[00:00:09] mooeypoo: The Firefox guy showed me http://jsfiddle.net/sqwxgn5c/2/ [00:04:14] (03CR) 10Alex Monk: TextInputWidget: Determine validity on blur and change. (031 comment) [oojs/ui] - 10https://gerrit.wikimedia.org/r/161516 (https://bugzilla.wikimedia.org/70957) (owner: 10Alex Monk) [00:04:18] mooeypoo: So basically 44 + 1 + 44 is in fact 88 for sufficiently low values of 443 [00:04:20] *of 44 [00:07:07] 3MediaWiki extensions / 3TemplateData: TempateData: Add an "autofill" attribute for parameters which clients would automatically insert if none specified - 10https://bugzilla.wikimedia.org/51428 (10James Forrester) 5PATC>3RESO/FIX [00:11:24] (03PS5) 10Alex Monk: TextInputWidget: Determine validity on blur and change. [oojs/ui] - 10https://gerrit.wikimedia.org/r/161516 (https://bugzilla.wikimedia.org/70957) [00:13:13] (03PS2) 10Jforrester: [wip] Add 'autovalue' to TemplateData [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/146396 (https://bugzilla.wikimedia.org/52029) (owner: 10Mooeypoo) [00:13:19] (03CR) 10jenkins-bot: [V: 04-1] [wip] Add 'autovalue' to TemplateData [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/146396 (https://bugzilla.wikimedia.org/52029) (owner: 10Mooeypoo) [00:13:39] 3VisualEditor / 3Editing Tools: VisualEditor: Template parameters with "autofill" TemplateData status should be autofilled - 10https://bugzilla.wikimedia.org/52029 (10James Forrester) p:5Low>3High a:3Moriel Schottlender [00:13:52] mooeypoo: Are you working on 146396 already? [00:34:13] James_F, I was going to get back into it [00:34:44] just need to quickly rebase it [00:34:48] will do it after i finish dinner [00:36:02] * James_F nods. [00:36:05] No worries. :-) [00:49:24] (03PS3) 10Mooeypoo: [wip] Add 'autovalue' to TemplateData [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/146396 (https://bugzilla.wikimedia.org/52029) [00:52:25] (03PS4) 10Mooeypoo: Add 'autovalue' to TemplateData [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/146396 (https://bugzilla.wikimedia.org/52029) [00:52:35] James_F, rebased, and I just tested it -- works for me! [00:52:38] so I removed the WIP [00:52:45] Whee. [00:56:53] (03CR) 10Jforrester: [C: 032] Add 'autovalue' to TemplateData [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/146396 (https://bugzilla.wikimedia.org/52029) (owner: 10Mooeypoo) [00:57:01] woot [00:57:19] :) [00:57:24] 3VisualEditor / 3Editing Tools: VisualEditor: Template parameters with "autofill" TemplateData status should be autofilled - 10https://bugzilla.wikimedia.org/52029 (10James Forrester) 5PATC>3RESO/FIX [00:57:42] James_F, curious, has the VE team ever looked at casperjs for test automation as an alternative to the selenium approach? I know it'd be more browser-limited ... but JS-only+faster [00:58:13] Eloquence: I haven't (first I've heard of it). Possibly Krinkle has? [00:58:27] (03Merged) 10jenkins-bot: Add 'autovalue' to TemplateData [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/146396 (https://bugzilla.wikimedia.org/52029) (owner: 10Mooeypoo) [01:03:35] I've vaguely heard of it [01:03:40] But never used it [01:08:02] (03PS1) 10Alex Monk: Use LinkCache instead of an object to store hidden status in MWCategoryWidget [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/161637 [01:08:43] (03PS2) 10Alex Monk: Use LinkCache instead of an object in MWCategoryWidget to store hidden status (to be shared with the input widget) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/161637 [01:10:13] (03CR) 10jenkins-bot: [V: 04-1] Use LinkCache instead of an object in MWCategoryWidget to store hidden status (to be shared with the input widget) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/161637 (owner: 10Alex Monk) [01:12:13] (03PS3) 10Alex Monk: Use LinkCache instead of an object to store hidden status in MWCategoryWidget [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/161637 [01:12:36] alex@alex-desktop:/var/www/MediaWiki/core/extensions/VisualEditor (categorywidgets-cleanup)$ grep getCategoryItemFromValue modules/* -R [01:12:36] modules/ve-mw/ui/widgets/ve.ui.MWCategoryWidget.js: var item = this.input.getCategoryItemFromValue( this.input.getValue() ), [01:12:36] modules/ve-mw/ui/widgets/ve.ui.MWCategoryWidget.js: categoryItem = this.input.getCategoryItemFromValue( value ); [01:12:36] modules/ve-mw/ui/widgets/ve.ui.MWCategoryInputWidget.js:ve.ui.MWCategoryInputWidget.prototype.getCategoryItemFromValue = function ( value ) { [01:12:49] I wonder if I should move that in this commit... [01:13:43] hah [01:13:45] (03CR) 10jenkins-bot: [V: 04-1] Use LinkCache instead of an object to store hidden status in MWCategoryWidget [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/161637 (owner: 10Alex Monk) [01:14:01] Probably [01:14:04] Eloquence: I've heard of it [01:14:19] Urgh, PS3 overwrote the commit message change I made in PS2. [01:14:26] I mean maybe CategoryInputWidget could have getCategoryValue() but you definitely shouldn't be getting one value from this.input and then feeding it to another method on this.input [01:14:41] Eloquence: We're using it for the ve-dirtydiff bot [01:14:47] ah, neat [01:15:07] Eloquence: It's not a js-only browsers though, it actually goes on top of PhantomJS/WebKit [01:15:12] yeah I know [01:15:27] it's a bit limited .. but wondering if it'd be useful as a middle layer between unit tests and selenium tests for more teams .. [01:15:51] given execution speed and ease of development [01:16:36] probably better to invest in fixing up the selenium test execution first though [01:16:53] Eloquence: It wouldn't be suitable for integration tests since it only supports PhantomJS (highly outdated version of Safari, not even Chrome) and SlimerJS (current Firefox). No other browsers. [01:17:18] (03PS4) 10Alex Monk: Use LinkCache instead of an object in MWCategoryWidget to store hidden status (to be shared with the input widget) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/161637 [01:17:25] VE already technically isn't supposed to pass its unit tests in PhantomJS because we don't support Safari 3.0 [01:17:33] which is what the WebKit engine in PhantomJS is based on [01:18:04] Yeah and we hack around the fact that PhantomJS doesn't have Function.prototype.bind [01:18:48] I'd be interested in moving our QA/browser tests to using the WebDriver js protocol under nodejs with webdriver adapters, which can be bound to both locally installed browsers on the Jenkins servers (chromium, firefox), as well as remote browsers implementing the same protocol (SauceLabs, BrowserStack) [01:19:16] cutting out the current ruby stack essentially [01:19:47] S mentioned that the Ruby stack is a barrier to entry for devs on the Flow team writing their own tests, so they rely on Chris and team to do it instead [01:20:00] Yeah same here [01:20:03] Yep. [01:20:04] I have tried messing with it [01:20:14] And actually got it to run on my system at one point [01:20:31] There's enough documentation to sort of get it running, but not enough to help you avoid several pitfalls [01:20:32] There's no fundemental problem in its design, but I think there's just no justification for it to be a giant ruby stack. And a lot of the tests are overly specific thus causing a maintenance burdon. [01:20:47] Like, by default it'll run a test that goes to every single production (!) wiki and asserts that VE is enabled there [01:20:54] heh [01:21:14] but yeah, phantomjs being so out of date definitely makes it useless for our purposes .. at least til they switch to blink [01:21:50] I'm considering to implement a much lighter-weight and closer-to-the-metal solution that's runnable on each patch set using a scalable protocol that would run for developers locally in their installed Chrome/Firefox and same for each patch set on Jenkins, and then post-commit using a remote service for additional browser coverage via saucelabs [01:22:05] ohhhhh that'd be awesome [01:22:06] using webdriver navigation and tes scenarios written directly in nodejs [01:22:51] nodejs would be operating from Grunt, and controlling the local or remote browser. [01:23:27] I met up with a bunch of folks