[11:31:46] !change 89184 | Nemo_bis [11:31:47] Nemo_bis: https://gerrit.wikimedia.org/r/#q,89184,n,z [11:40:36] mutante: thanks but I was still collecting them [11:40:51] presumably some more will pop up and I'm still waiting for better labels [11:41:49] Nemo_bis: another patch later doesnt hurt, or did you mean -1 ? [11:42:22] can always be improved later :) [11:42:58] then in it goes .. [11:43:12] twss [11:49:03] twhs [12:35:20] apergos: About? [12:35:26] yes [12:35:30] what's up? [12:36:23] Slightly offtopic question.. [12:36:50] not going to answer boxers or briefs. :-P [12:37:01] It's more ontop than that ;D [12:37:49] If I made a copy of a file, and then transferred it (in this case, over 2GB sqlite file) to a remote server, and then wanted to sync across any and all changes since the copy... [12:37:59] Could rsync do this? Or would it need to transfer it all again? [12:39:16] rsync does some clever delta stuff I believe [12:39:45] how that works on one file, I don't know, time to rtfm [12:39:55] I was discussing this with someone else [12:40:16] --whole-file will turn off delta, otherwise it's enabled by default [12:40:24] 8.71 Mbps upload on my wifi connection. 36 Minutes and 12 Seconds [12:40:24] 0.78 Mbps on my ADSL connection - 6 Hours, 44 Minutes and 20 Seconds to upload 2.1GB [12:41:30] so, by default it should not transfer the whole thing? [12:41:46] right, unless you have small changes scattered finely throughout the file [12:42:10] I worked out that in the time transferring said file over my ADSL connection... I could make 4 or 5 return trips to the datacentre where the machine is [12:42:29] ouch [12:43:05] I guess I could test this locally (or at least, "locally" over my home network) [12:43:05] 2gb 6 hours is ridiculous [12:43:25] yes, might as well start there [12:43:33] keep copies of everything and start timing [12:44:21] if there are going to be lots of changes scattered through the file, maybe you want to transfer the whole thing [12:44:28] in which case -W might be the thing for you [12:44:40] worth timing that way too [12:45:22] The question then is how the sqlite file format works [12:45:41] Presumably it would be some sort of ordered additions to the file, plus some at the index location [12:46:09] no real idea there, it's a binary format but dunno what parts get shifted around [12:46:15] you could [12:46:29] once you have the new and the old, do some judicious cmps or something to see [12:58:10] Diff complains binary files differ [13:11:29] of course diff will do that [13:11:42] cmp will stop at the first byte, there is probably a clever way to [13:18:54] http://ganglia.wikimedia.org/latest/?r=hour&cs=&ce=&s=by+name&c=PDF%2520servers%2520pmtpa&tab=m&vn= problems with fdp? [13:19:00] s1 colour red o_O [13:19:10] thinking if you are clever rdiff might tell you what you want to know [13:21:36] You could do sillier things like record and repeat the transaction log, couldn't you? [13:21:59] If you know which queries are being used for UPDATE, etc., you could just do them in multiple places. [13:22:04] Poor man's replication. [13:24:19] :/ [13:25:05] You could also switch backends and use built-in replication. [13:25:13] Switch backends? [13:25:17] I'm not talking about mediawiki [13:25:23] And I'm not doing replication in that sense [13:25:40] Aren't you talking about keeping two SQLite files in synced across hosts? [13:25:47] in sync * [13:25:49] As a one off [13:25:57] Ah. [13:25:59] Uploading 2GB from home takes an age [13:26:12] Sure, but only one age. [13:26:17] I thought this was a continuing need. [13:26:22] No [13:26:27] * Elsie shrugs. Then wait? :-) [13:26:31] But by the time said file is uploaded, said file is out of date on the remote host [13:26:34] Or