If curl is used (which is preferred), nothing happens. Fixing this is a one line change. [22:29:19] Krenair: Hmm. [22:29:36] If the php one is used (fopen and friends), encoded content is returned. [22:30:25] Krenair: Awesome. :-) [22:30:26] Krenair: I think? [22:30:26] Right. [22:30:51] Well it seems to be fairly simple to fix... I hope no one else has written a backend for MWHttpRequest :) [22:32:06] James_F, when you have time, see the saga I wrote above about the thumbnail default size, etc. I am not sure there can be a good comprehensive solution without being able to get the default wiki- and user defined- size from the settings :\ [22:38:06] James_F, actually apparently the curl backend was sending identity, deflate, gzip... and my server was just using identity (AKA no encoding) [22:38:45] As far as I can tell anyway [22:40:39] haha nice [22:42:59] Oh great, servers can use multiple encodings at once... [23:00:32] (03PS1) 10Esanders: WIP failing converter tests [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/114355 [23:01:13] (03Abandoned) 10Esanders: WIP failing converter tests [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/94249 (owner: 10Esanders) [23:01:38] (03CR) 10jenkins-bot: [V: 04-1] WIP failing converter tests [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/114355 (owner: 10Esanders) [23:03:10] (03PS11) 10Catrope: Show preview of edit summary in review screen [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112490 (owner: 10Alex Monk) [23:03:13] (03PS1) 10Esanders: WIP some failing tests in fixUpInsertion [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/114357 [23:03:32] (03Abandoned) 10Esanders: WIP some failing tests in fixUpInsertion [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/94241 (owner: 10Esanders) [23:04:25] (03CR) 10jenkins-bot: [V: 04-1] WIP some failing tests in fixUpInsertion [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/114357 (owner: 10Esanders) [23:06:05] RoanKattouw, is there a way to force-feed Javascript a php-based $wgParameter ? [23:06:06] (03CR) 10Catrope: [C: 032] Show preview of edit summary in review screen [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112490 (owner: 10Alex Monk) [23:06:08] through the hooks or something? [23:06:14] Yes [23:06:18] Nothing else works... [23:06:27] oh, goodie. I can't find that in the docs, can you give me a direction? [23:06:40] mooeypoo: VisualEditorHooks.php function onMakeGlobalVariablesScript [23:06:42] However [23:06:48] Please be conservative with the amount of data you put in there [23:06:54] I suspect this is for the default image size stuff [23:06:57] yeah [23:07:03] I just want default size and user-default size [23:07:08] as in, wiki default and user default [23:07:12] Hmm, right ... [23:07:15] User default is tricky [23:07:20] (03Merged) 10jenkins-bot: Show preview of edit summary in review screen [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/112490 (owner: 10Alex Monk) [23:07:29] $wgUserOptions iirc [23:07:40] Sorry, onResourceLoaderGetConfigVars is where you should be [23:07:53] sorry, $wgDefaultUserOptions [23:07:59] ohhh ok. [23:08:13] There is already a defaultUserOptions section in there [23:08:17] what?? [23:08:17] But the user thing is harder [23:08:30] ok wait i'm looking at th file now [23:09:04] oh oh I see it [23:09:08] onResourceLoaderGetConfigVars has no page or user context [23:09:17] but it's partial. Okay, I can probably add in the image size [23:09:18] oh. [23:09:34] So for the user-default size, I think you'll just have to expose the array of image sizes and use mw.user.options.get( '...' ) to get the preference [23:09:46] wait, isn't wgDefaultUserOptions in general user related and not wiki-related ? [23:09:59] No, it's the default value of every user pref [23:10:07] I see, so that's the wiki default [23:10:12] As in, if the user doesn't express a preference, that's what the value will be [23:10:14] Yeah [23:10:19] * mooeypoo nods [23:10:26] ok I didn't know of mw.user.optios.get() [23:10:29] The user's actual preferences are overrides backfilled by the defaults [23:10:36] And mw.user.options.get() handles that for you [23:10:47] that might solve things. But just for the record: Documentation is crap. [23:11:14] Yeah :( [23:11:26] OK so now James_F is telling me we should deliberately ignore the user preference [23:11:34] Which would make things easier [23:11:36] oh [23:11:43] no user preference? just wiki preference? [23:11:50] Because rather than sending the array of sizes to the client plus an index in that array, we can just do that lookup server-side [23:11:52] wiki default, I mean [23:11:56] * mooeypoo nods [23:12:36] RoanKattouw: Hence my comment about adding a new MW API call. [23:12:41] The argument being that the wiki default is what Joe Average Reader is gonna see [23:12:46] James_F: Comment where now? [23:12:50] RoanKattouw, if I add to the 'defaultUserOptions' array in .hooks.php how do I get it from javascript? mw.config.get ? [23:13:01] RoanKattouw: ^^^^ [23:13:08] (Sorry for being gnomic.) [23:13:09] mooeypoo: mw.config.get( 'wgVisualEditorConfig' ).defaultUserOptions [23:13:15] awesomesauce [23:13:17] thanks! [23:13:26] mooeypoo: But look a the context to see if wgVisualEditorConfig is already been retrieved nearby [23:13:38] James_F: Please ignore what I said about cURL earlier. I think I've managed to confuse myself about how it's working... [23:13:43] * mooeypoo nods [23:13:44] (03PS1) 10Esanders: Fix SurfaceWidget documentation [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/114359 [23:13:52] Krenair: Always. :-) [23:13:53] ok that could possibly solve most problems [23:14:05] I wish i could bookmark IRC [23:14:28] .... :D [23:15:08] (03PS1) 10Esanders: Always pass a document model to the surface widget [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/114362 [23:15:12] (03PS2) 10Esanders: Fix SurfaceWidget documentation [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/114359 [23:15:16] mooeypoo: Anyway, in general, for information being exported from PHP to JS, that's generally in .hooks.php [23:15:22] If core doesn't already export it [23:15:57] (03PS3) 10Esanders: Fix SurfaceWidget documentation [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/114359 [23:15:59] (03PS4) 10Catrope: Fix ce#getOffset when cursor is to the left of an annotation. [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/114000 (owner: 10Esanders) [23:17:51] edsanders: Hmm, so this getOffset change is confusing me [23:17:58] That function is generally very confusing and poorly documented [23:18:02] yup [23:18:04] But here's one thing that gets me: [23:18:10] you changed the direction to be fully recursive [23:18:25] And you changed the getOffsetFromTextNode() call to take direction into account for the second parameter [23:18:40] But you did *not* do this for the recursive call, that one always gets domOffset=0 [23:18:41] right [23:19:02] Which ... I suppose is fine? [23:19:08] Because domOffset === 0 is a special case? [23:19:33] what's important is the 'node' you select [23:19:41] once you're inside that node, you can go to 0 [23:19:49] as long as you remember which direction you came from [23:20:00] Right, yeah I think I see that now [23:20:15] although perhaps there is a problem with FooBar ? [23:20:16] I'm filing a tech debt bug about burning this function down to the ground [23:20:49] or rather FooBarBaz [23:22:38] There shouldn't be two adjacent nodes, in general [23:23:51] (03CR) 10Catrope: [C: 04-1] Fix ce#getOffset when cursor is to the left of an annotation. (031 comment) [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/114000 (owner: 10Esanders) [23:24:00] Code looks fine to me sans a typo in the docs [23:24:19] But if you're concerned about some of those b/i/u cases, maybe put in some more test cases for them? [23:28:06] (03CR) 10Catrope: [C: 04-1] Allow getAnnotationsFromOffset to ignore close elements, (032 comments) [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/113538 (owner: 10Esanders) [23:28:38] (03CR) 10Catrope: [C: 032] Fix shadow position in standalone toolbar [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/113345 (owner: 10Esanders) [23:29:17] RoanKattouw, okay, [23:29:39]

