[00:33:44] Anyone seen "Call to undefined method ElasticaConnection::__construct() in /vagrant/mediawiki/extensions/Flow/includes/Search/Connection.php" lately? [00:33:57] I'm getting it when I run all Flow PHPUnit tests locally. /cc ebernhardson [00:34:17] This is with e759d9601f0c0a54ade74fe844970fb6af605c37 of Flow. [00:37:09] matt_flaschen: make sure you have latest Elastica extension? we refactored that recently [00:37:17] probably Elastica and CirrusSearch to be safe [00:37:26] * ebernhardson|awy is killing singletons [00:37:34] it makes multi-cluster operations harder than they should be :) [00:39:50] ebernhardson|awy, that could be it, but I tried to check: [00:40:26] hmm, actually your right that probably doesn't work. sec [00:40:26] ebernhardson|awy, I'm at latest of CirrusSearch and Elastica extension. [00:41:28] (03PS1) 10EBernhardson: ElsaticaConnection doesnt have a constructor [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246134 [00:41:34] ^^ should be safe. That class never had a useful constructor anyways (it simply existed but was empty and protected to enforce the singleton status) [00:44:45] (03CR) 10jenkins-bot: [V: 04-1] ElsaticaConnection doesnt have a constructor [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246134 (owner: 10EBernhardson) [00:44:57] :S [00:45:19] oh, qunit. i saw mentino thats been flaky lately [00:45:26] (03CR) 10EBernhardson: "recheck" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246134 (owner: 10EBernhardson) [00:47:23] Thanks, ebernhardson|awy. I wonder why this isn't failing in Jenkins. [00:47:59] (03CR) 10Mattflaschen: [C: 032 V: 032] "Forcing due to unrelated QUnit." [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246134 (owner: 10EBernhardson) [00:48:21] (03PS2) 10Mattflaschen: Fix history pagination and give user the number of entries they requested [extensions/Flow] - 10https://gerrit.wikimedia.org/r/245621 (https://phabricator.wikimedia.org/T112230) [00:48:34] matt_flaschen: integration/config in jjb/mediawiki-extensions.yaml. CirrusSearch and Elastica arn't listed as test requirements so it's probably skipped? [00:50:56] (03CR) 10jenkins-bot: [V: 04-1] Fix history pagination and give user the number of entries they requested [extensions/Flow] - 10https://gerrit.wikimedia.org/r/245621 (https://phabricator.wikimedia.org/T112230) (owner: 10Mattflaschen) [00:50:59] (03Merged) 10jenkins-bot: ElsaticaConnection doesnt have a constructor [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246134 (owner: 10EBernhardson) [00:57:18] 6Collaboration-Team-Backlog, 10Echo: HTML email notifications caught by spam filters - https://phabricator.wikimedia.org/T54915#1724097 (10Tbayer) See also {T115416} [01:14:38] (03PS3) 10Mattflaschen: Fix history pagination and give user the number of entries they requested [extensions/Flow] - 10https://gerrit.wikimedia.org/r/245621 (https://phabricator.wikimedia.org/T112230) [01:17:55] * matt_flaschen exhales [01:22:18] (03PS4) 10Mattflaschen: Fix history pagination and give user the number of entries they requested [extensions/Flow] - 10https://gerrit.wikimedia.org/r/245621 (https://phabricator.wikimedia.org/T112230) [01:25:18] (03CR) 10jenkins-bot: [V: 04-1] Fix history pagination and give user the number of entries they requested [extensions/Flow] - 10https://gerrit.wikimedia.org/r/245621 (https://phabricator.wikimedia.org/T112230) (owner: 10Mattflaschen) [02:45:04] (03PS5) 10Mattflaschen: Fix history pagination and give user the number of entries they requested [extensions/Flow] - 10https://gerrit.wikimedia.org/r/245621 (https://phabricator.wikimedia.org/T112230) [02:46:15] (03PS6) 10Mattflaschen: Fix history pagination and give user the number of entries they requested [extensions/Flow] - 10https://gerrit.wikimedia.org/r/245621 (https://phabricator.wikimedia.org/T112230) [08:41:11] 6Collaboration-Team-Backlog, 10Flow, 10Parsoid, 10RESTBase: Flow posts being serialized from HTML -> WT without providing Parsoid data-parsoid attributes? - https://phabricator.wikimedia.org/T115236#1724469 (10mobrovac) Hm, no, scratch that, stashing wouldn't help you actually as you main point is you're m... [12:05:59] 6Collaboration-Team-Backlog, 10Flow, 6Community-Liaison, 7Documentation: Create a general public documentation for Flow - https://phabricator.wikimedia.org/T111367#1724832 (10Trizek-WMF) [12:13:25] (03CR) 10Sbisson: [C: 032] "Roan says so" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/245880 (owner: 10Sbisson) [12:16:50] (03Merged) 10jenkins-bot: Browser tests: Use button text instead of title [extensions/Flow] - 10https://gerrit.wikimedia.org/r/245880 (owner: 10Sbisson) [12:53:35] (03PS1) 10Sbisson: Browser tests: Wait for a button to exist before trying to click it [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246212 [13:06:46] 6Collaboration-Team-Backlog, 10Flow, 10Parsoid, 10RESTBase: Flow posts being serialized from HTML -> WT without providing Parsoid data-parsoid attributes? - https://phabricator.wikimedia.org/T115236#1724950 (10mobrovac) >>! In T115236#1724854, @ssastry wrote: > If this turns out to be the best way of doing... [13:09:17] 6Collaboration-Team-Backlog, 10Flow, 10Parsoid, 10RESTBase: Flow posts being serialized from HTML -> WT without providing Parsoid data-parsoid attributes? - https://phabricator.wikimedia.org/T115236#1724961 (10ssastry) >>! In T115236#1724950, @mobrovac wrote: >>>! In T115236#1724854, @ssastry wrote: >> If... [13:29:05] 6Collaboration-Team-Backlog, 10Flow, 10Parsoid, 10RESTBase: Flow posts being serialized from HTML -> WT without providing Parsoid data-parsoid attributes? - https://phabricator.wikimedia.org/T115236#1725026 (10mobrovac) >>! In T115236#1724961, @ssastry wrote: > It doesn't have to. You could return the data... [13:31:35] 6Collaboration-Team-Backlog, 10Flow, 10Parsoid, 10RESTBase: Flow posts being serialized from HTML -> WT without providing Parsoid data-parsoid attributes? - https://phabricator.wikimedia.org/T115236#1725053 (10mobrovac) Hm, now I see Arlo's email about getting data-parsoid in-lined in Parsoid's v3. @ssastr... [14:16:06] hey matt_flaschen, is the coworking space forcing you into a "daytime" schedule? ;) [14:17:25] stephanebisson, I kind of wanted to do it anyway, but a little bit. :) After I get a key, I can do whatever hours I want, but I'm going to try to keep an early schedule. [14:18:19] matt_flaschen: Good. But you'll get more questions from me, be warned [14:18:32] stephanebisson, and vice versa. [14:19:17] matt_flaschen: I'm investigating the null edit on zh.wp. flow_revision.rev_content contains 'DB://cluster25/7720491'. How/where can I access this data? [14:20:52] ^ halfak, is External Store accessible on stat1003? Sorry if I asked you this before. [14:21:13] matt_flaschen, as far as I know, we don't have a way to get at External Store. [14:21:17] stephanebisson, wait, do you have prod access now? [14:21:24] Or just bastion. [14:21:28] I use the API (with my advanced permissions) and the XML dumps to get at text data. [14:21:31] matt_flaschen: just bastion [14:21:45] Thanks halfak. [14:21:48] stephanebisson, I'll get it. [14:22:03] matt_flaschen: DB://cluster25/7720470 as well [14:22:09] Good luck stephanebisson [14:28:24] stephanebisson, sorry, I forgot how I did this last time. I thought I could just look up the credentials, but apparently not so much. [14:28:33] I know another way that should work, though. [14:32:55] That's not good. [14:33:13] stephanebisson, they're both false. [14:33:23] I'm going to look at the DB