[00:48:10] Would anyone be upset if I proposed turning off LiquidThreads on MediaWiki.org? [00:48:34] yes [00:48:46] unless you get community consenus [00:49:00] that's why it's called a 'proposal' :) [00:51:21] I'll take that as a no then [00:54:54] kaldari: until we have a replacement, yes [00:55:06] I use it pretty heavily [00:55:17] my issue is that it's bit rotting and I have to keep writing around it [00:56:04] well, then echo needs to get out the door [00:56:18] actually, that won't even help [00:56:27] we need *something* that provides sane talk pages [00:56:33] that's what I'm working on :) Killing LQT would make that faster. [00:56:47] well, that would be Flow [00:56:58] is it? [00:57:03] thats only for user talk pages [00:57:24] eventually it's supposed to be for everything [00:58:27] I'm not going back to wiki talk pages. you'll need to pry LQT from my cold dead hands if there isn't a reasonable replacement :) [00:58:41] fair enough [00:58:56] I'm using it as a support mechanism for like 5 extensions. I think I'd go insane [00:58:59] I just wish we had someone supporting it if we're going to actually use it [00:59:43] agreed [01:01:27] It seems like every time I check in something for Echo, someone is like "But you forgot about weird issue X for wikis that are running LQT" [01:01:51] sigh [01:04:59] job queue on enwiki is >1,000,000, which IIRC is unusually high; seems to be limited to enwiki, as other large wikis are not showing similar queue growth [01:13:42] kaldari: that sucks :( [01:14:08] oh well, I can deal with it [01:14:14] kaldari: I don't really have an alternative right now, except having people send me personal email [01:14:26] but then no one else gets to see that, so I get asked the same questions over and over [01:15:21] I just wish we wouldn't deploy half-finished software and never finish it [01:15:30] completely agree [01:15:42] I'd say go for consensus on mediawiki.org [01:15:52] I can live without LQT if I really need to [01:16:06] I'll sleep on it :) [01:16:25] one should never take action out of anger ;) [01:16:35] :D [01:17:37] I just wish we wouldn't deploy half-finished software and never finish it true [01:27:51] hello, an archeology question: How to check wich was and when was created the first article of pt.wp? (created in 2001) [01:30:58] if the history is not in the database now, then that information is probably lost forever [01:31:42] well, at least at the technical level [01:31:54] of course you could just find the person who created it and ask what it was [01:31:57] those where my suspicious. what about the dates regarding the artciles of fase I ? [01:32:42] have you checked archive.org? [01:32:44] well, there is a discussion on pt about that, it's known for ages wich page was, but checking dates, other pages seemed to be created first [01:32:58] yes, inconclusive by now [01:34:02] in usemod, old revisions were automatically deleted if they were older than a certain period [01:34:29] Alchimista: you could look at the old dumps [01:34:34] there was a copy hidden in the data directory but that copy was not migrated to phase 3 [01:34:52] then all copies of the original data directories were lost [01:35:20] except for a backup of the English Wikipedia from August 2001 which I found on sourceforge [01:35:38] Alchimista: The lowest page_id currently in the database is http://pt.wikipedia.org/wiki/Astronomia [01:35:44] on stats.wikimedia there is the zelteirgest, wich indicates the first article, should i assume it's reliable? [01:35:59] but it's page id is 220, so it's very unlikely it was the first [01:36:11] it was probably imported from some earlier set anyway [01:36:11] kaldari: yap, i know, i've searched by id and timestamp, but by *oral tradition* the first page was planeta [01:36:42] yes, from usemod [01:36:54] -> http://stats.wikimedia.org/EN/TablesWikipediaPT.htm#zeitgeist [01:38:05] Something weird is going on [01:38:50] oral tradition is