[00:00:58] (03CR) 10jenkins-bot: [V: 04-1] ActionSet: Add description for events and clarify method descriptions [oojs/ui] - 10https://gerrit.wikimedia.org/r/195489 (owner: 10Kmenger) [00:08:30] edsanders, are you awake or is your nickname deceptive? [00:11:38] (03CR) 10Kmenger: "I think the reason this fails jsduck is that I tried to link to the 'remove' method, but there is also a 'remove' event. Is there a way to" [oojs/ui] - 10https://gerrit.wikimedia.org/r/195489 (owner: 10Kmenger) [00:12:10] 10OOjs-UI, 10UI-Standardization, 7Design: Hover state on buttons - https://phabricator.wikimedia.org/T92049#1102394 (10violetto) 3NEW [00:14:35] (03Merged) 10jenkins-bot: Don't flicker "Manage TemplateData" button's disabledness state [extensions/TemplateData] - 10https://gerrit.wikimedia.org/r/195057 (https://phabricator.wikimedia.org/T91324) (owner: 10Bartosz Dziewoński) [00:15:16] 10OOjs-UI, 5OOjs-UI-0.9.0, 7Accessibility, 5WMF-deploy-2015-03-11_(1.25wmf21): PopupButtonWidget menu is not keyboard traversable/closeable - https://phabricator.wikimedia.org/T87695#1102406 (10Etonkovidova) Re-checked - the same as the fix in https://phabricator.wikimedia.org/T89810 - works in beta and t... [00:15:37] 10OOjs-UI, 5OOjs-UI-0.9.0, 7Accessibility, 7Verified-in-Phase0, 5WMF-deploy-2015-03-11_(1.25wmf21): PopupButtonWidget menu is not keyboard traversable/closeable - https://phabricator.wikimedia.org/T87695#1102407 (10Etonkovidova) [00:18:01] (03Merged) 10jenkins-bot: Use modifier functions keyed by name for queued modifications in TransactionProcessor [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195360 (owner: 10Catrope) [00:23:42] 10OOjs-UI, 5OOjs-UI-0.9.0, 5Patch-For-Review, 5WMF-deploy-2015-03-11_(1.25wmf21): Icon containers should always be the same size so we can use background-size: contain - https://phabricator.wikimedia.org/T85139#1102422 (10Etonkovidova) en.wiki: .oo-ui-popupToolGroup-handle .oo-ui-iconElement-icon { top: 0... [00:38:25] 10VisualEditor, 10VisualEditor-ContentEditable, 10VisualEditor-MediaWiki, 7Browser-Support-Firefox, 3VisualEditor 2014/15 Q3 blockers: Initiating VisualEditor via "Edit" button Firefox scrolls to bottom of page - https://phabricator.wikimedia.org/T90420#1102482 (10Mooeypoo) After a bit of debugging, I be... [00:44:27] 10VisualEditor, 10WikiEditor, 6Collaboration-Team, 10Flow, 7Design: Editing toolbar(s) for Flow (VE and/or wikitext toolbar) - https://phabricator.wikimedia.org/T78346#1102492 (10Mattflaschen) VE basically has two types of dialogs: * Inspector (the one you see when you click the link icon) * Dialog (e.g... [00:55:19] James_F, I looked at WikiEditor's special characters (https://phabricator.wikimedia.org/T91608 ) [00:55:29] Krenair: Oh, nice. [00:55:32] however I came across this: https://github.com/wikimedia/mediawiki-extensions-WikiEditor/blob/master/modules/jquery.wikiEditor.toolbar.config.js#L581-L643 [00:55:44] Krenair: Yeah, it's… a mess. [00:56:06] James_F: are you up for deploying the 'edit (beta)' change? [00:56:09] or letting me deploy it, rather? [00:56:09] our special character system just handles characters, not those objects [00:56:15] ori: Scheduled for tomorrow morning. [00:56:19] oh, cool [00:56:23] nice! [00:56:28] Krenair: Yeah, we'll have to simplify, I think. [00:56:54] Krenair: What do you think would be best? [00:57:26] I have no idea how important those entries are [00:57:41] * James_F nods. [00:57:46] But I'm tempted to just export the (vast majority) easy ones [00:57:51] I see them as nice-to-haves rather than… yeah. :-) [00:58:11] Maybe we can do the inserts and later do the encapsulates if we think it's needed. [00:58:43] Or should we convert this to be a JSON module we can depend on in VE and WE alike? [00:58:51] (Rather than copy them.) [00:59:56] https://phabricator.wikimedia.org/P382 [01:00:08] a JSON module? [01:00:22] jquery module? [01:00:42] Krenair: A JSON file in a RL module, sorry, being imprecise. :-0 [01:02:02] VE core pulls them out of a message, currently :/ [01:02:20] So does VE-MW. [01:02:23] We can over-ride though. [01:02:29] MediaWiki:Visualeditor-specialcharinspector-characterlist-insert. [01:02:51] But let's take the opportunity to use a new message for the first set (per-wiki configuration of common ones) [01:03:26] VE-MW is involved? [01:03:38] looks to me like core pulls visualeditor-specialcharinspector-characterlist-insert directly [01:03:50] That's done in platform. [01:03:52] Or target. [01:03:54] Or something. [01:03:59] There's a VE-MW-specific thing. [01:04:33] the code that connects ve.msg up to MW's message system? [01:05:25] Maybe it's as simple as that. [01:05:32] Our other .json messages are rather nicer. [01:05:36] We should follow suit. [01:06:54] 10VisualEditor, 10VisualEditor-DataModel, 10VisualEditor-Tables, 3VisualEditor 2014/15 Q3 blockers, and 2 others: [Regression wmf20] Table gets broken while changing the cell type after a merge cell operation - https://phabricator.wikimedia.org/T91831#1102522 (10Ryasmeen) checked the fix in Betalabs and test2 [01:07:10] 10VisualEditor, 10VisualEditor-DataModel, 10VisualEditor-Tables, 7Verified-in-Phase0, and 3 others: [Regression wmf20] Table gets broken while changing the cell type after a merge cell operation - https://phabricator.wikimedia.org/T91831#1102523 (10Ryasmeen) [01:07:12] So... [01:07:17] Move WikiEditor's character data to a separate RL module somehow. [01:07:20] Import it in VE-MW [01:07:56] Override VE core's special character system with our own to provide characters from that [01:07:58] ? [01:08:16] Yeah. [01:08:24] Or ignore me and pick a better solution. :-) [01:08:58] So that we don't end up with a ridiculous Visualeditor-specialcharinspector-characterlist-insert [01:09:04] Well. [01:09:20] We'll still need MediaWiki:Visualeditor-specialcharacters-common.json or something. [01:09:33] But we won't want more than ~20 characters in there. [01:09:40] Certainly not 20,000. :-) [01:10:15] So what would we keep in the message and what would we keep in the RL module? [01:11:15] Just the stuff that core already has in the message? [01:11:40] I was thinking that the module would have all those lists in P382. [01:12:08] And that the very first section would be wiki-specifyable "Common characters" created from MediaWiki:Visualeditor-specialcharacters-common.json if it exists. [01:12:17] Maybe this should be in VE-core. [01:12:48] (So wikis without a MediaWiki:Visualeditor-specialcharacters-common.json would have no "Common characters" first group, just Latin, Latin extended, IPA, Symbols, …). [01:13:23] Right so the message would just be an extra character group on the top for customisation by the wiki? [01:13:44] Yup. [01:13:52] One group, not multiple. [01:13:57] So simpler and easier for them to edit. [01:14:10] And we should absolutely definitely set it as JSON content type on the wiki. :-) [01:14:46] I don't know if we can do that with messages? Hm. [01:14:57] We do for the other ones. [01:15:10] yeah, just wondering how at the moment [01:15:18] Compare http://en.wikipedia.beta.wmflabs.org/wiki/MediaWiki:Visualeditor-cite-tool-definition.json and http://en.wikipedia.beta.wmflabs.org/wiki/MediaWiki:Visualeditor-specialcharinspector-characterlist-insert [01:15:26] I think we just add it to ::Hooks. [01:15:33] (03CR) 10Mattflaschen: [C: 04-1] "Still some small issues. I'll go ahead and fix these up then merge." (034 comments) [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/189333 (owner: 10Paladox) [01:16:04] right, it's in hooks [01:16:15] but with a capital letter at the beginning so my grep was failing :/ [01:16:43] * James_F runs away. [01:16:49] Krenair: Thanks so much. :-) [01:16:55] yeah it just sets the default model for Visualeditor-cite-tool-definition.json and Visualeditor-specialcharinspector-characterlist-insert to JSON [01:17:41] will put notes on the task [01:20:41] 10VisualEditor: Two progress bar appearing on the Autofill dialog - https://phabricator.wikimedia.org/T92055#1102525 (10Ryasmeen) 3NEW [01:21:49] 10VisualEditor: Two progress bars are appearing on the Autofill ciation dialog - https://phabricator.wikimedia.org/T92055#1102532 (10Ryasmeen) [01:23:58] 10VisualEditor, 10VisualEditor-MediaWiki, 3VisualEditor 2014/15 Q3 blockers: Import WikiEditor's list into the Special Character inserter - https://phabricator.wikimedia.org/T91608#1102537 (10Krenair) I made a quick script to pull the characters that WikiEditor currently has: {P382} Note the typeof check -... [01:24:47] 10VisualEditor, 10VisualEditor-MediaWiki, 10WikiEditor, 3VisualEditor 2014/15 Q3 blockers: Import WikiEditor's list into the Special Character inserter - https://phabricator.wikimedia.org/T91608#1102538 (10Krenair) [01:44:58] (03PS1) 10Krinkle: mw.ViewPageTarget: Keep pageview content hidden until surface teardown is ready [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195496 (https://phabricator.wikimedia.org/T91442) [01:51:35] (03PS2) 10Krinkle: mw.ViewPageTarget: Hide ve-ce-surface during deactivation [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195496 (https://phabricator.wikimedia.org/T91442) [01:53:12] (03PS1) 10Krinkle: mw.ViewPageTarget: Cache '#mw-imagepage-content' query [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195497 [01:54:23] 10VisualEditor, 10Continuous-Integration, 3VisualEditor 2014/15 Q3 blockers: Concurrent builds using local Chromium/Firefox browsers on Linux host fail - https://phabricator.wikimedia.org/T90673#1102552 (10Krinkle) [01:57:14] 10OOjs-UI, 10UI-Standardization, 7Design: Implement frameless popupButtonWidget - https://phabricator.wikimedia.org/T91857#1102554 (10violetto) [01:58:25] 10OOjs-UI, 10UI-Standardization, 7Design: Implement tipsy - https://phabricator.wikimedia.org/T91856#1102555 (10violetto) [01:59:16] (03PS4) 10Physikerwelt: Prototypical direct support of XQuery expressions [extensions/MathSearch] - 10https://gerrit.wikimedia.org/r/194233 [02:07:01] (03CR) 10Alex Monk: mw.ViewPageTarget: Cache '#mw-imagepage-content' query (031 comment) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195497 (owner: 10Krinkle) [02:11:37] (03CR) 10Krinkle: mw.ViewPageTarget: Cache '#mw-imagepage-content' query (031 comment) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195497 (owner: 10Krinkle) [02:23:09] (03CR) 10Alex Monk: [C: 032] mw.ViewPageTarget: Cache '#mw-imagepage-content' query [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195497 (owner: 10Krinkle) [02:25:12] (03Merged) 10jenkins-bot: mw.ViewPageTarget: Cache '#mw-imagepage-content' query [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195497 (owner: 10Krinkle) [02:40:25] 10VisualEditor, 10VisualEditor-MediaWiki, 3VisualEditor 2014/15 Q3 blockers: VisualEditor: Set up mass round-trip testing infrastructure on real content - https://phabricator.wikimedia.org/T52513#1102564 (10Krinkle) We currently have https://github.com/wikimedia/ve-dirtydiffbot, which: * For 720 random en.wi... [03:01:21] 10OOjs-UI: Make OOjs UI LabelWidget flaggable so text colors can be used, the way ButtonElement does - https://phabricator.wikimedia.org/T92024#1102568 (10Jay8g) [03:01:22] 10OOjs-UI: Make OOjs UI IconWidget flaggable so icon variants can be used, the way ButtonElement does - https://phabricator.wikimedia.org/T92023#1102569 (10Jay8g) [04:47:13] (03PS29) 10Mattflaschen: Convert .css to .less [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/189333 (owner: 10Paladox) [04:47:21] (03CR) 10Mattflaschen: Convert .css to .less (034 comments) [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/189333 (owner: 10Paladox) [04:47:28] (03CR) 10Mattflaschen: [C: 032] Convert .css to .less [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/189333 (owner: 10Paladox) [04:47:33] (03Merged) 10jenkins-bot: Convert .css to .less [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/189333 (owner: 10Paladox) [04:47:45] (03CR) 10Mattflaschen: "Thanks for the patch." [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/189333 (owner: 10Paladox) [05:54:20] (03PS4) 10Gerrit Patch Uploader: Rework color definitions in LESS [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/190175 [05:54:23] (03CR) 10Gerrit Patch Uploader: "This commit was uploaded using the Gerrit Patch Uploader [1]." [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/190175 (owner: 10Gerrit Patch Uploader) [06:10:20] (03PS3) 10Gerrit Patch Uploader: Use background-image instead of for spinner [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/184301 [06:10:23] (03CR) 10Gerrit Patch Uploader: "This commit was uploaded using the Gerrit Patch Uploader [1]." [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/184301 (owner: 10Gerrit Patch Uploader) [06:24:46] 10MediaWiki-Page-editing: editing page that is highly transcluded leads to read timeout - https://phabricator.wikimedia.org/T14814#1103525 (10BBlack) 5Open>3Invalid a:3BBlack I stumbled on this while running a related query. This bug report refers to infrastructure (and code) that has completely changed i... [07:27:06] (03PS14) 10Mvolz: Fix acceptLanguage commit a30d6be [services/citoid] - 10https://gerrit.wikimedia.org/r/195025 [07:32:29] (03CR) 10Raimond Spekking: [C: 031] "i18n review" [extensions/MathSearch] - 10https://gerrit.wikimedia.org/r/194233 (owner: 10Physikerwelt) [07:49:33] (03CR) 10Paladox: "Thanks." [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/189333 (owner: 10Paladox) [09:46:29] (03CR) 10Catrope: "Does using #method-remove instead of #remove work?" [oojs/ui] - 10https://gerrit.wikimedia.org/r/195489 (owner: 10Kmenger) [09:56:41] (03CR) 10Catrope: [C: 032] mw.ViewPageTarget: Hide ve-ce-surface during deactivation [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195496 (https://phabricator.wikimedia.org/T91442) (owner: 10Krinkle) [09:58:53] (03Merged) 10jenkins-bot: mw.ViewPageTarget: Hide ve-ce-surface during deactivation [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195496 (https://phabricator.wikimedia.org/T91442) (owner: 10Krinkle) [10:10:08] (03PS1) 10Paladox: Add svg to LESS file [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/195529 (https://phabricator.wikimedia.org/T37342) [10:30:36] (03PS2) 10Paladox: Add svg to LESS file [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/195529 (https://phabricator.wikimedia.org/T37342) [10:30:53] (03PS3) 10Paladox: Add svg to LESS file [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/195529 (https://phabricator.wikimedia.org/T37342) [10:39:09] (03PS4) 10Paladox: Add svg to LESS file [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/195529 (https://phabricator.wikimedia.org/T37342) [10:39:18] (03PS5) 10Paladox: Add svg to LESS file [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/195529 (https://phabricator.wikimedia.org/T37342) [10:40:38] (03PS6) 10Paladox: Add svg to LESS file [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/195529 (https://phabricator.wikimedia.org/T37342) [10:40:48] (03PS7) 10Paladox: Add svg to LESS file [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/195529 (https://phabricator.wikimedia.org/T37342) [10:41:18] (03PS8) 10Paladox: Add svg to LESS file [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/195529 (https://phabricator.wikimedia.org/T37342) [10:43:39] 10WikiEditor: Provide dialog to enter and edit templates - https://phabricator.wikimedia.org/T59232#1103975 (10Schnark) >>! In T59232#610546, @Aklapper wrote: > Michael: How is/was this request different from the (unmaintained) template editor which was just removed from WikiEditor recently? https://gerrit.wikim... [10:43:44] (03Abandoned) 10Paladox: Convert .css to .less [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/181790 (owner: 10Paladox) [10:44:01] (03Abandoned) 10Paladox: Adds svg to css file [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/181788 (https://phabricator.wikimedia.org/T37342) (owner: 10Paladox) [10:44:31] (03CR) 10Paladox: "@Mattflaschen please review." [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/195529 (https://phabricator.wikimedia.org/T37342) (owner: 10Paladox) [10:44:45] (03PS40) 10Paladox: Re add svg images [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/181786 (https://phabricator.wikimedia.org/T37342) [10:45:15] (03CR) 10Paladox: "@Mattflaschen please merge." [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/181786 (https://phabricator.wikimedia.org/T37342) (owner: 10Paladox) [11:04:50] 10MediaWiki-Page-editing, 10MediaWiki-General-or-Unknown: Include wikieditor in core - https://phabricator.wikimedia.org/T91872#1103994 (10Aklapper) 5stalled>3Open [11:09:29] (03PS2) 10Catrope: When replacing e.g. heading1 with heading2, use attribute changes [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195242 [11:09:31] (03PS2) 10Catrope: Add test for heading1->heading2 content branch conversion [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195241 [11:11:22] (03CR) 10Catrope: "PS2: Rebased off of controversial ce.View#initialize stack. Note that this means that this commit no longer benefits from https://gerrit.w" [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195242 (owner: 10Catrope) [11:15:08] 10OOjs-UI, 10UI-Standardization, 7Design: PopupButtonWidget should be shown at hover not on-click. - https://phabricator.wikimedia.org/T88630#1104020 (10Pginer-WMF) @violetto it was a generic concern, similar to the one in the "hover tunnels" section about [[ http://uxmovement.com/navigation/why-hover-menus-... [12:48:33] (03CR) 10Esanders: [C: 032] Remove useless ve-ce-mw*Node classes from GCN subclasses [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195223 (owner: 10Catrope) [12:50:43] (03Merged) 10jenkins-bot: Remove useless ve-ce-mw*Node classes from GCN subclasses [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195223 (owner: 10Catrope) [12:53:59] RoanKattouw, does https://gerrit.wikimedia.org/r/#/c/195242/2 make the table cell span fix redundant too? [12:54:25] edsanders: No, it's almost the opposite [12:54:41] Your table style dropdown already works by processing an attribute change transaction, right? [12:54:48] yes [12:55:00] Something like retain N, changeAttribute 'style' from 'data' to 'header', retain M [12:55:22] So before 195242, the format dropdown used replace transactions for everything [12:55:27] Including heading1->heading2 [12:56:05] e.g. retain N, replace { type: 'heading', attributes: { level: 1 } } with { type: 'heading', attributes: { level: 2 } }, retain L, replace { type: '/heading' } with { type: '/heading' } (yes really), retain < [12:56:08] *retain M [12:57:16] So that caused originalDomElements to be lost, any other attributes that might have existed to be lost, whitespace information to be lost, and new dm.Nodes and ce.Nodes to be constructed from scratch [12:57:34] This is bad because of all the information it throws away [12:57:42] (https://gerrit.wikimedia.org/r/#/c/195241/2/tests/dm/ve.dm.Transaction.test.js asserts the previous, bad behavior) [12:58:16] So 195242 makes the format dropdown use an attribute change transaction instead of a replace transaction if the type isn't changed [12:58:27] i.e. heading->paragraph will still be a replace, but headingN->headingM will not be [12:59:22] It also does a subset comparison of the attributes rather than a full comparison, so trying to convert something to { type: 'heading', attributes: { level: 1 } } will be a no-op if it's already a heading and already has level=1 set, whereas previously it was a no-op only if it was already a heading and its attributes object was *exactly* { heading: 1 } [12:59:41] (Now it checks that the existing attributes object is a superset of { heading: 1 } ) [13:00:53] (03CR) 10Esanders: [C: 032] Add test for heading1->heading2 content branch conversion [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195241 (owner: 10Catrope) [13:01:01] (03CR) 10Esanders: [C: 032] When replacing e.g. heading1 with heading2, use attribute changes [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195242 (owner: 10Catrope) [13:03:22] (03Merged) 10jenkins-bot: Add test for heading1->heading2 content branch conversion [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195241 (owner: 10Catrope) [13:03:58] (03Merged) 10jenkins-bot: When replacing e.g. heading1 with heading2, use attribute changes [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195242 (owner: 10Catrope) [13:23:33] I'm trying to remember why I did it, but what does this fix: https://gerrit.wikimedia.org/r/#/c/195046/1 ? [13:23:59] MatmaRex [13:25:59] (03CR) 10Esanders: "Why? We need to make sure the toolbar is detached from the last surface whenever it is re-setup with a new one." [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195046 (owner: 10Bartosz Dziewoński) [13:28:48] edsanders: dependency of https://gerrit.wikimedia.org/r/195048 [13:30:12] edsanders: there is a very unpleasant overlap between #initialize and #setup in that class [13:50:52] (03CR) 10Catrope: [C: 032] Schema:Edit instrumentation [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/191221 (https://phabricator.wikimedia.org/T88027) (owner: 10Alex Monk) [13:50:55] (03CR) 10jenkins-bot: [V: 04-1] Schema:Edit instrumentation [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/191221 (https://phabricator.wikimedia.org/T88027) (owner: 10Alex Monk) [13:51:04] grmbl [13:56:50] (03PS17) 10Catrope: Schema:Edit instrumentation [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/191221 (https://phabricator.wikimedia.org/T88027) (owner: 10Alex Monk) [13:59:15] (03CR) 10Catrope: [C: 032] Schema:Edit instrumentation [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/191221 (https://phabricator.wikimedia.org/T88027) (owner: 10Alex Monk) [13:59:20] (03Merged) 10jenkins-bot: Schema:Edit instrumentation [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/191221 (https://phabricator.wikimedia.org/T88027) (owner: 10Alex Monk) [14:01:01] (03CR) 10Bartosz Dziewoński: "But we shouldn't unhook the scroll and resize events on window then, should we? Perhaps this should be rearranged differently, though…" [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195046 (owner: 10Bartosz Dziewoński) [14:11:17] (03PS2) 10Bartosz Dziewoński: ve.ui.Toolbar: Do not remove event listeners in #setup [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195046 [14:11:19] (03PS2) 10Bartosz Dziewoński: Do not call Toolbar#initialize twice during editor setup [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195048 [14:24:10] (03CR) 10Esanders: "The window scroll events would fail if there was no surface to measure" [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195046 (owner: 10Bartosz Dziewoński) [14:25:38] (03PS3) 10Bartosz Dziewoński: ve.ui.Toolbar: Do not remove window event listeners in #setup [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195046 [14:26:56] 10VisualEditor, 10VisualEditor-ContentEditable, 10VisualEditor-CopyPaste, 3VisualEditor 2014/15 Q3 blockers: VisualEditor: Would be nice if copying/pasting kept working after a fatal - https://phabricator.wikimedia.org/T68229#1104403 (10Catrope) [14:30:04] (03CR) 10Bartosz Dziewoński: "Hmm… so, toolbar floating must be enabled in #setup, and not #initialize like now?" [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195046 (owner: 10Bartosz Dziewoński) [14:30:45] (03PS1) 10Catrope: ce.RangeState: Check this.node for falseyness, not just null [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195563 [14:34:32] 10VisualEditor, 10VisualEditor-CopyPaste: Cross-VE paste of reference by itself splits paragraph - https://phabricator.wikimedia.org/T92271#1104433 (10Catrope) 3NEW [14:39:26] ugh [14:39:34] we setup() the VE toolbar before we initialize() it [14:39:37] 10VisualEditor, 10VisualEditor-CopyPaste, 10VisualEditor-MediaWiki-References: Copy/pasting deletes references, can't save - https://phabricator.wikimedia.org/T89146#1104467 (10Catrope) I can't reproduce this. I tried copying the Efficacité section from the article you linked into your sandbox page, both at... [14:42:25] lolwut [14:42:29] setup before initialize? [14:42:46] Hmm I guess you're right [14:43:00] setupToolbar runs before initializeToolbar, I remember this now from when I put in instrumentation points for those [14:43:08] setup runs before attachment, initialize runs after [14:43:10] RoanKattouw: Heya [14:44:10] Hey [14:44:45] ugh [14:44:49] * MatmaRex stabs [14:46:28] (03CR) 10Catrope: "@James: This commit does not address any of those things. Once T90304 is fixed (which can be done independently of this commit, but relies" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/193026 (https://phabricator.wikimedia.org/T53569) (owner: 10Catrope) [14:47:47] 10VisualEditor, 10VisualEditor-Tables: [Regression wmf20] Cursor traverses through the table unexpectedly with right/left arrow keys and throws following error in the console "Uncaught TypeError: Cannot read property 'start' of null" - https://phabricator.wikimedia.org/T91726#1104497 (10Esanders) Can't reprodu... [14:47:52] (03PS1) 10Esanders: Unbreak cursoring inside table cells [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195564 [14:48:25] RoanKattouw, this is important: https://gerrit.wikimedia.org/r/#/c/195564/ [14:48:40] RoanKattouw: I'm stress testing Karma's canary build to see if the race condition is gone. [14:48:50] (03CR) 10Catrope: [C: 032] Unbreak cursoring inside table cells [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195564 (owner: 10Esanders) [14:49:13] Krinkle: Cool [14:49:37] RoanKattouw: Regarding targetLoader, where're we at? Are you blocked? [14:49:39] Krinkle: I'm starting on an Etherpad for TargetLoader work planning [14:49:55] I'm not blocked necessarily but I'll probably share it with you in a little bit and ask for input [14:50:06] And probably ask you to do some of the work [14:50:30] There's also https://phabricator.wikimedia.org/T91460 and its dependencies, just in case you have nothing to do :P [14:50:49] (03Merged) 10jenkins-bot: Unbreak cursoring inside table cells [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195564 (owner: 10Esanders) [14:57:11] RoanKattouw: btw, if you were using that instrumentation data for something, it's probably all wrong. [14:57:29] since we called #initialize twice, once in trace.setupToolbar and once in trace.initializeToolbar [14:57:35] lol [14:57:47] Well initializeToolbar was very very fast [14:57:55] setupToolbar takes 100ms or so [14:57:57] it hardly does anything [14:58:05] and no, it doesn't have guards against multiple calls [14:58:15] it just mostly keeps trucking [14:58:34] it would bind some window scroll events twice, but we were accidentally detaching these between the two calls [14:58:39] lol [14:59:40] i'm trying to untangle it. would it be a problem if trace.initializeToolbar disappeared and trace.setupToolbar started tracing initialize+setup, intentionally this time? [15:00:07] (03CR) 10Kmenger: "Thanks, Catrope! When I try that, the doc reads " #method-remove", no link." [oojs/ui] - 10https://gerrit.wikimedia.org/r/195489 (owner: 10Kmenger) [15:00:52] (03CR) 10Catrope: "Did you use {@link #method-remove removed} ?" [oojs/ui] - 10https://gerrit.wikimedia.org/r/195489 (owner: 10Kmenger) [15:02:58] 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#1104564 (10gpaumier) [15:11:50] Krinkle: https://etherpad.wikimedia.org/p/TargetLoader is my preliminary game plan, your input would be much appreciated [15:12:12] Krinkle: I hope that communicates my goals, strategy and open problems [15:12:23] If something needs clarification, please do ask [15:14:20] (03CR) 10Kmenger: "That works! Thanks." [oojs/ui] - 10https://gerrit.wikimedia.org/r/195489 (owner: 10Kmenger) [15:17:32] RoanKattouw: OK. So which class needs to be accessed by multiple modules? [15:18:16] Do you mean the question about "how do we not create a million modules?" ? [15:18:25] Yeah [15:18:28] Let's tackle that [15:18:35] Well, ve.init.mw.TargetLoader = .... ; requires ve.init.mw = {}; [15:18:42] which requires ve.init = {}; [15:18:48] I guess I could also rename / re-namespace TargetLoader [15:19:10] RoanKattouw: Does TargetLoader need any ve.