sneakernet! [13:27:08] Out of date, but you don't want them synchronized continuously. [13:27:23] You can always lock the DB. [13:27:28] Elsie: You're not being helpful. [13:27:30] but rsync should be smart enough for the diff once you uploaded it once [13:28:04] If the goal is to keep the two DBs in sync over a period of time, replicating the queries is surely the lowest overhead. [13:28:28] For a one-off sync, you may just have to be patient. Or get better Internet at home. [13:28:43] I wish the latter was true [13:28:53] true/doable [13:28:56] he already calculated sneaker net would be 4 times faster or so :) [13:29:10] It often is. [13:29:14] Reedy: Move! [13:29:32] Or just do everything on two servers. [13:29:43] Home connections are shitty everywhere. [13:29:46] Only if it's logged [13:29:49] That's relative [13:30:06] Incestuous. [13:30:12] "A 2M-10Mb/s burstable internet service starts from as little as £650/month (+VAT) and is typically around £1000/month (+VAT)" [13:30:33] That's about my only other choice for connections [13:30:40] upto 50 times more expensive [13:30:44] all .pdfs broken on commons, we need to purge every .pdf... :/ is the cache down ? [13:30:47] I'm always surprised when I read that someone has worse Internet options than the U.S. [13:30:56] *thumbnails [13:31:03] Don't the Asian markets have like five times the speed for a quarter of the cost? [13:31:32] South Korea has something crazy [13:31:52] They are a democracy. Just move there. Problem solved. [13:32:14] Looks like it took rsync about a minute to update the other local copy [13:32:54] http://fastnetnews.com/dslprime/42-d/4130-germanys-1-provider-government-owned-netcologne "Their current lead offer is 35 euro - about $50 - for 50 megabit down and voice calls." in 2011 [13:33:35] Apparently the UK recently jumped a lot in the league tables [13:33:38] I don't believe it [13:33:45] BT won't do anything without handouts [13:45:28] hi Reedy you might know this off the top of your head and save me some investigation: we recently moved some functions from the Vector skin into core, among them the expand/collapse thing for print/export. is it possible that we have both the old Vector skin functions *and* the new core expand/collapse functions running simultaneously on beta labs? I'm seeing some funny behavior on beta that might be explained by that. [14:03:21] i've had 100 Mbps for ~ 40 eur a month for years now [14:03:27] it's actually getting a bit slow ;) [14:04:41] they do 500 Mbps elsewhere now [14:08:15] * paravoid punches mark [16:25:12] @seenrx [Jj][Aa]n[-_][Dd]ud[íi]k [16:25:13] Base: I have never seen [Jj][Aa]n[-_][Dd]ud[íi]k [16:41:00] Reedy [~quassel@2001:41c8:1:63d3::2c] увійшов до кімнати. [16:41:00] muhahahaha [16:58:35] ^^ evil [16:59:07] Base: what is pdpc ? [17:30:05] Steinsplitter: hmm? [17:30:09] where? [17:31:06] https://www.google.com/search?q=pdpc+&ie=utf-8&oe=utf-8&aq=t&rls=org.mozilla:ru:official&client=firefox-beta ? [18:59:12] mutante: it works ;) http://blog.gmane.org/gmane.org.wikimedia.wikimania.committee it's so easy to get a list added on time for the first message without nasty archive imports :) [19:05:51] hello, I got some weird errors while saving a large wikipage on nl-wiki [19:06:00] https://nl.wikipedia.org/w/index.php?title=Lijst_van_rijksmonumenten_in_Den_Haag&diff=39159006&oldid=39158982 [19:06:03] https://nl.wikipedia.org/w/index.php?title=Lijst_van_rijksmonumenten_in_Den_Haag&diff=39158982&oldid=39158820 [19:06:14] those 2 edits gave me a "bad getaway error" [19:06:19] I thought they failed [19:06:30] Looks like they worked [19:06:32] but the edits saved [19:06:34] That page is pretty big [19:06:40] yes but it gave me errors [19:06:51] I know, I'm trying to reduce the