[11:40:45] legoktm said i should go here to ask about AFT5 on enwiki [11:41:15] https://bugzilla.wikimedia.org/show_bug.cgi?id=43099 [11:42:01] its broken and probably needs a rollback [11:42:21] yeah. and i'm a bit worried about the ramifications of there being a portion of the project where people can comment, but no one can moderate what they say [11:43:29] unless the feedback oversight function works through some other channel, the I assume oversight doesn't work either for feedback, at the moment? [14:17:41] when i make a comment responding to someone elses comment in gerrit, its showing up as (draft) [14:17:46] how do i fix that? [14:19:01] legoktm: click "Review" below that patchset [14:19:06] and submit a review, it may be even empty [14:19:47] for inline comments right? [14:19:52] yup [14:19:55] ok [14:20:07] awesome thanks :) [14:20:18] gerrit is full of "intuitive" features like this :/ [14:30:36] It's useful to let you make all your comments and amend them etc.. [14:31:05] but it's hidden under a "Review" button. [14:49:51] gerrit is down? [14:50:22] back now [14:50:27] legoktm: it was for a few second, back up for me now [14:50:30] seconds* [15:18:30] could somebody *please* look at https://bugzilla.wikimedia.org/show_bug.cgi?id=41832 ? a flaggedrevs-related caching bug that's plaguing wikipedias which have them enabled :/ [15:18:48] MatmaRex: have you asked andre__ to take a look? that's his job you know [15:18:54] basically, images display wrong or not at all on stable versions, and many wikis have the stable version set as default for anon users... [15:19:09] sumanah: he's cc'd. [15:19:15] and commented on the bug. [15:19:31] MatmaRex: he's cc'd on a lot of things. if you think something is particularly urgent right now, try pinging him in IRC or email. [15:21:11] MatmaRex: yeah, please feel free to ping me. I'm basically CC'ed on everything, and I admit that I don't always understand how important stuff is :) [15:21:40] MatmaRex: that bug had a truly abominably useless summary. I've now improved it to clarify what the real impact of the bug is [15:21:41] wondering though who to ask to take a look at this :-/ [15:22:19] sumanah: thank you, good point [15:22:21] andre__: Aaron Schulz is the main person with FlaggedRevs strength. Do you know how to check in Gerrit to see who else had patchsets merged into a repo? [15:23:41] sumanah https://gerrit.wikimedia.org/r/gitweb?p=mediawiki/extensions/FlaggedRevs.git;a=shortlog [15:24:01] but I always forget how to exclude the l10n bot [15:24:06] yup! :-) so Krenair might have some free time [15:24:30] example Gerrit search, andre__ https://gerrit.wikimedia.org/r/#/q/status:merged+project:mediawiki/extensions/ProofreadPage+-owner:L10n-bot,n,z [15:24:39] so you add "-owner:L10n-bot" [15:25:04] hmm, I tried ;-owner:L10n-bot [15:25:52] so it's + [15:26:15] matthiasmullie: Krenair - would either of you have time to look at https://bugzilla.wikimedia.org/show_bug.cgi?id=41832 , a FlaggedRevs problem that's affecting anonymous readers by cutting off images or displaying the wrong images? [15:27:02] MatmaRex: so, the next time you want attention on a bug, try (a) making its summary clear, so that people understand why it's a high-impact bug (b) pinging andre__ specifically in IRC [15:27:12] or email [15:28:05] and of course you could just do what Andre & I just did, identifying people who have worked on that codebase and might have time to fix the relevant issue [15:32:50] and that makes it so annoying: https://gerrit.wikimedia.org/r/gitweb?p=mediawiki/extensions/FlaggedRevs.git;a=shortlog works. Adding ;-owner:L10n-bot or +-owner:L10n-bot of course doesn't. [15:33:07] Gerrit. Where I not very often manage to find what I want. [15:33:08] we will be switching to gitblit at some point in the future [15:33:28] a consistent URL parameter scheme would be nice, whatever