directly. [14:33:59] matt_flaschen: the values are 'false'? It's supposed to be 3-4 sentences and a few links. [14:36:08] stephanebisson, yeah, ExternalStore::fetchFromURL didn't retrieve it. I could be doing something wrong, looking now. [14:37:05] stephanebisson, are they both from zhwiki? [14:37:26] matt_flaschen: should be [14:38:30] matt_flaschen: they both have rev_user_wiki=zhwiki [14:41:42] I can't remember how the gzip flag works. [14:42:05] the workflow they belong to has workflow_wiki=zhwiki [14:49:27] PMed. [14:50:16] Summary: i forgot that they're stored per-wiki, so if you access the DB directly using credentials from https://phabricator.wikimedia.org/diffusion/OMWC/browse/master/wmf-config/db-eqiad.php;753d5edd6c0c05a70dc7e8cde0c242077207a144$488 (which I tried first), you need to first use zhwiki (or whatever it is). [14:50:30] But it's also compressed, so if you need more than checking it it's there, you should use the shell. [14:50:40] mwscript eval.php --wiki=zhwiki [14:51:18] var_export( \Revision::decompressRevisionText( ExternalStore::fetchFromURL( 'DB://cluster25/7720491' ), explode( ',', 'utf-8,gzip,html,external' ) ) ); [14:51:40] Again, using the shell it has to be the right wiki. [14:51:58] Those flags are from rev_flags (probably the same for all, though). [14:53:23] stephanebisson, it looks like a change to how Parsoid deserialized/canonicalized the link (mediawikiwiki vs. mw). [14:53:44] matt_flaschen: where do you run those commands from? Can I do it from bastion? [14:54:30] stephanebisson, probably not, but I didn't check. [14:54:41] mwscript eval.php --wiki=zhwiki [14:54:45] would open a PHP shell if it works. [14:55:22] Sorry that took so long. [14:56:01] matt_flaschen: no worries, thanks a lot [14:56:19] I tried to do it the simple way, which ended up being the not-working way. [14:57:09] It doesn't work for me on bast1001. [14:57:32] matt_flaschen: so do you think it's a parsoid change that happened in the 3 minutes between those revisions? [14:57:55] stephanebisson, probably not. Let me check if it's a dirty diff. [14:59:40] stephanebisson, yeah. [15:00:17] Not sure why though. [15:01:00] Going to http://parsoid-lb.eqiad.wikimedia.org/ , choosing "HTML DOM -> WikiText form", putting in the first one, copying resulting wikitext, and going to "WikiText -> HTML DOM form" shows the same. [15:01:35] stephanebisson, but like you said, that just raises the question of why it was different the first time. Especially because we round-trip incoming HTML already. [15:03:47] stephanebisson, hypothesis: It was originally submitted as wikitext, so wasn't round-tripped, just converted to HTML. And the HTML->wikitext transform here is invalid. [15:06:35] stephanebisson, actually, this is probably caused by https://phabricator.wikimedia.org/T115236#1718692 since that HTML suspiciously has no data-parsoid. [15:07:01] Just the IDs which is consistent with what they're talking about on the bug. [15:07:17] possible [15:08:19] I'm wondering if we should use the diff engine to figure out if 2 revs are the same (to block null edits) instead of string comparison... [15:09:26] I think it would prevent these kind of bugs but is it overkill? [15:10:02] (03PS8) 10Matthias Mullie: Dump Flow data [extensions/Flow] - 10https://gerrit.wikimedia.org/r/242569 (https://phabricator.wikimedia.org/T89398) [15:10:04] I'm 99.9% sure that's what it is. stephanebisson, we already have the wikitext on submission (either because they submitted wikitext, or because we decode HTML->wikitext (we already do this for security reasons)). However, we would have to decode the prior version to wikitext to do a wikitext compare. [15:10:39] It's a judgment call, but I don't think it's necessary. Better to fix the Parsoid issues as they crop up. [15:11:16] now we compare using the storage format, should we always use wikitext? [15:12:28] (03PS3) 10Matthias Mullie: [WIP] Import Flow data dump [extensions/Flow] - 10https://gerrit.wikimedia.org/r/245501 (https://phabricator.wikimedia.org/T114703) [15:13:05] stephanebisson, one sec, checking something. [15:13:57] (03PS9) 10Matthias Mullie: Dump Flow data [extensions/Flow] - 10https://gerrit.wikimedia.org/r/242569 (https://phabricator.wikimedia.org/T89398) [15:14:03] (03PS4) 10Matthias Mullie: [WIP] Import Flow data dump [extensions/Flow] - 10https://gerrit.wikimedia.org/r/245501 (https://phabricator.wikimedia.org/T114703) [15:15:06] stephanebisson, never mind, we probably already have to decode the old version's HTML to wikitext in prod due to https://phabricator.wikimedia.org/diffusion/EFLW/browse/master/includes/SpamFilter/ConfirmEdit.php;0de48459a43f9290ee44580b54b3ffbb568cfef8$24 [15:15:17] Note how it's unconditional as long as there is an old revision which there should be here. [15:15:23] (03PS5) 10Matthias Mullie: Import Flow data dump [extensions/Flow] - 10https://gerrit.wikimedia.org/r/245501 (https://phabricator.wikimedia.org/T114703) [15:15:24] So that's a valid solution that shouldn't cause a performance regression. [15:16:05] 3Collaboration-Team-Current, 10Flow, 10Datasets-General-or-Unknown, 7WorkType-Maintenance: Create XML Schema for Flow dump format(s) - https://phabricator.wikimedia.org/T112799#1725298 (10matthiasmullie) a:3matthiasmullie [15:16:22] stephanebisson, except the diff itself, but you could consider a straight wikitext identical check too (but that will also work better if data-parsoid is passed through). [15:16:24] (03CR) 10Paladox: "recheck" [extensions/Echo] - 10https://gerrit.wikimedia.org/r/242934 (owner: 10Siebrand) [15:17:03] 6Collaboration-Team-Backlog, 10Flow, 10Parsoid, 10RESTBase, 7Regression: Flow posts being serialized from HTML -> WT without providing Parsoid data-parsoid attributes? - https://phabricator.wikimedia.org/T115236#1725300 (10Mattflaschen) [15:17:15] (03CR) 10jenkins-bot: [V: 04-1] Dump Flow data [extensions/Flow] - 10https://gerrit.wikimedia.org/r/242569 (https://phabricator.wikimedia.org/T89398) (owner: 10Matthias Mullie) [15:18:37] (03CR) 10jenkins-bot: [V: 04-1] Import Flow data dump [extensions/Flow] - 10https://gerrit.wikimedia.org/r/245501 (https://phabricator.wikimedia.org/T114703) (owner: 10Matthias Mullie) [15:22:56] 3Collaboration-Team-Current, 10Flow, 5Patch-For-Review, 7WorkType-NewFunctionality: Null edits should not be saved in posts, topic titles, descriptions, summaries - https://phabricator.wikimedia.org/T60975#1725332 (10SBisson) @mattflaschen is 99.9% sure this is caused by {T115236} Will do comparison in wi... [15:25:53] (03PS9) 10Paladox: Add PHP CodeSniffer tests [extensions/Echo] - 10https://gerrit.wikimedia.org/r/242934 (owner: 10Siebrand) [15:26:31] 3Collaboration-Team-Current, 10Flow, 5Patch-For-Review, 7WorkType-NewFunctionality: Null edits should not be saved in posts, topic titles, descriptions, summaries - https://phabricator.wikimedia.org/T60975#1725349 (10SBisson) The revisions have slightly different html representations. One has `title="mw:ta... [15:30:07] (03PS1) 10Sbisson: Check for null edits using wikitext format [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246249 (https://phabricator.wikimedia.org/T60975) [15:33:26] (03PS10) 10Paladox: Add PHP CodeSniffer tests [extensions/Echo] - 10https://gerrit.wikimedia.org/r/242934 (owner: 10Siebrand) [15:34:34] (03CR) 10jenkins-bot: [V: 04-1] Check for null edits using wikitext format [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246249 (https://phabricator.wikimedia.org/T60975) (owner: 10Sbisson) [15:38:12] (03CR) 10Paladox: [C: 031] "Passes