[00:00:12] (03PS1) 10Catrope: Add .jshintrc [oojs/ui] - 10https://gerrit.wikimedia.org/r/94286 [00:00:32] RoanKattouw: yessir, what's up? [00:00:37] (03CR) 10Legoktm: [C: 032] Remove exclusive module group to reduce HTTP request fragmentation [extensions/Echo] - 10https://gerrit.wikimedia.org/r/93053 (owner: 10Krinkle) [00:00:54] greg-g: Just an EventLogging thingy [00:01:06] ori-l: Do you need EventLogging changes on your end? Wanna share an LD window? [00:01:11] like the last one that killed the cache? :) [00:01:17] (03CR) 10Parent5446: [C: 04-1] "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/94258 (owner: 10Chad) [00:01:23] No, just adding a bunch of events to VE, on my end [00:01:29] Don't know what ori-l would have to do on his end [00:01:35] * greg-g nods [00:01:45] (03PS2) 10EBernhardson: Require core edit permission to perform write actions [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94103 [00:01:53] RoanKattouw: what's a more human readable summary of what's going out? [00:02:05] (03Merged) 10jenkins-bot: Remove exclusive module group to reduce HTTP request fragmentation [extensions/Echo] - 10https://gerrit.wikimedia.org/r/93053 (owner: 10Krinkle) [00:02:08] (03CR) 10jenkins-bot: [V: 04-1] Don't use wfIsWindows() in CDB writers [core] - 10https://gerrit.wikimedia.org/r/94268 (owner: 10Chad) [00:02:10] (sorry, I couldn't manage it on first try) [00:02:15] Krinkle: thanks [00:02:21] (03CR) 10Parent5446: "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/94274 (owner: 10Bartosz Dziewoński) [00:02:33] (03CR) 10Chad: "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/94258 (owner: 10Chad) [00:02:50] greg-g: So, currently VE EventLogging instruments two things, speed of load request and speed of save request. With this change we'll instrument more like 10 things, mostly around how much time various parts of the saving process take and how much time the user spends in the save dialog [00:03:00] sweet [00:03:02] thanks [00:03:07] * greg-g adds to deploy calendar [00:03:11] RoanKattouw: go forth [00:03:15] Excellent [00:03:44] Thanks man [00:04:08] (03CR) 10Parent5446: "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/94258 (owner: 10Chad) [00:04:23] (03CR) 10Bartosz Dziewoński: "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/94274 (owner: 10Bartosz Dziewoński) [00:04:49] (03PS1) 10Catrope: Update VisualEditor to wmf3 branch for cherry-pick [core] (wmf/1.23wmf3) - 10https://gerrit.wikimedia.org/r/94287 [00:05:30] (03CR) 10Catrope: [C: 032 V: 032] Update VisualEditor to wmf3 branch for cherry-pick [core] (wmf/1.23wmf3) - 10https://gerrit.wikimedia.org/r/94287 (owner: 10Catrope) [00:08:19] (03CR) 10Chad: "Everything about QuickTemplate makes me cry :(" [core] - 10https://gerrit.wikimedia.org/r/94248 (owner: 10Jdlrobson) [00:11:01] (03CR) 10Bartosz Dziewoński: "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/94248 (owner: 10Jdlrobson) [00:13:16] Krinkle: Argh, jslint is non-voting [00:13:20] Krinkle: https://gerrit.wikimedia.org/r/#/c/94286/ [00:14:02] (03CR) 10Catrope: [C: 04-2] "jshint failed (but Jenkins is misconfigured so it didn't V-1)" [oojs/ui] - 10https://gerrit.wikimedia.org/r/94286 (owner: 10Catrope) [00:14:07] Right, this is one of the few that are non-voting by default. Add voting: true; in zuul-config. [00:14:14] I'll do it [00:15:12] Yeah please do [00:15:15] I have no idea where to add that [00:15:38] (03PS1) 10Krinkle: Make oojs jshint voting [integration/zuul-config] - 10https://gerrit.wikimedia.org/r/94302 [00:15:42] Now you do :) [00:15:52] (03CR) 10Krinkle: [C: 032] Make oojs jshint voting [integration/zuul-config] - 10https://gerrit.wikimedia.org/r/94302 (owner: 10Krinkle) [00:18:28] (03PS1) 10EBernhardson: Prevent rendering topics like posts [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94304 [00:20:34] (03Merged) 10jenkins-bot: Make oojs jshint voting [integration/zuul-config] - 10https://gerrit.wikimedia.org/r/94302 (owner: 10Krinkle) [00:27:43] (03PS1) 10Aaron Schulz: Speed up thumbnail purging a bit for stash script [core] - 10https://gerrit.wikimedia.org/r/94305 [00:30:43] (03PS8) 10EBernhardson: Moderate a post or topic [extensions/Flow] - 10https://gerrit.wikimedia.org/r/91135 [00:30:45] (03CR) 10jenkins-bot: [V: 04-1] Moderate a post or topic [extensions/Flow] - 10https://gerrit.wikimedia.org/r/91135 (owner: 10EBernhardson) [00:32:07] (03PS9) 10EBernhardson: Moderate a post or topic [extensions/Flow] - 10https://gerrit.wikimedia.org/r/91135 [00:35:49] (03PS1) 10Yuvipanda: Add APISandbox Rolee [vagrant] - 10https://gerrit.wikimedia.org/r/94306 [00:36:41] (03PS2) 10Yuvipanda: Add APISandbox Role [vagrant] - 10https://gerrit.wikimedia.org/r/94306 [00:39:33] ^d: https://gerrit.wikimedia.org/r/#/c/94305/ 1 line [00:40:38] YuviPanda: It would be cool if there was some kind of meta-extension thing so you could do "vagrant install ApiSandbox" and it just did it. [00:40:52] greg-g: was there a deployment today? [00:41:01] (03CR) 10Chad: [C: 032] Speed up thumbnail purging a bit for stash script [core] - 10https://gerrit.wikimedia.org/r/94305 (owner: 10Aaron Schulz) [00:41:04] legoktm: yeah, me and ori-l talked about it, and I think it was considered a bit overkill, and put on backburner [00:41:16] legoktm: i'm working on a simple wikibase repo one now :D [00:41:19] ooh :D [00:41:57] api sandbox would be a good thing to have in vagrant with wikibase [00:42:01] for developers [00:42:04] yeah [00:42:10] * aude uses my bot to test the api [00:42:16] (03PS5) 10EBernhardson: Copy tweaks [extensions/Flow] - 10https://gerrit.wikimedia.org/r/93768 (owner: 10Matthias Mullie) [00:42:18] not everyone can do that :) [00:43:56] (03Merged) 10jenkins-bot: Speed up thumbnail purging a bit for stash script [core] - 10https://gerrit.wikimedia.org/r/94305 (owner: 10Aaron Schulz) [00:44:29] aude: heh! with vagrant it's a simple 'enable-role apisandbox' then 'vagrant provision' :D [00:45:50] YuviPanda: cool [00:52:37] (03PS3) 10Chad: Give the search page a little UI love [core] - 10https://gerrit.wikimedia.org/r/94276 [00:53:08] (03CR) 10Chad: "PS3 gives a little more padding to the right side to keep that (X) from rubbing up against the border." [core] - 10https://gerrit.wikimedia.org/r/94276 (owner: 10Chad) [00:58:10] (03CR) 10Catrope: "recheck" [oojs/ui] - 10https://gerrit.wikimedia.org/r/94286 (owner: 10Catrope) [00:58:14] (03CR) 10jenkins-bot: [V: 04-1] Add .jshintrc file [oojs/ui] - 10https://gerrit.wikimedia.org/r/94286 (owner: 10Catrope) [00:58:55] (03PS3) 10Catrope: Add .jshintrc [oojs/ui] - 10https://gerrit.wikimedia.org/r/94286 [01:00:51] (03PS2) 10TTO: Stop using prettifyIP for user links [core] - 10https://gerrit.wikimedia.org/r/92263 [01:05:24] hey roan, got a sec? [01:06:10] ori-l: Sure [01:07:09] it'd be nice to have !logs of git rebases / commits on tin, so i added a post-commit / post-rewrite / post-merge hook that !logs 'git show-branch --merge-base @{upstream}' [01:07:34] i'm not sure if that's the best way to get the right commit, and i'm also not sure how to get the starting point [01:08:12] there's pre-commit / pre-rebase / pre-rewrite but i am pretty confused about what the state of the repo is [01:09:12] OK, so [01:09:24] What is the output of show-branch, a hash? [01:09:33] Or is the commit summary included? [01:09:53] it's a hash, but i actually get the change ID by running a couple of additional commands [01:10:26] Do we have a cache for article data (page properties, restrictions, etc)? [01:10:37] this is whta i have: https://gist.github.com/atdt/7364641#file-git-log-hook-sh [01:10:37] It seems like we should do, but on my wiki I'm getting 29 queries per page load [01:11:36] git-deploy / trebuchet would replace that at some point presumably but it seemed like a good stopgap [01:12:02] ori-l: So we definitely cannot do commit summaries because of security patches [01:12:12] And I'm wary of doing hashes / change-IDs for that same reason [01:12:23] security patches won't show up in git show-branch --merge-base @{upstream} [01:12:30] that's why i didn't just git rev-parse HEAD [01:12:33] aha [01:12:56] I see, nifty [01:13:31] So it gives you the latest commit that's both in HEAD and origin/ ? [01:13:59] I think so, but I was hoping you'd know this extensively and would thus be able to sanity-check :) [01:14:30] heh sorry no I'm not personally familiar with show-bracnh [01:14:37] But a quick test on tin seems to indicate it works [01:14:58] k cool thanks for verifying [01:15:27] Also maybe the post-merge hook should be ZOMG Y U CREATE MERGE COMMIT [01:15:42] security patches [01:15:55] We tell people to use rebases for that [01:17:00] Also, we should ask Chris how comfortable he is revealing the existence of undeployed patches [01:17:09] why would this? [01:17:15] reveal their existence, I mean [01:17:25] The fact that someone did a rebase or merge flags that [01:17:37] but that's not !logged [01:17:43] I guess you're right though that it doesn't reveal what the action was [01:17:50] And the upstream hash will be the same [01:18:08] yeah [01:18:17] well, [01:18:23] when chris actually commits the patch on tin [01:18:29] there would be a duplicate !log [01:18:29] Yes, I was about to say [01:18:32] We can't do post-commit [01:18:40] because we never want to reveal that a local commit happened [01:18:47] Also, we have to filter duplicate !logs in general [01:19:00] Consider git pull; oh crap merge commit; git pull --rebase [01:19:09] people do local commits in wmf-config tho [01:19:27] Yeah but they git-review them [01:19:37] And as soon as they are pulled back down the upstream does change [01:19:47] we've disabled push for that repo AFAIK [01:20:24] I guess the change would not be recorded until the next pull [01:20:29] so that's suboptimal [01:20:49] But you already have this problem because you're comparing to origin/master [01:21:05] And at the time they make the local commit, origin/master doesn't have it, and so it will produce (or attempt to produce) a duplicate [01:21:27] It won't ever actually be able to discover the "correct" hash until a fetch/pull