tool it is. [15:33:36] sumanah: thanks [15:33:50] would gitblit replace gerrit? [15:34:05] gitblit replaces gitweb [15:34:36] ah, I see. [15:35:12] gerrit search works, including excluding certain owners. Gitweb is much less flexible [15:35:30] <^demon> The gitblit search is pretty nice. It's lucene-backed. [15:35:36] Hello. [15:36:06] andre__: could you find someone from operations to deploy https://gerrit.wikimedia.org/r/38631 ? [15:36:23] This is a throttle rule for an event at Santiago this week-end. [15:36:30] Dereckson, just ask here. [15:36:43] or in #wikimedia-operations even [15:36:52] it's not that I'd have more luck than you :) [15:37:16] andre__: you could file an RT ticket. [15:37:32] for deploying throttle changes? we've never done that before. [15:38:32] plus anybody can file an RT ticket. [15:38:35] it does not require me. [15:38:56] <^demon> I'll do it. [15:39:01] email ops-requests@rt.wikimedia.org , done :) [15:39:01] <^demon> No need for RT, this isn't ops. [15:40:32] there is some procedure for those [15:40:51] <^demon> The throttles? They just go through BZ and any shell deploys them. [15:41:10] yes, [15:41:24] there's also 'don't want til the last minute' [15:41:40] I dunno if we have written down anywhere any guidance for this sort of stuff [15:41:42] prolly not [15:41:44] <^demon> Yeah, this is for tomorrow and was in Gerrit since yesterday. [15:41:50] worksforme [15:41:51] <^demon> I don't consider this too last minute. [15:41:57] if I'm really expected to copy stuff from Bugzilla to RT from time to time, I think I'll spend then next 20min to write a script instead. [15:42:02] no, last min is literally 'in a half hour' [15:42:05] <^demon> :) [15:42:23] you don't need to do that for me [15:42:31] <^demon> Dereckson: Deployed, btw. [15:42:45] if by some chance it's a bug I would get, just assign it to me [15:42:47] in bz [15:43:44] yay. thanks [15:43:51] Thank you ^demon. [15:44:26] Alas, it's one of the earliest request this last months. [15:44:33] thes [15:44:34] e [15:45:47] Dereckson: https://meta.wikimedia.org/wiki/Mass_account_creation#Requesting_temporary_lift_of_IP_cap & http://wikitech.wikimedia.org/view/Increasing_account_creation_threshold -- can you edit those to make it any clearer how to get this done? [15:46:19] Ok. [15:48:21] <^demon> Nobody has reviewed https://gerrit.wikimedia.org/r/#/c/38655/ yet :( [15:50:09] ^demon: -1 for non-sematic markup [15:50:24] <^demon> Pfft. [15:51:55] ^demon: ha! [15:52:40] ^demon: people are too busy to celebrate their last (fill here some favourite activity) before the end of the world [15:53:17] Done on meta. [15:54:05] Well... I'm simplifying the list of info asked too. [16:00:37] ^demon: does sync-common-all do sync-dblist? [16:00:50] <^demon> Probably. [16:01:03] <^demon> But sync-dblist is easy, can't ever really hurt to run that. [16:05:01] ^demon- That's a good one! But now I have to go listen to that song. Also, -1 for no i18n. [16:05:10] ha [16:06:12] Done on wikitech. too (were mostly comprehensive there). [16:08:24] thanks Dereckson [16:10:05] Dereckson: what does "The list of wikis you wish to allow account creation on (think to side wikis like Commons)" mean? "think to side wikis" isn't clear.... [16:13:00] so whoever asks me next to copy a ticket from BZ to RT for whatever reasons, cloning is now just one click for me. Cheers. [16:13:10] and now back to work. [16:14:34] yay [16:15:56] in Bugzilla, https://rt.wikimedia.org/Ticket/Create.html?Queue=15&Subject=" + document.getElementById("short_desc_nonedit_display").innerHTML + "&Object-RT::Ticket--CustomField-2-Value=" + bugId + "&Content=Forwarding%20from%20" + document.URL + "%20.