now." [extensions/Echo] - 10https://gerrit.wikimedia.org/r/242934 (owner: 10Siebrand) [15:43:51] stephanebisson, have you been getting many EAGAIN with Vagrant? I always got them occasionally, mostly on saving my editor in a file that was shared with Vagrant via NFS. But now it's more often, and even for reads (which if they happened before were rare). [16:11:06] matt_flaschen: never [16:26:53] (03CR) 10Sbisson: "recheck" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246249 (https://phabricator.wikimedia.org/T60975) (owner: 10Sbisson) [16:43:38] 3Collaboration-Team-Current, 10Flow, 5Patch-For-Review, 5WMF-deploy-2015-10-06_(1.27.0-wmf.2), and 2 others: Impossible to edit Flow board description - https://phabricator.wikimedia.org/T115339#1725675 (10Etonkovidova) Checked in testwiki, zhwiki, mediawiki - in Safari 9.0, FF 41, Chrome 45. [16:46:20] 3Collaboration-Team-Current, 10Flow, 10BetaFeatures: Activate Flow as a beta feature on zh.wp - https://phabricator.wikimedia.org/T115100#1725680 (10Etonkovidova) Checked in zhwiki. {F2720921} [16:55:01] (03CR) 10Mooeypoo: "recheck" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246127 (owner: 10Mooeypoo) [16:57:46] 3Collaboration-Team-Current, 10Flow, 5Patch-For-Review, 5WMF-deploy-2015-10-06_(1.27.0-wmf.2), 5WMF-deploy-2015-10-13_(1.27.0-wmf.3): Browse topics fly-out does not work . - https://phabricator.wikimedia.org/T115265#1725714 (10Etonkovidova) Checked in mediawiki and testwiki. [17:07:09] (03CR) 10Catrope: [C: 032] Display clean titles for categories [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246127 (owner: 10Mooeypoo) [17:10:24] (03Merged) 10jenkins-bot: Display clean titles for categories [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246127 (owner: 10Mooeypoo) [17:12:08] (03PS2) 10Paladox: Update grunt-jscs to 2.1.0 [extensions/Echo] - 10https://gerrit.wikimedia.org/r/243503 [17:14:05] (03CR) 10jenkins-bot: [V: 04-1] Update grunt-jscs to 2.1.0 [extensions/Echo] - 10https://gerrit.wikimedia.org/r/243503 (owner: 10Paladox) [17:16:20] (03CR) 10Catrope: [C: 031] Check for null edits using wikitext format [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246249 (https://phabricator.wikimedia.org/T60975) (owner: 10Sbisson) [17:17:34] 3Collaboration-Team-Current, 10Flow: [betalabs] this.board.setCategoriesFromObject is not a function error - any Talk page is inaccessible - https://phabricator.wikimedia.org/T115501#1725794 (10Etonkovidova) 3NEW [17:18:17] (03PS1) 10Sbisson: [WIP] DO NOT MERGE [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246269 [17:19:29] stephanebisson, do you file share with Vagrant using NFS? [17:19:49] matt_flaschen: I think so [17:20:24] It would probably prompt you about the exports file (for sudo access) regularly if using NFS. [17:20:57] yes, it does [17:21:18] everytime I do `vagrant git-update` [17:21:26] no, vagrant provision [17:21:58] (03CR) 10Sbisson: "recheck" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246269 (owner: 10Sbisson) [17:22:03] Okay. It stopped for time being, so back-burnering it. If the EAGAIN keeps bugging me, will revisit. [17:25:03] 6Collaboration-Team-Backlog, 10Flow: Flow not warning me when I leave an edit page with unsaved changes - https://phabricator.wikimedia.org/T96531#1725821 (10Quiddity) p:5Normal>3High This is happening reproducibly in Firefox and Chromium. Raising priority, as it is very frustrating to lose partially writt... [17:25:44] (03PS2) 10Sbisson: [WIP] DO NOT MERGE [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246269 [17:27:50] * ebernhardson merges [17:28:02] Noooo [17:29:16] stephanebisson, you shouldn't have made it so ambiguous whether to merge. ;) [17:29:41] (03CR) 10jenkins-bot: [V: 04-1] [WIP] DO NOT MERGE [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246269 (owner: 10Sbisson) [17:29:57] (03CR) 10Sbisson: "recheck" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246269 (owner: 10Sbisson) [17:31:32] I'm waiting on hangouts to let me in [17:31:48] mlitn, did you mean to call me? [17:31:58] no [17:32:07] hey everyone [17:32:08] I try to get in via hangouts app and fucked up :p [17:32:51] 6Collaboration-Team-Backlog, 10Flow: Flow not warning me when I leave an edit page with unsaved changes - https://phabricator.wikimedia.org/T96531#1725843 (10Quiddity) [17:33:50] Yeah, I was able to do that a long time ago, but I don't remember the incantation (actually, I think the incantation stopped working a while ago) [17:37:45] (03PS3) 10Paladox: Update grunt-jscs to 2.1.0 [extensions/Echo] - 10https://gerrit.wikimedia.org/r/243503 [17:39:09] 3Collaboration-Team-Current, 10Flow: [betalabs] this.board.setCategoriesFromObject is not a function error - any Talk page is inaccessible - https://phabricator.wikimedia.org/T115501#1725864 (10Catrope) I don't see this on http://en.wikipedia.beta.wmflabs.org/wiki/Talk:Flow or http://en.wikipedia.beta.wmflabs.... [17:39:29] (03CR) 10jenkins-bot: [V: 04-1] Update grunt-jscs to 2.1.0 [extensions/Echo] - 10https://gerrit.wikimedia.org/r/243503 (owner: 10Paladox) [17:42:53] (03PS4) 10Paladox: Update grunt-jscs to 2.1.0 [extensions/Echo] - 10https://gerrit.wikimedia.org/r/243503 [17:46:17] (03CR) 10jenkins-bot: [V: 04-1] Update grunt-jscs to 2.1.0 [extensions/Echo] - 10https://gerrit.wikimedia.org/r/243503 (owner: 10Paladox) [17:46:21] stephanebisson, see this https://gerrit.wikimedia.org/r/#/c/246095/ [17:47:16] mooeypoo: see my comment in https://gerrit.wikimedia.org/r/#/c/246095/ [17:47:51] stephanebisson, wait, you reviewed? what comment? [17:48:10] mooeypoo: "Should there be a change to flow-initialize in this patch to move the creation of the description widget before the model initialization?" [17:48:17] oh shoo I missed that, sorry [17:48:54] stephanebisson, didn't we change that already? there was a quickfix that was cherry picked [17:49:32] (03PS5) 10Paladox: Update grunt-jscs to 2.1.0 [extensions/Echo] - 10https://gerrit.wikimedia.org/r/243503 [17:49:47] Also, with this, it's less critical, theoretically. Even if we're starting before the model is initialized, there's an event listener. If we start after, then the widget pulls the correct information for 'isEditable' [17:49:53] mooeypoo: I did a quick hack to create the desc widget AFTER model initialization. Isn't the point of this patch to move it back where it was and have the editable event-driven? [17:50:23] stephanebisson, it can work either way, I figured I won't touch the flow initialize stuff until I actually refactor it to be cleaner in general [17:50:59] But I can also move it up in this commit if it makes more sense [17:51:20] (03CR) 10jenkins-bot: [V: 04-1] Update grunt-jscs to 2.1.0 [extensions/Echo] - 10https://gerrit.wikimedia.org/r/243503 (owner: 10Paladox) [17:51:24] I thint it makes more sense [17:51:31] ok, I can move it [17:51:49] do you have time to help me with browser tests now, though? I can do the othre thing later less urgently [17:51:55] (can you -1 though?) [17:51:57] mooeypoo: yes [17:52:10] -1 so no one merges it and I don't forget to make that change [17:52:51] (03CR) 10Paladox: "recheck" [extensions/Echo] - 10https://gerrit.wikimedia.org/r/243503 (owner: 10Paladox) [17:53:17] (03CR) 10Sbisson: [C: 04-1] "I think that would be great to have the desc widget created before the model initialization so it is 1 less hack to the architecture and i" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246095 (https://phabricator.wikimedia.org/T115339) (owner: 10Mooeypoo) [17:53:22] voila [17:53:26] okay, bah, browser tests. They all