occurs [01:22:23] well, it grabs the change id [01:22:56] perhaps it should also echo the change id to /var/log/logged-commits and decline to !log if the change id appears in the log [01:23:06] Right [01:23:14] So filter by Change-Id instead [01:23:29] yeah, i think that works [01:23:34] But you'd still have to go back to the common ancestor in origin though right? [01:23:38] To avoid logging security commits [01:24:14] yep [01:25:16] <^d> RoanKattouw: So, the rebase thing shouldn't happen anymore if someone will review my change. [01:25:22] <^d> I want to make pull rebase by default. [01:25:29] Aweesome [01:25:39] <^d> Really, people should never blindly pull, but I know for a fact they do, so I want better behavior. [01:26:09] <^d> Anyway, https://gerrit.wikimedia.org/r/#/c/94301/ since it's on topic. [01:26:24] <^d> I tried to fix this awhile ago, but it's stupid because you have to config before checkout. [01:26:46] +2ed [01:27:55] <^d> So yeah, now git pull == git pull -r going forward. [01:27:59] <^d> :) [01:28:04] yay [01:29:05] Does Wikimedia's gerrit not support all of the REST API? [01:29:08] https://gerrit.wikimedia.org/r/changes/I838b546caa3f5887e6bd312017c1fa71d3c3463a/in doesn't work [01:29:17] https://gerrit-review.googlesource.com/Documentation/rest-api-changes.html#get-included-in [01:29:21] <^d> legoktm: We're a bit behind :( [01:29:30] hrm... [01:29:45] <^d> We're currently running something resembling 2.7-rc2. [01:30:02] <^d> We have the included documentation though, so you can see what we *do* support. [01:30:03] Would be nice if their documentation had @since tags. [01:30:08] ah [01:30:10] <^d> https://gerrit.wikimedia.org/r/Documentation/index.html [01:30:13] thanks [01:30:15] <^d> yw [01:30:38] yeah, not listed there. [01:30:54] <^d> Maybe we'll upgrade when 2.8 goes final. [01:31:06] OK who's been +2ing things in mediawiki-config without deploying them [01:31:14] I just pulled and got a couple of other commits riding along [01:31:33] * RoanKattouw glares at Reedy [01:31:37] (03CR) 10EBernhardson: [C: 032] (bug 56126) Inconsistent replying behavior [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94124 (owner: 10Matthias Mullie) [01:31:49] (03Merged) 10jenkins-bot: (bug 56126) Inconsistent replying behavior [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94124 (owner: 10Matthias Mullie) [01:32:13] I think I'll just hack it up using github's API. [01:33:04] RoanKattouw: For Oris change there's nothing to deploy [01:33:41] Uh [01:33:52] For 1 of oris changes there's nothing to deploy [01:33:54] For one of his changes there isn't [01:33:57] And for Chad's there isn't either [01:34:10] But I just synced InitSettings for the wm2014 logo, and bits/static-current for the symlink change [01:34:30] Alright all done [01:34:31] wm2014 logo was deployed before it was committed anyway [01:34:37] and jenkins was being sloooooooow [01:34:40] ha [01:34:42] thanks [01:35:05] But you didn't commit it locally then? Or maybe git's output was just confusing in telling me that commit was part of the update [01:35:40] I thought I had [01:35:58] Hmm, must be git smoking crack then [01:36:09] And that one commit of Ori's was +2ed by himself [01:36:16] * RoanKattouw apologizes for wrongly blaming Reedy [01:37:17] heh [01:37:40] (03CR) 10BryanDavis: [C: 032] Add APISandbox Role [vagrant] - 10https://gerrit.wikimedia.org/r/94306 (owner: 10Yuvipanda) [01:37:43] (03Merged) 10jenkins-bot: Add APISandbox Role [vagrant] - 10https://gerrit.wikimedia.org/r/94306 (owner: 10Yuvipanda) [01:43:54] (03CR) 10Krinkle: [C: 031] Add .jshintrc [oojs/ui] - 10https://gerrit.wikimedia.org/r/94286 (owner: 10Catrope) [01:45:45] (03PS1) 10MarkAHershberger: Make it possible to install extensions using Composer [core] (REL1_20) - 10https://gerrit.wikimedia.org/r/94310 [01:46:26] (03CR) 10jenkins-bot: [V: 04-1] Make it possible to install extensions using Composer [core] (REL1_20) - 10https://gerrit.wikimedia.org/r/94310 (owner: 10MarkAHershberger) [01:46:33] (03Abandoned) 10MarkAHershberger: Make it possible to install extensions using Composer [core] (REL1_20) - 10https://gerrit.wikimedia.org/r/94310 (owner: 10MarkAHershberger) [01:56:56] (03PS2) 10MarkAHershberger: Make it possible to install extensions using Composer [core] (REL1_22) - 10https://gerrit.wikimedia.org/r/93120 [01:58:55] (03CR) 10MarkAHershberger: [C: 032] "Thanks, all." [core] (REL1_22) - 10https://gerrit.wikimedia.org/r/94119 (owner: 10Hashar) [02:01:00] (03CR) 10MarkAHershberger: [C: 032] ":)" [core] (REL1_22) - 10https://gerrit.wikimedia.org/r/93704 (owner: 10Hashar) [02:03:06] (03CR) 10MarkAHershberger: [C: 032] ":)" [core] (REL1_22) - 10https://gerrit.wikimedia.org/r/93703 (owner: 10Hashar) [02:05:45] (03CR) 10MarkAHershberger: [C: 032] UploadStash::removeFileNoAuth shouldn't need auth [core] (REL1_22) - 10https://gerrit.wikimedia.org/r/92545 (owner: 10Reedy) [02:07:11] (03Merged) 10jenkins-bot: MWException: Cleanup exception message output [core] (REL1_22) - 10https://gerrit.wikimedia.org/r/93704 (owner: 10Hashar) [02:07:12] (03Merged) 10jenkins-bot: redact exception traces and abstract getTrace [core] (REL1_22) - 10https://gerrit.wikimedia.org/r/93703 (owner: 10Hashar) [02:09:00] (03CR) 10MarkAHershberger: [C: 032] FormatJson: Remove whitespace from empty arrays and objects [core] (REL1_22) - 10https://gerrit.wikimedia.org/r/92489 (owner: 10PleaseStand) [02:22:00] YuviPanda, bd808|BUFFER : my first labs-vagrant provision last night failed, see https://wikitech.wikimedia.org/wiki/Talk:Labs-vagrant. I'm re-provisioning [02:24:43] spagewmf: you do need to se sudo with the labs-vagrant command [02:24:50] *use [02:25:05] This is my "howto": https://www.mediawiki.org/wiki/User:BDavis_%28WMF%29/Notes/Labs-vagrant [02:26:06] bd808 yes I did, `sudo labs-vagrant provision` fails with same git clone returned 128. [02:26:21] thx, I'll check out your "howto" [02:27:09] The NFS homedirs part is not strictly needed, just makes them a little faster than the default storage. [02:31:00] bd808 heh, I was just about to about that. What I did was very close to your instructions. http://stackoverflow.com/questions/9617336/how-to-resolve-git-did-not-exit-cleanly-exit-code-128-error-on-tortoisegit suggests error 128 is a problem with ssh keys, sounds plausible [02:35:27] /home/vagrant has no .ssh and /root/.ssh is empty, maybe that's part of the problem. But the clone commands are using https. [02:37:33] spagewmf: if it cloned via https and that's in the remote url in .git/config , maybe you have to change it to ssh [02:37:50] had that before, cloned using https then wanted to push.. had to edit config [02:39:04] well, there's nothing in /home/vagrant/.git or /root/.git and the commands are git clone https. Maybe it's a permission problem. [02:39:28] (03PS4) 10Chad: Clean up CDB classes [core] - 10https://gerrit.wikimedia.org/r/94258 [02:39:29] (03PS5) 10Chad: Don't use wfIsWindows() in CDB writers [core] - 10https://gerrit.wikimedia.org/r/94268 [02:50:31] mutante, bd808: what is the permissions of /home/vagrant on your working labs-vagrant instance? Mine is root root 755 [02:50:39] thx for your help BTW! [02:53:21] aha, immediately before the git clone failures there's a "notice: /Stage[main]/Mediawiki::Parsoid/Git::Clone[mediawiki/extensions/Parsoid]/Exec[git clone mediawiki/extensions/Parsoid]/returns: fatal: Could not change back to '/home/spage': Permission denied" [02:53:36] sudo everything [02:53:44] spagewmf: I've seen a similar error before [02:54:04] I never figured out the cause of it [02:54:18] spagewmf: sorry, i dont have a vagrant instance, that was just a general git thing [02:54:26] try running the command again; that has worked for me in the past [02:54:59] legoktm: I'm running `sudo labs-vagrant provision`. bd808 I get those git clone failures every time. I'm going to change my home dir to 755 and see what happens [02:55:54] /home/legoktm is drwx-----x , I've never seen that one before :) [02:56:00] lolwut. [02:56:17] what instance is that on? [02:56:36] projectstorage.pmtpa.wmnet:/editor-engagement-home on /home type fuse.glusterfs (rw,default_permissions,allow_other,max_read=131072) [02:57:09] on ee-flow and ee-flow-extra [02:58:16] huh. [02:58:26] is that linked to other instances in the ee group? [02:58:40] I might have fiddled with it on some other instance, but I don't recall ever logging into ee-flow-* [02:59:58] (03PS1) 10MarkAHershberger: Clean up the release notes and reorder them a bit. [core] (REL1_22) - 10https://gerrit.wikimedia.org/r/94314 [03:00:09] /home/yuvipanda is 755, maybe that's why this works for him. I changed my /home dir to 755 and no more git clone failures from `sudo labs-vagrant provision` [03:00:37] (03CR) 10jenkins-bot: [V: 04-1] Clean up the release notes and reorder them a bit. [core] (REL1_22) - 10https://gerrit.wikimedia.org/r/94314 (owner: 10MarkAHershberger) [03:00:41] legoktm it's a mounted directory, /home/legoktm will have those perms everywhere [03:01:13] ah [03:01:19] yeah i probably screwed with it somewhere else [03:10:46] MelBee, what permissions should I use on my labs instance home directory? [03:11:42] you shouldn't put anything into your homedir for the most part [03:12:59] Ryan_Lane: yes, EE (Core features) use nifty /data/ project. But labs-vagrant provision wants access to my homedir, it seems. [03:13:17] YuviPanda: ^^ ?? [03:13:25] gross [03:13:37] I don't think it's purposeful. [03:15:00] `fatal: Could not change back to '/home/spage'` sounds like something that happens because sudo isn't wiping out $HOME [03:17:01] I'm not on a computer with access to my labs key at the moment or I'd try `HOME=/root sudo labs-vagrant provision` to see if that makes a difference [04:29:45] (03PS5) 10Chad: Clean up CDB classes [core] - 10https://gerrit.wikimedia.org/r/94258 [04:29:46] (03PS6) 10Chad: Don't use wfIsWindows() in CDB writers [core] - 10https://gerrit.wikimedia.org/r/94268 [04:30:17] (03CR) 10GICodeWarrior: [C: 04-1] "(1 comment)" [extensions/CategoryTree] - 10https://gerrit.wikimedia.org/r/94099 (owner: 10Navdeep