%0D%0A%0D%0A%0D%0A" + document.getElementById("comment_text_0").innerHTML for whoever is bored. [16:17:03] andre__: maybe you could note that someplace onwiki? [16:18:32] aharoni: hey, do you have a moment to talk about non-English Wikisources? [16:18:47] ין [16:18:52] that was non-English. [16:18:55] hi [16:19:05] yes [16:20:27] aharoni: https://bugzilla.wikimedia.org/show_bug.cgi?id=41324 [16:24:58] aharoni: how high-priority should that be? [16:30:41] looking [16:30:59] thanks [16:31:09] andre__: ^ maybe you could follow up on this? I need to break in a few minutes [16:32:07] I tend to think about all things related to Collection as not-high-priority, but maybe that's just me. [16:32:33] sure, I'll try [16:32:52] aharoni: sumanah: wikisource people use Collection particularly extensively, i think. [16:33:11] and particularly on these big entire-book-on-one-page things [16:33:17] Maybe, maybe not. [16:33:43] i've been asked about issues related to these pages myself, and i don't even edit ws :) [16:33:47] (not this one, though) [16:33:50] I'm not sure at all, for example, that collection handles local styles well, and on Wikisource they are heavily customized (at least English). [17:00:22] aharoni: FYI, I cannot reproduce bug 41324 anyway (mentioned 40min ago) [17:04:53] aharoni: the Collection extension is crucial for partnerships with libraries, schools, publishers, ... [17:54:11] http://www.netways.de/en/osdc/osdc2013/cfp/ Open Source Data Center conference, CfP closes 31 Dec. Conference will be April 17-18, 2013 in Nuremberg, Germany [18:00:27] Nemo_bis: I understand in principle, but I don't know how much is it actually used. If you actually saw it in action and you know for fact that a lot of people use it and need it, then you win :) [18:02:44] aharoni: do you want emails of a network of libraries and schools who are considering ita requirement for a collaboration? ;) [18:03:34] OK, I'll believe :) [18:03:37] brb [18:03:45] I was just wandering this morning if our WMIT member provided them with the ePubs they wanted for testing [18:23:52] Reedy: any idea why test2 would have a clicktracking error causing UW to hang? https://bugzilla.wikimedia.org/show_bug.cgi?id=43125 [18:24:43] more to the point: http://test2.wikipedia.org/wiki/Special:UploadWizard [18:25:15] Because the extension seemingly isn't enabled for some reason [18:25:27] oh.... [18:25:27] Uncaught Error: Unknown dependency: jquery.clickTracking [18:25:45] <^demon> I thought the click tracking dependency was supposed to be removed from UW? [18:26:04] Though, neither is it on commons [18:26:46] might still be useful to back test2 to 1.21wmf5 as a quick and dirty debug step [18:26:59] ^demon: it is [18:27:12] like I say, WFM on commons [18:27:27] <^demon> Rolling test2 back to 1.21wmf5 would undo a lot of wikidata stuff. [18:27:51] I wouldn't be suprised if E3 stuff broke it, tbh [18:28:58] Vector, E3Experiments and AFT/AFTv5 [18:29:27] ori-l: ^^ [18:29:42] looks like click tracking is removed from test2 [18:29:49] yeah, timingwise, that seems about right [18:30:05] the automated test run from yesterday worked [18:30:43] easy fixed then [18:31:18] yay browser tests [18:33:09] shouldn't click tracking be listed in https://test2.wikipedia.org/wiki/Special:Version?debug=true ? [18:33:23] or was it removed for some reason? or hidden? [18:33:30] or i'm blind :) [18:34:09] Click Tracking (Version 0.1.1)(c02ea00)Click tracking for tracking events that do not cause a page refresh [18:34:12] it's on enwiki [18:44:49] robla: I'm about to bring up https://bugzilla.wikimedia.org/show_bug.cgi?id=42152 " [Regression] All query pages (maintenance reports) updates are broken since 1.21wmf3" in the managers' meeting [18:45:10] test2 appears happier now [18:45:53] okee doke [19:49:56] ^demon: :-D Have you sorted out the ssh connection between manganese and gallium ? [19:50:44] Reedy: gah, just got in [19:50:53] <^demon> hashar: No, I haven't. Let's figure it out. [19:51:07] btw: spagewmf & superm401 are also e3 devs, i am no longer the single culpable person [19:51:13] the guilt is now thoroughly diffuse [19:54:03] <^demon> Ryan_Lane: hashar and I have been trying to get replication -> gallium working, but I keep getting "reject HostKey: gallium.wikimedia.org." [19:54:20] ^demon: I am not sure how i can help :/ [19:54:37] <^demon> The gerritslave user exists on gallium, I can manually `ssh -i gerritslave@gallium`, and the fingerprint is in gerrit2's known_hosts. [19:54:39] beside googling for that issue [19:54:48] <^demon> Ryan_Lane: Do you remember what we did last time we had this problem? [19:56:46] <^demon> hashar: Also, do you want to replicate the integration/* repos to github? [19:56:53] <^demon> That's trivial to setup. [19:57:23] ^demon: I don't care about github, but maybe some people might be interested :- [19:57:31] * valhallasw would be interested :-) [19:57:41] ^demon: if it is easy to setup and does not add to much hassle, set it up [19:58:03] imo gitweb is quite... frustrating. It's handy to have a good web interface to browse the repository [19:58:12] ^demon: just made sure it was in the user's trust [19:58:12] especially if you don't know for sure what you are looking for [19:58:41] valhallasw: I think we have a plan to use gitolite or something similar that would run on a dedicated server. [19:58:42] <^demon> valhallasw: This is why gitweb is going away. [19:58:43] <^demon> It sucks. [19:58:48] <^demon> gitblit. [19:58:57] <^demon> But no, it won't be on a dedicated box. Permissions problems. [19:59:15] <^demon> Ryan_Lane: You're saying to do that, or that you just did that? [19:59:25] cool! :-) [20:00:03] hashar: is it possible to do loops in Jenkins, other than in the bash scripts? I'm trying to extend the extension tests to run for different branches of mediawiki-core [20:00:19] valhallasw: oh [20:00:41] one option would be to let the looping to zuul instead, and to start multiple jobs in jenkins [20:00:56] or possibly just having several jenkins jobs, one for each branch [20:01:10] valhallasw: I am using a python script to generate jobs : Jenkins Job Builder. It let you define templates for jobs, you then define a set of parameters and the script apply them to the job template. [20:01:27] valhallasw: so we could expand all the current extensions jobs to run tests against master, REL1_19 and REL1_20 [20:01:33] ^demon: that's what I did last time [20:01:43] <^demon> How? :) [20:01:45] yes, that would be relatively simple (to just generate three jobs per extension) [20:01:47] valhallasw: so you would end up with a job per branch :) (what you just said) [20:01:53] ^demon: I su'd to gerrit2 [20:01:57] then ssh'd to the system [20:02:23] <^demon> Ah, I did that. That's what I meant by "fingerprint is in known_hosts" [20:02:35] valhallasw: of course some extension @ master might not be supporting all core branches :/ [20:03:28] hashar: of course. but it's a pain to test that manually, so I think it would be especially good to have it in jenkins [20:03:35] valhallasw: definitely. [20:03:49] <^demon> Ryan_Lane: Herp derp. ssh gallium != ssh gallium.wikimedia.org [20:04:01] valhallasw: maybe we could run the tests against REL1_20 / REL1_19 and ignore the result by default (aka not mark Verified: -1 if such job fail) [20:04:08] in my 'perfect world', the extension would just have a file 'test-branches' with contents 'master REL1_19 REL1_20' etc [20:04:48] so if the extension wants to test against, say 1.15, that would also be possible (although I have no idea why anyone would want to keep a 1.15 wiki around) [20:05:22] Running 1.15 dh [20:05:29] oops [20:06:25] valhallasw: anyway that is not really high priority for now. [20:06:42] <^demon> Ryan_Lane: ssh working now, thanks. [20:06:43] valhallasw: I would like us to actually test ALL extensions instead of just a few [20:06:49] ^demon: Ryan_Lane: \O/ [20:06:50] <^demon> hashar: Ok, so ssh works. Just getting "repo DNE" [20:07:02] DNE ? [20:07:08] hashar: and what is keeping you from doing that? [20:07:23] valhallasw: still have a few issues with the current extensions. [20:07:31] valhallasw: I also need a system to handle dependencies. [20:07:33] <^demon> hashar: does not exist. [20:07:41] valhallasw: some extensions requires other extensions to run their tests [20:07:53] <^demon> eg: [2012-12-14 20:06:07,413] ERROR com.google.gerrit.server.git.PushReplication : Cannot replicate to gerritslave@gallium.wikimedia.org:/var/lib/git/All-Projects.git; repository not found [20:08:10] ^demon: indeed that file does not exist [20:08:17] ^demon: isn't it to create them for us ? [20:08:28] ^demon: isn't it supposed to create the repo for us ? [20:08:55] <^demon> Thought so. [20:09:01] <^demon> Not sure why it's not...lemme find out. [20:09:55] hashar: do you want to manage them in JBB or let the extensions manage it theirselves? [20:17:09] valhallasw: ideally the master of an extension should publish which minimal version of mediawiki it supports [20:17:21] valhallasw: then extensions should have branches for older mediawiki versions [20:17:34] valhallasw: and yeah if possible I would like to manage that with JJB [20:52:55] hashar: cool, the testextensions-{branch} jobs are working. Of the tests are not working because they use stuff that is too new :-) [21:01:20] valhallasw: are you using JJB already ? :) [21:01:40] (just remembered about your vagrant image) [21:01:44] hashar: yes! https://gerrit.wikimedia.org/r/#/c/38790/ :-) [21:02:04] not with vagrant yet, but the hackish approach is working well enough [21:02:06] awesome [21:03:00] valhallasw: I will look at that next week ) [21:03:10] sure! [21:03:32] valhallasw: also need to have a look at your vagrant install. not going to happen before january though [21:03:48] valhallasw: then we might want to mail everyone about it so they can play with Jenkins on their local machine. [21:04:05] by the way - is the zuul config generated, or is is kept manually? there's a lot of repetition in there [21:04:20] valhallasw: manually [21:04:33] valhallasw: it could use a templating system indeed. [21:04:39] hashar: I mailed about my quick setup yesterday (or the day before) to wikitech-l [21:04:58] but a vagrant image combining jenkins and zuul would be, well, awesome [21:05:20] valhallasw: would also need Gerrit in the loop [21:05:29] valhallasw: Gerrit is not fully puppetized yet though :/ [21:07:19] valhallasw: https://gerrit.wikimedia.org/r/#/c/38790/ makes a lot of sense. [21:07:27] valhallasw: will probably deploy it next week. [21:08:20] cool! [21:08:50] of course, the zuul config will need some adaptation too... hmm. [21:09:20] wouldn't it make more sense to define the jobs in the zuul config and then letting JBB generate them from that list? [21:09:43] I'll fiddle around with that brainfart :-) [21:11:59] valhallasw: you are volunteer aren't you ? [21:13:11] yes :-) [21:13:25] I just like structuring stuff [21:13:53] I love cathedrals myself but my brain is a bazaar [21:14:16] your brain is bizarre :P [21:14:22] j/k [21:16:05] Hehe. I actually quite like stuff like this because it allows for relatively fast payoff - which is somewhat different from my day job (physics research) [21:17:19] valhallasw: I have added a quick status update for the continuous integration project : https://www.mediawiki.org/wiki/Continuous_integration/status#2012-12-14 [21:17:57] valhallasw: I choose CS for two reason: you (usually) don't kill people when doing mistakes. Mistakes are most always your entire fault. [21:18:12] entire / own [21:18:14] :D [21:18:18] also http://www.mithral.com/~beberg/manifesto.html [21:18:31] which I have read when I was still a child [21:18:56] The doc that got children of the 80's into hacking... :) [21:18:59] that text probably saved my life [21:19:50] I should find out