if that helps [04:29:45] And that's all I really see (of any importance) in the mw-content-text [04:46:28] Anyone? [04:58:56] hey robla [05:10:49] howdy [05:15:52] robla: sent some scrollback in /msg [05:16:14] idk how prevalent it is but it certainly doesn't seem to be going away [05:17:11] anyway, it's at least 3 days old so it's not a huge emergency. just thought you might have some hints about who to poke or where to look [05:56:48] jeremyb: only just now sitting down to the computer in earnest [05:59:29] * robla can't repro search issue as described [06:08:49] robla: i've been assuming it somehow effects some subset of user agents. anyway, >=6 independant reports over ~3 days sounds notable to me [06:15:35] jeremyb: one of the people says they're running IE8. I don't have access to that right now. I wonder if the people hitting problems are on older versions of IE [06:16:04] if they are, it'd be interesting to know if they have anything in their javascript console [06:49:30] TimStarling: sorry to nag, wondering if you could once again glance at https://gerrit.wikimedia.org/r/#/c/33157/2/EventLogging.module.php. not urgent. [08:11:33] hi zeljkof [08:11:59] morning ori-l [11:17:50] lol, Wikia fills the sites with nice popups to warn you of maintenance windows [11:18:16] but then when the DB is read-only instead of showing normal warning on action=edit they let you try and save it and throw you fatal errors [14:41:20] QueenOfFrance: "welcoming users with our without a bot flag unless they make substantial edits is stupid"... Thanks! [14:43:09] Sorry I typed this by mistake, I meant to write it in #wikimedia-stewards :) [15:22:08] !g test [15:22:25] !gerrit test [15:22:33] hello guys [15:22:36] I'm Stefan Petrea, external contractor working on WMF Analytics team [15:22:38] is there a reliable way to get hostname of stat1 and all prod machines ? [15:23:16] I guess there are racktables for this, but I don't really see into production stuff, you may ask in #wikimedia-operations [15:23:35] 3rd channel I've been directed to [15:23:40] heh :D [15:23:42] ok, I'm gonna ask [15:23:53] people from -operations are in here as well, just don't watch that much [15:24:06] petan: thanks [15:24:09] yw [15:24:34] petan: it's much better than other large-scale organizations I've seen where I had to phone 7 people to get an answer [15:24:50] well, did you get answer yet? :P [15:24:57] petan: I will [15:24:57] we will see... [15:24:59] well, WMF is not a large scale org :p [15:25:03] :D [15:25:26] Nemo_bis: actually there are a lot of people working on WMF as far as I can see [15:25:43] average_drifter: less than two hundreds AFAIK [15:25:53] Nemo_bis: contributors included ? [15:26:00] there are lot of people working on wikimedia project who are not WMF as well [15:26:21] https://wikimediafoundation.org/wiki/Template:STAFF-COUNT [15:26:35] average_drifter: well contributors are not WMF :) [15:39:13] what is the goal once you have the list? (this will help me understand what info source might be useful for your purposes) [15:39:28] average_drifter: [15:40:25] apergos: well I don't need a list of the machines [15:40:55] ok.. what are you trying to get exactly? [15:41:03] apergos: I just want, when I'm on a machine, I need to know if there's a uniform way of answering the question "what is the name of this machine ? is this prod or test ?" [15:41:11] ah [15:41:38] so "what is the name of this machine" when you are on it, is 'hostname' for the short (not fqdn) name [15:42:19] and whether something is used for production or test is a matter of what's installed on it, [15:42:27] so you could consider lookig at /etc/motd [15:42:42] which on most hosts will ahve a line or two describing what the host does [15:42:44] for example: [15:42:52] ms-be6 is a Wikimedia Swift pmtpa production cluster (role::swift::pmtpa-prod) [15:42:53] etc. [15:43:45] can I grep for "production" in the motd ? [15:43:50] paravoid: ^^ [15:44:05] paravoid: I am in need of a way of finding out if a machine is prod or not [15:44:08] paravoid: hi :) [15:44:09] I wouldn't count on that [15:44:16] what do you mean by "prod" or "test"? [15:44:23] paravoid: prod = production [15:44:29] paravoid: test = non-production [15:44:33] didn't the analytics assign which servers to be prod and which ones to be test? [15:44:38] *analytics team [15:44:38] please accept these "ad-hoc definitions" [15:44:50] there's /etc/wikimedia-realm [15:44:55] paravoid: yeah, that file is cool, I saw it [15:44:57] that's going to have the values of "production" and "labs" [15:45:02] was just wondering if it's on all machines so I can use it [15:45:32] it's in the base classes, so yes [15:45:38] paravoid: he is referring to analytics machines, not the cluster (I believe) [15:45:50] that's awesome, I'm gonna use that file [15:45:52] but yeah, it depends on your definition of "test" and "non-production" [15:45:52] Hydriz: I'm refering to stat1, build1, build2. But there may be other machines as well [15:45:56] the analytics cluster is production [15:46:05] basically everything that is not a labs VM is "production" [15:46:12] heh [15:46:40] problem is, different people/teams mean different things when they tell if something's production or not. [15:46:51] if you need some way to distinguish between analytics hosts doing different things, the team will have to add something to make that distinction [16:47:01] andre__: can you confirm the bug about incubator:Special:Preferences ? [16:48:46] PiRQuadratus, yes I could reproduce it [17:33:29] hi! [17:33:36] is it Internal error [17:33:36] Jump to: navigation, search [17:33:36] [5a906823] 2012-11-13 17:32:59: Fatal exception of type MWException [17:33:36] ok? [17:33:51] it is Special:Preferences @ incubator [17:34:12] Base: I already reported [17:34:19] https://bugzilla.wikimedia.org/show_bug.cgi?id=42068 [17:34:27] :) [17:39:37] ok :) [17:42:01] https://bugzilla.wikimedia.org/show_bug.cgi?id=42068 Could someone look this error up in the logs? [17:43:20] Can another sysop see if this happens? [17:43:28] Maybe it's non-sysop-only or something [17:45:03] MF-Warburg: for what it's worth, last night I asked Romaine to reproduce it -- [s]he was unable to [17:45:09] I have a vague remembering that it was related to the usual LinkCache bug [17:45:48] PiRQuadratus: well, he's also a sysop - while Merlissimo is not [17:46:04] MF-Warburg: seems to point toward sysop-only, no? [17:46:50] shall I sysop you to see what happens or what? :P [17:47:41] it depends on language [17:47:47] try another uselang [17:48:00] iirc [17:48:23] hm maybe not this one [17:48:57] MF-Warburg: i am using language "en" and test wiki "maintanace" [17:59:40] Oh I set uselangage throw api now I can understand interface ^) [17:59:43] *:) [18:00:00] change=language%3Duk [18:00:16] what is format for test wiki setting? [18:04:17] https://incubator.wikimedia.org/wiki/Special:Preferences?uselang=de doesn't work any more than English... [18:05:51] Nemo_bis: which language do you suggest? [18:06:05] PiRQuadratus: vo of course [18:06:26] Pöl ninik... [18:06:29] but as I said above that was another bug probably [18:07:04] Nemo_bis: Tio estas Volapukaĵo. [18:08:01] our favourite Wikipedia [18:09:43] what is a progress? any ideas how to fix error? [18:13:02] Base: https://bugzilla.wikimedia.org/show_bug.cgi?id=41337 [18:13:52] and when Gerrit change #31321 will be done? [18:52:23] * Nemo_bis hugs apergos [18:52:27] thanks for the interwikis! [18:52:33] you test em? [18:52:42] I'm pretty sure that's not the officially approved method [18:52:53] I'm also pretty sure that there is no third party method that works now [18:53:27] now that we have the steps written down I can ping someone like reedy and we can figure out what needs to be cleaned up to make this work for non wmf sites [19:50:36] Reedy (or anyone else): trace needed: https://bugzilla.wikimedia.org/show_bug.cgi?id=42079 [19:51:43] Argh [19:51:45] Ping! [19:51:51] pong [19:52:37] apergos: ? [19:53:08] yeah, I got interwiki shortcuts working on wikitech but I am quite sure it was not the officially authorized way to do it [19:53:18] ooh [19:53:28] I think I told mutante to just copy it from fenari last time... [19:53:30] https://bugzilla.wikimedia.org/show_bug.cgi?id=41786 [19:53:34] yeah but that was't enough [19:53:51] when you have a bit of spare time feel free to read what I did (it's n the bug comments near the bottom) [19:54:14] Yeah, I've literally just got in from driving back from Antwerp [19:54:15] maybe there is a better approach, or maybe there is something that could be cleaned up in core to make this less painful for third parties [19:54:26] yeah it's o rush, just stick it in the pile somewhere [19:54:29] *no [19:55:11] hoo: attached [19:55:45] And it is a dupe of 22068 [19:55:48] *42068 [19:56:48] Reedy: Ah yes... I'm going to post the trace over there... seems to be the incubator extension [19:56:59] I posted it to both [19:57:01] ah, you bet me :P [19:57:32] beat [19:57:42] bet-ting involves money ;) [21:26:47] Platonides: do you remember if we ever found out the reasons of this very mysterious technical problem? https://en.wikipedia.org/?diff=502495682&oldid=502494882 [21:29:59] Nemo_bis: We can't reset his password [21:30:03] No one can verify who he is [21:32:41] Reedy: nor what he voted [21:32:49] or if his hand was raised by someone behind him [21:32:57] wearing an invisibility coat [21:36:20] hi Reedy, see pm [21:37:51] robla: IE8/IE7 search issue fyi: https://bugzilla.wikimedia.org/show_bug.cgi?id=42082 [21:40:40] anomie: ^ [21:46:31] chrismcmahon- After a quick test, it seems IE8 only in compatibility mode? [21:47:28] anomie: I think I saw the issue in normal mode in IE8 via crossbrowsertesting.com, but not sure if CBT is 100% reliable, I can double-check if you want [21:48:48] chrismcmahon- Please do [21:52:28] anomie: you're right, IE8 on crossbrowsertesting automatically put me in compatibility mode (argh, phooey), with that off, search seems normal. so it is IE7 and IE8 in CM [22:06:26] syncing PrivateSettings.php [22:29:21] andre__: did you get OTRS access? [22:29:45] I replied to Maggie on that ... was just waiting on another note from her [22:29:53] okey [22:30:14] I can create it now though - just a matter of what access [22:30:27] He'll get the tech stuff now though, just need to know what his username is [22:31:03] !log restarting gmetad on nickel [22:31:09] Logged the message, Mistress of the network gear. [22:33:53] RD: so, info-en && info-en::tech-issues ? [22:34:07] Yeah, info-en F [22:34:23] I asked her whether WTS access was needed [22:34:28] (Wikimedia Technical Staff) [22:34:36] I rather large package that contains a lot of irrelevant things. [22:34:49] RD: right, idk [22:34:58] s/I/A [22:51:14] chrismcmahon- How was this not broken until just a few days ago? [22:51:53] anomie: it may have been broken earlier than that, but the first reports I know of are from last week [22:52:49] chrismcmahon- It looks to me like it should have been broken at least since February, or since "SimpleSearch" was turned on (whichever is later). [22:53:08] anomie: I put it on the browser-test backlog (we're still working on the browser-test architecture, but forging ahead as we can) [22:53:13] anomie: http://www.mediawiki.org/wiki/QA/test_backlog#Backlog [22:53:40] anomie: not that many IE7 users, and those don't give us much feedback. thus the need for some tests [22:54:58] chrismcmahon- https://gerrit.wikimedia.org/r/33287 [22:56:10] thanks anomie, nice catch [22:57:57] An alternative fix would be to add a JavaScript hack that runs on those browsers and explicitly sets type="submit". Not sure if we have such hacks already, I'd hate to add the first. [23:03:11] anomie: there's a changeset on gerrit that seem to fix a related issue? https://gerrit.wikimedia.org/r/#/c/30819/4 [23:03:14] seems* [23:03:29] sitting for two weeks... [23:03:32] with three +1s... [23:05:07] MatmaRex- Good catch [23:07:05] * anomie goes to review that now [23:10:16] * anomie merges changeset 30819 [23:12:16] so it took two weeks for somebody who can and is willing to merge such changes to look at it, at which point it was reviewed in 10 minutes. [23:12:31] there's something deeply wrong here :/ [23:14:46] MatmaRex_- Well, it helps that I did just about the exact same thing in my patch a few minutes earlier. But yeah, it all depends on someone with +2 to look at it and feel confident enough to take responsibility for merging it. [23:17:48] MatmaRex_- Which is why sometimes you'll see someone with +2 only give it a +1, they want someone else to look it over before it gets merged. [23:21:38] anomie: yeah, except i've been told by people with merge privileges that they mostly don't look at changesets which already have a +/-1 [23:21:43] anomie: But for some things that doesn't seem to happen... eg. nobody except of Krinkl.e yet took a look at https://gerrit.wikimedia.org/r/26440 (AJAX patrolling). This can be annoying, as I wont to go on developing other things, while not having to many things open in gerrit which I have to worry about [23:21:48] so it gets a +1 or two [23:21:50] and then rots [23:21:53] for weeks [23:22:29] hoo: see, your code introduces some new non-trivial code [23:22:30] s/wont/don't want/ [23:22:43] but changes like the one that started this disucssion we're having now [23:22:49] they are trivial [23:23:01] this one has a commit message longer than the changed code, with justification and documentation [23:23:02] MatmaRex_: Trivial changes are lying around for ages as well [23:23:29] * Krenair is tempted to take the opportunity to moan about the oversight migration script review [23:23:43] hoo: my own oldest change is, i thnik, from september 7th [23:23:49] MatmaRex_- That seems odd to me. I'd prefer to see a few +1s on something I might merge. [23:23:53] although it went through an iteration or two in the meantime [23:24:14] anomie: well, then feel free to look at my dashboard, lots of +1s there [23:24:17] Mine is "Sep 16, 2012 10:32 PM" [23:25:24] hoo- I've looked at it, and it looks good to me, but since Krinkle has been reviewing it I didn't do anything to it yet. I see he +1ed it now, which makes me feel good about it. [23:26:57] anomie: Why didn't you give any feedback if you already had a look? This way it seems like the change is just lying there till Krinkl.e has time to take a look [23:27:08] (no offense, just asking) [23:27:47] hoo- I had nothing to add this morning when I looked, and I hoped Krinkle would approve it within the day. [23:28:32] I have to admit that I do this sometimes as well, especially with changes which affect parts I'm not used to much (eg. new API modules) [23:30:26] thanks, anomie :)