Bagga) [04:30:18] (03PS5) 10Chad: Update docs in CdbPHP.php [core] - 10https://gerrit.wikimedia.org/r/94267 (owner: 10Siebrand) [04:35:58] (03PS5) 10Tim Starling: filebackend: improved "adviseStat" performance [core] - 10https://gerrit.wikimedia.org/r/94237 (owner: 10Aaron Schulz) [04:36:08] (03CR) 10Tim Starling: [C: 032] filebackend: improved "adviseStat" performance [core] - 10https://gerrit.wikimedia.org/r/94237 (owner: 10Aaron Schulz) [04:40:50] (03Merged) 10jenkins-bot: filebackend: improved "adviseStat" performance [core] - 10https://gerrit.wikimedia.org/r/94237 (owner: 10Aaron Schulz) [04:48:06] (03PS3) 10Tim Starling: Work around HHVM bug with stdin/stdout in proc_open() [core] - 10https://gerrit.wikimedia.org/r/94065 (owner: 10Chad) [04:48:11] (03CR) 10Tim Starling: [C: 032] Work around HHVM bug with stdin/stdout in proc_open() [core] - 10https://gerrit.wikimedia.org/r/94065 (owner: 10Chad) [04:51:25] ow [04:51:27] spagewmf: hey! [04:51:31] spagewmf: sudo -u vagrant -s [04:51:32] then do stuff/ [04:51:33] ? [04:51:39] spagewmf: I thought I fixed the home issue [04:52:48] (03Merged) 10jenkins-bot: Work around HHVM bug with stdin/stdout in proc_open() [core] - 10https://gerrit.wikimedia.org/r/94065 (owner: 10Chad) [04:58:02] (03CR) 10Tim Starling: [C: 04-1] "(2 comments)" [core] - 10https://gerrit.wikimedia.org/r/87482 (owner: 10saper) [05:06:04] (03PS3) 10Tim Starling: Allow ParserLimitReportFormat to change $value [core] - 10https://gerrit.wikimedia.org/r/80412 (owner: 10Anomie) [05:06:12] (03CR) 10Tim Starling: [C: 032] Allow ParserLimitReportFormat to change $value [core] - 10https://gerrit.wikimedia.org/r/80412 (owner: 10Anomie) [05:10:50] (03Merged) 10jenkins-bot: Allow ParserLimitReportFormat to change $value [core] - 10https://gerrit.wikimedia.org/r/80412 (owner: 10Anomie) [05:14:00] greg-g: did you see my pings about deployment dashboard? [05:17:22] (03PS9) 10Tim Starling: Use the new limit report hooks [extensions/Scribunto] - 10https://gerrit.wikimedia.org/r/53804 (owner: 10Anomie) [05:18:05] Ryan_Lane: yeah, just did [05:18:13] sounds like a great option [05:18:40] it was my original intention for the data, but a cli was quicker/easier to build [05:19:57] bd808: thanks for merging it! [05:20:45] Ryan_Lane: cool, I'm copy/pasting your raw thoughts from -operations to the wiki so I don't forget tomorrow :) [05:20:53] (03PS4) 10Tim Starling: Expose mw.log data on preview [extensions/Scribunto] - 10https://gerrit.wikimedia.org/r/55602 (owner: 10Anomie) [05:23:47] now, to some non-work computer stuff... [05:23:56] YuviPanda: Np [05:33:41] (03PS1) 10Pastakhov: add date/time functions (some of these) [extensions/Foxway] (develop) - 10https://gerrit.wikimedia.org/r/94317 [05:35:12] (03CR) 10Pastakhov: [C: 032] add date/time functions (some of these) [extensions/Foxway] (develop) - 10https://gerrit.wikimedia.org/r/94317 (owner: 10Pastakhov) [05:35:19] (03Merged) 10jenkins-bot: add date/time functions (some of these) [extensions/Foxway] (develop) - 10https://gerrit.wikimedia.org/r/94317 (owner: 10Pastakhov) [05:55:10] (03PS1) 10EBernhardson: Fix the IndexTest [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94318 [07:20:20] (03CR) 10EBernhardson: "(8 comments)" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94102 (owner: 10Werdna) [07:36:56] (03CR) 10EBernhardson: "I believe the changes in .flow-topic-reply-container were attempting to fix the issue where, with VE disabled, the wikitext editor changes" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94165 (owner: 10Matthias Mullie) [07:44:02] (03CR) 10Nikerabbit: "How is one supposed to override the debug callback now?" [core] - 10https://gerrit.wikimedia.org/r/94258 (owner: 10Chad) [07:47:56] (03CR) 10Robert Vogel: "Seems I've waited for too long to review this. :) Thanks a lot!" [extensions/BlueSpiceExtensions] - 10https://gerrit.wikimedia.org/r/94082 (owner: 10Aude) [08:16:24] (03PS3) 10Hashar: Make it possible to install extensions using Composer [core] (REL1_22) - 10https://gerrit.wikimedia.org/r/93120 (owner: 10MarkAHershberger) [08:16:44] (03CR) 10Hashar: [C: 032] Make it possible to install extensions using Composer [core] (REL1_22) - 10https://gerrit.wikimedia.org/r/93120 (owner: 10MarkAHershberger) [08:21:27] (03Merged) 10jenkins-bot: Make it possible to install extensions using Composer [core] (REL1_22) - 10https://gerrit.wikimedia.org/r/93120 (owner: 10MarkAHershberger) [09:02:28] (03CR) 10Nemo bis: "(1 comment)" [extensions/CentralAuth] - 10https://gerrit.wikimedia.org/r/7274 (owner: 10Szymon Świerkosz) [09:04:50] (03PS10) 10Nemo bis: Implement global user CSS/JS module [extensions/CentralAuth] - 10https://gerrit.wikimedia.org/r/7274 (owner: 10Szymon Świerkosz) [09:08:04] (03CR) 10Nemo bis: "I tweaked the second part of the commit message but I don't understand the first part:" [extensions/CentralAuth] - 10https://gerrit.wikimedia.org/r/7274 (owner: 10Szymon Świerkosz) [09:17:57] (03PS2) 10Navdeep Bagga: Category Tree Extension: Fix CSS according to the vector layout [extensions/CategoryTree] - 10https://gerrit.wikimedia.org/r/94099 [09:43:46] (03CR) 10Mark Bergsma: [C: 031] "Awesome, thanks." [core] - 10https://gerrit.wikimedia.org/r/94186 (owner: 10BryanDavis) [10:01:46] (03PS1) 10Tpt: (bug 56669) fix failure when $links = array() and $params = null [extensions/ProofreadPage] - 10https://gerrit.wikimedia.org/r/94331 [10:20:43] (03CR) 10Nemo bis: "Tim, have you tested it more? The committer says: "My commit message could be wrong as Tim said. But the commit fixed our problem, the dat" [core] - 10https://gerrit.wikimedia.org/r/59046 (owner: 10Yosnoop) [11:04:34] (03CR) 10Springle: [C: 031] "Seems fine. Changing the names in production should definitely be done for consistency." [extensions/Echo] - 10https://gerrit.wikimedia.org/r/40095 (owner: 10Dereckson) [12:02:36] (03CR) 10Helder.wiki: "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/77256 (owner: 10Krinkle) [12:06:14] y'know, i thought that tonight i wouldn't end up waking up at a silly hour [12:06:19] turns out that i was wrong [12:23:24] (03PS2) 10Matthias Mullie: auto-expand text-area as content grows [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94285 [12:25:32] (03PS1) 10Matthias Mullie: Fix element that seems to have changed in VE [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94347 [12:35:17] (03PS2) 10Matthias Mullie: Design changes [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94165 [12:39:01] (03CR) 10Matthias Mullie: "New patch set moved the new topic icon to the left." [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94165 (owner: 10Matthias Mullie) [13:06:27] (03CR) 10Daniel Kinzler: [C: 032] "works for me." [core] - 10https://gerrit.wikimedia.org/r/93737 (owner: 10Aude) [13:08:06] (03PS1) 10Mwjames: Asserts if a stream wrapper 'uploadsource' already exists [core] - 10https://gerrit.wikimedia.org/r/94351 [13:09:44] (03Merged) 10jenkins-bot: Add convenience maintenance script wrapper [core] - 10https://gerrit.wikimedia.org/r/93737 (owner: 10Aude) [13:52:04] (03PS1) 10Sakshi: Modified File [extensions/CSS] - 10https://gerrit.wikimedia.org/r/94353 [13:53:24] (03PS2) 10Umherirrender: API: Fix possible errors to avoid Unknown error [extensions/CheckUser] - 10https://gerrit.wikimedia.org/r/92193 [13:53:42] (03PS2) 10Sakshi: Modified File [extensions/CSS] - 10https://gerrit.wikimedia.org/r/94353 [14:29:08] (03Draft1) 10Aude: Improve variable naming in EnhancedChangesList [core] - 10https://gerrit.wikimedia.org/r/94357 [14:31:21] (03CR) 10Aude: "still open to suggestions on improving the tests :)" [core] - 10https://gerrit.wikimedia.org/r/93611 (owner: 10Aude) [14:32:15] (03CR) 10Aude: "if it makes reviewing easier, you can try rebasing this against Ia6a371d (tests)" [core] - 10https://gerrit.wikimedia.org/r/94357 (owner: 10Aude) [14:32:25] (03PS6) 10Matthias Mullie: Copy tweaks [extensions/Flow] - 10https://gerrit.wikimedia.org/r/93768 [14:33:55] (03PS1) 10Matthias Mullie: Never use getModeratedByUserText in history, always revision usertext [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94358 [14:34:52] (03PS1) 10Aude: Add visibility and setters for RCCacheEntry variables [core] - 10https://gerrit.wikimedia.org/r/94360 [14:36:30] (03CR) 10Krinkle: [C: 032] Move files from modules/oojs-ui to the repository root [oojs/ui] - 10https://gerrit.wikimedia.org/r/94264 (owner: 10Catrope) [14:36:34] (03PS1) 10Umherirrender: Let mw.util.addCSS also return a CSSStyleSheet in IE8 [core] - 10https://gerrit.wikimedia.org/r/94361 [14:38:16] (03CR) 10jenkins-bot: [V: 04-1] Move files from modules/oojs-ui to the repository root [oojs/ui] - 10https://gerrit.wikimedia.org/r/94264 (owner: 10Catrope) [14:39:00] (03CR) 10Krinkle: [C: 032 V: 032] "In addition this also grants JenkinsBot the right to merge at all. That's why the gate-and-submit pipeline didn't finish as expected" [oojs/ui] (refs/meta/config) - 10https://gerrit.wikimedia.org/r/94279 (owner: 10Krinkle) [14:40:12] (03CR) 10Krinkle: [C: 032 V: 032] Restrict Verified/Submit to JenkinsBot and l10n-bot [oojs/core] (refs/meta/config) - 10https://gerrit.wikimedia.org/r/94278 (owner: 10Krinkle) [14:40:59] (03CR) 10Krinkle: [C: 032] Add .jshintrc [oojs/ui] - 10https://gerrit.wikimedia.org/r/94286 (owner: 10Catrope) [14:41:11] (03Merged) 10jenkins-bot: Add .jshintrc [oojs/ui] - 10https://gerrit.wikimedia.org/r/94286 (owner: 10Catrope) [14:44:26] (03CR) 10PleaseStand: [C: 031] "This is one of the changes I made in I7daa7e41ce275c105805104872fb79ef43bc0a52." [core] - 10https://gerrit.wikimedia.org/r/93777 (owner: 10Umherirrender) [14:48:03] (03CR) 10Bartosz Dziewoński: [C: 032] Fix call to function applyPatch in MysqlUpdater [core] - 10https://gerrit.wikimedia.org/r/93777 (owner: 10Umherirrender) [14:48:33] (03PS1) 10Bartosz Dziewoński: Fix call to function applyPatch in MysqlUpdater [core] (REL1_22) - 10https://gerrit.wikimedia.org/r/94364 [14:49:00] (03CR) 10Bartosz Dziewoński: [C: 032] "Backporting, this fixes a silly fatal when upgrading from ancient MW versions." [core] (REL1_22) - 10https://gerrit.wikimedia.org/r/94364 (owner: 10Bartosz Dziewoński) [14:51:16] (03Merged) 10jenkins-bot: Fix call to function applyPatch in MysqlUpdater [core] - 10https://gerrit.wikimedia.org/r/93777 (owner: 10Umherirrender) [14:54:52] (03Merged) 10jenkins-bot: Fix call to function applyPatch in MysqlUpdater [core] (REL1_22) - 10https://gerrit.wikimedia.org/r/94364 (owner: 10Bartosz Dziewoński) [15:04:51] MatmaRex: hi [15:05:58] (03PS7) 10Matthias Mullie: Copy tweaks [extensions/Flow] - 10https://gerrit.wikimedia.org/r/93768 [15:06:00] (03PS5) 10Matthias Mullie: Move output functionality from AbstractRevision to Templating [extensions/Flow] - 10https://gerrit.wikimedia.org/r/93953 [15:06:42] hi liangent [15:07:08] MatmaRex: [jquery.makecollapsible] is it possible to add [show/hide] text to custom or premade toggles? [15:08:03] (03CR) 10Matthias Mullie: "Siebrand: made it more verbose (Username suppressed) and italic (for now; TBD with design): http://www.mullie.eu/temp/93953-2.png" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/93953 (owner: 10Matthias Mullie) [15:08:23] liangent: no, i don't think you can do that in wikitext [15:08:25] (did I ask this before? anyway I can't find such a feature in jquery.makeCollapsible.js [15:09:06] you probably could make it happen by calling $().makeCollapsbile() from JS with some options, but it's not possible otherwise [15:09:58] I can't find any path in this way either [15:10:24] if I pass in a toggleText in options, it dumbly replace text in the first [15:11:24] ah. hm. :( [15:11:40] and if I abuse this feature to add a , should I set linksPassthru ? [15:12:22] if I set that, clicking on that won't collapse/expand the box [15:12:39] if not, it doesn't preventDefault&stopPropagation [15:13:54] linksPassthru: true makes clicks on elements use the native behavior (go to link's href) instead of toggling things [15:14:13] if the is just a placeholder (e.g. href="#"), then linksPassthru should be false [15:14:46] it is useful e.g. to apply to navboxes if you want to make the top bar toggle the entire navbox, but the links to be clickable [15:15:54] MatmaRex: ok but then I can't use other links in my own togglers [15:16:10] (03PS1) 10Matthias Mullie: Fix parameter for a couple of msg; should be creator name, not revision user [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94368 [15:16:42] liangent: well, yeah [15:16:45] MatmaRex: actually it might better to have $textContainer = $toggle.find( '> a, .mw-collapsible-toggle-text' ); instead of $textContainer = $toggle.find( '> a' ); [15:17:12] or maybe $textContainer = $toggle.find( '> a, .mw-collapsible-toggle-text' ).not( '.mw-collapsible .mw-collapsible-toggle-text' ); [15:17:16] liangent: sorry, i don't have time to play with that right now – patches with tests welcome :) [15:27:50] (03PS1) 10Phuedx: Add the gettingstarted tour from the GuidedTour extension. [extensions/GettingStarted] - 10https://gerrit.wikimedia.org/r/94369 [15:28:24] (03PS1) 10Phuedx: Remove the gettingstarted tour, modules, messages, and message documentation. [extensions/GuidedTour] - 10https://gerrit.wikimedia.org/r/94370 [15:49:09] (03PS1) 10Manybubbles: Term containing * match against unstemmed text [extensions/CirrusSearch] - 10https://gerrit.wikimedia.org/r/94373 [15:50:18] (03PS1) 10Manybubbles: Tests for term containing * match unstemmed text [extensions/CirrusSearch] - 10https://gerrit.wikimedia.org/r/94374 [15:54:40] (03CR) 10Bartosz Dziewoński: [C: 032] Improve variable naming in EnhancedChangesList [core] - 10https://gerrit.wikimedia.org/r/94357 (owner: 10Aude) [15:55:53] (03CR) 10Dereckson: [C: 031] "Verified per http://msdn.microsoft.com/en-us/library/ms535871.aspx, this remark being the more pertinent content:" [core] - 10https://gerrit.wikimedia.org/r/94361 (owner: 10Umherirrender) [15:56:14] (03CR) 10Matthias Mullie: "(3 comments)" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94103 (owner: 10EBernhardson) [15:57:54] (03Merged) 10jenkins-bot: Improve variable naming in EnhancedChangesList [core] - 10https://gerrit.wikimedia.org/r/94357 (owner: 10Aude) [15:59:41] (03CR) 10Bartosz Dziewoński: [C: 032] "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/94361 (owner: 10Umherirrender) [16:02:42] (03Merged) 10jenkins-bot: Let mw.util.addCSS also return a CSSStyleSheet in IE8 [core] - 10https://gerrit.wikimedia.org/r/94361 (owner: 10Umherirrender) [16:13:30] (03CR) 10Matthias Mullie: [C: 032] Fix the IndexTest [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94318 (owner: 10EBernhardson) [16:13:46] (03Merged) 10jenkins-bot: Fix the IndexTest [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94318 (owner: 10EBernhardson) [16:13:57] (03CR) 10Matthias Mullie: [C: 032] Prevent rendering topics like posts [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94304 (owner: 10EBernhardson) [16:14:07] (03Merged) 10jenkins-bot: Prevent rendering topics like posts [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94304 (owner: 10EBernhardson) [16:31:34] (03CR) 10Rtdwivedi: [C: 031] (bug 56669) fix failure when $links = array() and $params = null [extensions/ProofreadPage] - 10https://gerrit.wikimedia.org/r/94331 (owner: 10Tpt) [16:35:34] (03CR) 10Krinkle: "recheck" [oojs/ui] - 10https://gerrit.wikimedia.org/r/94264 (owner: 10Catrope) [16:49:06] (03CR) 10Tpt: [C: 032] "Reviewed by Aarti" [extensions/ProofreadPage] - 10https://gerrit.wikimedia.org/r/94331 (owner: 10Tpt) [16:49:13] (03Merged) 10jenkins-bot: (bug 56669) fix failure when $links = array() and $params = null [extensions/ProofreadPage] - 10https://gerrit.wikimedia.org/r/94331 (owner: 10Tpt) [17:09:23] (03PS1) 10Umherirrender: Remove sql patches for old table 'validate' [core] - 10https://gerrit.wikimedia.org/r/94384 [17:13:39] (03PS2) 10Manybubbles: Term containing * match against unstemmed text [extensions/CirrusSearch] - 10https://gerrit.wikimedia.org/r/94373 [17:15:33] I am struggling with the bit of addNewAccountInternal that does this: if ( !$wgAuth->addUser( $u, $this->mPassword, $this->mEmail, $this->mRealName ) ) [17:16:17] In my extensions (OpenStackManager + LDAPAuth) there are a variety of failure conditions for that, but at the moment it can only return one useless error message. [17:16:38] Would it be reasonable to add a validate hook somewhere in there? [17:16:56] (I ask on the off-chance that there is already a hookable validation stage that I'm missing) [17:23:21] bah, this thing where all my exceptions just say foo( string, string ) is really annoying :P [17:33:12] (03PS2) 10Jforrester: Move files from modules/oojs-ui to the repository root [oojs/ui] - 10https://gerrit.wikimedia.org/r/94264 (owner: 10Catrope) [17:33:55] (03CR) 10Chad: "I hadn't really thought about it since my use-case involves copy+pasting this to a new location and changing 2 lines of code." [core] - 10https://gerrit.wikimedia.org/r/94258 (owner: 10Chad) [17:34:04] (03CR) 10Jforrester: [C: 032] "Per Timo." [oojs/ui] - 10https://gerrit.wikimedia.org/r/94264 (owner: 10Catrope) [17:34:08] (03Merged) 10jenkins-bot: Move files from modules/oojs-ui to the repository root [oojs/ui] - 10https://gerrit.wikimedia.org/r/94264 (owner: 10Catrope) [17:36:07] brion, can you advise me about addNewAccountInternal? I'm trying to spruce up error reporting for new wikitech accounts. [17:36:25] i can tell you that it's horrible and should die [17:36:41] Well… yes. [17:36:46] But, in the short-run… :) [17:37:14] :D [17:37:24] i can attempt to help decipher it but i make no guarantees [17:37:27] wikitech adds an extra field to the creation page, 'shell account name' [17:37:40] Right now the validation of that happens inside this call: if ( !$wgAuth->addUser( $u, $this->mPassword, $this->mEmail, $this->mRealName ) ) [17:38:05] But that means that if you provide a bogus shell name you get an unhelpful failure message. [17:38:35] I'm wondering if I should add a validation hook so that plugins can check extra args [17:38:43] …or if there's something like that already that I've overlooked. [17:38:53] that sounds like it might be good... but yeah i'm not too sure the current state of it [17:39:29] OK -- if it doesn't strike you as definitely crazy then I'll write a patch and see if it floats. [17:39:38] Are there other folks besides you I should ping for reviews? [17:40:07] brion: As your attorney, I advise you to take a hit out of the little brown bottle in my shaving kit. [17:40:54] don't stop here, this is bat country [17:41:33] * andrewbogott codes during ether binge [17:43:14] YuviPanda: How does one configure grrrit-wm? [17:44:52] (03Abandoned) 10Jforrester: Update VisualEditor to tip of wmf2 branch for cherry-pick [core] (wmf/1.23wmf2) - 10https://gerrit.wikimedia.org/r/92915 (owner: 10Catrope) [17:46:01] * James_F wonders if the ever-helpful marktraceur might know. [17:46:04] Denied. [17:46:15] Configure it to do what? [17:46:43] Reedy: Stalk another few repos in #mediawiki-visualeditor. [17:46:56] Reedy: There's a lack of config documentation that I can find, at least. [17:47:06] * marktraceur um [17:47:16] * marktraceur looks [17:47:56] labs/tools/grrrit.git is the repo [17:48:09] (03PS1) 10IAlex: Correct tooltip of "Next n results" on query special pages. [core] - 10https://gerrit.wikimedia.org/r/94386 [17:48:12] Oh, it's a gerrit push? Fun. [17:48:15] James_F: Which channels do you want? [17:48:18] Er [17:48:19] Repos [17:48:20] Sorry [17:48:27] * marktraceur drinks moar coffee [17:48:34] marktraceur: oojs/*, or failing that, oojs/core and oojs/ui [17:48:51] wildcards are hard [17:48:54] marktraceur: Also, oojs/core and oojs/ui now exist. Commit bringing them into MW core as libraries forthcoming. [17:49:01] marktraceur: Yay, etc. [17:49:01] Wooooooo [17:49:09] * marktraceur psyched like nobody's business [17:49:25] (03PS1) 10Manybubbles: Use filtered query rather than results filtering [extensions/CirrusSearch] - 10https://gerrit.wikimedia.org/r/94387 [17:50:42] (03CR) 10Bartosz Dziewoński: [C: 032] "Derp." [core] - 10https://gerrit.wikimedia.org/r/94386 (owner: 10IAlex) [17:51:13] Hmm. I might move the multimedia extensions into our