size now [19:06:52] basvb: {{worksforme}} [19:07:33] It was above the max transclusion size [19:07:36] now it's just below [19:07:48] It's bigger than Obama on enwiki! [19:08:22] yes, we've even bigger lists like this on nl-wiki [19:08:34] that's what you get with cities with 10.000 monuments :p [19:08:52] ok I'm trying another edit now [19:10:29] "504 Gateway Time-out" [19:10:49] could any of you try to do an edit on that page? (just add "test" to the bottom or something) [19:10:58] "nginx/1.1.19" [19:11:13] basvb: it's normal to get 504 errors when editing [19:11:27] hmm, doesn't seem ok to me Nemo_bis :P ? [19:11:29] (yes, people say it isn't for them but I've always had such errors, routinely) [19:11:49] indeed, never had this before [19:11:50] you can't stress poor mediawiki too much [19:12:01] ok, but the size of the article is a factor in it? [19:12:30] usually it's more how many templates with weird recursions and parserfunctions there are [19:13:12] e.g. here some 30-40 kB would be enough https://meta.wikimedia.org/wiki/User:Cbrown1023/Logos?action=history [19:13:34] ok, this has a lot of templates [19:13:46] but I've been working on the lists (also the big ones) with this template a lot [19:13:53] and never have had this error [19:15:03] basvb: aren't you hiding us something? ;) https://nl.wikipedia.org/w/index.php?namespace=10&tagfilter=&target=Lijst+van+rijksmonumenten+in+Den+Haag&title=Speciaal%3ARecenteWijzigingenGelinkt [19:15:38] do you think that has something to do with it [19:15:39] ah you only added to a comment, but still [19:15:47] my edit is a non included comment [19:15:50] dunno, it's possible [19:15:54] the other one is the link again [19:16:04] but we've had that link before like that [19:16:46] 600 B multiplied by hundreds calls are quite a lot, so it's at least possible they might impact something [19:17:56] but it's noinclude ? [19:18:52] the wiki isn't made for these big lists, but so many people use this information [19:19:33] everybody who want to know something about dutch monuments uses wikipedia, even the people from the monuments organisation :P [19:20:16] there is a similar bug about timeouts on a DR log page on Commons [19:21:05] I hope it will restore itself then, we've got a editathon around this list tomorrow [19:21:34] don't hope [19:21:59] :p [19:22:02] I don't expect [19:23:20] or do you have any good advise Nemo_bis ? [19:25:50] I'm now testing on a 3% smaller similar list [19:27:44] that gives me the more regular error [19:27:47] "Our servers are currently experiencing a technical problem. This is probably temporary and should be fixed soon. Please try again in a few minutes." [19:28:06] also saved [19:28:09] weird [19:32:33] basvb: known issue when editing pages that large [19:33:10] weird that it didn't happen before (i've been editing these lists around max transclusion size quite a lot) [19:33:59] you could consider https://bugzilla.wikimedia.org/show_bug.cgi?id=49734 the relevant bug [19:36:32] ok I reverted the template edits from yesterday [19:36:35] but that didn't help [19:37:30] they say it's frequent now [19:39:49] If only this graph lasted a bit longer https://gdash.wikimedia.org/dashboards/reqerror/ [19:39:50] ok I will report that I have the problem as well [19:43:08] 1000's of people get these errors? [19:45:12] nah [20:00:04] thx [20:32:32] evening... [20:32:53] some problem with pt.wiki user rights: https://pt.wikipedia.org/wiki/Especial:Privil%C3%A9gios [20:33:18] a few rights disappeared [20:34:34] Teles: hmm, didn't you guys just have flaggedrevs enabled? [20:35:14] yeah... I've just mentioned that here https://bugzilla.wikimedia.org/show_bug.cgi?id=54828 [20:35:15] yup [20:35:26] but I don't know if they are related [20:36:15] # Remove 'editor', 'reviewer' and 'autoreview' groups [20:36:17] unset( $wgGroupPermissions['editor'], $wgGroupPermissions['reviewer'], $wgGroupPermissions['autoreview'] ); [20:36:27] doesn't pt.wp use 'editor' for something already? [20:36:44] Teles: which rights disappeared? that one? ^ [20:37:02] Teles: (that page you linked is only accessible to sysops and above, btw) [20:37:13] Reedy: ^ [20:37:24] danilo.mac provided that config [20:37:26] https://bugzilla.wikimedia.org/show_bug.cgi?id=54828#c5 [20:38:10] I was a bit confused by the array_diff [20:38:30] can't remember if we use it elsewhere [20:39:01] I'm sure we use array_diff somewhere [20:39:04] MatmaRex: sysop,deleter, bureaucrat, interface editor are the ones I couldn't see anymore [20:40:50] hm, it seems like some wikis use 'editor' for various things, but not pt [20:41:38] ah [20:41:39] $wgAddGroups['bureaucrat'] = array_diff( $wgAddGroups['sysop'], array( 'reviewer' ) ); [20:41:46] well, that explains everything. :D [20:42:41] o_0 [20:43:15] Reedy: s/sysop/bureaucrat/ in this line and the next one, will you? :D [20:43:54] Hi, I made a mistake in the configuration of FlaggedRevs in ptwikis, can sameone correct it to me? [20:44:52] * Reedy mashes his keyboard [20:45:17] or is better I create a bug? [20:45:24] danilo_: Reedy's on it [20:45:37] assuming you mean the "$wgAddGroups['bureaucrat'] = array_diff( $wgAddGroups['sysop'], array( 'reviewer' ) );" part ;) [20:45:49] already should be fixed [20:47:27] yes, in copy paste I forgot to change to 'bureaucrat', thank you [20:49:59] Fixed? [20:50:47] Teles: Fixed? [20:50:58] Reedy, back to normal :) [20:51:21] tks, guys [20:52:35] is it difficult to make those rights be set in alphabethical order? [20:52:57] no.. [20:53:21] Hope I am not asking too much :) [20:53:33] danilo_: does "correct it to me" mean "for me"/to my benefit? we'd say "correggermelo" in Italian, do you have the same in pt? [20:53:37] Are they sorted normally? [20:54:17] yes. Just like before [20:55:12] sob sob sob https://en.wiktionary.org/wiki/mihi [20:56:30] Nemo_bis: my english is bad, sorry... in portuguese "to" and "for" is the same word: "para" [20:57:02] danilo_: don't be sorry, I like to learn other languages' peculiarities via English ;) [20:58:05] What I'm wondering is if you have ethic dative too https://en.wikipedia.org/wiki/Dative#Latin [20:59:41] Nemo_bis: "can sameone correct for me" in portuguese is "algu�m poderia corrigir para mim" [21:02:15] is that just "in my stead"? [21:02:51] anyway not ethic dative, too bad ;) thanks [21:03:51] Nemo_bis: yes, just like "in my stead" [21:48:54] why is the option in the preferences gone to set the sidebar uncollapsable? [21:49:06] and how to get that back? [21:50:38] Romaine: it's gone because of https://bugzilla.wikimedia.org/show_bug.cgi?id=45051 and https://bugzilla.wikimedia.org/show_bug.cgi?id=46512 and the patches there [21:50:53] Romaine: and i'm afraid you can't, just uncollapse them? the state is remembered across page loads [21:50:57] afk [21:51:59] if you visit a lot of wiki's this is so much annoying and no useful at all [21:52:47] also I get people on my back as tech ambassador as again the technical people have removed a usefull part of vector and get annoyed again [21:54:37] but my queston remains, how to disable that collapsing sidebar as a lot of users has set in their preferecnes? [21:55:38] I assume this can be done in vector.css [22:00:28] MatmaRex: any idea how? [22:04:02] my solution to everything is always "use monoboook" [22:04:57] :S [22:05:52] that is not a sollution [22:06:58] :( [22:08:16] The navigation was so clean with that option enabled… :( [22:08:41] what code should a user add to his vector.css to restore the previous