still fail for me. I followed your recommendations in the email, but nothing changed. [17:53:29] stephanebisson, awesome, thanks [17:53:43] stephanebisson, should I try running the tests outside of the vm ? [17:53:57] wait [17:54:09] that's nuts [17:54:16] yes. [17:54:23] Welcome to my computer. Feel at home. [17:54:32] let me try again one more time. [17:54:34] you're sure you've unset all MEDIAWIKI_* ? [17:55:25] uhm, maybe not? [17:55:47] MEDIAWIKI_USER and MEDIAWIKI_URL [17:55:52] (03PS1) 10Catrope: CI test, do not merge [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246277 [17:56:06] MEDIAWIKI_API_URL, MEDIAWIKI_PASSWORD [17:56:54] and then `export MEDIAWIKI_ENVIRONMENT=local` [17:57:21] (03PS6) 10Paladox: Update grunt-jscs to 2.1.0 [extensions/Echo] - 10https://gerrit.wikimedia.org/r/243503 [17:57:33] (03Abandoned) 10Sbisson: [WIP] DO NOT MERGE [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246269 (owner: 10Sbisson) [17:58:34] (03PS7) 10Paladox: Update grunt-jscs to 2.1.0 [extensions/Echo] - 10https://gerrit.wikimedia.org/r/243503 [17:58:57] (03CR) 10jenkins-bot: [V: 04-1] CI test, do not merge [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246277 (owner: 10Catrope) [17:59:40] Nothing from Scrum of Scrums that we don't already know about. [18:00:50] (03CR) 10jenkins-bot: [V: 04-1] Update grunt-jscs to 2.1.0 [extensions/Echo] - 10https://gerrit.wikimedia.org/r/243503 (owner: 10Paladox) [18:02:56] (03CR) 10Paladox: "recheck" [extensions/Echo] - 10https://gerrit.wikimedia.org/r/243503 (owner: 10Paladox) [18:03:36] (03PS8) 10Paladox: Update grunt-jscs to 2.1.0 [extensions/Echo] - 10https://gerrit.wikimedia.org/r/243503 [18:04:39] etonkovidova: You're still in the hangout :) [18:04:57] RoanKattouw: ok [18:09:11] eh, stephanebisson I think I got disconnected. Do you want me to show you my screen? It's still failing. [18:09:52] Hey d3r1ck [18:10:16] mooeypoo: what is the first failing test? [18:10:17] stephanebisson, it's like the browser is waiting for the interaction that never comes. It's failing on timing out, even though I've increased the timeout to 30 [18:10:37] stephanebisson, the first test :\ features/action_menu_permalink.feature:10 [18:10:45] When I select the permalink for the first post of the first topic # features/step_definitions/action_menu_permalink_steps.rb:8 [18:10:45] timed out after 30 seconds, waiting for {:css=>".flow-topic .flow-post .flow-menu-js-drop a", :index=>0} to become present (Watir::Wait::TimeoutError) [18:11:23] ok, we'll work on this one [18:11:53] stephanebisson, wait [18:12:54] crap. I ran 'bundle install' and now everything's failing. http://pastebin.com/QH7BwXDt [18:13:08] stephanebisson, ... it's like my computer is cursed or something [18:13:43] seems like my webdriver might need an update? [18:14:28] stephanebisson, the update seemed to have upgraded my web drivers, though. The output is: [18:14:29] Installing selenium-webdriver 2.48.1 (was 2.47.1) [18:14:29] Installing watir-webdriver 0.9.0 (was 0.8.0) [18:14:35] ... what is going on :\ [18:17:51] I have the same error [18:18:06] So it's not just my computer! PROGRESS! [18:18:32] maybe marxarelli can assist? [18:19:46] mooeypoo: first, git checkout the Gemfile.lock that you have just modified and refrain from doing `bundle update` [18:20:05] bundle install is what you want [18:20:20] bundle install installs what is specified in the lock file [18:20:58] bundle update gets the new versions in accordance with the gemfile and creates a new lock file - you shouldn't have to do that [18:21:02] (03CR) 10Catrope: "recheck" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246277 (owner: 10Catrope) [18:21:20] what stephanebisson said :) [18:22:00] 3Collaboration-Team-Current, 10Flow: Opening VE produces dmDoc.getNearestCursorOffset is not a function - https://phabricator.wikimedia.org/T115507#1725990 (10Etonkovidova) 3NEW [18:22:57] stephanebisson, okay now it works [18:23:05] stephanebisson, but it fails the test [18:23:53] mooeypoo: so it opens an old-looking firefox with an X server, right? [18:24:23] old-looking yes. Not sure what x server is? [18:24:23] it hangs there and the test times out? [18:24:27] yes [18:24:52] does it leave an X11 app open after? [18:25:11] It passes the background conditions okay, but fails on the first scenario condition ... seems to have a problem specifically with selecting things maybe? [18:25:13] no, it closes it [18:25:24] nevermind, you're on linux, I don't know what it would look like [18:25:37] stephanebisson, ... you prefer I test on Windows? [18:25:58] mooeypoo: I'm trying to help you.. don't make things worst ;) [18:26:01] haha [18:26:10] fair enough :P [18:26:37] let's give it a shot on your host to see if we can learn something [18:26:50] I think I had a problem running on my host because of the webdriver installation [18:27:00] which is why I ended up doing everything on the VM, bt I can try again [18:27:25] so, go to your Flow dir [18:27:31] `ruby -v` [18:28:27] stephanebisson, just tried running the test on the host: it's running, but with same failure. [18:29:01] ruby -v is "ruby 2.2.1p85 (2015-02-26 revision 49769) [x86_64-linux]" [18:29:29] 3Collaboration-Team-Current, 10Flow: [betalabs]Opening VE produces dmDoc.getNearestCursorOffset is not a function - https://phabricator.wikimedia.org/T115507#1726013 (10Etonkovidova) [18:29:35] you did all the prep? (bundle install, unset old variables, env=local)? [18:29:53] I forgot to unset. Let me do this again. [18:30:56] Same issue. [18:31:38] stephanebisson, just to verify, i'm doing these steps: http://pastebin.com/KLS3FbMq [18:31:38] 3Collaboration-Team-Current, 10Flow, 5Patch-For-Review, 5WMF-deploy-2015-10-13_(1.27.0-wmf.3), 5WMF-deploy-2015-10-20_(1.27.0-wmf.4): Resolve topic does not start summarize - https://phabricator.wikimedia.org/T115337#1726021 (10Etonkovidova) Checked in betalabs. [18:32:00] mooeypoo: all good [18:32:06] stephanebisson, still failing :\ [18:32:09] on timing out. [18:32:17] mooeypoo: export BROWSER=chrome [18:33:22] Chrome opened up, but same failure. [18:33:38] stephanebisson, your fix to how the menu items are detected was merged, right? I saw it [18:33:53] mooeypoo: yes [18:33:56] I am on master, so that shouldn't be the problem... it just seems like it doesn't even recognize what to click on at all [18:34:06] the element is not present [18:34:09] mooeypoo: your setup is all up to date, right? [18:34:21] I think so, hang on one second, I'm checking something [18:34:37] do you `vagrant git-update` once in a while? [18:34:54] I did, but I will do that again now [18:35:09] stephanebisson, I just opened my local wiki in chrome again, and just to make sure, I tested if $( '.flow-topic .flow-post .flow-menu-js-drop a' ) exists [18:35:10] and it does [18:35:23] let's try with features/new_topic.feature:8 [18:35:28] AND it's visible. So the test not recognizing it is weird [18:36:17] stephanebisson, that passes [18:36:27] stephanebisson, let me run vagrant git-update to make sure [18:36:51] what about description.feature:17 ? [18:39:47] stephanebisson, hang on, I'm running git-update [18:43:33] 3Collaboration-Team-Current, 10Echo, 6Editing-Analysis: Measure the notification types that are most abundantly received at 5 sample wikis - https://phabricator.wikimedia.org/T113664#1726042 (10Catrope) Anonymized data shouldn't be blocked on the analytics-store task, should it? You wouldn't have to join aga... [18:44:08] mooeypoo, it should open up a normal Window on your main system, except that it will