* ? [15:19:16] Or can it be like init init [15:19:21] It probably could be [15:19:26] It needs very basic things, let me check [15:19:43] It seems to only need ve.track [15:19:51] Which even init init depends on [15:20:16] So that's ve.js (which is just ve = {}; and ve.now), and ve.track.js (ve.track, ve.trackSubscribe), none of the stuff in ve.utils.js [15:20:26] I also made TargetLoader not use oojs on purpose [15:20:31] So we could make it like init init I guess [15:20:53] Cool [15:20:55] That would probably be more sensible than what I did actually [15:21:09] Does TargetLoader need to be extended or referenced by other classes, or can it be local? [15:21:32] Then it wouldn't need a public namespace [15:22:05] I started out making it public because I wasn't sure, but I don't think I discovered a use case for it needing to be public [15:22:15] Other than addPlugin but that's already routed through init init anyway [15:22:31] The way I did that should also be cleaned up a bit [15:22:47] To the tune of mw.libs.ve.addPlugin = targetLoader.addPlugin; or something [15:23:18] (03CR) 10Jforrester: "Is this the fix for T91726?" [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195564 (owner: 10Esanders) [15:23:24] RoanKattouw: Hm.. yeah [15:23:42] RoanKattouw: So plugins need to wait for targetLoader to call addPlugin? [15:23:55] Maybe make a [].push faux buffer in init init so it's available in the top module. [15:23:59] Oh right, yes [15:24:01] Of course [15:24:10] That's pretty much what I did [15:24:27] I mean, I just have init init maintain an array and call targetLoader.addPlugins( plugins ) [15:24:33] ok [15:24:48] So maybe TargetLoader can have a simpler API because it's really only going to be given its set of plugins once [15:25:15] (Plugins calling addPlugin() is not a reasonable use case, because plugin functions can already return a promise to achieve the same outcome) [15:26:32] 10VisualEditor, 10VisualEditor-MediaWiki, 5WMF-deploy-2015-03-11_(1.25wmf21): The transition while switching between Readmode and VE is now not smooth, new content is inserted before the old content is removed leading to a flicker - https://phabricator.wikimedia.org/T91442#1104659 (10Jdforrester-WMF) 5Open... [15:29:47] 10VisualEditor, 10Continuous-Integration, 3VisualEditor 2014/15 Q3 blockers: Concurrent builds using local Chromium/Firefox browsers on Linux host fail - https://phabricator.wikimedia.org/T90673#1104669 (10Krinkle) Per the suggestion from upstream ([karma-users thread](https://groups.google.com/forum/#!msg/k... [15:35:12] (03CR) 10Alexandros Kosiaris: "@Mvolz, is that enough for citoid to use zotero.svc.eqiad.wmnet ?" [services/citoid/deploy] - 10https://gerrit.wikimedia.org/r/195333 (https://phabricator.wikimedia.org/T89873) (owner: 10Alexandros Kosiaris) [15:45:54] 10VisualEditor, 10Wikimedia-Site-requests, 3VisualEditor 2014/15 Q3 blockers, 5WMF-deploy-2015-03-11_(1.25wmf21), 7user-notice: Disable "Beta" suffixes on tabs for VisualEditor opt-in wikis - https://phabricator.wikimedia.org/T60583#1104720 (10Jdforrester-WMF) 5Open>3Resolved [15:46:05] 10VisualEditor, 10VisualEditor-Initialisation, 10VisualEditor-Performance, 3VisualEditor 2014/15 Q3 blockers, 5WMF-deploy-2015-03-11_(1.25wmf21): VisualEditor "beta" appended label causes edit tab to flicker - https://phabricator.wikimedia.org/T89668#1104722 (10Jdforrester-WMF) [15:46:30] 10VisualEditor, 10VisualEditor-Initialisation, 10VisualEditor-Performance, 3VisualEditor 2014/15 Q3 blockers, 5WMF-deploy-2015-03-11_(1.25wmf21): VisualEditor "beta" appended label causes edit tab to flicker - https://phabricator.wikimedia.org/T89668#1104726 (10Jdforrester-WMF) 5Open>3Resolved [15:47:56] 10VisualEditor, 10VisualEditor-Performance, 10WikiEditor, 10Analytics, and 3 others: Apply Schema:Edit instrumentation to WikiEditor - https://phabricator.wikimedia.org/T88027#1104728 (10Jdforrester-WMF) 5Open>3Resolved [15:48:24] (03PS1) 10Krinkle: [BREAKING CHANGE] ve.utils: Remove ve.getObjectKeys in favour of Object.keys [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195575 [15:49:46] Krenair: Do you know who is planning to merge https://gerrit.wikimedia.org/r/#/c/193826/ and when? Is everyone awkwardly waiting for someone else? :-) [15:53:37] Oh right [15:53:50] That's the core dependency of the VE-MW commit that I was going to ask Krenair to review [15:55:24] (03CR) 10Catrope: [C: 04-2] "Needs cleanup change in VE-MW to be written and merged first" [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195575 (owner: 10Krinkle) [15:55:28] RoanKattouw: Yeah. [15:55:52] edsanders: Which deployment branches are affected by https://gerrit.wikimedia.org/r/#/c/195564/ ? [15:55:59] edsanders: AKA in which commit did this break? [15:56:13] 10OOjs-UI, 10VisualEditor: VisualEditor's and OOjs UI's toolbars are conceptually incompatible - https://phabricator.wikimedia.org/T92282#1104751 (10matmarex) [15:56:14] 10OOjs-UI, 10VisualEditor: VisualEditor's and OOjs UI's toolbars are conceptually incompatible - https://phabricator.wikimedia.org/T92282#1104752 (10Jdforrester-WMF) [15:56:25] RoanKattouw, https://gerrit.wikimedia.org/r/#/c/194352/ [15:56:54] (03CR) 10Bartosz Dziewoński: "That is actually impossible now because the toolbars are terrible. Filed T92282." [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195046 (owner: 10Bartosz Dziewoński) [15:57:38] (https://phabricator.wikimedia.org/T92282 may or may not be relevant to the target rewrite, RoanKattouw) [15:57:59] Hmm [15:58:37] Maybe, but I think I can avoid it [15:59:36] James_F: Sounds like we'll need a wmf20 cherry-pick for https://gerrit.wikimedia.org/r/#/c/195564/ [16:00:39] Is that a "yes, it does fix T91726 which was a wmf20 regression"? [16:00:48] Or "No, but it's still a wmf20 regression"? [16:02:02] James_F: Ed says "I can't reproduce T91726 after my fix so maybe" [16:02:09] Ha. [16:02:14] OK. [16:02:24] (03PS1) 10Jforrester: Unbreak cursoring inside table cells [VisualEditor/VisualEditor] (wmf/1.25wmf20) - 10https://gerrit.wikimedia.org/r/195578 [16:02:33] (03CR) 10Jforrester: [C: 032] Unbreak cursoring inside table cells [VisualEditor/VisualEditor] (wmf/1.25wmf20) - 10https://gerrit.wikimedia.org/r/195578 (owner: 10Jforrester) [16:03:05] 10VisualEditor, 10VisualEditor-Tables, 3VisualEditor 2014/15 Q3 blockers, 5WMF-deploy-2015-03-04_(1.25wmf20), 5WMF-deploy-2015-03-11_(1.25wmf21): [Regression wmf20] Cursor traverses through the table unexpectedly with right/left arrow keys and throws follo... - https://phabricator.wikimedia.org/T91726#1104776 [16:05:06] (03Merged) 10jenkins-bot: Unbreak cursoring inside table cells [VisualEditor/VisualEditor] (wmf/1.25wmf20) - 10https://gerrit.wikimedia.org/r/195578 (owner: 10Jforrester) [16:06:19] edsanders: Did you ever get a PIN? [16:06:26] (for the powpow conference call?) [16:06:38] I got the US number to work via hangouts, but I don't see a pin anywhere in the email thread [16:06:39] (03PS1) 10Jforrester: Update VE core to c0d60b6 for cherry-pick [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195579 [16:06:44] (03CR) 10jenkins-bot: [V: 04-1] Update VE core to c0d60b6 for cherry-pick [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195579 (owner: 10Jforrester) [16:06:53] Bah. [16:07:36] Krinkle: yes [16:07:47] in a separate email somewhere [16:07:55] 10VisualEditor, 10VisualEditor-EditingTools: VisualEditor: Open formula inspector breaks when clicking onto a link/reference/image/template/… - https://phabricator.wikimedia.org/T58644#1104784 (10Esanders) Appears to work now [16:07:55] it should be in the calendar event too [16:08:01] Bah. git sucks. [16:08:32] (03PS1) 10Jforrester: Update VE core to c0d60b6 for cherry-pick [extensions/VisualEditor] (wmf/1.25wmf20) - 10https://gerrit.wikimedia.org/r/195580 [16:08:46] (03Abandoned) 10Jforrester: Update VE core to c0d60b6 for cherry-pick [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195579 (owner: 10Jforrester) [16:10:00] (03CR) 10jenkins-bot: [V: 04-1] Update VE core to c0d60b6 for cherry-pick [extensions/VisualEditor] (wmf/1.25wmf20) - 10https://gerrit.wikimedia.org/r/195580 (owner: 10Jforrester) [16:10:00] 10VisualEditor, 10Flow, 10LiquidThreads, 3Collaboration-Team-Sprint-R-2015-03-11: R6. LQT conversion test for Mediawiki.org VE/Feedback - https://phabricator.wikimedia.org/T90025#1051694 (10matthiasmullie) This one had an error: `[2015-03-10 15:16:51] Imported 309 items, failed 1` I've filed T92284 to fi... [16:10:23] 10VisualEditor, 10Flow, 10LiquidThreads, 3Collaboration-Team-Sprint-R-2015-03-11: R6. LQT conversion test for Mediawiki.org VE/Feedback - https://phabricator.wikimedia.org/T90025#1104812 (10matthiasmullie) a:5EBernhardson>3matthiasmullie [16:10:31] * James_F rolls his eyes. [16:12:31] (03CR) 10Mvolz: "Put the comment above rather than below, changed name of 'test doi' to 'doi' (redundant), and added more info to describes string param in" [services/citoid] - 10https://gerrit.wikimedia.org/r/195025 (owner: 10Mvolz) [16:14:23] 10VisualEditor, 10VisualEditor-Tables, 3VisualEditor 2014/15 Q3 blockers, 5WMF-deploy-2015-03-04_(1.25wmf20), 5WMF-deploy-2015-03-11_(1.25wmf21): [Regression wmf20] Cursor traverses through the table unexpectedly with right/left arrow keys and throws follo... - https://phabricator.wikimedia.org/T91726#1104823 [16:14:55] (03PS1) 10Krinkle: [BREAKING CHANGE] ve.utils: Remove ve.indexOf in favour of Array#indexOf [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195582 [16:15:46] (03CR) 10Catrope: [C: 04-2] "There are still occurrences in VE-MW, those need to be ported first before merging this." [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195582 (owner: 10Krinkle) [16:15:48] omg, this call system is hilarious. It verbalises all commands. So when someone leaves the call, the call is muted for 10 seconds as a computer voice says "participant has left the call" [16:16:09] Oh yeah the thing we used in the pre-SIP era was kind of like that [16:16:15] 10VisualEditor-MediaWiki-Plugins, 10MediaWiki-extensions-ProofreadPage, 6Wikisource: Create a VisualEditor plugin to integrate with ProofreadPage - https://phabricator.wikimedia.org/T48580#1104861 (10Aklapper) [16:16:16] * James_F nods. [16:16:19] 10MediaWiki-Page-editing, 6Wikisource: [Regression] Dynamically added textarea within the editform no longer work in the classic edit toolbar - https://phabricator.wikimedia.org/T41268#1104859 (10Aklapper) [16:16:19] 10MediaWiki-Page-editing, 6Wikisource, 7Design, 5MW-1.23-release, 7Regression: Revert bug 27860 or add configuration variable to restore "Mark all edits minor by default" preference behaviour - https://phabricator.wikimedia.org/T44008#1104858 (10Aklapper) [16:16:23] " has left the meeting" [16:16:36] 10WikiEditor, 6Wikisource: Advanced edit toolbar doesn't allow changing the width of the editor textbox on Wikisource projects. - https://phabricator.wikimedia.org/T38230#1104876 (10Aklapper) [16:17:09] (03CR) 10jenkins-bot: [V: 04-1] [BREAKING CHANGE] ve.utils: Remove ve.indexOf in favour of Array#indexOf [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195582 (owner: 10Krinkle) [16:17:40] 10MediaWiki-Page-editing, 6Wikisource: toolbar and header/footer from proofread page on wikisource - https://phabricator.wikimedia.org/T33566#1104938 (10Aklapper) [16:18:07] (03PS9) 10Paladox: Add svg to LESS file [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/195529 (https://phabricator.wikimedia.org/T37342) [16:22:45] (03CR) 10Krinkle: "Looks like that failure is genuine. rules.blacklist is undefined in the tests and it was previously being masked by jQuery." [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195582 (owner: 10Krinkle) [16:28:13] (03PS1) 10Krinkle: Remove use of ve.indexOf and ve.getObjectKeys [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195586 [16:32:53] (03CR) 10Catrope: [C: 032] Remove use of ve.indexOf and ve.getObjectKeys [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195586 (owner: 10Krinkle) [16:33:05] edsanders: What's wrong with https://gerrit.wikimedia.org/r/#/c/195361/ or are you going to merge it? [16:33:33] (03PS2) 10Kmenger: ActionSet: Add description for events and clarify method descriptions [oojs/ui] - 10https://gerrit.wikimedia.org/r/195489 [16:34:17] (03CR) 10Catrope: [C: 032] [BREAKING CHANGE] ve.utils: Remove ve.getObjectKeys in favour of Object.keys [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195575 (owner: 10Krinkle) [16:34:57] (03Merged) 10jenkins-bot: Remove use of ve.indexOf and ve.getObjectKeys [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195586 (owner: 10Krinkle) [16:35:41] (03CR) 10jenkins-bot: [V: 04-1] ActionSet: Add description for events and clarify method descriptions [oojs/ui] - 10https://gerrit.wikimedia.org/r/195489 (owner: 10Kmenger) [16:36:21] (03Merged) 10jenkins-bot: [BREAKING CHANGE] ve.utils: Remove ve.getObjectKeys in favour of Object.keys [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195575 (owner: 10Krinkle) [16:36:23] (03CR) 10Catrope: [C: 04-1] [BREAKING CHANGE] ve.utils: Remove ve.indexOf in favour of