FooBarBaz

doesn't work as expected.... [23:30:17] ,0 and ,0 give offsets of 7 instead of 1 [23:30:36] * edsanders removes adjacent 's [23:30:59] (03Merged) 10jenkins-bot: Fix shadow position in standalone toolbar [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/113345 (owner: 10Esanders) [23:32:57] Okay:

FooBarBaz

[23:33:13] ,0 gives 7 (exected 1) [23:33:36] ,0 also gives 7 (expected 4) [23:34:03] (03CR) 10Catrope: [C: 032] Fix layout of keyboard shortcuts page [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/109483 (owner: 10Esanders) [23:34:04] (03CR) 10jenkins-bot: [V: 04-1] Fix layout of keyboard shortcuts page [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/109483 (owner: 10Esanders) [23:36:05] (03PS4) 10Catrope: Fix layout of keyboard shortcuts page [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/109483 (owner: 10Esanders) [23:37:15] (03CR) 10Catrope: [C: 032] Fix layout of keyboard shortcuts page [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/109483 (owner: 10Esanders) [23:37:42] RoanKattouw, I think the problem is node = $domNode.contents().last()[0]; [23:37:51] edsanders: BTW mind answering my questions on https://gerrit.wikimedia.org/r/#/c/113538/1/modules/ve/dm/lineardata/ve.dm.ElementLinearData.js so I can merge it? [23:37:52] shouldn't that be first? [23:37:56] (Don't worry about the typo, I'll fix it) [23:38:00] Yeah that one did seem suspect [23:38:08] Does it break when you use first? [23:38:14] (03Merged) 10jenkins-bot: Fix layout of keyboard shortcuts page [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/109483 (owner: 10Esanders) [23:38:49] Especially since domOffset === 0, I would expect first there [23:39:16] But I have NFI what most of this code is doing, so I'm planning to rewrite it [23:39:21] 30 days and 1 hour from now :) [23:39:32] (03CR) 10Esanders: Allow getAnnotationsFromOffset to ignore close elements, (031 comment) [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/113538 (owner: 10Esanders) [23:40:12] RoanKattouw, given how few test cases there are, it 'working' isn't very assuring [23:41:11] True [23:41:17] so first() fixes those two cases ^^, but exposes the change of direction bug I was worried about. heading in the right direction though... [23:41:23] Right [23:41:47] (03CR) 10Krinkle: "I noticed these suspicious changes when generating a diff of the build (e.g. lib/oojs-ui.svg.css in latest master of VE compared to the di" [oojs/ui] - 10https://gerrit.wikimedia.org/r/114093 (owner: 10Trevor Parscal) [23:41:52] (03PS2) 10Catrope: Allow getAnnotationsFromOffset to ignore close elements, [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/113538 (owner: 10Esanders) [23:42:25] * RoanKattouw hugs Krinkle [23:42:34] Krinkle: Thank God there's an automated way to check the integrity of that LESSification change [23:42:39] That makes me really happy [23:42:41] specifically ,2 jumps into ,0 [23:43:03] then finishes at the end of the instead of the end of the [23:43:10] (03CR) 10Catrope: [C: 032] Allow getAnnotationsFromOffset to ignore close elements, [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/113538 (owner: 10Esanders) [23:43:14] (test case:

FooBarBaz

) [23:44:13] (03Merged) 10jenkins-bot: Allow getAnnotationsFromOffset to ignore close elements, [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/113538 (owner: 10Esanders) [23:45:06] Krinkle: https://gerrit.wikimedia.org/r/111620 [23:45:10] (03PS4) 10Krinkle: [WIP] Convert styles to LESS and implement grunt task [oojs/ui] - 10https://gerrit.wikimedia.org/r/114093 (owner: 10Trevor Parscal) [23:45:15] Finally getting 'round to starting on the sprint thing [23:45:16] RoanKattouw, how do I find the last offset in a node? [23:45:19] edsanders: Wait, ,2 is at the very end of the , right? [23:45:27] As the has 2 children [23:45:37] yes [23:45:40] (03CR) 10Krinkle: "Squashed I361cf95f1aa71 into this one since it has to be one atomic change in order to work and be testable." [oojs/ui] - 10https://gerrit.wikimedia.org/r/114093 (owner: 10Trevor Parscal) [23:45:41] (is it node.children.length?) [23:45:52] (03Abandoned) 10Krinkle: [WIP] build: Implement less compile and dist in Grunt [oojs/ui] - 10https://gerrit.wikimedia.org/r/114170 (owner: 10Krinkle) [23:45:58] edsanders: node.childNodes.length [23:45:59] (or childNodes?) [23:46:03] (node.children excludes text nodes) [23:46:04] ok [23:46:37] So presumably if you're at the end that's a special case and you need to descend into child i-1, domOffset last, or something? [23:46:43] I seem to recall there was code for this once [23:46:57] Maybe it was in another function but I think it might have been in a previous version of this function [23:49:57] (03CR) 10Catrope: [C: 032] Fix SurfaceWidget documentation [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/114359 (owner: 10Esanders) [23:50:04] (03CR) 10Catrope: [C: 032] Always pass a document model to the surface widget [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/114362 (owner: 10Esanders) [23:50:58] (03Merged) 10jenkins-bot: Fix SurfaceWidget documentation [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/114359 (owner: 10Esanders) [23:51:17] (03Merged) 10jenkins-bot: Always pass a document model to the surface widget [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/114362 (owner: 10Esanders) [23:52:16] RoanKattouw, ugh, this has thrown up another problem we were masking [23:52:22] Yay. :-( [23:52:30]

Foo

[23:53:23] (03PS8) 10Catrope: Add MW-specific keyboard shortcut help [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/113118 (owner: 10Esanders) [23:53:25] starting from

| you get to | [23:53:41] Ahm, starting from |

surely? [23:53:46]

,2 ? [23:53:50] yeah [23:54:07] but then the next step should be , i.e. not descending [23:54:36] (btw this is an not in the DM, i.e. a shield) [23:54:47] Wait, it's a shield? [23:54:58] Then we're already inside something that has .data( 'view' ) [23:55:00] yup [23:55:01] And descending is pointless [23:55:10] How can the cursor even *get there*? [23:55:28] Isn't there code that catches this case by looking for cE=false ancestors? [23:56:27] so this is the real test case:

FooBarBaz

[23:56:27] 'expected': [ [23:56:35] (03CR) 10Catrope: [C: 032] Add MW-specific keyboard shortcut help [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/113118 (owner: 10Esanders) [23:56:58] OK [23:57:03] So the is in the , not outside it [23:57:06] That makes more sense [23:57:31] ,1 -> ,2 -> ,0 [23:57:48] (03Merged) 10jenkins-bot: Add MW-specific keyboard shortcut help [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/113118 (owner: 10Esanders) [23:57:54] Ahm, wat [23:57:58] The has .data('view') [23:58:05] So it shouldn't descend [23:58:23] ...but it only checks for that if domOffset === 0 [23:58:26] yup [23:58:37] should we always check that? [23:59:03] No, only if domOffset === 0 or domOffset === last [23:59:16] You should be able to ask what's in the middle of some node [23:59:53] ok