channel, too. [17:51:15] Different commit. [17:52:49] (03PS1) 10MaxSem: Remove insane version check [extensions/Collection] - 10https://gerrit.wikimedia.org/r/94389 [17:53:10] marktraceur: ε> [17:53:37] (03Merged) 10jenkins-bot: Correct tooltip of "Next n results" on query special pages. [core] - 10https://gerrit.wikimedia.org/r/94386 (owner: 10IAlex) [17:53:47] Heh, yup [17:59:39] (03CR) 10Ori.livneh: [C: 032] "LGTM." [core] - 10https://gerrit.wikimedia.org/r/94186 (owner: 10BryanDavis) [18:00:34] (03PS6) 10Chad: Clean up CDB classes [core] - 10https://gerrit.wikimedia.org/r/94258 [18:00:35] (03PS7) 10Chad: Don't use wfIsWindows() in CDB writers [core] - 10https://gerrit.wikimedia.org/r/94268 [18:01:38] (03PS6) 10Chad: Update docs in CdbPHP.php [core] - 10https://gerrit.wikimedia.org/r/94267 (owner: 10Siebrand) [18:03:13] <^d> cdb patches need review :) [18:03:31] hmmm [18:03:33] All patches need review, ^d [18:03:34] (03Merged) 10jenkins-bot: Support CIDR ranges in $wgSquidServersNoPurge [core] - 10https://gerrit.wikimedia.org/r/94186 (owner: 10BryanDavis) [18:03:37] i think i'll add a new Linker method. [18:03:38] * marktraceur shakes head condescendingly [18:03:47] in addition to the 70-odd already there. [18:03:52] <^d> marktraceur: I reviewed lots of patches yesterday :) [18:04:19] turns out we have no convenient way to generate an error message that you must be logged in, with a link to special:userlogin with returnto. [18:06:35] hm, actually, i could just use Linker::specialLink with some parameters [18:09:29] (03PS1) 10EBernhardson: Pretend, for parsoid, that we are always on Main_Page [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94391 [18:09:35] (03PS1) 10BryanDavis: Support CIDR ranges in $wgSquidServersNoPurge [core] (wmf/1.23wmf3) - 10https://gerrit.wikimedia.org/r/94392 [18:14:15] (03CR) 10EBernhardson: [C: 032] Pretend, for parsoid, that we are always on Main_Page [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94391 (owner: 10EBernhardson) [18:14:25] (03Merged) 10jenkins-bot: Pretend, for parsoid, that we are always on Main_Page [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94391 (owner: 10EBernhardson) [18:14:46] (03CR) 10Swalling: [C: 031] "Thanks for the quick followup Chad." [core] - 10https://gerrit.wikimedia.org/r/94276 (owner: 10Chad) [18:16:18] (03PS1) 10IAlex: Correct tooltip of "Next n results" on query special pages. [core] (REL1_22) - 10https://gerrit.wikimedia.org/r/94393 [18:17:33] (03CR) 10Bartosz Dziewoński: "New screenshot plz? :)" [core] - 10https://gerrit.wikimedia.org/r/94276 (owner: 10Chad) [18:19:48] (03PS1) 10MaxSem: Honor $wgExtensionAssetsPath [extensions/Collection] - 10https://gerrit.wikimedia.org/r/94394 [18:21:25] <^d> MatmaRex: Yeah, sec. Editing now and will upload to mw.org [18:21:56] (03PS1) 10EBernhardson: Blank string conversion is implicitly a blank string [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94395 [18:22:12] (03CR) 10EBernhardson: [C: 032] Blank string conversion is implicitly a blank string [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94395 (owner: 10EBernhardson) [18:22:21] (03Merged) 10jenkins-bot: Blank string conversion is implicitly a blank string [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94395 (owner: 10EBernhardson) [18:30:33] (03PS1) 10BryanDavis: Support CIDR ranges in $wgSquidServersNoPurge [core] (wmf/1.23wmf2) - 10https://gerrit.wikimedia.org/r/94396 [18:30:46] (03PS1) 10Aaron Schulz: Made Lua code in JobQueueRedis more readable with unpack() [core] - 10https://gerrit.wikimedia.org/r/94397 [18:31:50] <^d> MatmaRex: Well, I tried to upload the file then the site crashed. Priorities :p [18:32:31] heh [18:33:02] (03CR) 10Aaron Schulz: "Works fine in eval (delayed and undelayed jobs, dups and non-dups) and passes queue tests" [core] - 10https://gerrit.wikimedia.org/r/94397 (owner: 10Aaron Schulz) [18:33:13] hm, site crash == bad [18:33:31] ^d: Stop uploading exes. You're breaking stuff [18:33:57] <^d> How do I share screenshots other than .exe form? [18:34:11] use .msi instead [18:34:23] or .scr, that's for screenshots [18:40:37] (03CR) 10Anomie: [C: 031] "Seems sensible to me. Haven't tested." [extensions/Collection] - 10https://gerrit.wikimedia.org/r/94389 (owner: 10MaxSem) [18:41:33] (03PS1) 10Anomie: Cleanup watchlist preference usage [core] - 10https://gerrit.wikimedia.org/r/94399 [18:46:15] (03CR) 10Anomie: [C: 031] "Looks good to me. Haven't tested." [extensions/Collection] - 10https://gerrit.wikimedia.org/r/94394 (owner: 10MaxSem) [18:50:25] (03CR) 10Chad: "New screenshot uploaded at the same URL." [core] - 10https://gerrit.wikimedia.org/r/94276 (owner: 10Chad) [18:50:30] huh. [18:50:31] wtf. [18:50:45] special:changeemail is accessible to anons if the POST the request. [18:50:48] if they* [18:51:00] o_0 [18:51:11] <^d> How would an anonymous user change their e-mail if they couldn't reach the page? [18:51:32] awww, but they can't change their email [18:51:35] because no password is valid [18:52:36] MatmaRex: Yeah, they wouldn't have anything to change... but yeah, it should bail out earlier [18:53:26] and changepassword, too [18:53:27] heh [18:53:38] i am cleaning up the logic for checking if the user is logged in on various pages [18:53:52] each does it differently [18:54:05] special:resettokens, too, and this time it's my fault :( [18:55:23] <^d> Patch incoming for Special:ChangeEmail [18:55:29] ^d: noooo [18:55:30] don't [18:55:34] i'll have to rebase my shit [18:55:35] :D [18:55:41] <^d> You'll live :) [18:55:54] ^d: i'll abandon yours. :P [18:56:06] <^d> I'll restore it! [18:56:25] hmmmmmm [18:56:51] (03PS1) 10Chad: Anons should not be able to use the change e-mail form at all [core] - 10https://gerrit.wikimedia.org/r/94400 [18:56:59] is there are circumstance where anon users could meaningfully access special:changepassword or changeemail? [18:57:10] because they both check if the user is logged in in a few places [18:57:14] <^d> Change email no way. [18:57:29] <^d> Change password...maybe? csteipp? [18:57:38] it can [18:57:44] an, no, just password does it [18:57:48] change password for anons [18:57:57] since name/password is given [18:58:02] The password reset, yes. The change password no. [18:58:10] (I can never remember which is called which...) [18:58:17] <^d> Nor can I. [18:58:24] <^d> It's confused me since forever. [18:58:57] Special:ChangePassword is accessed by anons [18:59:07] (password resets) [18:59:45] ugh, i am more confused now than i was when i asked. [18:59:51] you fix it, guys. :P [19:01:02] And.. Special:PasswordReset is accessed by anons too. So yeah, both need anon access. [19:01:31] <^d> Change e-mail doesn't! Go go gadget mergetool! [19:03:19] <^d> MatmaRex: Whatcha think of the new screenshot? [19:03:41] ^d: i am currently ears-deep in exception handling code, brb [19:03:50] <^d> k :) [19:18:46] (03PS1) 10leucosticte: Bug: 42594 Set $wgNoFollowLinks = false by default [core] - 10https://gerrit.wikimedia.org/r/94401 [19:19:40] (03PS2) 10Dereckson: Bug: 42594 Set $wgNoFollowLinks = false by default [core] - 10https://gerrit.wikimedia.org/r/94401 (owner: 10leucosticte) [19:21:20] (03CR) 10jenkins-bot: [V: 04-1] Set $wgNoFollowLinks to false by default [core] - 10https://gerrit.wikimedia.org/r/94401 (owner: 10leucosticte) [19:23:36] (03PS1) 10Aaron Schulz: Clarify thumbnail log failure entries when the local file copy failed [core] - 10https://gerrit.wikimedia.org/r/94402 [19:24:25] csteipp: so, wanna look at Chad's https://gerrit.wikimedia.org/r/#/c/94400/ very quickly? :) [19:25:10] <^d> He watched me write the patch in person :p [19:25:26] (03CR) 10CSteipp: [C: 031] "I think this is the right direction" [core] - 10https://gerrit.wikimedia.org/r/94400 (owner: 10Chad) [19:26:12] ^d: well, he didn't merge it yet, and i could base my patch off yours [19:26:46] MatmaRex: I'll merge it when I have time to review it... I'm just in the middle of another patch atm [19:27:03] I'd say it's safe to use it as a base [19:27:17] <^d> self-merge time! gogogo! [19:30:28] (03CR) 10Dereckson: [C: 04-1] "There are some tests which depend of this setting." [core] - 10https://gerrit.wikimedia.org/r/94401 (owner: 10leucosticte) [19:31:10] (03PS1) 10Legoktm: Use timestamp of page move instead of creation date [extensions/PageTriage] - 10https://gerrit.wikimedia.org/r/94403 [19:35:02] (03CR) 10Dereckson: "Well... according the name of the test, you can just remove the nofollow. The purpose isn't to test link output but "description" of exter" [core] - 10https://gerrit.wikimedia.org/r/94401 (owner: 10leucosticte) [19:36:38] (03CR) 10EBernhardson: "(3 comments)" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94103 (owner: 10EBernhardson) [19:37:29] (03PS1) 10CSteipp: Reuse more understandable message [extensions/OAuth] - 10https://gerrit.wikimedia.org/r/94405 [19:37:57] (03PS1) 10Bartosz Dziewoński: Consistently handle anonymous users on logged-in-only special pages [core] - 10https://gerrit.wikimedia.org/r/94406 [19:38:03] ^d: ^ [19:38:35] AaronSchulz: easy OAuth patch https://gerrit.wikimedia.org/r/#/c/94405 [19:39:39] (03CR) 10Dereckson: [C: 04-1] "(2 comments)" [core] - 10https://gerrit.wikimedia.org/r/94402 (owner: 10Aaron Schulz) [19:41:03] <^d> MatmaRex: Looks cool :) [19:41:28] so many files changed. [19:41:40] and the worst thing, the next result is more code! [19:41:42] (03PS2) 10Dereckson: Remove sql patches for old table 'validate' [core] - 10https://gerrit.wikimedia.org/r/94384 (owner: 10Umherirrender) [19:41:46] the net result* [19:42:27] (03PS3) 10Dereckson: Remove SQL patches for old table 'validate' [core] - 10https://gerrit.wikimedia.org/r/94384 (owner: 10Umherirrender) [19:45:03] (03PS2) 10Dereckson: Clarify thumbnail log failure entries when the local file copy failed [core] - 10https://gerrit.wikimedia.org/r/94402 (owner: 10Aaron Schulz) [19:45:56] (03CR) 10Dereckson: "PS2: improving ponctuation (cf. PS1 inline comments)" [core] - 10https://gerrit.wikimedia.org/r/94402 (owner: 10Aaron Schulz) [19:47:51] (03CR) 10Dereckson: "(This is punctuation in English, ponctuation is in