most likely the best source of information [01:39:16] like I said, old revisions were deleted, so if other pages were apparently created earlier, that is only because they were not edited for a while [01:39:29] the oldest page on the english wikipedia according to phase 3 data is UuU [01:39:51] just because it was created early and not edited, so usemod didn't delete the original history entry [01:40:23] it's acepted to be planeta, but how the discussion returned, all i wanna do is to explain what happened and wich info is *reliable* [01:41:56] have I explained it well enough already? [01:42:28] sure. all i needed was a confirmation. i'll post locally. [01:45:22] anyone? any idea what's going on with enwiki job queue? [01:47:20] the expiry period would have been 14 days, if that's of interest [01:47:53] that's what $KeepDays was set to in the english wikipedia backup [02:14:55] TimStarling: i've posted already, thanks for the info, and explanation [02:15:26] no problem [02:52:26] How do you change the interwiki map of a wiki? [02:52:48] e.g. switch from Wikipedia standard interwiki map to the wikisource one [03:24:31] (There's only one interwiki map for Wikimedia wikis. It's hosted at Meta-Wiki.) [03:24:40] not true! [03:25:41] 0) wikisource treats en: differently than wikipedia [03:25:42] jeremyb: What's not true? [03:26:03] en: is relative not absolute [03:26:12] each wiki has it's own map [03:26:25] its [03:26:29] And I'm not sure what you're talking about. [03:26:50] 1) some are not even based on that central map. see labsconsole and wikitech [03:27:00] anyway, /me runs away [03:27:10] 04 02:52:25 < Dereckson> How do you change the interwiki map of a wiki? [03:27:10] 04 02:52:47 < Dereckson> e.g. switch from Wikipedia standard interwiki map to the wikisource one [03:27:17] 04 03:24:31 < Susan> (There's only one interwiki map for Wikimedia wikis. It's hosted at Meta-Wiki.) [03:27:37] labsconsole and wikitech are irrelevant. [03:27:46] And I still have no idea what the hell you're talking about. [03:32:52] Susan: http://dpaste.com/hold/862915/ [03:33:20] jeremyb: wikitech is not even in the cluster, last I recall it was hosted on some weird linode, tho I think that got changed [03:33:32] s/weird // [03:33:36] i doubt it changed [03:34:01] although i know it wasn't something everyone liked [03:34:20] QueenOfFrance: labsconsole isn't on the cluster either [05:31:08] .config wgExtraNamespaces plwiki [05:31:09] NS_PROJECT_TALK => Dyskusja_Wikipedii / NS_USER => Wikipedysta / NS_USER_TALK => Dyskusja_wikipedysty / 100 => Portal / 101 => Dyskusja_portalu / 102 => Wikiprojekt / 103 => Dyskusja_wikiprojektu] [05:31:16] .config wgExtraNamespaces plwiki 102 [05:31:17] Wikiprojekt] [05:31:59] Currently, this is not UTF-8 compliant. It will be at Nasqueron crashes, it has been recompiled to have a UTF-8 support in its message queue. [05:33:27] I coded this feature, when I were in flaggedrevs.php and wanted to have a namespace number. [05:34:03] Don't know if it will be often useful, but enjoy. [05:34:12] hmmm I didn't test boolean value [05:34:19] .config wmgEnableLandingCheck default [05:34:19] ] [05:34:26] .config wmgEnableLandingCheck testwiki [05:34:28] 1] [05:36:48] .config wmgEnableLandingCheck default [05:36:49] false [05:36:52] Better. [05:37:39] So if you need any command related to Gerrit or to InitialiseSettings.php parsing, just ask and I'll add that to Nasqueron. [08:41:08] Does anyone know the irc.wikimedia channel name for wikidata? [08:42:03] if you issue /list on that network it should give you a listing of all channels [08:42:32] right. [08:43:18] #wikidata.wikipedia is the answer. [08:44:09] sigh [08:44:16] Nemo_bis: i know >.