Array#indexOf (031 comment) [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195582 (owner: 10Krinkle) [16:37:32] (03PS2) 10Krinkle: [BREAKING CHANGE] ve.utils: Remove ve.indexOf in favour of Array#indexOf [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195582 [16:38:17] James_F: Re the copy paste exceptions bug, do you know of a clear, reproducible path to a copy paste exception? [16:38:24] I kind of need that to even start on that [16:38:32] RoanKattouw: No. :-( [16:38:48] RoanKattouw: Get a large Google Docs page and try to paste it in. It'll probably fail. [16:38:54] RoanKattouw: Not a minimal case, though. [16:38:54] haha [16:38:59] I tried pasting entire MW pages [16:39:10] But I'll try gdocs [16:39:21] Also because I suspect https://phabricator.wikimedia.org/T85577 isn't a real bug [16:39:30] No, MWParser and Parsoid HTML both go in cleanly. [16:39:34] (03CR) 10jenkins-bot: [V: 04-1] [BREAKING CHANGE] ve.utils: Remove ve.indexOf in favour of Array#indexOf [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195582 (owner: 10Krinkle) [16:41:46] (03CR) 10Esanders: [C: 032] Make transaction processing exception-safe [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195361 (https://phabricator.wikimedia.org/T70892) (owner: 10Catrope) [16:41:52] Woo. [16:41:55] Yay [16:42:05] edsanders: Thank you. :-) [16:42:54] lol, I pasted in the entire team travel planning spreadsheet for Lyon [16:42:58] It took a few seconds but it pasted fine [16:43:21] RoanKattouw: Switch to https://gerrit.wikimedia.org/r/#/c/193026/ for now? [16:43:40] If we don't have reproduction steps… [16:44:04] (03Merged) 10jenkins-bot: Make transaction processing exception-safe [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195361 (https://phabricator.wikimedia.org/T70892) (owner: 10Catrope) [16:44:09] (03PS3) 10Kmenger: ActionSet: Add description for events and clarify method descriptions [oojs/ui] - 10https://gerrit.wikimedia.org/r/195489 [16:44:48] James_F: Yeah I'm switching to that [16:44:54] Already did a little bit of planning together with Krinkle [16:45:05] (03PS1) 10Jforrester: Update VE core submodule to master (e85ed49) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195592 [16:45:12] * James_F saw. [16:45:25] RoanKattouw: ^^^ +2 for Beta Cluster testing? [16:45:27] Krinkle: When you're done messing with relatively minor things like ve.indexOf, could you work on the jquery.client thing that's assigned to you? https://phabricator.wikimedia.org/T91460 [16:45:34] * James_F is a bit nervous this close to the cut. [16:45:35] RoanKattouw: Yeah [16:45:49] (03CR) 10Catrope: [C: 032] Update VE core submodule to master (e85ed49) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195592 (owner: 10Jforrester) [16:45:52] Thanks. [16:46:51] (03CR) 10Krinkle: "ve.init.Target.static.importRules is lacking a 'blacklist' property. And it features an 'external' property that doesn't seem to used by a" [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195582 (owner: 10Krinkle) [16:48:31] (03Merged) 10jenkins-bot: Update VE core submodule to master (e85ed49) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195592 (owner: 10Jforrester) [16:48:32] Krinkle: Ahm, let me see there [16:48:40] I recently removed unused things from importRules [16:49:25] Oh yeah removeStyles was unused [16:49:53] Krinkle: https://gerrit.wikimedia.org/r/#/c/190628/ also removed unused stuff from there [16:50:00] Krinkle: Ed will know all about importRules [16:50:13] Re the external thing, James_F might know because I think he touched some of that code [16:50:18] Specifically the rename from pasteRules to importRules [16:55:08] James_F: Yeah, I'm making the git repo now. Preserving history (only a few dozen commits) [16:55:18] Krinkle: Thanks. [16:55:28] Krinkle: Feel free to come up with a better name, BTW. [16:55:39] Krinkle: Also, is it MIT licensed? If not, we can't use in VE-core… [16:56:06] YAY I GOT IT TO CRASH [16:56:13] RoanKattouw: Woo. How? [16:56:24] facebook.com ; Ctrl+A, Ctrl+C ; local VE; Ctrl+V --> exception [16:56:32] James_F: Yeah, it'll have mediwiki's license initially. The first issue I'll open is for authors to sign MIT [16:56:34] RoanKattouw: Ha. [16:56:51] Krinkle: Meh. Who are the substantive authors? [16:56:59] What are you guys putting in a repo? [16:57:09] RoanKattouw: jquery.client. [16:57:13] https://github.com/wikimedia/mediawiki/commits/master/resources/jquery/jquery.client.js [16:57:15] RoanKattouw: It's currently embedded with VE. [16:57:16] https://github.com/wikimedia/mediawiki/commits/master/resources/src/jquery/jquery.client.js [16:57:21] Oooh right [16:57:35] OK, Trevor/Roan/Timo/Ed/James/Bartosz. [16:57:36] It goes back further in svn, but that's in the Roan/Trevor branch [16:57:44] I think we can convince that lot to relicense. [16:57:56] hm? [16:58:20] oh, yes, MIT. [16:58:23] MatmaRex: Are you OK with us re-licensing jquery.client as MIT so we can use it in VE? [16:58:25] Good. [16:58:29] I am. [16:58:35] edsanders, eh, the scroll in firefox happens after this.emit( 'select', this.selection.clone() ); in ve.dm.Surface#setSelection - but I am losing the track at that point. I put debugger statements on any place I could find that listens and responds to 'select' events from the surface model, none trigger, so I'm clearly missing something... What picks that event up on ve load? [16:58:54] Yeah MIT is fine [16:59:24] Krinkle: Sounds like we've got 100% sign-off. [17:00:32] James_F, are you coming to standup? [17:02:17] mooeypoo: Lila meeting running over. [17:03:12] OK, yes, Services want us to merge https://gerrit.wikimedia.org/r/#/c/194338/ if we're happy with it. [17:03:42] mooeypoo: Is the SF conference room joining the standup? [17:03:47] Oh there you are [17:08:23] lol, it's calling .getClonedElement() on a ve.dm.DocumentNode [17:17:25] RoanKattouw: Hm.. filter-branch sucks for this. I'm instead doing 'git show :path > new/repo/jquery.client.js', and then added my mediawiki core clone as remote and cherry-picked each relevant commit. Eventhough there is no rename in the history to root/jquery.client.js, cherry-picked applies it to the right file. [17:17:43] Quite nice :) [17:18:09] Have to do a few 'git rm *other files*' in between for conflicts, but other than that easy peasy [17:20:08] TrevorP|Away: https://phabricator.wikimedia.org/T91152 is what i need you to comment on [17:25:41] anyone knows if there's a good way / trick to catch whatever is listening to an event, assuming I have no idea *what* listens to that event? [17:26:06] in the Firefox debugger. This is getting frustrating. [17:28:28] (03CR) 10Trevor Parscal: [C: 032] ve.ui.FormatTool: Set deactivateOnSelect = false [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195056 (https://phabricator.wikimedia.org/T91827) (owner: 10Bartosz Dziewoński) [17:29:12] (03PS4) 10Trevor Parscal: ve.ui.Toolbar: Do not remove window event listeners in #setup [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195046 (owner: 10Bartosz Dziewoński) [17:29:23] mooeypoo: OOjs event? a conditional breakpoint somewhere in EventEmitter, probably? [17:29:33] (03PS5) 10Trevor Parscal: ve.ui.Toolbar: Do not remove window event listeners in #setup [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195046 (owner: 10Bartosz Dziewoński) [17:29:58] lol, dude, why the hell is Toolbar using #detach instead of #teardown [17:30:00] :( [17:30:11] TrevorParscal: the toolbar is fubar [17:30:13] i filed a bug [17:30:17] cool [17:30:20] i would actually like to get ed to sign off on those [17:30:25] or whoever touched toolbars last [17:30:32] (03CR) 10Esanders: "'external' is used in ve.ce.Surface. It's so you can apply different sanitization rules to external/internal sources. For the blacklist pr" [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195582 (owner: 10Krinkle) [17:30:34] (03Merged) 10jenkins-bot: ve.ui.FormatTool: Set deactivateOnSelect = false [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195056 (https://phabricator.wikimedia.org/T91827) (owner: 10Bartosz Dziewoński) [17:30:34] VE toolbars* [17:30:36] (03CR) 10Trevor Parscal: [C: 032] ve.ui.Toolbar: Do not remove window event listeners in #setup [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195046 (owner: 10Bartosz Dziewoński) [17:30:50] (03CR) 10Trevor Parscal: [C: 032] Toolbar: Tighten whitespace on narrow displays [oojs/ui] - 10https://gerrit.wikimedia.org/r/193282 (https://phabricator.wikimedia.org/T52227) (owner: 10Bartosz Dziewoński) [17:30:58] (03PS2) 10Trevor Parscal: ve.init.mw.ViewPageTarget: Recalculate toolbar sizes when adding "Save" button [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195049 (https://phabricator.wikimedia.org/T52227) (owner: 10Bartosz Dziewoński) [17:31:07] TrevorParscal: https://phabricator.wikimedia.org/T92282 is about toolbars being a huge mess [17:31:40] hmmm [17:31:43] (03CR) 10Trevor Parscal: [C: 032] ve.init.mw.ViewPageTarget: Recalculate toolbar sizes when adding "Save" button [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195049 (https://phabricator.wikimedia.org/T52227) (owner: 10Bartosz Dziewoński) [17:32:46] (03Merged) 10jenkins-bot: ve.ui.Toolbar: Do not remove window event listeners in #setup [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195046 (owner: 10Bartosz Dziewoński) [17:32:58] (03Merged) 10jenkins-bot: Toolbar: Tighten whitespace on narrow displays [oojs/ui] - 10https://gerrit.wikimedia.org/r/193282 (https://phabricator.wikimedia.org/T52227) (owner: 10Bartosz Dziewoński) [17:34:37] James_F, TrevorParscal, RoanKattouw_away: Noam is pretty sick so I'm taking today off. I'll be on IRC sporadically. I'll crunch perf numbers per James's request. [17:34:47] ori: Thanks. [17:35:38] ori: bummer man, hope he gets well soon [17:35:42] ori: And my sympathies. [17:37:02] Krenair: I assume https://gerrit.wikimedia.org/r/#/c/194338/ isn't too risky and can merge (it'll be switched off in production until after testing), but I leave that call to you. [17:38:05] (03CR) 10Jforrester: [C: 04-2] "OOjs UI dependency not in a released version yet, let alone pulled through." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195049 (https://phabricator.wikimedia.org/T52227) (owner: 10Bartosz Dziewoński) [17:38:58] (03CR) 10Bartosz Dziewoński: "Well, it wouldn't break anything." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195049 (https://phabricator.wikimedia.org/T52227) (owner: 10Bartosz Dziewoński) [17:39:38] MatmaRex: Sure, but its parent commit isn't merged yet so… [17:42:32] man, our catch-alls in toolbars are evil. [17:42:45] also, wikibugs is down, otherwise it would report me griping about it on phab. [17:44:08] James_F, still trying to deal with my inbox :( [17:44:17] Krenair: Yeah, no worries. :-) [17:45:34] Krenair: laudable, but ultimately futile. :> [17:45:48] * James_F grins. [17:47:44] (03CR) 10Mobrovac: [C: 032] Fix acceptLanguage commit a30d6be [services/citoid] - 10https://gerrit.wikimedia.org/r/195025 (owner: 10Mvolz) [17:47:46] (03PS1) 10Bartosz Dziewoński: ve.ui.MoreTextStyleTool: Set autoAddToCatchall = false; [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195606 (https://phabricator.wikimedia.org/T91826) [17:48:04] MatmaRex: Hmm; I tested before and that wasn't needed… [17:48:11] Oooooh. [17:48:17] That's what caused the insert menu issue? [17:48:18] Bah. [17:48:24] yeah [17:48:27] (03CR) 10Jforrester: [C: 032] ve.ui.MoreTextStyleTool: Set autoAddToCatchall = false; [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195606 (https://phabricator.wikimedia.org/T91826) (owner: 10Bartosz Dziewoński) [17:48:31] the tool does appear in it, it's just not visible [17:48:36] by accident [17:48:37] :-D [17:48:48] because it's hacked up [17:48:59] James_F: apologies if you guys are already triaging this, but in case you haven't seen it: I just got this on enwiki: http://i.imgur.com/F7FwLP9.png [17:49:28] ori: Eurgh. I got one of those during the deploy and thereafter they went. [17:49:37] ori: Possibly just cached JS at your end? [17:49:40] still there on hard refresh [17:49:56] Darn [17:49:57] cleared localstorage and browser cache [17:50:06] Not there for me. [17:50:31] ok, wtf. it *just* went away. [17:50:34] In either of my logins. [17:50:34] * ori boggles. [17:50:36] Ha. [17:50:36] (03Merged) 10jenkins-bot: ve.ui.MoreTextStyleTool: Set autoAddToCatchall = false; [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195606 (https://phabricator.wikimedia.org/T91826) (owner: 10Bartosz Dziewoński) [17:50:39] ori: "Magic". [17:52:25] (03PS7) 10Mvolz: Update citoid to master and localsettings.js [services/citoid/deploy] - 10https://gerrit.wikimedia.org/r/194536 (https://phabricator.wikimedia.org/T89756) [17:53:04] James_F: btw, why isn't 'Edit' before 'Edit source'? [17:54:00] ori: It's secondary status whilst it's in Beta. [17:54:12] ori: Soon enough there'll only be one tab, so… [17:54:36] (03CR) 10Mvolz: "Ready for review/merge." [services/citoid/deploy] - 10https://gerrit.wikimedia.org/r/194536 (https://phabricator.wikimedia.org/T89756) (owner: 10Mvolz) [17:54:48] beta opt-in is a signal that the user is interested in trying VE out, so making it primary seems warranted to me [17:55:44] Sure, but sooner it won't be opt-in for the majority of accounts with it. [17:56:32] James_F: Only one tab? [17:56:46] Rosencrantz: Yup. Lots forthcoming. [17:57:01] Rosencrantz: (Months, not weeks.) [17:57:22] Ahh. So is that getting rid of the source edit altogether or is it just moving somewhere else? [17:57:41] (03CR) 10Mvolz: "@Akosiaris, you know better than I do! @mobrovac ?" [services/citoid/deploy] - 10https://gerrit.wikimedia.org/r/195333 (https://phabricator.wikimedia.org/T89873) (owner: 10Alexandros Kosiaris) [17:58:10] Rosencrantz: It's embedding them together in the editor, with full switching back and forth and 'sticky' memory of which one you prefer (so no endless opting out) [17:58:27] Oooooooh ,nice! [18:00:26] So with full switching back, does that mean's it would keep track of wikitext changes and convert them when switching to VE? [18:00:32] err switching to full visual [18:04:13] afaik that is still a limitation of Parsoid [18:04:35] switching from ve to wikitext works, but the opposite has limitations on tracking changes still [18:06:55] I don't quite understand that's a limitation of parsoid. Wouldn't the biggest issue be tracking that a changed happened inside the textbox field? Or is that and possible bad wikitext getting sent to parsoid? [18:07:01] *how that's [18:08:10] RoanKattouw: do you recall the details? something to do with losing some of the meta information on the HTML tags? [18:12:11] MatmaRex: how's it going? [18:12:26] sup [18:12:56] TrevorParscal: i want to know your thoughts on https://phabricator.wikimedia.org/T91152 i [18:13:52] (03CR) 10Trevor Parscal: [C: 032] Do not call Toolbar#initialize twice during editor setup [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195048 (owner: 10Bartosz Dziewoński) [18:16:11] (03Merged) 10jenkins-bot: Do not call Toolbar#initialize twice during editor setup [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195048 (owner: 10Bartosz Dziewoński) [18:18:02] MatmaRex: was looking at this: https://phabricator.wikimedia.org/T91152 [18:18:07] where are we at there? [18:18:48] TrevorParscal: yeah. does what i say there make sense? [18:19:44] haven't started working on it yet because it looks like a time sink and pain, but i think that needs to be done [18:19:53] (03CR) 10Alexandros Kosiaris: "Not sure I actually do. I am not really familiar with citoid's configuration. Zotero exists on zotero.svc.eqiad.wmnet for sure, hence my c" [services/citoid/deploy] - 10https://gerrit.wikimedia.org/r/195333 (https://phabricator.wikimedia.org/T89873) (owner: 10Alexandros Kosiaris) [18:19:56] I think you are making sense [18:20:14] TrevorParscal: and related question. what should dictate the font-size of OOUI widgets in MediaWiki – the OOUI theme, or the MediaWiki skin? [18:20:22] I just want to see us remove all base font-size adjustments from OOjs UI and rely on integration code to size things appropriately [18:20:38] the skin [18:20:51] (right now it's basically neither, with VisualEditor's integration patching over the biggest issues) [18:21:07] MediaWiki skins need an overlay div at the end with font-size adjustment equivalent to the body text size [18:21:09] right. so, we should start using 0.875em for everything on Vector, rather than 0.8em? [18:21:48] yeah, well I think the problem started when the typography refresh was botched and they did an incomplete changeover from 0.8 to 0.875 [18:22:05] but also, hardcoding base font sizes in OOjs UI is crap [18:22:10] and are we going to do anything about how that affects icons etc.? they're SVG, but still aligned to pixel grid [18:23:14] what should be happening is that the icons fit perfectly @ 0.875em [18:24:03] in cases where people use the library at 1em, they can provide differently fit icons (3rd party use) or are probably on a high-dpi screen anyway (mobile) [18:24:19] i bet the "interns" are going to have fun re-aligning the fifty-odd icons again [18:25:17] no, they don't need to be realigned [18:25:19] but yeah, we definitely need to state somewhere what font-size each OOUI theme expects you to use for your website [18:25:28] yeah [18:26:08] omg I found it [18:26:14] * mooeypoo shoots Firefox in triumph [18:26:25] mooeypoo: Success? [18:26:26] $( rangeSelection.start.node ).closest( '[contenteditable=true]' ).focus(); <-- THIS. [18:26:38] This scrolls the page down. [18:26:47] * mooeypoo shakes her fists to the air [18:26:48] TrevorParscal: i think the wikifont icons' pixel sizes were adjusted to 0.8em? are we going to ignore grid alignment, make them smaller in comparison to text (=keep current size), or just disregard aligning? [18:27:11] (that's assuming we use 0.875 for Vector, and that Vector is our primary skin-target) [18:27:23] we use 0.875 for OOUI widgets in Vector [18:27:25] * [18:28:58] hm, so it seems like this actually focuses on the NODE in rangeSelection.start.node rather than the content editable [18:29:09] and the node at that point in a big document is somewhere in the middle. [18:37:23] [19:26] mooeypoo $( rangeSelection.start.node ).closest( '[contenteditable=true]' ).focus(); <-- THIS. [18:37:45] RoanKattouw, it seems to not do the "closest" very well [18:37:46] mooeypoo: What are the values of rangeSelection.start.node and $( rangeSelection.start.node ).closest( '[contenteditable=true]' ) ? [18:37:59] Are they the same? :O [18:37:59] instead, we get the element for the node in rangeSelection.start.node [18:38:17] $( rangeSelection.start.node ).closest( '[contentEditable=true]' ) [18:38:23] Object { 0: , length: 1, prevObject: Object, context: #text "Some test cases for VisualEditor." } [18:38:27] ^ see? [18:38:36] that's not the content editable itself, it's a node inside it [18:38:36] Well yes [18:38:39] That's the documentNode div [18:38:47] Wait is that not the one that has cE=true on it? [18:38:59] it's inside the contentEditable [18:39:08] (03CR) 10Mobrovac: [C: 04-1] "I don't like localsettings.js being here at all. It should be puppetised instead, IMHO. We already have operations/puppet/modules/citoid, " (031 comment) [services/citoid/deploy] - 10https://gerrit.wikimedia.org/r/194536 (https://phabricator.wikimedia.org/T89756) (owner: 10Mvolz) [18:39:11] No, it is the contentEditable [18:39:20] RoanKattouw, it is too. [18:39:21] i am confused [18:39:36]
[18:39:44] RoanKattouw, I'm using this page as wikitext: https://www.mediawiki.org/wiki/VisualEditor:Basic_example_worksheet [18:40:12] OK [18:40:16] What was rangeSelection.start.node ? [18:40:44] #text "Some test cases for VisualEditor." [18:40:48] which is correct [18:40:55] OK, yeah [18:41:02] But it's weird that focusing it would scroll to the bottom [18:41:10] not to the bottom, halfway down [18:41:13] Or does it only scroll halfway down? [18:41:15] not EVEN bottom [18:41:15] Oh, hmm [18:41:26] (03CR) 10Alexandros Kosiaris: "@Mobrovac, yup, you are absolutely right about localsettings.js" (031 comment) [services/citoid/deploy] - 10https://gerrit.wikimedia.org/r/194536 (https://phabricator.wikimedia.org/T89756) (owner: 10Mvolz) [18:42:00] RoanKattouw, it scrolls halfway down to a duplicate of the page [18:42:21] RoanKattouw, it seems that the way we're initializing the content editable on firefox, it comes after the actual content -- is that making sense? And only later we bring it up? [18:42:26] In that page, I hve *double* page [18:43:06] one after the other, same content, only I get focused on the latter half at this point (and I can see in debug mode the nodes are marked) [18:43:07] Ooooooh [18:43:18] so then it scrolls to the contentEditable [18:43:26] mooeypoo: Are you saying you have two copies of the page, and it scrolls down to the bottom one, because that one is editable? [18:43:27] but the contentEditable is down the page [18:43:31] yes [18:43:39] RoanKattouw, I can share my screen if you want to see it [18:43:46] No it's OK I understand the bug now [18:43:58] Looking to see if this is what Krinkle touched today [18:44:10] It's not a bug from today [18:44:23] No that was about deactivation [18:44:33] No, I merged something of his today that fixed something like this [18:44:33] https://phabricator.wikimedia.org/T90420 [18:44:36] But not precisely this [18:44:54] > bash: if [[ "blabla"=="somethingelse" ]]; then echo fail; fi [18:45:06] Always returns true [18:45:11] only works when there are spaces around == [18:45:14] because you know, bash arguments [18:45:18] Lovely [18:45:22] :-) [18:47:27] MatmaRex: let me look at the icon sizing stuff [18:47:34] RoanKattouw, I documented what I found, I hope taht makes sense https://phabricator.wikimedia.org/T90420 [18:48:03] I am not sure how to fix this. Can we have this focus wait until we're at the actual end of the process and the CE is at the top? [18:48:27] That's what I'm looking at [18:48:55] that also explains why it wouldn't happen if you directly go to the page with ?veaction=edit [18:50:17] MatmaRex: ok, so here's the trick, the icons don't have to change size, we just adjust their span boxes to still end up being 1:1 px ratio at @0.875 [18:51:13] mooeypoo: How does that explain it? That's what still puzzles me [18:51:27] alright. that'd be nice indeed. [18:51:29] mooeypoo: I've responded on the task (wikibugs bot is AWOL) [18:51:53] because it's a 1px font size change, the relative size of the icons and the text will not be a problem [18:51:58] RoanKattouw, maybe I misunderstanding how VE loads, but I thought that if you go directly, it has nothing yet to overlay over? it rebuilds while loading the page, which would make the CE already at the top of the page [18:52:02] or ... am I wrong? [18:52:02] it will be inperceptable [18:52:05] No that's wrong [18:52:12] My best guess is that the edit link has focus [18:52:24] after you click it [18:52:33] hm [18:53:17] MatmaRex: if the base font is 0.875em, the icon boxes should be 1.7em (from testing, not calculating) [18:53:33] instead of 0.8em base and 1.875em icon boxes (current) [18:53:36] (03CR) 10Mobrovac: "Ok, will send a patch your way today or tomorrow for ops/puppet for this." [services/citoid/deploy] - 10https://gerrit.wikimedia.org/r/194536 (https://phabricator.wikimedia.org/T89756) (owner: 10Mvolz) [18:54:09] OK the stack trace says it comes from the surface focus call in onSurfaceReady, as I had hoped [18:54:22] >> 0.875 * 1.7 [18:54:23] MatmaRex: (number) 1.4875 [18:54:25] >> 0.8 * 1.875 [18:54:26] MatmaRex: (number) 1.5 [18:54:35] close enough. ;) [18:54:41] >> 0.875 * 1.725 [18:54:42] TrevorParscal: (number) 1.5093750000000001 [18:54:56] mooeypoo: since you're working with FF - in FF the toolbar does not stay in focus when the page scrolls down... something of the recent commits did it? [18:55:22] Yup it's the edit tab having focus [18:55:37] >> 1.5 / 0.875 [18:55:38] TrevorParscal: (number) 1.7142857142857142 [18:55:43] that's the EXACT number [18:55:51] mooeypoo: If you click Edit, then click in the blank space between the Discussion and Read tabs, then by the time the editor loads nothing will have focus, and I confirmed the page doesn't scroll in that case [18:56:01] yeah [18:56:13] i think that at least Chrome truncates at 2 decimal places [18:56:14] and what a lovely number it is [18:56:37] TrevorParscal: Have I ever told you how that number is related to my favorite number? :D [18:56:38] i wonder if it'll look good, but hopefully [18:56:43] 1.71 should be fine, we can add a comment above it saying: // 1.5 / 0.875 [18:56:48] (I'm not even joking, I can tell immediately that that number is 1 5/7) [18:56:50] RoanKattouw, hm, interesting. Hang on, I'm working on the call stack for you [18:56:56] mooeypoo: nm I got it already [18:56:57] i'm looking at it in a browser, works great [18:57:18] oh, neat [18:57:23] mooeypoo: Sorry for asking you to do it, I realized I could easily have done that myself so I did it [18:57:31] TrevorParscal: by the way, have you looked at how our SVG icons render in IE? [18:57:40] no [18:57:44] i'm guessing badly [18:57:59] mooeypoo: Does moving $( 'html' ).removeClass( 've-activating' ).addClass( 've-active' ); to right before the surface focus call in onSurfaceReady fix it? [18:58:23] That in itself isn't a clean way to fix it, but if it does, all you have to do is pick Timo's brain about a clean way to do it [18:58:29] TrevorParscal: http://i.imgur.com/zO7ZwMu.png [18:58:57] MatmaRex: What about with the new icon packs? [18:59:18] it's the background-size: contain; [18:59:27] James_F: about the same, presumably [18:59:35] yeah, I was going to suggest taking a screenshot without that rule [19:00:37] (and the em sizing computing to 23.93px rather than 24) [19:00:49] RoanKattouw, ... I just noticed you said nevermind. Well, the stack trace is in the bug report :P [19:02:31] mooeypoo: Didyou see my suggestion for a potential fix? [19:02:33] RoanKattouw, wait, you mean inside ve.init.mw.ViewPageTarget.prototype.activate ? the call already inside the $.when() ? [19:02:37] Yeah [19:02:41] The one in the $.when() [19:02:51] Move that to right before .focus() is called in onSurfaceReady [19:03:01] (Which as you may notice is before activatingDeferred is resolved, so is earlier) [19:03:53] yeah [19:05:20] that works, RoanKattouw [19:05:27] but does it make sense? [19:05:34] Argh, I accidentally closed the tab where I had reproduced the Facebook paste thing [19:05:37] And now I can't reproduce it any more [19:05:54] mooeypoo: It's more likely that moving the focus call is the right thing to do [19:06:01] But Krinkle is the expert on this, you should talk to him [19:06:25] I was about to say, shouldn't we moive the focus into the $.when [19:08:26] Probably [19:08:35] I don't know what adverse affects that might have [19:08:37] You should ask Krinkle and edsanders|away [19:08:51] RoanKattouw, I'm trying it now in any case [19:08:54] Or put it in Gerrit and get them to review :) [19:08:59] exactly my thought :P [19:09:10] that bug was annoying, firefox kept crashing on the debugger stuff [19:09:41] okay, one adverse effect I already see makes no sense at all [19:10:09] the page loaded perfectly, but the spinner is still there [19:10:13] wtf [19:10:14] which makes no sense [19:10:24] Let me clear cache and try again [19:10:32] Did you put the addClass/removeClass thing back in its original place? [19:10:36] yes [19:11:13] $.when( this.activatingDeferred, this.toolbarSetupDeferred ).always( function () { [19:11:13] $( 'html' ).removeClass( 've-activating' ).addClass( 've-active' ); [19:11:13] this.getSurface().getView().focus(); [19:11:13] } ); [19:11:20] O.o why would the spinner not vanish... [19:11:34] wait [19:11:37] i'm an idiot [19:11:40] this [19:11:40] "this" [19:11:41] yes yes [19:12:21] hoorah! [19:14:40] (03CR) 10Trevor Parscal: [C: 032] Reduce complexity around onbeforeunload handler [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/177107 (owner: 10Catrope) [19:17:36] (03Merged) 10jenkins-bot: Reduce complexity around onbeforeunload handler [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/177107 (owner: 10Catrope) [19:18:05] [13jquery-client] 15Krinkle created 06master (+30 new commits): 02https://github.com/wikimedia/jquery-client/compare/eb02d625d933^...f7b09f24deea [19:18:05] 13jquery-client/06master 14eb02d62 15Trevor Parscal: initial commit... [19:18:05] 13jquery-client/06master 145a56009 15Trevor Parscal: fixed comment [19:18:05] 13jquery-client/06master 143259ac5 15Trevor Parscal: Changed profile property names - this should be the stable API from now on. [19:18:33] RoanKattouw: This is what I ended up using: https://gist.github.com/Krinkle/d0f31ac974e23f3b0f69 [19:18:39] (03PS1) 10Mooeypoo: Focus the surface after VE is already active [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195631 (https://phabricator.wikimedia.org/T90420) [19:19:19] (03CR) 10Trevor Parscal: [C: 04-1] "This causes the newly inserted reference list to be scrolled to, which may be WAY on the bottom of the page." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/149117 (https://bugzilla.wikimedia.org/54906) (owner: 10Alex Monk) [19:19:39] Krinkle, can you make sure this bug fix isn't going to destroy VE's initialization horribly? From testing it seems to work, but I don't know if there might be other effects...? https://gerrit.wikimedia.org/r/#/c/195631/ [19:20:15] InezK: https://gerrit.wikimedia.org/r/#/c/191658/ [19:20:27] this is going to cause an issue for you guys, no? [19:20:49] I don't know when you guys forked, but you probably don't have this new hook [19:21:00] how much of our integration code are you guys still using? [19:21:01] mooeypoo: checking [19:21:10] Krinkle, thanks! [19:21:20] TrevorParscal: Oh, whoops, that needs to land today. [19:21:42] (03CR) 10Trevor Parscal: "I'm checking with Wikia folks, because I worry this could cause problems for them and I want to give them warning." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/191658 (owner: 10Alex Monk) [19:22:01] i'm not saying it's bad [19:22:12] TrevorParscal: We are still using most of your integration code, with just small Wikia specific overwrites. [19:22:13] I just thought reaching out and mentioning it to Inez would be useful [19:22:19] (03PS2) 10Krinkle: Focus the surface after VE is already active [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195631 (https://phabricator.wikimedia.org/T90420) (owner: 10Mooeypoo) [19:22:20] it might not affect them at all [19:22:28] TrevorParscal: *nods* [19:23:05] (03CR) 10Krinkle: [C: 031] "I can reproduce the bug on latest master. Confirmed this change fixes it for me." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195631 (https://phabricator.wikimedia.org/T90420) (owner: 10Mooeypoo) [19:23:30] InezK: so, do you have SkinEditSectionLinks in your fork of MediaWiki? [19:23:42] James_F: Boostrapping lint and testing now [19:24:27] (03CR) 10Bartosz Dziewoński: "Needs rebase, by the way." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/191658 (owner: 10Alex Monk) [19:24:56] we don't [19:25:19] we handle loading VE for sections completely in JS - we attach click handlers [19:25:26] we don't modify URLs generated in PHP [19:25:43] Krinkle: Thanks. [19:33:23] (03PS7) 10Jforrester: Use new SkinEditSectionLinks hook instead of DoEditSectionLink [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/191658 (owner: 10Alex Monk) [19:34:20] (03CR) 10jenkins-bot: [V: 04-1] Use new SkinEditSectionLinks hook instead of DoEditSectionLink [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/191658 (owner: 10Alex Monk) [19:35:10] (03PS1) 10Catrope: Remove unused iconModuleStyles property [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195639 [19:35:47] lunch, brb [19:36:14] (03PS8) 10Jforrester: Use new SkinEditSectionLinks hook instead of DoEditSectionLink [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/191658 (owner: 10Alex Monk) [19:36:45] (03CR) 10Jforrester: "PS7 is a rebase; PS8 is a James-is-a-moron fix-up. :-)" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/191658 (owner: 10Alex Monk) [19:38:22] RoanKattouw: I assume iconModuleStyles wouldn't be useful in OOjs UI icon pack land? [19:40:18] (03CR) 10Jforrester: "Would it be helpful for the commit to use the same format we use for VE sub-module pulls? See e.g. https://gerrit.wikimedia.org/r/#/c/1955" [services/citoid/deploy] - 10https://gerrit.wikimedia.org/r/194536 (https://phabricator.wikimedia.org/T89756) (owner: 10Mvolz) [19:40:22] I don't know enough about whatever the plans are for OOUI icon support to know that [19:40:29] RoanKattouw: Sure. :-) [19:40:51] But that's a valid question, and that commit should be reviewed by someone who knows the answer to that question [19:41:16] * James_F nods. [19:41:23] * James_F throws it at MatmaRex. [19:46:03] i'm afraid i don't know what are the plans either. the finish line keeps moving. [19:47:12] that property could potentially be useful, but it's not like it'll be difficult to reimplement it then. [19:47:50] My changes make it ever so slightly tricker and the feature was unused, so I decided to kill it rather than bother reimplementing it [19:48:04] yeah, let's do it. [19:48:06] (03CR) 10Bartosz Dziewoński: [C: 032] Remove unused iconModuleStyles property [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195639 (owner: 10Catrope) [19:50:15] (03Merged) 10jenkins-bot: Remove unused iconModuleStyles property [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195639 (owner: 10Catrope) [19:55:30] James_F: Text for copyright line? [19:55:42] Another mit-license team? [19:55:45] * @license The MIT License (MIT); see LICENSE.txt [19:55:49] Or whatever. [19:56:06] James_F: That's only sufficient for distribution of the repo. Not the file itself. [19:56:12] Given it's for distribution in a single file… yeah. [19:56:15] I meant the license file. [19:56:24] Copyright <> Names.. team.. ? [19:56:34] Team. [19:56:39] jQuery Client Team? [19:56:41] Too many names to list comfortably. [19:56:51] I'll add AUTHORS.txt, then. [19:56:56] Can we rename it MediaWiki.Client? [19:57:12] Why "MediaWiki" ? [19:57:33] Why jQuery? [19:57:41] The name if the plugin. [19:57:41] (03CR) 10Trevor Parscal: [C: 032] Use new SkinEditSectionLinks hook instead of DoEditSectionLink [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/191658 (owner: 10Alex Monk) [19:57:47] Thanks TrevorParscal. [19:57:52] yup [19:57:59] Krinkle: That's exactly what I'm suggesting we fix. [19:58:16] James_F: The plugin doesn't relate to MediaWiki. We could make it Wikimedia-related though [19:58:18] (the team name at least) [19:58:25] Wikimedia.client? [19:58:27] * James_F shrugs. [19:58:34] We're not talking about https://code.google.com/p/dnd-file-upload/source/browse/trunk/jquery.client.js?r=4 [19:58:41] We're not talking about http://www.stoimen.com/blog/2009/07/16/jquery-browser-and-os-detection-plugin/ [19:59:04] I mean, e.g. Copyright Wikimedia Engineering and other contributors. [19:59:10] That could work. [19:59:11] But that's just wrong [19:59:19] We don't do CLA [19:59:34] (03CR) 10Alex Monk: [C: 04-1] "Should update the wfUseMW call in VisualEditor.hooks.php" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194338 (https://phabricator.wikimedia.org/T89066) (owner: 10Mobrovac) [19:59:50] (03Merged) 10jenkins-bot: Use new SkinEditSectionLinks hook instead of DoEditSectionLink [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/191658 (owner: 10Alex Monk) [20:00:37] Krinkle: Then FFS call it a team like I suggest 4 minutes ago. :-P [20:00:44] Krenair: Want me to do that so you have a clean +2? [20:00:56] Haven't checked the rest of it yet :p [20:00:58] but okay [20:01:02] * James_F grins. [20:01:17] Copyright 2011-2015 jquery-client maintainers and other contributors. [20:01:33] (03PS5) 10Jforrester: Use the RESTBase back-end if available [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194338 (https://phabricator.wikimedia.org/T89066) (owner: 10Mobrovac) [20:01:49] (03CR) 10Jforrester: "Done." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194338 (https://phabricator.wikimedia.org/T89066) (owner: 10Mobrovac) [20:05:01] 2010 * [20:05:58] James_F: merge https://gerrit.wikimedia.org/r/195647 to fix https://phabricator.wikimedia.org/T92305 ? [20:06:25] Oh, was just about to do that. [20:09:06] No wikibugs and no grrrit-wm makes this a dull channel. :-) [20:15:27] https://phabricator.wikimedia.org/T92310 [20:16:22] Krinkle: Component is not a project type. [20:16:37] https://www.mediawiki.org/wiki/Phabricator/Creating_and_renaming_projects#Type_of_project [20:16:50] Oh. [20:16:53] Well that's silly. [20:17:17] Anyway, done. [20:17:37] James_F: The one-line description will be duplicated in 4 places. Any changes you recommend? [20:17:48] WFM. [20:17:49] (readme, package.json, readme, phabricator, mw.org) [20:17:54] OK [20:19:02] [13jquery-client] 15Krinkle created 06mit-license (+1 new commit): 02https://github.com/wikimedia/jquery-client/commit/a13ccaccf9c2 [20:19:02] 13jquery-client/06mit-license 14a13ccac 15Timo Tijhof: Re-license under MIT [20:27:23] (03PS1) 10Bartosz Dziewoński: Update VE core submodule to master (ad44a9b) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195647 [20:27:29] (03CR) 10Jforrester: [C: 032] Update VE core submodule to master (ad44a9b) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195647 (owner: 10Bartosz Dziewoński) [20:27:43] Late much. [20:27:43] (03CR) 10Bartosz Dziewoński: "I'll poke at the icons." [oojs/ui] - 10https://gerrit.wikimedia.org/r/188289 (owner: 10Jforrester) [20:27:46] (03PS10) 10Bartosz Dziewoński: MediaWiki Theme: Add "Wikicon" icons [oojs/ui] - 10https://gerrit.wikimedia.org/r/188289 (owner: 10Jforrester) [20:27:49] (03Merged) 10jenkins-bot: Update VE core submodule to master (ad44a9b) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195647 (owner: 10Bartosz Dziewoński) [20:27:59] (03CR) 10Legoktm: [C: 04-1] Use the RESTBase back-end if available (033 comments) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194338 (https://phabricator.wikimedia.org/T89066) (owner: 10Mobrovac) [20:31:41] (03CR) 10Mobrovac: Use the RESTBase back-end if available (033 comments) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194338 (https://phabricator.wikimedia.org/T89066) (owner: 10Mobrovac) [20:34:18] (03CR) 10Alex Monk: Use the RESTBase back-end if available (031 comment) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194338 (https://phabricator.wikimedia.org/T89066) (owner: 10Mobrovac) [20:37:12] (03CR) 10Jforrester: [C: 031] ce.RangeState: Check this.node for falseyness, not just null [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195563 (owner: 10Catrope) [20:37:56] Jforrester: "Autofill citations by URL or DOI" Lookup - never returns 'not found/error' response? [20:38:45] (03CR) 10Mobrovac: Use the RESTBase back-end if available (031 comment) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194338 (https://phabricator.wikimedia.org/T89066) (owner: 10Mobrovac) [20:39:20] etonkovi_: Yes? [20:39:34] etonkovi_: mooeypoo is the best to ask if mvolz isn't here. [20:39:59] Jforrester: no error responses for lookup ... [20:40:00] ok [20:41:20] mooeypoo: "Autofill citations by URL or DOI" Lookup - never returns 'not found/error' responses? [20:48:33] Krinkle, I added you to https://phabricator.wikimedia.org/T91608 because I wanted to hear what you thought about putting that data in a file on the server somewhere, then copying it down via RL somehow [20:54:33] Krenair: James_F: Hm.. wasn't Ed working on that in SF? [21:02:13] Krinkle: I'd like to make es5-shim part of the default JS payload in core. Has this been discussed already? [21:05:08] ori: Yes and no [21:05:32] ori: I've considered it, but unless we're worried about bug in modern browsers, we should only load it when needed. [21:05:45] Which the skipFunction takes care of [21:06:18] ori: Note that ES5 includes JSON, but es5-shim does not. [21:06:22] We have a separate 'json' shim. [21:06:40] ori: I'm interested though. Why is using a dependency not sufficient? [21:07:05] As long as nothing uses it in top modules, why would we want to make it load that from the top? [21:07:30] i want to use it in mediawiki.js, and encourage its usage everywhere [21:08:00] ori: Hm.. that's tricky because mediawiki can't conditionally load dependencies for itself. [21:08:14] So we'd download it always with mediawiki and jquery, which is quite a lot of wasted bits. [21:08:21] it's a little like jquery: if you make jquery optional, then people get fancy and try to use the DOM API directly, which is typically not warranted [21:08:39] by loading it everywhere unconditionally, it becomes idiomatic [21:08:43] I disagree with that comparison. es5-shim is not optional like that [21:09:13] i know; i thought about the size, and it's not a lot. it's 6.3k minified and gzipped [21:09:18] It's optional because it's built-in to most browsers. [21:09:23] Not because it's not used everywhere. [21:10:12] i don't think the skip function is worth it, in this case [21:10:17] i can benchmark it [21:10:25] How? [21:10:38] vbench