have an @ sign or something like that in the title to indicate it's running on a different computer (in the VM). [18:47:01] stephanebisson: Behold https://gerrit.wikimedia.org/r/#/c/246277/ (with thanks to legoktm for merging and deploying my change) [18:47:07] (03Abandoned) 10Catrope: CI test, do not merge [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246277 (owner: 10Catrope) [18:49:05] stephanebisson, http://pastebin.com/xfh1LpeB [18:49:21] stephanebisson, description.feature:17 fails again on timeout [18:51:07] (03CR) 10Sbisson: "recheck" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246249 (https://phabricator.wikimedia.org/T60975) (owner: 10Sbisson) [18:51:48] mooeypoo: when youlook at the browser window, does it look like the editor is open or not? [18:52:55] it opens it at first, writes the "first version" and saves [18:53:06] then gets stuck and doesn't even open the editor for the second version [18:54:06] stephanebisson, I was trying to figure out where it got stuck, but it seems page.description.edit is not clicked [18:54:35] mooeypoo: you're still on the host with chrome, right? [18:55:12] stephanebisson, yep [18:56:08] stephanebisson, I'm a little confused with abstract_flow_page.rb ... where is description.editor defined? description seems to be page_section( ... ) [18:56:37] oh, in components [18:56:43] I see it [18:57:38] mooeypoo: `gem list webdriver` [18:58:43] stephanebisson, okay, I changed " a(:edit, text: "Edit description")" in BoardDescription to "a(:edit, css: ".flow-ui-boardDescriptionWidget-editButton a")" and now it passes [19:01:24] omg everything passes! [19:01:30] wtf [19:01:33] stephanebisson: mooeypoo says she's staring at her screen and not touching anything because everything is passing and she doesn't understand why [19:01:47] 3Collaboration-Team-Current, 10Echo, 10MediaWiki-extensions-PageCuration, 5Patch-For-Review, 5WMF-deploy-2015-10-13_(1.27.0-wmf.3): PageTriage's "patrolled" notifications are jargony and scary - https://phabricator.wikimedia.org/T114415#1726061 (10Etonkovidova) "The page [page] was reviewed by [username]... [19:01:53] mooeypoo: just by changing this selector?? [19:02:18] "I know!!!!" [19:02:25] * RoanKattouw feels like a court scribe or something [19:02:41] Having just taken notes in a meeting as well [19:15:20] stephanebisson, I ran tests through the VM, and [19:15:21] 51 scenarios (2 failed, 49 passed) [19:15:21] 303 steps (2 failed, 4 skipped, 297 passed) [19:15:23] !!!! [19:15:37] I am sure it's not the selector correction, it can't be. [19:16:11] It's probably a good change for that selector in general, but it can't be what fixed everything... maybe it was the complete update? I saw your email, I'll add that to my bashrc file and do that daily. [19:16:32] what is still failing? [19:16:46] stephanebisson, matt_flaschen also, related to this, when I run 'vagrant up' the machine loads and then asks for a password *every single time*... is there a way to avoid that? [19:16:55] stephanebisson, cucumber features/opt_in.feature:38 # Scenario: Re-opt-in [19:16:56] cucumber features/thank.feature:12 # Scenario: Thank the user [19:17:03] the second one failed for wrong password in the second user [19:17:09] mooeypoo, the exports password? [19:17:40] matt_flaschen, I think the ssh password? I need to redo it to check. It's asking for a password for my current super user [19:17:49] not the vagrant password [19:17:52] mooeypoo, I think it's for NFS. Stephane and I were talking about it earlier. [19:18:03] mooeypoo: yes, with the 'local' env you have to create the 2nd user once (Selenium Flow user 2 / vagrant) [19:18:07] I'm sure there's a way to make it not prompt. [19:18:42] You can even turn off ssh passwords altogether (not advised), but sudo also allows making some limited actions possible passwordless. I saw something about that a while back but I never tried it. [19:18:50] stephanebisson, should I commit the css selector change? I think it's probably safer to use regardless of whether that was the solution or not? or do you want to continue using it as text selector? [19:19:18] matt_flaschen, if it's a big hassle, nevermind. I just thought I could be lazier and just let vagrant load without interruption [19:20:01] mooeypoo: I care mostly about the tests passing on your box [19:20:29] mooeypoo: if you revert it, does it still pass? [19:20:36] was going to check now [19:20:41] it will be REALLY REALLY weird if it doesn't [19:20:48] now that your setup is "awake" [19:21:44] stephanebisson, okay, first test passed, I ^C after that. Let me do this: I'll run all tests again outputting to a file while I go get lunch [19:21:50] then I can actually look at what failed properly [19:21:58] alright [19:22:13] mooeypoo, I don't know that it's a big hassle, I just haven't looked into it (recently at least). Can you forward me that browser test/setup email if I'm not already a recipient? [19:22:19] If stephanebisson doesn't mind. [19:22:52] matt_flaschen, the email was about an updating-all-of-vagrant-before-loading-it function in bash [19:23:07] We should write an email to everyone on how to properly install browser tests, though [19:23:27] ok I"m running to get lunch and activating tests. See you after the fold [19:23:46] Yeah, I think I mostly know the first part (even if I don't always do it because it's super-slow), but I do need a refresh on the right way to run browser tests locally. [19:24:36] matt_flaschen: do you want to do this now? [19:25:11] stephanebisson, no, I should finish topic title first. [19:25:37] You could email me (link is fine if there are working instructions somewhere) and I could try it some morning then ping you with questions. [19:26:02] Would be good to have up to date docs on MediaWiki.org if you have time to fix/write them. [19:26:24] There is https://www.mediawiki.org/wiki/Quality_Assurance/Browser_testing/Running_tests but I have no idea if it's up to date. [19:26:52] matt_flaschen: you could have a new browser test as part of "link in topic title"... [19:27:00] I'll look at the instructions [19:27:14] Yeah, that's a good idea. I'll write one for this. [19:27:36] I still want to get it working first, though. There's a PHPUnit failure I'm already fixing. [19:34:51] marxarelli: Is there a page with up to date instructions about MEDIAWIKI_ENVIRONMENT? These two still suggest setting the MEDIAWIKI_* variables one by one: https://www.mediawiki.org/wiki/Quality_Assurance/Browser_testing/Running_tests and https://www.mediawiki.org/wiki/Quality_Assurance/Browser_testing/Setup_instructions [19:46:24] stephanebisson, RoanKattouw_away, so I think I want to provide a way to get "the wikitext format" for a given revision. [19:46:50] I say "the wikitext format", not just 'wikitext', because my WIP introduces 'topic-title-wikitext' because it's really a different format with different allowed syntax. [19:47:14] I have the following possible ideas, but I'm also open to suggestions: [19:47:29] 1. getContentInWikitextFormat() [19:47:44] 2. getContentInCanonicalSourceFormat [19:47:57] 3. getContentInCanonicalFormat (don't really like this one) [19:48:35] 4. getContent( 'wikitext' ) (this is the only one of these methods that actually already exists), but change the behavior so if you call getContent( 'wikitext' ) and it's a topic title, it will just return to you the 'topic-title-wikitext' version. [19:49:11] I think I favor #1, but not overwhelmingly. [19:56:09] matt_flaschen: I like #4 [19:56:30] but I'm probably missing something here. Are we