at the jQuery Conference last week in Chicago about how we want to move forward there. Embracing webdriver in js was a desirable solution there as well. [01:23:34] For jQuery UI integration testing [01:24:05] *nod* [01:47:33] (03CR) 10Legoktm: "Should this go in a separate repo? It's slightly funny that we're putting PHP in a repo that has "js" in its name." [oojs/ui] - 10https://gerrit.wikimedia.org/r/160996 (owner: 10Trevor Parscal) [01:49:56] (03CR) 10Catrope: "We discussed this, and Trevor said he didn't want to split this into a separate repo because OOphpUI and OOjsUI share the same stylesheets" [oojs/ui] - 10https://gerrit.wikimedia.org/r/160996 (owner: 10Trevor Parscal) [01:53:35] RoanKattouw: okay, that's reasonable. I'm just concerned that later on it might make it more difficult to distribute properly, but I'm not sure if it eventually is supposed to become a proper PHP library with autoloading/composer support/etc. [01:54:09] Well the next big thing there is to take PHP-generated HTML and use JS to bring it to life on the client [01:54:27] So then we'd have the PHP and JS versions of OOUI working in concert [01:54:38] But yeah distribution is something we should figure out [01:58:36] like, should MW be responsible for adding all the classes to $wgAutoloadClasses? or can we just require_once('oojsui/php/autoload.php'); ? [01:58:57] Right [01:59:01] We haven't really figured that out yet [02:01:38] :) [02:41:03] (03CR) 10Krinkle: "The autovalue property is optional and defaults to null, not string. Take care that this only works with the latest version of the Templat" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/146396 (https://bugzilla.wikimedia.org/52029) (owner: 10Mooeypoo) [02:41:07] mooeypoo: ^ [03:32:22] 3MediaWiki / 3Page editing: Revert bug 27860 or add configuration variable to restore "Mark all edits minor by default" preference behaviour - 10https://bugzilla.wikimedia.org/42008#c11 (10George Orwell III) (In reply to Nemo from comment #9) > Dunno what qualifies for "strongly care" but I still think this... [05:52:07] 3VisualEditor / 3Data Model: VisualEditor: In the Re-use reference dialog, search is not returning correct results on the second initialization for a specific case - 10https://bugzilla.wikimedia.org/70886#c14 (10Ritu Swain) (In reply to Rummana Yasmeen from comment #13) > The reason it is showing the other c... [12:02:56] 3VisualEditor: VisualEditor: Some comments are not displayed - 10https://bugzilla.wikimedia.org/71085 (10Elitre) 3NEW p:3Unprio s:3normal a:3None See the 6 hidden comments I placed in https://en.wikipedia.org/w/index.php?title=User:Elitre_(WMF)/SandboxVE&action=edit . Only 2 of them are visible when VE... [12:14:37] (03CR) 10Siebrand: "i18n/L10n reviewed." [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/159312 (owner: 10Oliverb) [13:49:23] (03CR) 10Mooeypoo: "Yes; OOui's InputWidget's 'setValue' method first passes the value to a 'sanitizeValue' method that sanitizes 'null' and 'undefined' value" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/146396 (https://bugzilla.wikimedia.org/52029) (owner: 10Mooeypoo) [13:50:07] Krinkle|detached, ^^ [15:21:22] (03PS1) 10Esanders: Tweak traverseUpstream documentation and code style [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/161668 [17:53:00] (03PS1) 10Seb35: Check the existence of BetaFeatures before adding the 'visualeditor-enable' preference [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/161685 (https://bugzilla.wikimedia.org/71088) [17:53:13] (03CR) 10jenkins-bot: [V: 04-1] Check the existence of BetaFeatures before adding the 'visualeditor-enable' preference [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/161685 (https://bugzilla.wikimedia.org/71088) (owner: 10Seb35) [17:57:07] (03CR) 10Bartosz Dziewoński: [C: 04-1] Check the existence of BetaFeatures before adding the 'visualeditor-enable' preference (031 comment) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/161685 (https://bugzilla.wikimedia.org/71088) (owner: 10Seb35) [18:07:37] (03PS2) 10Seb35: Check the existence of BetaFeatures before adding the 'visualeditor-enable' preference [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/161685 (https://bugzilla.wikimedia.org/71088) [18:10:14] (03CR) 10Seb35: "You’re right. It was added in I76ba788751dacc24cdbc667b4d24f1656f0ade44." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/161685 (https://bugzilla.wikimedia.org/71088) (owner: 10Seb35) [18:10:51] (03CR) 10Bartosz Dziewoński: [C: 032] Check the existence of BetaFeatures before adding the 'visualeditor-enable' preference [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/161685 (https://bugzilla.wikimedia.org/71088) (owner: 10Seb35) [18:12:24] (03Merged) 10jenkins-bot: Check the existence of BetaFeatures before adding the 'visualeditor-enable' preference [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/161685 (https://bugzilla.wikimedia.org/71088) (owner: 10Seb35) [19:25:06] 3MediaWiki / 3Page editing: Collapse templates in editor - 10https://bugzilla.wikimedia.org/28684 (10Jackmcbarn) [20:52:05] MatmaRex: Hmm… [20:53:12] MatmaRex: https://gerrit.wikimedia.org/r/#/c/161685/ looks wrong. [20:53:32] how so? [20:54:46] That code isn't meant to be affected by BF [20:54:56] The BF section should instead also check that condition. [20:55:25] On wikis where VE is on by default, BF is active but isn't in BF. [20:55:40] BF is exclusively for opt-in features, never opt-out ones. [20:55:56] (03PS1) 10Jforrester: Revert "Check the existence of BetaFeatures before adding the 'visualeditor-enable' preference" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/161750 [20:56:22] Want to +2 my revert? [20:57:05] Self +2'ing reverts is so tacky. :-) [20:57:13] (thinking) [20:58:32] Instead this change should be made to the onGetBetaPreferences block. [20:58:46] If it's needed at all – doesn't the other commit fix this anyway? [21:00:52] 3MediaWiki / 3ResourceLoader: CSSMin: Don't generate double rules for IE < 8 when embedding SVG files - 10https://bugzilla.wikimedia.org/71003 (10Krinkle) [21:00:58] (stil thinking, give me a minute. :) ) [21:01:23] * James_F grins. [21:01:57] James_F: per the bug, we can't rely on the order in which the two hooks run, so the original check was not correct [21:02:17] MatmaRex: The original check pre-dates BF by a few years, I believe. [21:02:20] James_F: neither the original nor current is per se needed, thanks to https://gerrit.wikimedia.org/r/161685 in betafeatures [21:02:36] hm, or maybe not [21:02:37] wrong [21:02:37] MatmaRex: Also, we totally can demand correct order in loading. We do it for other things. :-) [21:02:46] *a* check is still needed. [21:02:52] either way [21:03:15] we want to add that preference if BetaFeatures is not present, not if a preference for that already exists [21:03:42] We want to add a preference if the preference doesn't already exist. [21:03:45] so i don't understand why you're reverting [21:03:50] huh? [21:03:52] that's weird. [21:04:05] The BetaFeatures code needs to not add a preference if the default user preference is on. [21:04:11] Because in that case it's not BF-able. [21:04:24] (That's what takes it out of BF status.) [21:05:24] James_F: the current code adds the BF preference even if that is the case, as far as i can see. [21:05:30] If the preference is set true by default, we show a /different/ preference instead, and don't show a BF preference. [21:05:59] visualeditor-betatempdisable [21:06:08] we always hook the hook, and there are no ifs for pref values inside it. [21:07:03] MatmaRex: And yet: https://pl.wikipedia.org/wiki/Specjalna:Preferencje#mw-prefsection-betafeatures [21:07:30] Hmm. Did I do this in config instead? [21:07:31] * James_F looks. [21:07:43] Aha, yes. [21:07:57] In that case I put visualeditor-enable as a hidden preference. [21:08:07] So BF ignores it, as does regular preferences. [21:08:31] ugh. [21:08:41] So… [21:08:44] Hmm. [21:09:30] James_F: i can +2 the revert if it makes someone's life easier, and if you distill this conversation into a commit message :) [21:09:40] * James_F ponders. [21:11:22] (03Abandoned) 