with network throughput throttling [21:10:42] We should optimise for the common case and the future. [21:10:53] Which does not feature ES5 [21:11:08] I imagine in a year or less we'll simply drop support for non ES5 browsers in the startup module and not have the shim. [21:11:36] ok, but why not load it in the interim? [21:11:47] i think the benefit provided by the optimisation here is intangible, and the cost substantial, because it means you always have to be deliberate and careful about using ES5 [21:12:17] the fact that it has the status of an optional library prevents developers from thinking of it as part of the set of idioms provided by the language [21:12:18] What's the benefit? Anything using json or es5-shim only has to indicate a dependency. It's the same as with mw.user, util or mw.api And those are quite idomatic and have no shortage of people using them everywhere. [21:13:03] Its' only the bootstrapping code that a small set of people maintain (mediawiki.js) that's affected in this case. [21:14:13] And while it's not a reason againt Es5 in general, I feel like mediawiki.js should not have to use ES5 features and not using them might encourage better performant code in that hot piece of code. [21:14:22] What would you like to use for example? [21:14:29] because as a developer, if you're working on a module, perhaps it occurs to you to use Array.prototype.some (for example), but if it means stopping what you're doing and editing the module declaration to add an additional dependency, you just go "eh, fuck it" and use a for loop. [21:15:21] I think it's more likely people will use it anyway but forget to add a dependency. And it'll work in most cases thanks to race conditions. That's what I've seen so far with modules that were so common. [21:15:37] both in gadgets and in extensions. [21:15:50] all the more reason to make it a non-issue by loading it by default [21:15:54] Used to be every other week, now like once a month. [21:16:16] i think the mistake here is more innocent than using mw.api without declaring a dependency, because es5 is not a "library", it's just a set of language features [21:16:48] It's a former library some hipsters added to the standard so it's available in browsers. :P [21:18:21] https://docs.google.com/presentation/d/1MtDBNTH1g7CZzhwlJ1raEJagA8qM3uoV7ta6i66bO2M/present?slide=id.g3eb97ca8f_1170 [21:18:31] see the lower graph, "highest correlation to render time" [21:18:50] the correlation of JS xfer size to render time is miniscule [21:19:28] How do you pronounce xfer? [21:19:29] the browser already has an open connection to fetch jquery &c.; adding 6k to that payload won't measurably affect page performance, imo [21:19:37] oh it's just short for "transfer" :) [21:19:57] Yeah, I figured, but don't get the x [21:20:19] yeah, it's a funny quirk of English, i hadn't noticed it before [21:20:36] k [21:20:36] What is this based on? [21:20:45] (the slide) [21:20:45] science! [21:20:50] j/k, i dunno [21:20:58] I imagine on mobile an extra 7kb in the blocking request could be significant [21:21:31] i mean, i think it's worth being careful, and it's worth doing some measurements [21:21:43] etonkovi_, it shouldn't return not found, no, because if it's not found, the tool makes it into a plain website citation. [21:21:56] but before i go about collecting measurements, i wanted to get a sanity-check from you [21:38:37] ori: I'm okay with adding it to the startup payload. [21:39:18] Note that we might be able to keep the conditional. The startup payload is a document.write (yuck), but that could be fragmented with our current es5-shim check in an if-conditional [21:39:27] It would add 2 lines of code including a load.php url. [21:39:33] But totally doable if we want to [21:39:55] Krinkle: He never did a commit, just chat. [21:39:59] oh, interesting. I hadn't thought of that. [21:40:00] ori: The reason I'd like this is that es5-shim can be quite aggressive when it comes to overriding native methods. It even overrides methods in latest versions of many modern browsers. [21:40:42] So loading it unconditionally means it *will* overwrite methods. Arguably for the better, but I'm not sure we want to go there. [21:41:06] it would change behaviour anyway. [21:41:26] James_F: Who did? [21:41:30] Krinkle: Ed. [21:41:47] James_F: I saw him writing code and re-arranging the file and all. [21:42:11] I was sitting next to it and developing side-by-side for a few minutes [21:42:14] to him* [21:42:38] Krinkle: Oh. Well, I asked him two weeks ago and he said nothing… [21:42:45] OK [21:43:45] so instead of if ( isCompatible() ) { document.write( /* script tag */ ); }, it would be if ( isCompatible() ) { document.write( needsShim() ? scriptTagWithShimInURL : scriptTagWithoutShim ); [21:43:51] that's what you have in mind, right? [21:46:47] don't you need to account for 3, not 2 scenarios, "isCompatible()", "isCompatibleWithShim()" and "else" [21:47:30] this has three scenarios, it just uses a ternary to branch rather than a nested if [21:48:50] no, that's isCompatible ? assume it's only incompatible in a way that can be shimmed but should still get js : assume all is well [21:49:16] so, isCompatible has to include shimmable [21:49:29] which means you have to make isCompatible() a superset of needsShim() [21:49:50] ori: Yeah, something like that [21:49:55] so... i guess it could work, but it seems confusing that isCompatible returns true when in fact it is only compatible with a shim [21:50:24] it should probably just be one function with three possible return values [21:50:48] TrevorParscal: No, isCompatible is our base support. The browser blacklist. [21:50:50] ryasmeen: https://phabricator.wikimedia.org/maniphest/query/XBDbYc6t9vOs/#R and https://phabricator.wikimedia.org/P386 [21:51:09] TrevorParscal: That one has no else-branch. If not supported we load nothing. And subsequent script tags wrap if(window.mw) [21:51:40] Krinkle: right, so isCompatible returns true, even when it's only compatible when shimmed, so it lies [21:51:48] TrevorParscal: We'd split the url written by 2 scenarios. Either jquery/mediawiki or jquery/es5-shim/mediawiki [21:52:09] TrevorParscal: I don't see what you mean. [21:52:12] ryasmeen: And https://github.com/wikimedia/oojs-ui/blob/v0.9.0/History.md#v090--2014-03-04 https://phabricator.wikimedia.org/maniphest/query/oZ8hL0ntjCI0/#R [21:52:36] Krinkle: I +1 (in a bit more formal language) :D [21:52:36] Krinkle: isCompatible() ? then choose whether to shim or not : do nothing [21:52:52] means isCompatible() returns true, in cases where it's not actually compatible [21:52:52] TrevorParscal: Our browser support doesn't change based on this. [21:53:09] We are compatible with those browsers. [21:53:14] We just use a shim for some extra methods. [21:53:29] We also load jQuery. Are we not compatible with browsers that don't ship jQuery in binary? [21:53:31] ok, well it seems confusing to me to write it that way, but I'll leave it to you [21:54:27] TrevorParscal: It could use documentation, but the definition of compatible here is whether we support the browser. [21:54:41] with regards to its JavaScript engine and DOM etc. [21:55:13] mooeypoo: https://phabricator.wikimedia.org/diffusion/GOJU/history/master/ [21:55:16] We still don't support IE6, IE7, Firefox 1 etc. [21:56:25] Firefox 0.9.0 support or die trying! [22:06:17] TrevorParscal: You having fun in your merge-spree? :-) [22:06:35] trying to tackle what I can [22:06:37] good time [22:06:39] *times [22:06:47] :-) [22:06:57] TrevorParscal: Can I steal you for bug fixes in your own code? [22:07:22] TrevorParscal: https://phabricator.wikimedia.org/maniphest/query/HTnRkZY1LxXJ/#R [22:09:03] sure [22:09:25] TrevorParscal: And did you not-merge https://gerrit.wikimedia.org/r/#/c/194197/ intentionally? You +2'ed a patch sitting on top of it. [22:09:54] i was testing that before you interrupted me :) [22:10:14] Ha. :-) [22:10:29] This is what happens when we have no bot to ping me. :-) [22:14:52] mooeypoo: https://gerrit.wikimedia.org/r/195666 (!) [22:18:21] \o/ [22:21:36] TrevorParscal, GroupWidget doesn't extend Widget, but does extend GroupElement which doesn't extend Element... and the result is that it cannot be 'toggle()' [22:21:56] GroupWidget and GroupElement are mixins [22:21:56] TrevorParscal, I need this specifically for Citoid, but before I extend and do my own thing -- should I add this functionality to GroupWidget? [22:22:00] * Krenair kicks grrrit-wm [22:22:13] they are intended to be used in specific ways [22:22:17] TrevorParscal, oh wait [22:22:22] and thus, you can assume they have been used correctly [22:22:36] and use things like toggle() in GroupWidget [22:22:41] hm, okay so I have to create my own widget that will mixin GroupWidget [22:23:03] GroupWidget is not a concrete class, so yes [22:23:08] yeah [22:23:09] Krenair: Whee. [22:23:26] hence, line #7, @abstract [22:24:28] TrevorParscal, yes, sorry, i missed that [22:24:30] GroupWidget is meant to be used with ItemWidget so that when the GroupWidget is disabled, the ItemWidget children are too [22:24:53] otherwise, if you don't need that sort of functionality, just mixin GroupElement for basic add/remove/clear items stuff [22:26:23] (03Merged) 10jenkins-bot: ce.RangeState: Check this.node for falseyness, not just null [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195563 (owner: 10Catrope) [22:26:25] (03PS11) 10Bartosz Dziewoński: MediaWiki Theme: Add "Wikicon" icons [oojs/ui] - 10https://gerrit.wikimedia.org/r/188289 (owner: 10Jforrester) [22:26:27] (03CR) 10Bartosz Dziewoński: "They should all rasterize correctly now." [oojs/ui] - 10https://gerrit.wikimedia.org/r/188289 (owner: 10Jforrester) [22:26:31] (03CR) 10Bartosz Dziewoński: "(Some of the SVG files are pretty messy, but I am definitely not planning to edit 200 files manually to tweak stuff.)" [oojs/ui] - 10https://gerrit.wikimedia.org/r/188289 (owner: 10Jforrester) [22:26:59] (03CR) 10Trevor Parscal: [C: 032] Focus the surface after VE is already active [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195631 (https://phabricator.wikimedia.org/T90420) (owner: 10Mooeypoo) [22:27:21] (03Merged) 10jenkins-bot: Focus the surface after VE is already active [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195631 (https://phabricator.wikimedia.org/T90420) (owner: 10Mooeypoo) [22:27:31] (03PS8) 10Trevor Parscal: Use widgets for wiki-page-name and wiki-user-name parameters [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/169936 (https://bugzilla.wikimedia.org/53613) (owner: 10Alex Monk) [22:27:33] (03CR) 10jenkins-bot: [V: 04-1] Use widgets for wiki-page-name and wiki-user-name parameters [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/169936 (https://bugzilla.wikimedia.org/53613) (owner: 10Alex Monk) [22:27:35] (03CR) 10Mobrovac: "@Jforrester very much so! Please illuminate us" [services/citoid/deploy] - 10https://gerrit.wikimedia.org/r/194536 (https://phabricator.wikimedia.org/T89756) (owner: 10Mvolz) [22:28:00] mooeypoo: TrevorParscal: i filed a bug to kill these two, by the way. :) [22:28:16] James_F, gwicke: So now we need restbase config [22:28:29] (03PS1) 10Cmcmahon: [BrowserTest] fix edit test [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195683 [22:28:33] (03CR) 10Trevor Parscal: [C: 04-1] ActionSet: Add description for events and clarify method descriptions (031 comment) [oojs/ui] - 10https://gerrit.wikimedia.org/r/195489 (owner: 10Kmenger) [22:28:37] (03CR) 10Trevor Parscal: [C: 032] Icon width should only be applied if there is an icon [oojs/ui] - 10https://gerrit.wikimedia.org/r/195320 (owner: 10Mooeypoo) [22:28:40] (03CR) 10Trevor Parscal: [C: 032] ButtonElement: Use #setButtonElement correctly [oojs/ui] - 10https://gerrit.wikimedia.org/r/192182 (owner: 10Krinkle) [22:28:41] (03Merged) 10jenkins-bot: Icon width should only be applied if there is an icon [oojs/ui] - 10https://gerrit.wikimedia.org/r/195320 (owner: 10Mooeypoo) [22:28:48] (03Merged) 10jenkins-bot: ButtonElement: Use #setButtonElement correctly [oojs/ui] - 10https://gerrit.wikimedia.org/r/192182 (owner: 10Krinkle) [22:28:57] (03Merged) 10jenkins-bot: PanelLayout: Add 'framed' config option [oojs/ui] - 10https://gerrit.wikimedia.org/r/194976 (https://phabricator.wikimedia.org/T91153) (owner: 10Bartosz Dziewoński) [22:29:01] (03CR) 10Jforrester: "Fine." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195049 (https://phabricator.wikimedia.org/T52227) (owner: 10Bartosz Dziewoński) [22:29:03] (03CR) 10Jforrester: [C: 032] "…" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195049 (https://phabricator.wikimedia.org/T52227) (owner: 10Bartosz Dziewoński) [22:29:06] (03CR) 10Jforrester: [C: 031] Use the RESTBase back-end if available [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194338 (https://phabricator.wikimedia.org/T89066) (owner: 10Mobrovac) [22:29:08] (03Merged) 10jenkins-bot: ve.init.mw.ViewPageTarget: Recalculate toolbar sizes when adding "Save" button [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195049 (https://phabricator.wikimedia.org/T52227) (owner: 10Bartosz Dziewoński) [22:29:15] MatmaRex, eh, kill them and replace them with what? [22:29:16] (03CR) 10Trevor Parscal: [C: 032] TextInputWidget: Adjust size and label on first focus, too [oojs/ui] - 10https://gerrit.wikimedia.org/r/194571 (https://phabricator.wikimedia.org/T89399) (owner: 10Bartosz Dziewoński) [22:29:25] (03PS1) 10TheDJ: AUTHORS: Add my own name [oojs/ui] - 10https://gerrit.wikimedia.org/r/195771 [22:29:28] MatmaRex, as in, what should I use if I need a group element widget? [22:30:27] Also, gwicke, I am on the "Editing" team, not the "MediaWiki Core" team :p [22:30:36] mooeypoo: just mixin GroupElement [22:30:39] in the container widget [22:30:45] mooeypoo: i don't understand? [22:30:56] yeah, just mixin GroupElement probably [22:31:02] she's saying, if we get rid of those classes, what should she use [22:31:33] nothing [22:31:33] ok so not the GroupWidget with itemWidgets, but the GroupElement mixin [22:31:37] ... nothing? [22:31:47] hehe what? [22:32:01] (03CR) 10Cmcmahon: [C: 032] [BrowserTest] fix edit test [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195683 (owner: 10Cmcmahon) [22:32:08] we could probably fold the functionality into GroupElement, implementing it with some events [22:32:17] Krenair: oops ;) [22:32:29] but even now you most likely don't need ItemWidget nor GroupWidget [22:32:30] MatmaRex, right, so use GroupElement, not GroupWidget -- but you're not going to get rid of *both* of those, right? [22:32:32] Krenair: it's all very fluid these days ;P [22:32:38] no, no [22:32:44] ok got scared for a second [22:32:44] mooeypoo: no, he's suggesting GroupWidget was a mistake [22:32:46] not GroupElement [22:32:49] i was to get rid of ItemWidget and GroupWidget :) [22:32:54] oh oh those two [22:32:57] okay, yeah [22:32:58] GroupElement is a stroke of genius, and ye shall bow down before it [22:33:04] lol [22:33:10] :D :D [22:33:17] I am relieved to hear I won't have to not use it [22:33:30] TrevorParscal: Points update: 1100 "done" (198 tickets), 426 "accepted" (25 tickets), 186 "polish" (14 tickets), and 12 dependencies. [22:34:08] (03PS1) 10Bartosz Dziewoński: Revert "Prevent form submission in FragmentInspector" [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195780 [22:34:09] (03Merged) 10jenkins-bot: [BrowserTest] fix edit test [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195683 (owner: 10Cmcmahon) [22:35:42] James_F: nice [22:36:07] so, we are something like 2/3 the way through with 3 or so weeks left [22:37:09] MatmaRex: https://gerrit.wikimedia.org/r/#/c/194197/4 either proves you are extremely smart, or extremely crazy [22:38:03] TrevorParscal: nah, just desperate [22:38:19] ha ha [22:38:56] so, basically you figured out a way to know when it's attached because presumably to attach it, that involves detaching it from it's parent, offscreen as it may be [22:39:22] a way to *usually* know when it's attached [22:39:31] and that's clever, but I didn't quite grasp how that also detects when it's been attached to something else offscreen, and then that has been attached to the document [22:39:43] any more clever DOM operations will break the approach [22:39:50] ok, so you knowlingly don't cover that case [22:40:03] TrevorParscal: i'm wondering if that's not overengineered, really. just calling #onElementAttach from a setTimeout(…, 0) would probably be just as effective in practice, since we basically always attach the widget somewhere right after creating it. [22:40:04] I suspect thought that the case I'm talking about isn't all that clever or strange [22:40:18] lol [22:40:20] yeah, so that one is covered i think [22:40:50] hmm [22:40:54] the first "offscreen" attachment is detected [22:41:09] then the code checks if the tree of $element is attached to the document root [22:41:27] if it is, that's over; but if it's not, it traverses up to the root of the tree, and now watches that instead of $element. [22:42:36] okay seriously, I need a gadget that slaps me over the head every time I forget the word "new" in front of OO.ui objects. [22:46:16] Hello. [22:53:41] hi James_F this terribly trivial report is now two weeks old and I wonder if you'd care to reject it. I assume it's a feature not a bug https://phabricator.wikimedia.org/T90654 [22:54:07] chrismcmahon: Oh. Yes. That's a feature. [22:54:23] chrismcmahon: Specifically, it doesn't work as a redirect unless it's at the top. :-) [22:54:36] James_F: I updated the test also: https://gerrit.wikimedia.org/r/#/c/195786/ [22:55:03] * James_F nods. [22:55:06] Want me to +2? [22:55:11] LGTM. [22:55:41] James_F: please, if you don't mind. Rummana was catching up on another project today. [22:55:46] * James_F nods. [22:59:43] chrismcmahon: Merged. [22:59:48] thanks James_F [22:59:59] Happy to help; sorry for slow response on that task. [23:00:05] now to hack the search results mess and we'll be back to green :-) [23:00:09] :-D [23:03:15] Krenair: Do you have any idea why https://gerrit.wikimedia.org/r/#/c/195580/ fails? I don't feel right push it to SWAT with failures, but… [23:03:49] 16:08:51 ./composer.json not found. [23:03:49] hn [23:03:50] hm* [23:04:14] Yeah. [23:04:23] karma times out [23:04:31] Always. Ignore that one. :-) [23:04:42] There's a reason it's non-voting. [23:04:49] About 200 points assigned to Krinkle to try to fix it. :-( [23:05:17] no idea. [23:05:23] * James_F nods. [23:05:25] 16:09:15 TypeError: 'undefined' is not a valid argument for 'instanceof' (evaluating 'targetFn.prototype instanceof originFn') TypeError: 'undefined' is not a valid argument for 'instanceof' (evaluating 'targetFn.prototype instanceof originFn') [23:05:30] That's a genuine error [23:05:34] missing class [23:05:37] that is being extended [23:05:53] https://integration.wikimedia.org/ci/job/mwext-VisualEditor-qunit/14275/console [23:06:01] OK, I'll ask RoanKattouw_away to look over it and push it out in the morning, and to hell with the wikis affected overnight. :-( [23:06:30] Krinkle: Sure, except I can't replicate the error at all locally, and no trace is given to work out what causes it. [23:06:40] I wasn't able to reproduce https://phabricator.wikimedia.org/T91943 exactly [23:06:44] but found a very similar error [23:07:02] James_F: You checked out that wmf branch of VisualEditor with the same submodule version and VE worked? [23:07:06] undoing changes to the group of a reference... does not work properly [23:07:16] Krinkle: I think so. [23:07:38] Krinkle: I might have been doing something wrong. [23:07:39] I can give it a try later, but I'm pretty sure it will fail and it will also fail in production. [23:07:57] Krinkle: You focus on the 10 things you've already got assigned you to. :-) [23:08:06] It's depending on a class that doesn't exist. Probably ve-core or oojs-ui related. [23:08:21] James_F: Yeah :) [23:08:31] Krinkle: It's a VE-core patch that worked fine in cherry-pick there. [23:08:39] Krinkle: I.e. I consider the test failure bollocks. [23:09:02] As for "./composer.json not found." - something will probably need backporting for that as well. Shuld be staight forward. or disable the check for wmf branches. Though that's no longer as possible as it used to be now that the jobs are consolidated. That exception was removed. [23:09:16] Helpful. [23:09:26] REL1_23 branches are just screwed? [23:09:33] No, those are excluded [23:09:47] but we don't want to exclude all new wmf branches from composer-test, just like we use npm-test [23:09:51] * James_F nods. [23:10:07] it's just that VE added it slightly after we added the jobs, and then it became voting, but the wmf branch is from before that. [23:10:53] Maybe we should have a 2 week rule about jobs being introduced before they're made voting. [23:11:12] But… eh. [23:15:11] Krinkle: https://phabricator.wikimedia.org/T91715 would be nice to get done sooner, FWIW. [23:15:39] why does undoing a change to the reference group not update the internallist... [23:17:31] Krenair: There's a bug about that. [23:17:39] at least two [23:18:40] :-) [23:27:10] James_F: nzr: http://i.imgur.com/VYJHafB.png https://gerrit.wikimedia.org/r/#/c/194753/ [23:27:35] nzr: I'm assuming you are Nirzar? sorry if you aren't [23:28:04] James_F: I'm ok with these changes, the line-height quirks go away and the buttons are less wide [23:28:08] your thoughts? [23:28:59] hmm... actually, it looks like the button labels are too high [23:32:26] TrevorParscal, you may like this change https://gerrit.wikimedia.org/r/#/c/195802/ [23:32:43] The bot is apparently dead... hm [23:32:49] James_F, edsanders: So I sort of got the reference group undo thing working... [23:33:26] TrevorParscal, it's the redesign of the citoid option widget to have an "Insert" button like the design specified. I wasn't sure how to get the solid button design, though -- isn't it "constructive"? [23:33:40] watch for attribute changes in MWReferenceNode, if key is listGroup then getInternalList().removeNode( from, listKey, listIndex, this ); getInternalList().removeNode( to, listKey, listIndex, this ); [23:34:08] but I realised that things like listKey and listIndex would change as well [23:34:15] mooeypoo: solid colored buttons are for primary actions [23:34:24] oh thaaaat's how you do it [23:34:25] so, if you want solid, add primary [23:34:45] TrevorParscal, okay, that's what the design pictures show, so I will add it in [23:34:48] but the thing is, if there actually are multiple, we probably should only set one to primary or none [23:34:54] kaity: ^^^ [23:35:03] TrevorParscal, oh, hm, right. [23:35:23] presumably primary suggests a singular action [23:35:33] but, technically we don't restrict this [23:35:44] TrevorParscal, I don't think it looks bad as "just" constructive [23:35:57] but that's for you and kaity to decide. I'll leave it as-is until you take a look [23:36:04] TrevorParscal: sorry whats the question? [23:36:17] kaity: if there are multiple results (the edge case) should we still display the insert buttons as primary? [23:36:36] it makes sense when there is one, for it to be primary [23:36:39] also, TrevorParscal we'll have to discuss how to handle "edit" action (I left the edit button out for now since the behavior is a bit more elaborate; do we open a new dialog? if so, do we still allow staging, or do we start over with the information that exists, etc) [23:36:42] but, when there are multiple... [23:36:57] mooeypoo: yeah, we can work on that in a follow up patch tomorrow [23:37:04] mooeypoo: ideally, it will stage [23:37:06] * mooeypoo nods [23:41:03] TrevorParscal: mooeypoo I spoke with May and we think its fine if there are multiple primary buttons in this edge case [23:41:19] it would be confusing to turn them into quiet buttons [23:41:33] not as clear what to do next [23:41:35] kaity: they wouldn't be quite, just normal, there are 3 modes [23:42:00] quiet (not framed), normal (framed) and primary (framed with a solid bold color in the background and white label text) [23:42:03] TrevorParscal: I think the insert button should always be primary [23:42:31] indeed, do you also think the insert buttonS should always be primary is the question [23:42:55] I'm hearing yes, I'm just making it clear that we are talking about the buttons going from primary to quiet, but from primary to normal [23:43:04] TrevorParscal: hey, sorry i was away [23:43:27] mooeypoo: I think you have your answer, let's go with all primary for now and adjust if it tests poorly [23:43:40] TrevorParscal: i am looking into the line height issue. it looks different in different browsers because of rounding of pixels [23:44:00] TrevorParscal: i will find a way [23:45:02] TrevorParscal, kaity in the current fix -- non-primary http://oi61.tinypic.com/24qpeth.jpg vs primary http://oi60.tinypic.com/2urxsno.jpg [23:45:04] nzr: cool man, as I said, I think it's wrong already, I just noticed it made it worse in chrome at least [23:45:10] I'm sure we can make it consistent somehow [23:45:23] TrevorParscal, kaity okay, I just saw the decision. I'll make them all primary. [23:45:32] TrevorParscal: yeah [23:46:00] mooeypoo: we can test to see if users know what to do, in the new design the lookup button becomes next, and is not visible when they are chosing to insert [23:46:09] sigh, no, my approach was not working [23:46:09] everything was still going out of sync in one way or another [23:46:11] lots of things will be different [23:46:25] nzr: thanks for looking into this man [23:46:56] TrevorParscal, okay, makes sense. I just pushed another patchset with buttons primary, we can adjust later. [23:50:01] mooeypoo: you saw how the new design separates the results and the lookup into screens and uses a next button in the top right to advance? [23:50:12] uh... no? [23:50:15] the next button is going to be an action, as will the back button [23:50:18] am I loking at an old design? [23:50:22] you should look at the task in phabricator [23:50:25] it has the latest [23:50:35] I was... hm [23:50:47] you are doing things toward the new design so far, but you should look at the new design to be sure your next step will also be in line [23:50:52] oh oh [23:51:00] okay I see it now [23:51:02] hm [23:51:31] TrevorParscal, can we have "pages" (OO.ui.PanelLayouts?) in an inspector with different modes? I wasn't sure if we can only do that to a dialog [23:52:30] hm, we should, there's an actionProcess and all. Just verifying. [23:52:36] mooeypoo: sure, just make sure they aren't expanded and that when we switch modes we update the inspector size [23:52:44] * mooeypoo nods [23:53:01] I'll do it in a separate commit to make things easier to review [23:53:07] yes, for sure [23:53:13] It shouldn't be too hard [23:53:13] there are lost of intermediate steps [23:53:29] you know how this stuff works better than most people, I'm sure you will have no trouble [23:53:34] just remember to use "new" :P [23:53:42] :D Thank you [23:53:55] edsanders, around? [23:54:03] hi [23:54:20] I'm trying to work out why undoing a ref group change is so broken [23:54:38] I think the change just isn't affecting the internal list [23:55:17] the dm internal list? [23:55:24] yes [23:56:02] have you made a failing unit test? [23:56:12] no [23:56:28] dm internal list may be garbage, but it does have unit tests [23:57:05] I think it's just not calling anything to update the internal list [23:59:42] TrevorParscal: what browser did you use to take the screenshots?