situation? [22:10:32] sorry, i don't have this off-hand, and right now i'm kinda busy [22:11:00] technically speaking, jdlrobson was the perpetrator of that change, i can only be accused of instigation :P [22:11:24] i heard my name [22:13:24] jdlrobson: why is the option in the preferences gone to set the sidebar uncollapsable? and how to get that back? [22:13:46] (i've handled the first question and left the second one for you :P ) [22:15:02] Romaine: the preference was removed as the usage of it was so low ( 0.008% on en.wikipedia) [22:15:17] This is the kind of thing that would make a nice gadget [22:15:30] sure, but then I need a code for that [22:15:46] I'm just reminding myself of it.. [22:15:50] i've been away from that code for a while [22:17:42] Romaine: seems as simple as $( '.collapsible-nav .body' ).show(); ? [22:18:02] i assume you only care about the default? [22:18:28] I have little knowledge about that [22:19:25] but setting up a gadget is simple [22:26:16] MatmaRex / Nemo_bis / jdlrobson / anyone: can you point me otherwise to the person who can help me with it? [22:28:22] and I would suggest when removing a feature should be kept in a page where also is mentioned how users can set this for themselves [22:45:19] anyone? [22:48:11] Hi Romaine. [22:48:13] Romaine: i'm sure nl.wp has some talented scripts writers who could emulate that :/ i'm currently preoccupied with something else [22:48:15] You should learn how to post to the wiki. [22:48:28] Not everybody wants to answer your questions immediately. [22:48:36] (and tbh, this is a really insignificant preference and insignificant change.) [22:49:04] we have unfortunately too little talented script writers, often requests are waiting weeks if not months to get performed [22:49:22] that final stage of vector cleanup was done too quickly, but oh well, shit happens, sorry about that [22:49:30] it's not like we took down the sites [22:49:45] Elsie: what should I post to what wiki? [22:51:07] Romaine: https://meta.wikimedia.org/wiki/Tech [22:51:11] There's a link in the channel topic. [22:51:13] Explain your problem. [22:51:16] And someone will respond. [22:51:47] if I see the reactions of people on the wiki, users who daily have to deal with unrequested changes, they get to often in the situation that something is changed without able to deal with it [22:52:11] this causes a bad reputation for developers [22:52:21] The developers don't care. [22:52:32] yes, that is a problem [22:52:39] Only for the non-developers. ;-) [22:52:54] no for the whole Wikimedia movement [22:53:12] the more end users get annoyed, the less the feel being welcome to contribute [22:53:13] That user preference was stupid. [22:53:32] Wikimedia/MediaWiki does a lot of stupid shit, but removing that user preference doesn't make the list. [22:53:32] * MatmaRex wonders if he's being considered a developer [22:53:39] It never should've been implemented. [22:53:44] It was preferences clutter. [22:53:46] And nobody used it. [22:53:54] Elsie: don't you know, once you implement something you've to support that forever. [22:53:58] in the eyes of regular users everyone in this channel is a developer :p [22:54:24] Even wm-bot. [22:55:11] (nobody told the communicaty how to name all the tech people) [22:58:36] How might the tech people name all the community people? [22:58:48] users [22:59:04] Elsie: there are some channels to do it [22:59:29] Elsie: yes, I can find a page where you can ask that question [22:59:47] https://meta.wikimedia.org/wiki/Global_message_delivery for example [22:59:47] I am sure there is on meta a page for that [23:00:03] but I am not here to discuss werther it is a good or bad idea, but to hear what code users can use to restore this setting