just talking about naming or is there something else? [19:58:07] matt_flaschen: how do youplan to present it? Is it gonna be an editor with a switcher to VE and all that, or just a regular text box that allows [[link]] on save? [20:06:06] 3Collaboration-Team-Current, 10Flow: Beta features: "Discussion" link should go to Talk:Flow - https://phabricator.wikimedia.org/T114197#1726189 (10SBisson) @pginer-wmf, @catrope, do you have an opinion on this? I would love to wrap up the remaining opt-in tickets in the near future. [20:07:28] 3Collaboration-Team-Current, 10Flow: [betalabs] this.board.setCategoriesFromObject is not a function error - any Talk page is inaccessible - https://phabricator.wikimedia.org/T115501#1726194 (10Etonkovidova) It turned out that the error was present only in Chrome and only in en.betawiki and only on my computer... [20:07:45] 3Collaboration-Team-Current, 10Flow: [betalabs] this.board.setCategoriesFromObject is not a function error - any Talk page is inaccessible - https://phabricator.wikimedia.org/T115501#1726196 (10Etonkovidova) 5Open>3Resolved [20:08:26] stephanebisson, note it's only the wikitext that is allowed in normal (not Flow) edit summaries. Basically just internal links in media links. [20:08:40] stephanebisson, thus, VE is out of the question, and I don't think we need an editor, at least for now. [20:08:43] Does that affect your answer? [20:08:51] 6Collaboration-Team-Backlog, 10Flow: [betalabs] this.board.setCategoriesFromObject is not a function error - any Talk page is inaccessible - https://phabricator.wikimedia.org/T115501#1725794 (10Etonkovidova) [20:09:06] matt_flaschen: no [20:09:33] Okay. Yeah, it's basically just naming, except with #4 it's also whether it has to take the incoming format name literally. [20:09:50] matt_flaschen: well, I won't argue for #4 [20:10:29] I'll read whatever you choose in the context of the whole patch and go from there [20:10:32] internal links and media links. [20:11:03] stephanebisson, I'll go with #1 for now, but reading it in context makes sense. [20:13:52] The WIP is up BTW, so you can pretty much see where I'm coming from already (https://gerrit.wikimedia.org/r/#/c/243066/1). But no need to review now. I also started fixing some issues with it already. [20:17:00] stephanebisson, all tests but 2 passed. The one with the bad password,though, what do I need to add to environments.yml? Can we add it to the repo or is it something just locally for me? It should work globally for all repo users, no? [20:17:20] 6Collaboration-Team-Backlog, 10Flow, 10MediaWiki-extensions-LiquidThreads: Removing no-content links to empty LQT archives - https://phabricator.wikimedia.org/T115513#1726231 (10Quiddity) 3NEW [20:17:55] stephanebisson, is it just a matter of actually creating that user? [20:18:11] mooeypoo: please do it manually for now: create user "Selenium Flow user 2" with password "vagrant" [20:18:15] (03PS21) 10Paladox: Add extension.json, empty PHP entry point [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) [20:19:05] stephanebisson, "Account creation error Username entered already in use. Please choose a different name." [20:19:07] :\ [20:19:13] (03CR) 10jenkins-bot: [V: 04-1] Add extension.json, empty PHP entry point [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) (owner: 10Paladox) [20:19:14] how did it not work with that password, then. hmm. [20:19:50] mooeypoo, you can change the password from the command line with changePassword.php. [20:19:52] hm. crap. it's the wrong password. how the hell do I reset a password without an email address :\ [20:19:53] oh [20:19:54] thanks [20:19:55] :D [20:20:30] If you do need to create a user in Puppet, I can do that (I also added support for assigning a group from Puppet, but that isn't merged yet...) [20:20:30] legoktm: Hmm. I feel we should convert EventLogging to use extension registration. [20:20:48] (03PS22) 10Paladox: Add extension.json, empty PHP entry point [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) [20:21:40] (03CR) 10jenkins-bot: [V: 04-1] Add extension.json, empty PHP entry point [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) (owner: 10Paladox) [20:23:17] James_F: We should, I'm not sure how well (if at all) we support conditional namespaces. [20:23:20] matt_flaschen: that would be helpful for the browser tests [20:23:33] legoktm: Ah. Callback? [20:24:29] Yeah, but I'd like to do it properly. [20:24:32] (03CR) 10Paladox: "recheck" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) (owner: 10Paladox) [20:24:42] legoktm: Yeah. [20:25:06] 3Collaboration-Team-Current, 10Flow: Beta features: "Discussion" link should go to Talk:Flow - https://phabricator.wikimedia.org/T114197#1726246 (10Catrope) @Jdforrester-WMF: You're our interim product support **and** the PM for BetaFeatures, what are your thoughts? [20:25:35] stephanebisson, yeah, that's why I started the patch to begin with (for oversight/suppress which some tests require/required). [20:25:36] (03PS23) 10Paladox: Add extension.json, empty PHP entry point [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) [20:25:50] 3Collaboration-Team-Current, 10Flow: Beta features: "Discussion" link should go to Talk:Flow - https://phabricator.wikimedia.org/T114197#1726253 (10Catrope) As for my opinion: if all other BetaFeatures just link to mw.org, I'd be happy to follow that precedent. [20:26:46] stephanebisson, if you can think of any other users/groups, could you file a bug and CC me? [20:27:13] I can't believe we're *still* talking about whether to link the Beta Feature to mediawiki.org [20:27:14] sure [20:27:54] (03CR) 10jenkins-bot: [V: 04-1] Add extension.json, empty PHP entry point [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) (owner: 10Paladox) [20:27:57] Can't we just do that now that Danny's no longer CC'ed to the bug (or PM), considering all the other BF's do... [20:28:20] (03PS24) 10Paladox: Add extension.json, empty PHP entry point [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) [20:28:31] 3Collaboration-Team-Current, 10Flow: Beta features: "Discussion" link should go to Talk:Flow - https://phabricator.wikimedia.org/T114197#1726256 (10Mattflaschen) >>! In T114197#1726253, @Catrope wrote: > As for my opinion: if all other BetaFeatures just link to mw.org, I'd be happy to follow that precedent. +1 [20:28:42] As I said on the task, I'd be happy to do whatever the precedent for BFs is [20:30:16] (03CR) 10jenkins-bot: [V: 04-1] Add extension.json, empty PHP entry point [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) (owner: 10Paladox) [20:31:37] (03PS25) 10Paladox: Add extension.json, empty PHP entry point [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) [20:34:01] (03CR) 10jenkins-bot: [V: 04-1] Add extension.json, empty PHP entry point [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) (owner: 10Paladox) [20:35:31] 6Collaboration-Team-Backlog, 10Flow: Make sure mw-vagrant Flow role creates users/groups required for Flow browser tests - https://phabricator.wikimedia.org/T115515#1726297 (10SBisson) 3NEW [20:36:10] matt_flaschen: ^ [20:37:55] 3Collaboration-Team-Current, 10Flow, 10MediaWiki-Vagrant, 5Patch-For-Review: Have 'flow-suppress' work on MediaWiki-Vagrant out of the box - https://phabricator.wikimedia.org/T113641#1726310 (10Mattflaschen) [20:38:18] 3Collaboration-Team-Current, 10Flow: Make sure mw-vagrant Flow role creates users/groups required for Flow browser tests - https://phabricator.wikimedia.org/T115515#1726297 (10Mattflaschen) [20:38:24] Thanks [20:42:08] 