10Jforrester: Revert "Check the existence of BetaFeatures before adding the 'visualeditor-enable' preference" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/161750 (owner: 10Jforrester) [21:11:24] 3MediaWiki / 3ResourceLoader: LESS compiler should preserve the position of CSSMin / CSSJanus annotations - 10https://bugzilla.wikimedia.org/54673 (10Bartosz Dziewoński) [21:11:24] 3MediaWiki / 3ResourceLoader: ResourceLoader: Remove custom LESS functions - 10https://bugzilla.wikimedia.org/67368 (10Bartosz Dziewoński) 5PATC>3RESO/FIX [21:11:39] 3MediaWiki / 3ResourceLoader: ResourceLoaderLESSFunctions::embed does not support CSSJanus flipping for RTL, affecting .background-image-svg() and others - 10https://bugzilla.wikimedia.org/66091#c12 (10Bartosz Dziewoński) *** Bug 66773 has been marked as a duplicate of this bug. *** [21:11:47] If we just broke everything, I'm blaming you. :-) [21:12:08] 3MediaWiki / 3ResourceLoader: ResourceLoaderLESSFunctions::embed does not support CSSJanus flipping for RTL, affecting .background-image-svg() and others - 10https://bugzilla.wikimedia.org/66091#c13 (10Bartosz Dziewoński) *** Bug 68326 has been marked as a duplicate of this bug. *** [21:12:08] 3MediaWiki / 3ResourceLoader: External link icon is not flipped in RTL - 10https://bugzilla.wikimedia.org/68326#c7 (10Bartosz Dziewoński) 5PATC>3RESO/DUP *** This bug has been marked as a duplicate of bug 66091 *** [21:12:11] (03PS7) 10Krinkle: build: Create minified JS and CSS distribution files [oojs/ui] - 10https://gerrit.wikimedia.org/r/157315 (owner: 10Jforrester) [21:12:22] 3MediaWiki / 3ResourceLoader: ResourceLoaderLESSFunctions::embed does not support CSSJanus flipping for RTL, affecting .background-image-svg() and others - 10https://bugzilla.wikimedia.org/66091#c14 (10Bartosz Dziewoński) 5PATC>3RESO/FIX Fixed and backported. [21:13:30] James_F: i admittedly haven't tested it with the awful and terrible WMF configs [21:13:39] * James_F grins [21:13:49] Hey, lego's working on Config. [21:13:51] Be nice. [21:13:57] heh [21:14:43] James_F: do we have wikis with the various VE configs on beta? [21:18:00] MatmaRex: There's only two main ones, and… no? [21:18:01] * James_F checks. [21:18:26] you've now made me worries that this change is going to break the world ;) [21:18:33] Aha, yes. he.wikipedia.beta.wmflabs.org [21:18:34] worried* [21:18:37] * James_F grins [21:18:41] It's unlikely. [21:20:29] MatmaRex: BTW: https://wikimediafoundation.org/w/index.php?diff=99555&oldid=99547 [21:21:15] :o [21:22:07] James_F: you forgot "(Contractor)"? not that it matters [21:22:19] Eh. Point. [21:22:58] Fixed. [22:29:08] (03CR) 10Krinkle: "recheck" [oojs/ui] - 10https://gerrit.wikimedia.org/r/160996 (owner: 10Trevor Parscal) [22:29:32] (03CR) 10Krinkle: "recheck" [oojs/ui] - 10https://gerrit.wikimedia.org/r/157315 (owner: 10Jforrester) [22:29:58] (03CR) 10jenkins-bot: [V: 04-1] [WIP] PHP Implementation of OOjs UI [oojs/ui] - 10https://gerrit.wikimedia.org/r/160996 (owner: 10Trevor Parscal) [22:31:25] (03CR) 10jenkins-bot: [V: 04-1] build: Create minified JS and CSS distribution files [oojs/ui] - 10https://gerrit.wikimedia.org/r/157315 (owner: 10Jforrester) [22:33:35] (03CR) 10Krinkle: "recheck" [oojs/ui] - 10https://gerrit.wikimedia.org/r/160996 (owner: 10Trevor Parscal) [22:33:44] (03CR) 10Krinkle: "recheck" [oojs/ui] - 10https://gerrit.wikimedia.org/r/157315 (owner: 10Jforrester) [22:37:32] (03CR) 10Krinkle: "recheck" [oojs/ui] - 10https://gerrit.wikimedia.org/r/157315 (owner: 10Jforrester) [22:44:14] (03CR) 10Krinkle: "recheck" [oojs/ui] - 10https://gerrit.wikimedia.org/r/160996 (owner: 10Trevor Parscal) [23:05:35] (03PS1) 10Jforrester: build: Split test into a lint sub-pipelin for quickly running those tests [oojs/ui] - 10https://gerrit.wikimedia.org/r/161760 [23:09:52] 3VisualEditor / 3Technical Debt: VisualEditor: Main application namespace should be capitalised and not 2 characters short (ve -> VisualEditor) - 10https://bugzilla.wikimedia.org/67642 (10James Forrester) 5NEW>3ASSI p:5Unprio>3Lowest s:5normal>3enhanc