> [08:44:42] aka https://bugzilla.wikimedia.org/show_bug.cgi?id=28276 [13:07:44] Hi [13:07:46] How I could download all images in a Category? [13:07:54] commons [13:15:08] wilfredor: i don't think there's an easy answer. should not be too hard to script that using some bot framework. possibly such a script exists somewhere... [13:15:33] interesting [13:16:50] I need download images for offline use [13:42:12] Reedy: ^demon poke [13:50:36] <^demon> aude: Howdy. Jenkins just merged 42037. [13:50:45] \o/ [13:50:47] yay [13:51:01] so, i'm tagging our extensions at where we want them [13:51:28] aude: what about the "everythign is english" bug? was the Sites issue causing that, or is that something else? [13:51:30] have that done in a few minutes and then, whenever a good time for you or reedy, try switching wikidata back to wmf7 [13:51:36] DanielK_WMDE: exactly [13:51:47] ok [13:51:56] <^demon> aude: You going to submit the change to 1.21wmf7? [13:52:18] hoo man's patch works nicely, although agree more protections against changes in serialization,e tc. is good too [13:52:25] ^demon: yes i can [13:52:37] <^demon> Okie dokie. I'll review as soon as it's in. [13:55:36] i have the cherry pick also [13:56:24] ah, you got it [13:56:45] ^demon: while you are around, could you help us to find bug 41586? [13:56:52] let me verify that it's still there [13:57:40] * aude confused, still need to cherry pick [13:59:34] ^demon: hm, seems to work for me now. strange [13:59:40] i'll poke it some more [13:59:54] <^demon> Was it broken on 1.21wmf6 or 7 or both? [13:59:56] <^demon> (before) [14:04:08] ^demon: wmf6 is okay [14:04:16] ^demon: wmf6 and before, never tried in 7 [14:04:30] <^demon> Hmm. Well it's working now. We'll check again with 7. [14:04:31] * aude tried with both [14:04:50] ok, there are my patches [14:05:02] https://gerrit.wikimedia.org/r/#/c/42253/ [14:05:07] hm, not even positive i tried in 6. it was still broken on december 14, didn't try after that. [14:06:03] well, the memcached needs to be kicked and the change forces the cache for sites stuff to be refreshed [14:06:23] when i switched my own instance between wmf6 and wmf7, saw the bug [14:06:37] just like happened on wednesday [14:12:28] <^demon> Ok, running scap. [14:12:47] yay! [14:15:00] <^demon> I can't wait for git-deploy. Anything's gotta be faster than scap. [14:15:07] :) [14:15:30] <^demon> It was l10ncache and rebuild-tex-every-scap that made it so slow. It used to not be this bad. [14:16:30] we'll be using git-deploy after the data center migration stuff? [14:16:53] <^demon> Around the same time. [14:17:01] ah, ok [14:17:08] texvc is pretty quick [14:17:11] <^demon> (Actually the plan I think it so to try git-deploy once before we changeover, to make sure it works) [14:17:21] That and you can spawn 10s of threads at a time and it doesn't really matter [14:17:22] sure [14:17:22] <^demon> Reedy: Yeah, tex isn't as bad. It was really the l10n cache. [14:20:31] <^demon> Reedy: /home/wikipedia/common/wmf-config/ExtensionMessages-1.21wmf7.php - permissions still borked? [14:20:32] <^demon> Bah [14:20:43] Oh, I never asked ops to fix it [14:20:53] You can ignore that unless your deploying a new extension ;) [14:23:17] <^demon> -rw-rw-r-- 1 reedy wikidev 20592 Dec 27 22:53 ExtensionMessages-1.21wmf7.php [14:23:20] <^demon> Why can't I write? [14:23:38] it's running as mwdeploy [14:23:39] chown mwdeploy /home/wikipedia/common/wmf-config/ExtensionMessages-1.21wmf7.php [14:23:42] ^ needs that [14:24:21] <^demon> oh bah [14:24:35] * aude surprised ^demon isn't root [14:32:41] Nikerabbit: hi! about the EditFilterMergedContent hook: the Translate extension will need a small update to work with I99a19c93. Want me to submit a patch? It should be simple, but I don't know enough about the Translate extension to test it properly. [14:33:14] I can test it if it even needs testing [14:34:22] Nikerabbit: ok, I'll change it. I think I'll make separate functions for EditFilterMerged and EditFilterMergedContent though, so keep things clean. [14:34:43] * DanielK_WMDE is still running pull for all extensions [14:40:25] aude: you know reading this kind of commit - https://gerrit.wikimedia.org/r/#/c/42232/1/wmf-config/InitialiseSettings.php - we can't ask "why ja. and the others?" [14:40:34] (from the editor perspective, of course this is different) [14:41:03] (the diff narrows what we see and highlights the fact, where in an editor, it's natural to focus on the line we edit, not the 10 lines surrounding) [14:41:11] DanielK_WMDE: About those hooks: Did you see https://gerrit.wikimedia.org/r/34900 ? [14:41:42] Dereckson: ? [14:43:39] Dereckson: should we put full bugzilla urls or just bug xxxxx? [14:44:37] <^demon> Man, I should've started a crossword puzzle or something. [14:44:40] <^demon> Scap is stupid slow. [14:44:49] :( [14:45:06] ^demon: and then the only thing left to do is update wikiversions? [14:45:10] <^demon> Yup. [14:45:13] ok [14:45:19] aude: just bug xxxxx [14:45:19] <^demon> That's a 2second change + sync-file. [14:45:32] Dereckson: ok [14:45:43] Took 55 minutes on wednesday :/ [14:46:01] aude: I will harmonize the comments for all the file when I'll be finished with the unit tests for this file. [14:47:17] Dereckson: nice [14:48:02] aude: by the way, I implemented a nice small IRC feature to get current config value: [14:48:06] .config wgAutoConfirmCount enwiki [14:48:07] 10 [14:48:56] Dereckson: :) [14:49:14] .config wgAutoConfirmCount jawiki [14:49:14] 10 [14:51:03] hoo: i think i looked at it back when john submitted it. it doesn't really solve the problem. i'll submit a new patch soon. [14:52:03] Look at the last comment, please [14:52:24] ( first one to Patch set 5 I meant) [14:53:26] hoo: but EditFilterMergedContent is called on new page creation?... [14:53:50] No, it's not [14:53:54] I even tested it [14:54:15] it's called in an if that's distinguishing new pages and edits [14:54:18] hoo: oh, right, sorry! i missed that originally, but fixed it now "by accident" in I99a19c93 [15:01:04] .config wgServerName enwiki [15:01:27] .config wgAutoConfirmCount enwiki [15:01:28] 10 [15:01:41] .config wgDBpassword enwiki [15:01:42] .config wgSiteName enwiki [15:02:06] It reads InitialiseSettings.php, so password information isn't available [15:02:08] pretty cool idea, though [15:02:12] yay! :) [15:02:51] My security issue were more in the PHP inline execution side. [15:03:09] .config ']; die("Quux"); [15:03:10] Not a valid setting: ']; [15:03:44] But regexpes took care of that. [15:04:07] .config wgSitename enwiki [15:04:37] Ah yes, when it doesn't return a value in this case, it's because it's not defined. en.wiki follows the default setting: [15:04:40] .config wgSitename wikipedia [15:04:48] andre__: only because I was told to ping you again ;) , http://lists.wikimedia.org/pipermail/wikitech-l/2012-November/064728.html is IMHO an important bug and there's been no progress/answer at all [15:04:51] which is also a defaullt ^^ [15:04:53] .config wgSitename default [15:04:54] Wikipedia [15:04:55] Dereckson: or stuff in commonsettings [15:05:21] Nemo_bis, ah thanks [15:05:56] Nemo_bis, you refer to https://bugzilla.wikimedia.org/show_bug.cgi?id=39667 ? [15:06:26] andre__: Well, https://bugzilla.wikimedia.org/show_bug.cgi?id=15434 which that bug is meant to solve [15:06:37] whatever the approach chosen [15:06:44] I see, so we agree :) [15:07:00] reedy@fenari:/home/wikipedia/common$ echo "var_dump( \$wgServer );" | mwscript eval.php enwiki [15:07:01] string(18) "//en.wikipedia.org" [15:07:05] ^ much easie r;) [15:07:43] Nemo_bis, so basically it's about pinging to get an answer on https://bugzilla.wikimedia.org/show_bug.cgi?id=39667#c14 right now? [15:07:43] who's managing the planet.wm? [15:07:52] Alchimista: mutante mostly [15:08:04] Won't be about for another 2-3 hours I think [15:08:25] ok, thanks. any idea when the new planet will se the sun light? [15:08:49] I think it's "soon", but don't quote me on it [15:09:31] i'll try to poke mutante later. pt.planet is not been updated [15:09:53] andre__: that, or the questions I asked in the thread, which elaborate that question :) [15:13:27] Alchimista: I am also interested in Planet stuff.... there are a lot of outstanding requests to be added to the English Planet [15:15:22] sumanah: there isn't any page or list for coordinating planet, isn't it? [15:16:28] there's nothing to coordinate :o [15:16:56] https://meta.wikimedia.org/wiki/Planet_Wikimedia [15:17:24] that's the only page i know :s [15:17:49] So, when you say "coordinating planet" what kinds of activities do you mean? [15:19:33] discussion about layout and translations, who can manage it (add nem blogs, or ask when something goes wrong), new improvments [15:21:18] I think that meta page is probably the best place to start that. Or the talk page. [16:02:33] Ryan_Lane: about the https://bugzilla.wikimedia.org/show_bug.cgi?id=39866 , an *anexo* is basicly a list or a more developed list, without being has complete in text development like an article [16:02:43] (at least on pt.wp) [16:03:57] You really mean Ryan L not K? [16:04:31] dammit, too many ryans :S [16:08:57] I just found an editlag of 5 minutes after block, I know that sometimes a blocked user can do an edit for 1 minute after blocking because of processingtime etc. but 5 mins? seems high for me [16:09:06] http://de.wikipedia.org/wiki/Spezial:Beitr%C3%A4ge/Mr._Mister_and_Ms._Misses [16:11:09] is there any chance to have some userpages with a *protection* like user .css or .js ones? i mean, user subpages that only sysops and the user can edit? [16:12:17] Alchimista: we use wikis . that everybody can edit a page is part of the concept [16:12:51] I hear you Alchimista [16:15:04] DaBPunkt: is related to this bug: https://bugzilla.wikimedia.org/show_bug.cgi?id=41706 [16:16:46] my bot helps with some tasks, so i've been using .css pages so that sysops could set the bot to perform a specific task, in that example, send a warning about a new unblock request. that way, only sysops could *activate* the bot, and the bot could mark the work has done [16:17:43] now that .css is giving that error, the page must be protected, but then my bot can't mark it as done, or in order to do it, it must have sysop flag :S [16:17:51] DaBPunkt, sumanah ^ [16:18:36] (well, it can be done, but people must add the template directly, without the simplicity of a page preload) [16:18:56] just check the history if the last editor was a sysop or the bot. If that's not the case, ignore the content or stop the bot – whatever suits you more [16:19:56] but then, if the last editor was not a sysop it won't run, and if a sysop checked the page and saw that the request was there already, he wont check who marked it [16:20:49] the bot could also revert the page [16:22:07] hum, or i could use the abusefilter.. i start to see new aproaches [16:22:20] * sumanah leaves for a little bit, still on freenode [16:31:18] Alchimista: ask sysop rights. Protect the page. [16:31:45] What you're asking is "I'm a trusted user and I would like..." [16:32:01] If you're a trusted user, stop all the pt. drama and ask sysop rights. [16:32:53] This is how bot owners do in other wikis. CommonsDelinker pages for example on Commons. [16:33:07] The alternative solution is to explain to your bot to read the history and see who did the edit. [16:33:12] Dereckson: sysop-rights for a bot? Only it can edit a controll-page? [16:33:21] DaBPunkt: no, for the human [16:33:45] now that .css is giving that error, the page must be protected, but then my bot can't mark it as done, or in order to do it, it must have sysop flag :S [16:33:49] Oh, it's the bot and not Alchimista who has to edit the pages? [16:33:57] Ok. [16:34:00] Dereckson: i'm sysop. i' tryng abusefilter, it provably would do the same thing withou changing anything else [16:34:18] * Damianz slaps Dereckson with a wet fish [18:53:45] What does it mean when the rev_page for a revision is set to 0? [19:03:25] <^demon> kaldari: They're orphaned :( https://bugzilla.wikimedia.org/show_bug.cgi?id=16674 [19:05:07] interesting [19:05:50] <^demon> Basically, it shouldn't ever happen (but has), and if it is actively happening, that's a bug. [19:09:38] <^demon> kaldari: re: bug 43625 and this question: http://p.defau.lt/?VJSIr6SJSTINDJA1iynmLQ [19:10:53] ^demon: Are you keylogging my computer ;) [19:11:28] <^demon> kaldari: No, I'm just extrapolating and trying to be helpful ;-) [19:12:41] ^demon: I ran that exact query a few seconds ago. Did you run it separately or did you somehow see mine? [19:13:01] <^demon> Nope, I did it myself. [19:13:06] lol [19:33:59] kaldari, is there any easy way to fix these revisions? maybe based on rev_parent_id's page? [19:34:33] I think some of them could be fixed with some detective work, but some can't [19:35:12] perhaps someone could write a maintanance script to fix the ones that can be fixed and delete the rest [19:35:43] <^demon> Deleting would be bad, we don't want to lose data :( [19:35:48] How old are these broken revisions? Is there any chance current code is causing it? [19:36:33] I didn't check, but that would be a good thing to investigate [19:36:53] (I have no sort of DB access, so can't check this sort of thing myself) [19:37:15] I might take a look if I have some time later today [19:37:39] <^demon> Most are pretty old. [19:38:03] Would be good to check if the numbers reported in the bug have grown since the comments were posted [19:38:51] <^demon> http://p.defau.lt/?bYAHMQ326JLDq2Hw3qoStA [19:41:53] They're all 54* revisions...? [20:40:23] ^demon, do those revisions have rev_parent_ids? [20:40:39] <^demon> lemme log back in and check, one moment. [20:41:16] <^demon> Nope, all 0. [20:43:05] <^demon> Granted, rev_parent_id isn't always set. There's 75636 rows with rev_parent_id = 0 on mediawikiwiki. [20:45:42] Yeah, it should be 0 for new page revisions, for example [20:48:08] <^demon> We could try seeing if any of those orphaned revs are *a* rev_parent_id for some rev. [20:48:28] <^demon> As in they're the initial revision, but are missing the rev_page [20:49:07] <^demon> select count(*) from revision where rev_parent_id in ( select rev_id from revision where rev_page = 0 ); [20:49:07] <^demon> 0 :( [22:09:39] if someone tries to get a new password for their account, how long should it take for the e-mail to be sent ? [22:14:08] hi Dragonfly6-7 -- I'd say that if it doesn't happen within 5 min then something is a bit wrong [22:14:43] sumanah - can you come on #wikipedia-en-help , then: [22:18:12] marktraceur: if you're about, I just discovered that upon doing "Upload a new version of this file", we invoke Special:Upload and not Special:UploadWizard. this is OK on commons I guess, but on test2wiki you'll get NACKed [22:18:57] chrismcmahon: Noted. I'm not sure what the procedure would be to fix it--I guess it's another config variable that UW doesn't mention. [22:19:24] marktraceur: is that worth a BZ ticket? or is it expected behavior? [22:20:24] chrismcmahon: Well, it's expected since we haven't configured it to do anything else [22:21:08] chrismcmahon: It may be a bug in core....maybe see if there's a hardcoded reference to Special:Upload and replace it with a lookup of the default upload URI? [22:21:55] marktraceur: Special:Upload takes args, e.g. http://commons.wikimedia.org/w/index.php?title=Special:Upload&wpDestFile=Chris_Dyer_%28Artist%29.jpeg&wpForReUpload=1 and I guess that might be an issue for UW [22:22:27] Hm, yeah [22:22:56] chrismcmahon: Especially since we don't upload new versions, we assume the user is uploading new images [22:23:11] Er files. Whatever. [22:23:13] edge cases ftw [22:23:42] * chrismcmahon guesses the simple thing to do is not deprecate Special:Upload on test2wiki the way it is now.  [22:25:08] but since I have no immediate use for that, I'll just leave it alone. thanks marktraceur [22:26:27] chrismcmahon: you got fenari access meanwhile [22:26:36] thanks mutante [23:45:57] mutante: around? http://pt.planet.wikimedia.org/ isn't beeng updated