6Collaboration-Team-Backlog, 10Flow: "Talk" tab stays red - https://phabricator.wikimedia.org/T115201#1726321 (10Alsee) I don't know if it is related, but in this comment: https://www.mediawiki.org/w/index.php?title=Topic:Sqd8qg7l54ih29v5&topic_showPostId=sqd8qgcfqx72z629#flow-post-sqd8qgcfqx72z629 Quiddity's... [20:44:06] 6Collaboration-Team-Backlog, 10Flow: "Talk" tab stays red - https://phabricator.wikimedia.org/T115201#1726333 (10Catrope) >>! In T115201#1726321, @Alsee wrote: > I don't know if it is related, but in this comment: https://www.mediawiki.org/w/index.php?title=Topic:Sqd8qg7l54ih29v5&topic_showPostId=sqd8qgcfqx72z... [20:44:16] This change will have some interesting Unforseen Consequences TM (nothing too serious). [20:48:27] 6Collaboration-Team-Backlog, 10Flow: "Talk" tab stays red - https://phabricator.wikimedia.org/T115201#1726356 (10Catrope) >>! In T115201#1726333, @Catrope wrote: >>>! In T115201#1726321, @Alsee wrote: >> I don't know if it is related, but in this comment: https://www.mediawiki.org/w/index.php?title=Topic:Sqd8q... [20:49:08] matt_flaschen: which change? [20:50:03] (03CR) 10Paladox: "recheck" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) (owner: 10Paladox) [20:50:57] stephanebisson, the topic title one. For example, the watchlist editor will display link markup if the topic title has it, but link to the actual topic, because the alternative of nested links is worse. [20:51:41] interesting... [20:56:34] 3Collaboration-Team-Current, 10Flow: [betalabs]Opening VE produces dmDoc.getNearestCursorOffset is not a function - https://phabricator.wikimedia.org/T115507#1726379 (10Catrope) Is this reproducible in non-debug mode? Because things tend to be pretty broken on beta in debug mode (for infrastructural reasons). [20:57:26] 6Collaboration-Team-Backlog, 10Flow: Delete unneeded LQT/Flow templates from wikis that didn't have LQT - https://phabricator.wikimedia.org/T115518#1726384 (10Quiddity) 3NEW [20:57:39] 6Collaboration-Team-Backlog, 10Flow: Add Wikidata items for Flow's templates - https://phabricator.wikimedia.org/T115519#1726394 (10Quiddity) 3NEW [20:57:44] 3Collaboration-Team-Current, 10Flow: Opt-out: Add template with a link to Flow Archive page - https://phabricator.wikimedia.org/T112957#1726404 (10SBisson) a:3SBisson [21:01:07] 6Collaboration-Team-Backlog, 10Flow: Delete unneeded LQT/Flow templates from wikis that didn't have LQT - https://phabricator.wikimedia.org/T115518#1726412 (10Mattflaschen) The script creates them all. If we do this, we should also have the script auto-detect whether LQT is installed, or they'll get regenerat... [21:04:17] (03CR) 10Paladox: "Not sure why test mediawiki-extensions-qunit is failing." [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) (owner: 10Paladox) [21:10:53] (03PS26) 10Paladox: Add extension.json, empty PHP entry point [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) [21:12:50] (03CR) 10jenkins-bot: [V: 04-1] Add extension.json, empty PHP entry point [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) (owner: 10Paladox) [21:16:49] (03CR) 10Paladox: "check experimental" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) (owner: 10Paladox) [21:17:32] 6Collaboration-Team-Backlog, 10Flow, 10Parsoid, 10RESTBase, 7Regression: Flow posts being serialized from HTML -> WT without providing Parsoid data-parsoid attributes? - https://phabricator.wikimedia.org/T115236#1726483 (10Arlolra) @mobrovac It's already available in v2 by requesting `html` instead of `p... [21:26:09] 6Collaboration-Team-Backlog, 10Flow, 10Parsoid, 10RESTBase, 7Regression: Flow posts being serialized from HTML -> WT without providing Parsoid data-parsoid attributes? - https://phabricator.wikimedia.org/T115236#1726500 (10Mattflaschen) @Arlolra is there a version of that URL using transform? We current... [21:29:09] (03PS27) 10Paladox: Add extension.json, empty PHP entry point [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) [21:30:46] (03CR) 10jenkins-bot: [V: 04-1] Add extension.json, empty PHP entry point [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) (owner: 10Paladox) [21:31:54] 6Collaboration-Team-Backlog, 10Flow, 10Parsoid, 10RESTBase, 7Regression: Flow posts being serialized from HTML -> WT without providing Parsoid data-parsoid attributes? - https://phabricator.wikimedia.org/T115236#1726527 (10ssastry) So, however this missing data-parsoid issue is fixed, there is one other... [21:33:53] 3Collaboration-Team-Current, 10Flow: 'undo' for Board description edits is not functioning - https://phabricator.wikimedia.org/T115525#1726533 (10Etonkovidova) 3NEW [21:38:25] (03PS28) 10Paladox: Add extension.json, empty PHP entry point [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) [21:39:50] (03PS29) 10Paladox: Add extension.json, empty PHP entry point [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) [21:40:09] 6Collaboration-Team-Backlog, 10Flow, 10Parsoid, 10RESTBase, 7Regression: Flow posts being serialized from HTML -> WT without providing Parsoid data-parsoid attributes? - https://phabricator.wikimedia.org/T115236#1726549 (10Mattflaschen) >>! In T115236#1726527, @ssastry wrote: > So, however this missing d... [21:40:28] 6Collaboration-Team-Backlog, 10Flow, 10Parsoid, 10RESTBase, 7Regression: Flow posts being serialized from HTML -> WT without providing Parsoid data-parsoid attributes? - https://phabricator.wikimedia.org/T115236#1726550 (10ssastry) >>! In T115236#1726500, @Mattflaschen wrote: > @Arlolra is there a versio... [21:41:27] 3Collaboration-Team-Current, 10Flow, 5Patch-For-Review, 5WMF-deploy-2015-10-13_(1.27.0-wmf.3): Investigate the uses of the old editor system in Flow, and clean up the uses in handlebars - https://phabricator.wikimedia.org/T114448#1726553 (10Etonkovidova) > Clicking "undo" from the history page Blocking iss... [21:41:43] 3Collaboration-Team-Current, 10Flow: 'undo' for Board description edits is not functioning - https://phabricator.wikimedia.org/T115525#1726533 (10Etonkovidova) [21:46:17] matt_flaschen: Could you take a quick look at https://gerrit.wikimedia.org/r/#/c/246249/ ? [21:46:23] It looks good to me, but I'd like a second opinion [21:48:00] (03CR) 10Catrope: [C: 032] Browser tests: add @integration to some tests [extensions/Flow] - 10https://gerrit.wikimedia.org/r/245973 (owner: 10Sbisson) [21:48:20] (03CR) 10Catrope: [C: 032] Browser tests: Wait for a button to exist before trying to click it [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246212 (owner: 10Sbisson) [21:48:57] 6Collaboration-Team-Backlog, 10Flow, 10MediaWiki-extensions-SpamBlacklist: Flow should not use core parser unconditionally for SpamBlacklist integration - https://phabricator.wikimedia.org/T115526#1726565 (10Mattflaschen) 3NEW [21:49:32] RoanKattouw, sure. [21:50:18] 3Collaboration-Team-Current, 10Echo: Make "mark as read" an explicit action on Special:Notifications - https://phabricator.wikimedia.org/T115528#1726587 (10Catrope) 3NEW [21:50:48] (03Merged) 10jenkins-bot: Browser tests: add @integration to some tests [extensions/Flow] - 10https://gerrit.wikimedia.org/r/245973 (owner: 10Sbisson) [21:50:51] (03Merged) 10jenkins-bot: Browser tests: Wait for a button to exist before trying to click it [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246212 (owner: 10Sbisson) [21:51:46] RoanKattouw, it is fine conceptually. It could potentially add a conversion to the POST path, but it does not look like it will in production due to the abuse filters (which need the old revision anyway). We talked about this a little