French, and I see the term doesn't always include the sentence start uppercase rule.)" [core] - 10https://gerrit.wikimedia.org/r/94402 (owner: 10Aaron Schulz) [20:00:54] (03PS1) 10Bartosz Dziewoński: Remove 'watchnologin' and 'prefsnologin' messages [core] - 10https://gerrit.wikimedia.org/r/94411 [20:03:35] (03CR) 10Aaron Schulz: [C: 032] Reuse more understandable message [extensions/OAuth] - 10https://gerrit.wikimedia.org/r/94405 (owner: 10CSteipp) [20:03:40] (03Merged) 10jenkins-bot: Reuse more understandable message [extensions/OAuth] - 10https://gerrit.wikimedia.org/r/94405 (owner: 10CSteipp) [20:06:48] (03CR) 10Mwalker: [C: 032] "works on my local" [extensions/Collection] - 10https://gerrit.wikimedia.org/r/94394 (owner: 10MaxSem) [20:07:38] (03Merged) 10jenkins-bot: Honor $wgExtensionAssetsPath [extensions/Collection] - 10https://gerrit.wikimedia.org/r/94394 (owner: 10MaxSem) [20:11:02] (03CR) 10Parent5446: [C: 032] Anons should not be able to use the change e-mail form at all [core] - 10https://gerrit.wikimedia.org/r/94400 (owner: 10Chad) [20:12:56] (03PS2) 10Aaron Schulz: Cleanup watchlist preference usage [core] - 10https://gerrit.wikimedia.org/r/94399 (owner: 10Anomie) [20:14:00] (03PS1) 10Umherirrender: migrate redirectToFragment() from wikibits.js to own module [core] - 10https://gerrit.wikimedia.org/r/94412 [20:14:27] (03Merged) 10jenkins-bot: Anons should not be able to use the change e-mail form at all [core] - 10https://gerrit.wikimedia.org/r/94400 (owner: 10Chad) [20:15:52] (03CR) 10Parent5446: "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/94406 (owner: 10Bartosz Dziewoński) [20:17:44] (03PS2) 10Umherirrender: Migrate redirectToFragment() from wikibits.js to own module [core] - 10https://gerrit.wikimedia.org/r/94412 [20:18:03] (03PS1) 10Bartosz Dziewoński: Canonicalize text and update docs of messages using 'loginreqlink' [core] - 10https://gerrit.wikimedia.org/r/94414 [20:18:17] (03CR) 10Umherirrender: "Patch Set 2: Removed debug code" [core] - 10https://gerrit.wikimedia.org/r/94412 (owner: 10Umherirrender) [20:20:05] so many patches. [20:23:09] (03PS1) 10CSteipp: Add hook to override message [extensions/OAuth] - 10https://gerrit.wikimedia.org/r/94416 [20:30:16] (03CR) 10Jakobj: [C: 031] "Works and follows the guidelines." [core] - 10https://gerrit.wikimedia.org/r/94150 (owner: 10Bartosz Dziewoński) [20:31:13] (03CR) 10Bartosz Dziewoński: "Cleaning up docs and messages in followup patch https://gerrit.wikimedia.org/r/#/c/94414/ ." [core] - 10https://gerrit.wikimedia.org/r/94406 (owner: 10Bartosz Dziewoński) [20:33:14] (03CR) 10Bartosz Dziewoński: "@Tyler:" [core] - 10https://gerrit.wikimedia.org/r/94406 (owner: 10Bartosz Dziewoński) [20:34:39] (03PS3) 10Umherirrender: Migrate redirectToFragment() from wikibits.js to own module [core] - 10https://gerrit.wikimedia.org/r/94412 [20:35:42] (03CR) 10Umherirrender: "Patch Set 3: Keep the old function for one version, because inline script calls to this function can stay in cache." [core] - 10https://gerrit.wikimedia.org/r/94412 (owner: 10Umherirrender) [20:36:32] (03CR) 10Bartosz Dziewoński: "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/94406 (owner: 10Bartosz Dziewoński) [20:36:57] (03PS2) 10Bartosz Dziewoński: Consistently handle anonymous users on logged-in-only special pages [core] - 10https://gerrit.wikimedia.org/r/94406 [20:40:04] (03CR) 10Bartosz Dziewoński: "Why change release notes here and not on master? Not sure if this matters." [core] (REL1_22) - 10https://gerrit.wikimedia.org/r/94393 (owner: 10IAlex) [20:46:25] hey AaronSchulz, i've been looking into the chunked upload api method. how can i get the api to temporarily accept .xml files without having to change the LocalSettings.php [20:49:22] (03CR) 10IAlex: "Because it's here it should be listed if the change is backported." [core] (REL1_22) - 10https://gerrit.wikimedia.org/r/94393 (owner: 10IAlex) [20:52:13] (03PS3) 10BryanDavis: Clarify thumbnail log failure entries when the local file copy failed [core] - 10https://gerrit.wikimedia.org/r/94402 (owner: 10Aaron Schulz) [20:53:54] (03CR) 10BryanDavis: [C: 032] Clarify thumbnail log failure entries when the local file copy failed [core] - 10https://gerrit.wikimedia.org/r/94402 (owner: 10Aaron Schulz) [20:56:41] (03CR) 10Bartosz Dziewoński: "Yes, but we currently also have RELEASE-NOTES-1.22 file on master… Paging release manager to sort this out. :P" [core] (REL1_22) - 10https://gerrit.wikimedia.org/r/94393 (owner: 10IAlex) [20:56:48] (03CR) 10Nikerabbit: [C: 031] Canonicalize text and update docs of messages using 'loginreqlink' [core] - 10https://gerrit.wikimedia.org/r/94414 (owner: 10Bartosz Dziewoński) [20:56:51] hexmode: ^^ [20:57:20] (03Merged) 10jenkins-bot: Clarify thumbnail log failure entries when the local file copy failed [core] - 10https://gerrit.wikimedia.org/r/94402 (owner: 10Aaron Schulz) [20:59:18] (03PS1) 10Amire80: Add beta preferences icons for Hebrew and Arabic [extensions/VectorBeta] - 10https://gerrit.wikimedia.org/r/94423 [21:02:22] hm, ^d [21:02:40] (03PS1) 10CSteipp: Update i18n from TWN comments [extensions/OAuth] - 10https://gerrit.wikimedia.org/r/94425 [21:02:44] ^d: re the sarch thing - i am pretty sure that the padding should go to mediawiki.ui itself [21:02:57] not just the search page styles [21:03:16] <^d> Reasonable suggestion :) [21:03:52] but the mediawiki.ui module is currently all messed up [21:04:11] and superm401 is supposedly rewriting it or somehting [21:05:39] <^d> Well, hopefully a rewrite would involve finding and updating existing usages :p [21:06:05] ^d: lol [21:06:07] https://gerrit.wikimedia.org/r/#/c/93584/ [21:06:16] meh, i'll just submit a followup [21:06:21] or no [21:06:23] i'll file a bug [21:07:41] hmmmmmm [21:07:52] .mw-ui-input has padding: 0.35em 0px 0.35em 0.5em; [21:08:07] so zeroed padding-right [21:08:09] that's stupid. [21:09:53] <^d> MatmaRex: Yeah, I saw that during implementation. [21:09:59] <^d> Which is *probably ok* for normal text fields. [21:10:06] <^d> But for input type=search, notsomuch :) [21:10:08] it's dumb for normal text fields too [21:10:16] 0.5em padding left, 0 right? [21:10:40] i'm filing a bug, to be fixed after that mess is rewritten [21:10:43] which should be very soon [21:12:43] <^d> :) [21:15:02] ^d: https://bugzilla.wikimedia.org/show_bug.cgi?id=56805 [21:16:38] ^d: also, one last thing! [21:17:07] ^d: mediawiki.special.search should be in addModuleStyles too, not just addModules [21:17:46] <^d> Not my fault :D [21:18:18] (03CR) 10Bartosz Dziewoński: [C: 04-1] "mediawiki.special.search should be in addModuleStyles too, not just addModules. Apart from that – sexy, LGTM. Filed bug 56805 about movin" [core] - 10https://gerrit.wikimedia.org/r/94276 (owner: 10Chad) [21:18:29] ^d: yours, you added basic styles to it [21:18:44] <^d> The file was already there :p [21:18:56] yeah, but it only included some shitty hack for IE6 [21:19:03] now it has proper styles [21:19:09] which shouldn't even be there [21:19:10] meh. [21:19:29] ^d: you can help matt and juliusz with the rewrite instead. :P [21:19:37] (03CR) 10Spage: [C: 04-1] "(1 comment)" [extensions/GuidedTour] - 10https://gerrit.wikimedia.org/r/94370 (owner: 10Phuedx) [21:19:38] <^d> Nooooooooooo [21:19:41] <^d> I'll just fix my patch :) [21:22:38] actually, wtf is that hack [21:24:44] i'm pretty sure it's not needed now [21:25:29] (03CR) 10Mwalker: [C: 032] Remove insane version check [extensions/Collection] - 10https://gerrit.wikimedia.org/r/94389 (owner: 10MaxSem) [21:25:35] (03Merged) 10jenkins-bot: Remove insane version check [extensions/Collection] - 10https://gerrit.wikimedia.org/r/94389 (owner: 10MaxSem) [21:25:37] <^d> MatmaRex: I definitely needed it in my local testing. [21:25:46] <^d> But maybe something changed in master? [21:26:34] i mean, the hack that was there before [21:43:24] marktraceur: Do you know how those grrrit changes are meant to be deployed? [21:44:15] Also, zuul seems broken in that it has +2ed but not merged those changes [21:46:14] hey AaronSchulz any thoughts on my earlier message? [21:47:42] MatmaRex, as far as I know, no one is planning to rewrite mediawiki.ui, at least not right now. [21:47:58] I was originally going to convert it from SASS to LESS, but jgonera beat me too it, so I'm reviewing that. [21:48:23] And there will be some UX changes that need to get implemented in core (e.g. new button semantics and UX). [21:48:26] * AaronSchulz was doing other stuff + a meeting [21:48:32] superm401: yeah, that's what i meant by rewrite there [21:48:39] Nothing to say you can't fix the padding, though. [21:49:01] i'd rather avoid rebase conflicts and/or the change getting lost in the process [21:50:18] <^d> We could land said fix in master before the rewrite is done. [21:50:31] <^d> (Then rewrite would just have to incorporate the change) [21:53:05] " i'd rather avoid rebase conflicts and/or the change getting lost in the process" [21:53:42] ^d: unless you're volunteering to do the change and the incorporation and coordinate with juliusz so he doesn't overwrite it, in which case go ahead [21:53:44] :) [21:56:18] <^d> Me? Volunteer? [21:56:20] <^d> You must be new here. [21:57:12] (03PS1) 10Aaron Schulz: Error handling tweaks in JobQueueFederated [core] - 10https://gerrit.wikimedia.org/r/94436 [21:57:33] (03CR) 10Aaron Schulz: [C: 032] Add hook to override message [extensions/OAuth] - 10https://gerrit.wikimedia.org/r/94416 (owner: 10CSteipp) [21:58:06] (03Merged) 10jenkins-bot: Add hook to override message [extensions/OAuth] - 10https://gerrit.wikimedia.org/r/94416 (owner: 10CSteipp) [21:59:02] dan-nl: might also want to ask csteipp [22:00:59] hey csteipp, i've been looking into the chunked upload api and have run into an issue. is there a way to get the api, via an ajax post, to temporarily accept .xml files without having to alter wgFileExtensions[]='.xml' in LocalSettings.php? [22:03:19] dan-nl: No, it shouldn't allow that [22:04:34] The probably we fixed a while back was that individual chunks could contain html, and were served from the temp directory, so big xss whole. [22:06:49] csteipp: so we should't be chunk uploading xml - correct? [22:07:56] csteipp: what about storing xml in the upload stash? [22:08:07] (03CR) 10Cscott: "The bar for merging should be much higher for changes which affect so much of the parserTests file. This change breaks Parsoid parserTest" [core] - 10https://gerrit.wikimedia.org/r/90696 (owner: 10Gerrit Patch Uploader) [22:19:57] (03PS1) 10Bsitu: Undefined $self variable in closure [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94442 [22:20:13] (03PS1) 10Cscott: Revert "Have list items occupy their own line" [core] - 10https://gerrit.wikimedia.org/r/94443 [22:21:46] csteipp: just to confirm, we should not be chunk uploading .xml files via the mw api - correct? [22:22:14] csteipp: is it okay to upload the xml file to the upload stash with the intent that they fall off after expiration and are never stored as a page? [22:23:19] (03CR) 10Aaron Schulz: [C: 032] Cleanup watchlist preference usage [core] - 10https://gerrit.wikimedia.org/r/94399 (owner: 10Anomie) [22:24:20] dan-nl: Correct, we shouldn't upload .xml files. We don't want xml files accessible to end users. [22:24:40] The api should block it, if it detects a mime type that sniffs as html [22:24:44] csteipp: and using the upload stash temporarily so that we can process the metadata? [22:25:08] I'd have to look into it. Are you using the upload api? [22:25:56] no, i'm using UploadStash [22:26:43] i'm temporarily allowing .xml and using the UploadStash->stashFile to store the xml file and then retrieve later and process the file [22:27:04] (03Merged) 10jenkins-bot: Cleanup watchlist preference usage [core] - 10https://gerrit.wikimedia.org/r/94399 (owner: 10Anomie) [22:27:07] How are you temporarily allowing xml? [22:27:30] the stashed file is never saved as a page and is assumed to be deleted when the expiry timestamp reaches its limit [22:28:12] * csteipp still not quite understanding [22:28:13] (03CR) 10MarkAHershberger: [C: 032] "The RELEASE-NOTES-1.22 on master should be updated, too, so the released tarball matches." [core] (REL1_22) - 10https://gerrit.wikimedia.org/r/94393 (owner: 10IAlex) [22:28:21] James_F: tell me and i'll do it [22:28:22] So you're using Special:UploadStash? [22:28:58] no i'm using UploadStash directly … looking for the code lines ... [22:30:56] YuviPanda: Already done - thanky, though. [22:31:18] YuviPanda: Oh, wait, no. [22:31:28] James_F: yeah, patches are there, but not merged [22:31:30] let me do that now [22:31:31] (03CR) 10Cscott: "Let's not +2 this until discussion happens on bug 39617. Ideally I'd like to have a CSS-based fix for 39617 in hand before merging this." [core] - 10https://gerrit.wikimedia.org/r/94443 (owner: 10Cscott) [22:31:52] YuviPanda: https://gerrit.wikimedia.org/r/#/c/94388/ and https://gerrit.wikimedia.org/r/#/c/94390/ aren't merging, though. Yay Jenkins. [22:32:09] csteipp: https://github.com/dan-nl/GWToolset/blob/master/includes/Handlers/UploadHandler.php#L232 [22:32:18] James_F: merged [22:32:23] James_F: now I just need to deploy it [22:32:24] moment [22:32:30] Thanks Yu [22:32:31] Thanks YuviPanda [22:32:50] csteipp: that's how i augment the wgFileExtensions … there's another method in there that restores the settings [22:33:14] csteipp: https://github.com/dan-nl/GWToolset/blob/master/includes/Handlers/UploadHandler.php#L567 [22:33:15] YuviPanda: Thanks! :-) [22:33:23] dan-nl: Yeah... that's not going to work [22:33:27] James_F: deployed. can you test in a second when it comes back? [22:33:30] csteipp: that's how i store the file in the stash [22:33:39] csteipp: oh, okay, why not? [22:33:47] James_F: 'tis back. test? [22:33:58] (03CR) 10Cscott: "oh, and it likely breaks parser tests in the ParserFunctions extension, Scribunto, etc, etc." [core] - 10https://gerrit.wikimedia.org/r/90696 (owner: 10Gerrit Patch Uploader) [22:34:48] James_F: hehe, I'm deploying and getting on a plane [22:35:28] marktraceur: RoanKattouw James_F there's also https://wikitech.wikimedia.org/wiki/Grrrit-wm on how to deploy, config, might be helpful in the future [22:36:09] YuviPanda: Yay, it works. :-) [22:36:12] YuviPanda: Thanks! [22:36:15] James_F: :) [22:36:15] Awesome, thanks [22:36:29] (03PS2) 10Kaldari: Adding a quotation marks message to core. [core] - 10https://gerrit.wikimedia.org/r/93880 [22:37:00] marktraceur: RoanKattouw if you have a toollabs account, I can add you guys to the project [22:37:05] the more people who are on it the better... [22:37:09] Rigth [22:37:14] I don't have one I don't think [22:38:21] YuviPanda: MarkTraceur as alwasy [22:38:23] always [22:38:35] marktraceur: doesn't exist on tools [22:38:39] hmm or maybe I do [22:39:00] ...wat [22:39:11] marktraceur: toollabs, not labs [22:39:15] Oh [22:39:19] No, eff toollabs [22:39:38] RoanKattouw: you don't [22:40:45] meh [22:40:53] marktraceur: okay, you've a toollabs account now [22:40:57] ... [22:40:58] RoanKattouw: can you tell me your labs username? [22:40:59] I thought I did [22:41:02] Catrope? [22:41:11] RoanKattouw: ah, nevermind. Was looking for Roan [22:41:12] I remember complaining to Coren about how 'become' worked a while ago [22:41:26] RoanKattouw: added you and marktraceur [22:41:33] Awesome [22:41:39] marktraceur: you should login to tools-login.wmflabs.org at some point so your homedir gets created [22:41:40] RoanKattouw: Sure, but you didn't complain justly. :-) [22:42:04] (03CR) 10Dereckson: "I concur with Bartosz for the optional." [core] - 10https://gerrit.wikimedia.org/r/93880 (owner: 10Kaldari) [22:42:15] That's true :) [22:42:18] YuviPanda: See also "eff toollabs" [22:42:20] My complains were misguided [22:42:27] You should be running that bot on a different labs instance [22:42:38] wat [22:42:39] no [22:42:45] and maintain the underlying infrastructure myself? [22:42:49] nothankyou [22:42:55] that's what we have Coren for :P [22:42:58] YuviPanda: You lazy bastard! [22:43:20] marktraceur: don't you mean 'you do not want to sysadmin when I do not need to!' bastard? [22:43:33] csteipp: what part of the code i showed you won't work? [22:43:52] dan-nl: Sorry, I was just talkign about this with aaron [22:43:58] np [22:44:09] (03CR) 10Kaldari: "I have no opinion on the optional-ness. Siebrand, what do you think? I've marked it as optional for now." [core] - 10https://gerrit.wikimedia.org/r/93880 (owner: 10Kaldari) [22:44:25] i'll wait [22:44:30] The issue is that the xml is publicly accessible. So attacker uploads their xml, sends the stash url to a victim, and owns the account [22:44:40] dan-nl: I talked to Chris about this IRL a bit...so for uploading the XML file I'd imagine it could be down with an upload by URL job, not using the stash but just some private FileBackend container. The client could poll for job completion before moving on (in JS) to the next step. The jobs of course could still access the private container. [22:44:59] YuviPanda: ...yes [22:45:01] as long as users never need to see the xml this works [22:45:22] the user won't need to see the xml [22:45:26] RoanKattouw: i used VE to make https://wikitech.wikimedia.org/w/index.php?title=Grrrit-wm&diff=88496&oldid=87837, it warned me of Corruption, even though there was none [22:45:27] weird [22:45:29] (it could still allow access with auth+streamFile() if needed) [22:45:45] but my understanding is that it is just thrown away after job usage [22:46:03] a script could also traverse the container and delete old files just like cleanupUploadStash [22:46:03] YuviPanda: Did you use Firefox with HTTPSEverywhere enabled? [22:46:15] YuviPanda: Also, you are encouraged to spell my name correctly [22:46:16] RoanKattouw: Firefox *without* HTTPSEverywhere [22:46:18] AaronSchulz: so you're suggesting turing it from push (upload by user) to pull (download by server) [22:46:20] well the last loop of cleanupUploadStash [22:46:21] RoanKattouw: gah! [22:46:30] bd808: that will handle larger files better, yes [22:47:00] pull from url … so the user needs to post their xml somewhere online … ? [22:47:06] and then refer to it? [22:47:08] RoanKattouw: fixed your name, again it complains of potential corruption [22:47:10] one would also have the option of making the job due Ranged GETs to chunk download if needed...but I bet that won't even be needed for now [22:47:12] AaronSchulz, csteipp: And also address the security concern of reflected XSS [22:47:21] RoanKattouw: this is Aurora, but wikitech is https only [22:47:23] Getting corruption warning in Chrome too [22:47:31] Hmm I wonder how up-to-date VE is on wikitech [22:47:38] it's out of band, IIRC [22:47:42] dan-nl: right [22:47:48] oh boy [22:48:01] "Version ede7dad" 19:57, 18 September 2013 [22:48:08] There's your answer [22:48:13] That's the one that has the pre corruption bug IIRC [22:48:33] * RoanKattouw updates VE on wikitex [22:48:35] dan-nl: I doubt museums would have a problem with this...random users could probably use labs or something [22:48:55] (03CR) 10EBernhardson: "I think what she was trying for is where the unfocused size:" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94165 (owner: 10Matthias Mullie) [22:49:31] RoanKattouw: hmm, so needs update, I guess [22:49:39] so i'm not 100% clear on the approach … the xml file would reside on the net … the extension would ask for a url to it … the extension would pull it from that url and store it in the FSBackend? [22:49:43] dan-nl: of course if the file is like <=20mb (not an exact number), you don't need jobs for these...though I don't know how often that would happen [22:50:08] well, some FileBackend...it would be SwiftFileBackend for us, but that should all be configurable [22:50:18] Argh [22:50:20] Well that backfired [22:50:26] because it depends on mw.Title [22:50:26] (see Math/Captcha/timeline extensions for examples) [22:50:32] isn't that also a public available temp directory just like the upload stash? [22:50:45] I'll have to schedule an update [22:50:59] it depends on the container, the "deleted" file container is not public nor are all containers for our private wikis (like officewiki) [22:51:12] from what i saw earlier … FSBackEnd creates a container in the images directory [22:51:21] greg-g: Could I have a deployment window to update wikitechwiki from its current version (1.22wmf18!) to something more recent like 1.23wfm3? [22:51:30] wikitech is out of band and we tend to forget about upgrading it [22:51:40] so, why not upload the xml file via a form and store it in the same place? [22:51:42] (03CR) 10EBernhardson: [C: 032] "all hail php5.3" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94442 (owner: 10Bsitu) [22:51:48] for private containers, you need to make sure the docs mention that the user should secure the path they configure it to use and also use the right params to FileBackend::prepare() [22:51:51] (03Merged) 10jenkins-bot: Undefined $self variable in closure [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94442 (owner: 10Bsitu) [22:52:02] RoanKattouw: yessir, who should be on point for those in the future, I assume Ryan was doing it before [22:52:24] RoanKattouw: fyi, i also added legoktm to grrrit-wm [22:52:51] <^d> RoanKattouw: Loop me in for that. [22:52:54] AaronSchulz: what's the issue with using the form to do the same? why only via a url? [22:52:55] greg-g: He was but irregularly [22:52:56] <^d> Might need to reindex Elastic. [22:53:00] ^d: OK [22:53:14] dan-nl: when you create the container using prepare() with the right flags it will try to make the container private (which might not work for non-apache or apache with htaccess files disabled) but it would work fine for us [22:53:17] greg-g: I don't know who should do that. Not me, I'm just doing it this one time because I don't want VE to be broken [22:53:25] RoanKattouw: :) fair [22:53:31] (this is why the extension docs should mention making sure the directory is private) [22:53:40] k [22:53:55] but what about using the form to upload the xml file to the same place? [22:54:15] we could offer both uploadbyurl and via the form [22:55:09] sure, as long as users know that by-URL is good for large files [22:55:25] those using 100mb files will need the URL option most likely [22:55:35] yes we'll mention in the form, but allow them to decide what they want to do [22:55:57] eventually we will want to allow them to hit an api to their data as well [22:56:19] access the xml data? [22:56:21] so the main thing is where we store the temporary xml … [22:56:50] so i need to look at FSBackEnd and how it works and how to prepare the container? [22:56:56] any docs on that? [22:57:11] I'd look at FileBackend.php which is mostly prologue comments [22:57:21] there is also a README file in that directory [22:57:27] k [22:58:06] a future concept is to allow a user to enter credentials to an external api that might send json or xml of their metadata and use that to process the batch upload [22:58:13] but that's later down the line [22:59:44] AaronSchulz, thanks for your help … unfortunately it means i need to re-design … again … [22:59:56] (03CR) 10GWicke: "It also takes away formatting options from the user the same way tidy does. See https://bugzilla.wikimedia.org/show_bug.cgi?id=39617#c17 f" [core] - 10https://gerrit.wikimedia.org/r/90696 (owner: 10Gerrit Patch Uploader) [23:01:14] i really wish this would have come up earlier when we first introduced the idea of the upload stash ... [23:01:30] (03PS2) 10GWicke: Revert "Have list items occupy their own line" [core] - 10https://gerrit.wikimedia.org/r/94443 (owner: 10Cscott) [23:01:45] (03CR) 10GWicke: [C: 031] Revert "Have list items occupy their own line" [core] - 10https://gerrit.wikimedia.org/r/94443 (owner: 10Cscott) [23:05:55] dan-nl: I think stash could work with a total xml whitelist...but it actually less code to just to the thing I mentioned above [23:06:24] you don't really need most of the stuff the stash could give anyway [23:06:51] plus the private container is more secure by design [23:06:58] i'm not familiar with FileBackend yet, so i'll have to read through it and sort out what to do ... [23:07:01] *do the thing [23:07:24] yes that part i understand [23:08:45] AaronSchulz: bd808 just asked me about the chunked upload … as far as i understand it, that option is out now … correct? [23:09:33] the extension would temporarily allow xml for the upload into the private FileBackend container [23:10:05] (03CR) 10EBernhardson: [C: 032] "Merging this, we can deal with the wikitext edit box size in a small additional patch" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94165 (owner: 10Matthias Mullie) [23:10:06] or download from a url into the private container [23:10:18] we can't do the chunked upload anyway [23:10:21] (03Merged) 10jenkins-bot: Design changes [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94165 (owner: 10Matthias Mullie) [23:10:30] because the api won't allow the xml [23:10:39] correct? [23:11:29] dan-nl: I think upload-by-url removes most need for chunking [23:11:36] If I'm following (and I may not be) Dan would just need to change from storing via UploadStash in the local repo to storing in a custom private container. [23:11:38] (for the xml) [23:11:52] that's what i understand [23:12:16] so anyway, the Math extension is a good example of FileBackend (Math.php use of $wgMathFileBackend and the MathTexvc.php code) [23:12:17] whether it's from an external url or from the form [23:12:28] right [23:12:43] thanks that will help … i'll take a look at it on monday [23:13:06] it's a bit late over here [23:14:00] to start in on it … thanks for your help AaronSchulz, csteipp, bd808 [23:15:24] dan-nl: No problem. Sorry I miss understood that during the review. [23:16:44] (03CR) 10Bartosz Dziewoński: [C: 04-2] "-2 per C. Scott. I find it had to believe that the supposed CSS solution was not tested in the past, let's wait for a reply." [core] - 10https://gerrit.wikimedia.org/r/94443 (owner: 10Cscott) [23:16:48] dan-nl: I can abandon https://gerrit.wikimedia.org/r/#/c/87607/ then [23:17:13] AaronSchulz: yes [23:17:45] (03Abandoned) 10Aaron Schulz: Added an expiry field to upload stash table [core] - 10https://gerrit.wikimedia.org/r/87607 (owner: 10Aaron Schulz) [23:17:53] the best code is no code [23:18:40] csteipp: i understand … what would have been good was to have some type of review process from the beginning of the project [23:19:32] i really needed some dedicated time from developers at the foundation at the start [23:21:29] csteipp: hopefully i can sort out the FileBackend solution early next week. that solution, using a private container works for you? [23:23:04] dan-nl: Yeah, what I'm hearing from Aaron, it sounds fine [23:23:30] good, then i'll focus on that on monday [23:24:37] thanks [23:30:25] (spage for Rumanah) YuviPanda or ori-l, if you accidentally check out an extension in /vagrant/mediawiki, is there a way to make vagrant "re-provision" mediawiki? [23:31:46] ryasmeen_: did you enable the extension? [23:32:45] yes (visualeditor), I think problem is doing a git fetch of a gerrit change to VE in the /vagrant/mediawiki directory [23:33:04] Okay, so why is the

from == headlines == without a class? [23:33:10] you can run 'vagrant disable-role visualeditor' [23:33:10] You'd think it would use .mw-headline. [23:33:13] #butitdoesnt [23:33:13] and then 'vagrant provision' [23:33:26] it won't wipe the VisualEditor directory from extensions/, but it'll stop it from loading [23:33:28] (03CR) 10Bsitu: "(2 comments)" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/91135 (owner: 10EBernhardson) [23:33:29] .mw-headline is on the inner span. [23:33:34] The

has no class. [23:34:39] ori-l: I set up Bitlbee today at work. [23:34:48] what's Bitlbee? [23:35:06] It lets you use an IRC client with IM networks. [23:35:07] Basically. [23:36:00] It's a bit trippy, but it works fairly well. [23:36:10] http://bitlbee.org/main.php/news.r.html [23:41:59] MatmaRex: Re. bug 54617, the "hodge-podge" is what we currently have, as exemplified by Tidy. [23:42:52] James_F: yes, i know, which is why i am trying to kill it. [23:43:20] MatmaRex: But you said that you don't think there's a hodge-podge going to be created. I'm arguing we already have it, and we're trying to get rid of it. [23:43:31] yes, i am trying. [23:43:53] comments like that one tend to make it harder, though. [23:44:13] greg-g: OK for an LD on Tuesday to enable VE on officewiki? [23:44:59] MatmaRex: The argument "we must keep supporting shared web hosts" kills lots of much better wiki advances because it so limits our technologies. [23:45:15] greg-g: ...which would have to be post-deploy train so that wmf3 will be on officewiki [23:45:21] RoanKattouw: LD. [23:45:27] RoanKattouw: The clue is in the request. :-P [23:45:34] :) [23:47:42] James_F: RoanKattouw weeee, yeppers [23:47:46] MatmaRex, isn't tidy currently installed via apt-get since i believe it is not bundled as part of mediawiki? [23:48:19] subbu: php has a tidy module, not sure how widely it's installed [23:48:25] i would prefer a bundled PHP library [23:48:51] James_F: let's rewrite mediawiki in ruby! [23:49:15] right .. what i was basically getting at was .. if tidy is installed via apt-get, then node.js can also be installed via apt-get. but anyway, this is part of the larger packaging discussions. [23:49:38] i imgaine you guys will all be having at the arch. summit :) [23:50:36] (03CR) 10Kaldari: [C: 032] DifferenceEngine minor fixes [core] - 10https://gerrit.wikimedia.org/r/92598 (owner: 10MaxSem) [23:53:35] (03Merged) 10jenkins-bot: DifferenceEngine minor fixes [core] - 10https://gerrit.wikimedia.org/r/92598 (owner: 10MaxSem) [23:53:38] MatmaRex: Let's not. :-) [23:53:55] greg-g: Ta. [23:54:48] MatmaRex: Re-Write in LOLCODE! [23:55:46] (03CR) 10CSteipp: "(2 comments)" [extensions/OAuth] - 10https://gerrit.wikimedia.org/r/93121 (owner: 10CSteipp) [23:58:32] greg-g: Wikitechwiki is missing the Thanks links, so <3 instead. [23:59:22] greg-g: So about that deploy window for wikitech upgrade, could you allocate me a window for that next week? [23:59:58] (03PS3) 10Bsitu: auto-expand text-area as content grows [extensions/Flow] - 10https://gerrit.wikimedia.org/r/94285 (owner: 10Matthias Mullie)