whether the author is still alive and meet him face to face [21:20:27] http://en.wikipedia.org/wiki/Hacker_manifesto ;-) [21:20:42] so the guy is called Loyd Blankenship and still alive [21:21:45] doh [21:21:48] that site is nice [21:22:00] do they have biography about everyone ? I should write my own [21:22:42] hmm [21:22:49] maybe I will visit Austin, TX [22:05:54] ^demon: if you succeed to unlock https://gerrit.wikimedia.org/r/#/c/38063/ situation, I can resubmit the change without dependency. I only added them to avoid a manual Git merge. [22:06:36] <^demon> Submit the parents ;-) [22:07:28] ^demon: this changeset could also serve as a dependencies-related error demo, as we've the missing changeset identified by Id or stuff like "Depends on patch set 3 of Ibade248e, however the current patch set is 7." [22:08:08] Well the parents are ready but I'm not sur they will be submitted very soon. [22:08:30] There are the initial configuration for es.wikivoyage and pt.wikivoyage. [22:50:47] is there any reason why TeX doesn't work at all on English Wiktionary? [22:51:47] it just gives me this: " Failed to parse (Missing texvc executable; please see math/README to configure.)" [22:51:53] but I wasn't sure if it's intentional [22:55:14] Liliana60: what page? [22:55:43] it happens on any page where

is used, here's one example: http://en.wiktionary.org/wiki/Kurepa_tree
[22:56:06] i vaguely recall last time that happened was because someone had brand new provisioned hosts (new OS major version maybe?) in rotation
[22:56:16] notpeter: ping? ^^
[22:57:36] > [22:57:58] purge didn't fix
[22:58:08] and now it's mw34
[22:59:11] both of those hosts have been in production for months now
[22:59:20] (at their current package/OS state)
[23:00:02] Liliana60: do you have any clue when it first broke?
[23:00:39] sadly no
[23:01:52] it was only reported recently but we don't use very often so it may be an older thing that no one noticed up to now
[23:02:44] afaict, mediawiki-math has only ever been installed on the pdf servers
[23:03:14] I can definitely put it on all of our app servers, but I'd like confirmation from someone in eng about this change
[23:03:34] ping Reedy RoanKattouw_away
[23:03:59] oorrrrrr... AaronSchulz 
[23:04:51] hmmm
[23:04:52] http://en.wiktionary.org/wiki/Venn_diagram
[23:04:58] on this page, the math displays correctly.
[23:05:42] would you be willing to make a bug and add me?
[23:05:53] if this is determined to just bea need for a package, I can take care of it
[23:06:09] maybe but I have yet to create a bugzilla account
[23:06:10] but I'd prefer someone with more mediawiki knowledge than I take a look at it first
[23:06:15] ah, ok :)
[23:06:18] then don't worry about it
[23:06:23] I cna
[23:06:25] *can
[23:07:21] holy crap.... nope. bugzilla is way complicated
[23:07:30] what is mediawiki-math?
[23:07:34] nfi
[23:07:36] * AaronSchulz  has not context
[23:07:42] if only SUL applied to bugzilla too
[23:08:28] what I know: some pages look pretty broken. (see above) it's asking for texvc
[23:08:41] notpeter: is it a package or something?
[23:08:43]!log synchronized payments to 4f5296e43b1c708
[23:08:44] it looks like that is provided by wikimedia-math package
[23:08:51] Logged the message, Master
[23:08:58] er,
[23:09:00] a pecl extension?
[23:09:02] mediawiki-math
[23:09:08] I thought math just shelled out to a binary
[23:09:16] looks like it
[23:09:22] seems to be all new pages, if I add to my userpage aaa it gives me the big red error too
[23:09:26] that was compiled on each box (formerly by scap)
[23:09:29] but old pages seem to still work, or something
[23:09:33] notpeter: is it all machines?
[23:09:45] AaronSchulz: please define "it"
[23:09:56] well there is some error going on right?
[23:10:00] yes
[23:10:04] it looks to not just be one box
[23:10:11] a subset of apaches?
[23:10:15] and 2 = all ;)
[23:10:21]