earlier. Do you want me to test? [21:52:04] Sure [21:52:11] Okay, one sec. [21:52:11] What do you mean by that? [21:52:18] "add a conversion to POST" [21:53:31] (03PS1) 10Paladox: Add support for running composer test [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246424 [21:53:40] RoanKattouw, i.e. when you actually save an edit, it could (in some setups) cause the prior revision to be converted to wikitext when maybe otherwise we wouldn't need to (so perf issue). But I think we're doing that conversion anyway in production [21:53:52] (03PS2) 10Paladox: Add support for running composer test [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246424 [21:53:55] Oh, I see [21:54:06] Because you already need to do that conversion for AbuseFilter [21:54:26] Is that conversion cached though? [21:54:28] (03CR) 10Paladox: "check experimental" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246424 (owner: 10Paladox) [21:54:45] RoanKattouw, right, for https://phabricator.wikimedia.org/diffusion/EFLW/browse/master/includes/SpamFilter/ConfirmEdit.php;0de48459a43f9290ee44580b54b3ffbb568cfef8$24 . [21:54:48] Yes, it's cached. [21:54:51] OK [21:59:48] (03PS3) 10Paladox: Add support for running composer test [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246424 [22:00:02] RoanKattouw, I'm not sure if there's a good "fails on master", "works with change" test (we still get data-parsoid locally), but I'm comfortable merging it unless you want to review further. [22:00:09] I did basic regression testing. [22:00:20] OK [22:00:21] (03CR) 10Paladox: "check experimental" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246424 (owner: 10Paladox) [22:00:25] Yeah I'm comfortable merging it too [22:00:48] Speaking of data-parsoid, I think I may have found a temporary workaround [22:00:50] testing now [22:05:46] (03PS1) 10Catrope: Temporarily disable RESTbase support to avoid data-parsoid issues [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246427 (https://phabricator.wikimedia.org/T115236) [22:06:53] (03PS2) 10Catrope: Temporarily disable RESTbase support to avoid data-parsoid issues [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246427 (https://phabricator.wikimedia.org/T115236) [22:07:12] (03CR) 10Paladox: "This passes." [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246424 (owner: 10Paladox) [22:08:05] (03PS4) 10Paladox: Add support for running composer test [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246424 [22:10:02] 3Collaboration-Team-Current, 10Flow, 7Blocked-on-Parsoid, 5Patch-For-Review: [Regression] In betalabs "Due to a technical error, this post could not be retrieved." for entries with triple curly brackets - https://phabricator.wikimedia.org/T112350#1726654 (10Catrope) a:3Catrope [22:14:51] (03CR) 10Mobrovac: [C: 031] Temporarily disable RESTbase support to avoid data-parsoid issues [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246427 (https://phabricator.wikimedia.org/T115236) (owner: 10Catrope) [22:22:22] 6Collaboration-Team-Backlog, 10Flow, 10Parsoid, 10RESTBase, and 2 others: Flow posts being serialized from HTML -> WT without providing Parsoid data-parsoid attributes? - https://phabricator.wikimedia.org/T115236#1726697 (10Catrope) >>! In T115236#1726640, @gerritbot wrote: > Change 246427 had a related pa... [22:23:10] (03CR) 10Mattflaschen: [C: 032] Check for null edits using wikitext format [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246249 (https://phabricator.wikimedia.org/T60975) (owner: 10Sbisson) [22:25:48] (03Merged) 10jenkins-bot: Check for null edits using wikitext format [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246249 (https://phabricator.wikimedia.org/T60975) (owner: 10Sbisson) [22:27:08] matt_flaschen: Also could you look at https://gerrit.wikimedia.org/r/246427 ? [22:35:08] RoanKattouw, looks fine. Did you test locally? [22:35:42] Yeah I did [22:35:46] I think right now Vagrant uses the else { version. [22:35:48] {{{foo}}} round-trips for me now, it 500ed before [22:35:51] On html2wt IIRC [22:36:02] It depends on whether you have a VRS config for RESTbase locally [22:36:09] which depends on whether you have RESTbase at all [22:37:12] RoanKattouw, yeah, I don't have the role enabled right now. But with your change with that role, it would still use the else {, since there is no VRS definition for Parsoid AFAICT. Going to merge without testing myself. [22:39:23] RoanKattouw, can you create a bug to add this stuff to the Parsoid Vagrant role, paste in the config you used locally, and CC me? [22:39:39] OK [22:39:49] Well that role also needs to install RESTbase then :) [22:40:11] (03CR) 10Mattflaschen: [C: 032] Temporarily disable RESTbase support to avoid data-parsoid issues [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246427 (https://phabricator.wikimedia.org/T115236) (owner: 10Catrope) [22:42:29] RoanKattouw, I thought in this mode it uses VirtualRestConfig but not Parsoid? [22:42:40] (03Merged) 10jenkins-bot: Temporarily disable RESTbase support to avoid data-parsoid issues [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246427 (https://phabricator.wikimedia.org/T115236) (owner: 10Catrope) [22:42:45] There is already a role that installs RESTBase with VirtualRestConfig, but your change would bypass it. [22:43:37] RoanKattouw, oh, whoops, I mean "VirtualRestConfig but not RESTBase"] [22:44:33] Oh OK [23:07:47] 3Collaboration-Team-Current, 10Flow, 5Patch-For-Review, 7Performance, 5WMF-deploy-2015-10-13_(1.27.0-wmf.3): Flow JS loading / deferred loading is very slow - https://phabricator.wikimedia.org/T108313#1726867 (10Etonkovidova) Checked in betalabs - DOMLoadEvent(and Load) - averages - seem to be lower th... [23:18:17] (03CR) 10Cscott: Temporarily disable RESTbase support to avoid data-parsoid issues (031 comment) [extensions/Flow] - 10https://gerrit.wikimedia.org/r/246427 (https://phabricator.wikimedia.org/T115236) (owner: 10Catrope) [23:21:51] matt_flaschen: 'Remove dead code to fix ResourcesTest::testFileExistence' for T113497 - so the references to png files were deleted? [23:22:14] matt_flaschen: even though the images exist in different locations? /mnt/srv/mediawiki/php-master/extensions/WikiEditor/modules/images/dialogs/insert-link-exists.png [23:23:57] etonkovidova, BRB. [23:28:53] Backl [23:30:37] etonkovidova, yeah, I don't think the CSS is referencing those in the WikiEditor location (it's trying to override them, but the replacements it provides don't actually exist. [23:31:34] etonkovidova, a path in CSS is always relative to the CSS file. [23:31:48] matt_flaschen: ok then - I was just making sure that we are not losing valuable unit tests :) [23:32:49] Didn't remove any tests. etonkovidova, if you want, you could test that functionality with LQT WikiEditor (if that functionalty even worked before the change). That I didn't test. [23:33:15] matt_flaschen: ok [23:33:17] The test basically caught the fact that the CSS was referencing an invalid file, so I removed the invalid references. [23:33:35] It's a new test which is why it caught it now. [23:33:45] matt_flaschen: I see [23:34:13] I'm going to go home. I'll be on later tonight. [23:48:29] 6Collaboration-Team-Backlog, 10Flow, 7Availability: DB master connections requested by Flow on GET/HEAD requests - https://phabricator.wikimedia.org/T94028#1726997 (10Catrope) [23:49:02] 6Collaboration-Team-Backlog, 10Flow, 7Availability: DB master connections requested by Flow on GET/HEAD requests - https://phabricator.wikimedia.org/T94028#1153037 (10Catrope)