[07:46:04] BadDesign: why are you downloading everything? [07:46:30] if it's a snapshot just use a recursive wget maybe? you could also pull them from archive.org [07:52:05] Nemo_bis: Using recursive wget it goes into other directories on the server side and it downloads other unwanted files [07:53:14] Nemo_bis: I want to count the number of page hits for each article on every month of every year there is data available for [07:54:40] BadDesign: you can exclude the unwanted dirs or use -np ... [07:54:53] and that in all languages? [07:56:28] Nemo_bis: I think they're only available for english [07:57:10] no [07:58:10] ah, no, I meant that there is only a single set of files [07:58:14] for all languages [07:59:18] I've tried to write a Python script to download all the files based on the template: url = "http://dumps.wikimedia.org/other/pagecounts-raw/{0}/{0}-{1}/pagecounts-{0}{1}{2}-{3}.gz" where # {0} = year, {1} = month, {2} = day [07:59:46] But the files with that unexpected 1 are not included so I think wget is my only alternative [07:59:52] if I can't figure out the pattern [08:04:12] forgot to mention, {3} will take values from 000000 to 230000 for each day of the month [12:31:21] need to PM a dev ASAP --> might be security issue [12:33:38] Reedy: can you help me with this please? [12:36:56] apergos: ? [12:37:19] I can hear it but I may not be able to do much about it [12:37:21] go ahed [12:43:27] mmm [12:43:37] matanya: best bet would be to file in bugzilla under "Security" (its a private restricted section) then to poke Reedy or apergos with the number if neither respond [12:44:02] Well, we've both now responded ;) [12:44:06] :-D [12:44:06] natually Reedy responds as I type [12:44:19] if it's a mw issue then /me punts to reedy :-D [12:44:44] heh [12:45:01] But we've got ops around too if necessary [12:45:09] matanya: feel free to PM me [12:45:40] apergos: tbh, just revert to 1.6 if thats the case >.> [12:45:54] seconded [12:46:53] :-D [12:47:28] thank you apergos and Reedy. I PMed reedy [12:47:36] cool [12:56:11] somebody changed ext. translate? [12:56:24] English pages now not at /en? [12:59:50] on commons? [12:59:55] or somewhere else? [17:03:09] If i update an i18n message from say [[User talk:$2#$3|talk page]] to [{{fullurl:User talk:$2#$3|redirect=no}} talk page] am i supposed to also update the other i18n messages? I'm thinking yes because the actual text isnt changing, but wanted to double check because last time i updated the other messages there was actually no need [17:03:58] nm found someone nearby to answer :) [17:04:27] :) [17:29:05] I've noticed that at least 2 users appear to have been affected by a block I made (on enwiki) of 127.0.0.1 [17:29:44] uhhh [17:29:51] I'm wondering why some people seem to be using that IP, that can't possibly be on localhost [17:29:58] right [17:30:03] might have to do with XFF [17:30:16] I'm, not sure. csteipp (just because of the XFF keyword) :) [17:30:19] ^^ [17:31:40] Prodego, well.. it could, but depends on the people [17:32:17] Prodego: Yes... that block could be blocking people who foward through their own proxy [17:32:55] We check for blocks on all ip's in their XFF chain [17:33:30] So if they're running their own proxy, or going through a poorly configured proxy, that could happen [17:33:41] csteipp: so what should be the best approach to that? I imagine we don't want them actually editing as 127.0.0.1. [17:33:56] It depends :) [17:34:04] It would be rather confusing :) [17:34:19] MediaWiki will use the IP that is "nearest" the WMF server [17:35:01] is that always true? I thought there were some cases where the prior XFF IP was used [17:35:05] So if they did make an edit, their "IP" shouldn't be 127.0.0.1, unless they are spoofing their XFF [17:35:38] Yep. The exception is if an IP in the list is "Trusted" by us (Ex:TrustedXFF) [17:35:57] Was this a recent change to start checking all XFF IPs, or has it always been that way? [17:36:48] Prodego: Recent change -- https://bugzilla.wikimedia.org/show_bug.cgi?id=23343 [17:38:35] csteipp: So this would apply to those without accounts equally as those who do have it. So unless 127.0.0.1 is completely unblocked, it would prevent anyone running their own proxy from editing while logged out. [17:39:02] Prodego: unblock it then? [17:39:22] Nemo_bis: right, that's what I'm thinking I should do ;) [17:39:24] Prodego: Yes, if there isn't a block on any IP that's "closer" [17:39:55] I would say yes, probably should remove that block [17:40:09] Ok, thanks for the info csteipp, it has been very helpful [17:40:09] Although I'm curious, what was the reason for it? [17:40:37] csteipp: there wasn't one, I was just testing something about blocks on IPs and I forgot to unblock it [17:40:49] since no one uses 127.0.0.1, it didn't matter until that change was made [17:41:29] Ha, makes sense :) [17:43:35] Its come to my attention that some fucktard removed from the DOM. Last time when Trevor Parscal changed it from (against my protest) to broke lots of things for many many many months. [17:44:27] FYI, TimStarling if you can help ^ [17:45:20] legoktm ^ above conversation may be interesting to you [17:49:20] Calling whoever did something a 'fucktard' is probably not the best way to get anyone to help you. [18:28:48] greg-g, just to nitpick slightly: "The contract will commence on a mutually agreeable date no later than July 8th, 2013 for a duration of 12 months. [18:28:49] The contract will end no later than July 8th, 2013 unless it is extended" [18:29:06] It ends on the day it begins? :P [18:31:45] lol [18:32:00] It actually contradicts itself [18:36:11] Krenair: hah, oops [18:41:17] Krenair: fixed [18:56:16] Can a shell user find the stack trace for this please? https://bugzilla.wikimedia.org/show_bug.cgi?id=48683 [19:09:51] Krenair: I can get one but I'll need the hash from the error message [19:10:39] Dispenser: Wait, a *comment* changed and it broke things? Whoever's relying on that must be doing weird stuff then [19:11:24] yes, wikimedians do all the time [19:11:29] Remind again how poorly MediaWiki is documented [19:12:48] So what was this thing and what used it? [19:13:26] Copying page contents without the interface elements [19:14:01] um [19:14:10] use action=render or something? [19:14:58] Or, you know, the ID of a DOM element somewhere [19:14:59] That has caching problems [19:16:41] Right because everyone parses the XML^H^H^H (hopefully well formed) HTML5 tree instead of just doing fast string manipulation [19:17:50] starting scap [19:17:55] Greetings [19:17:58] hey Dispenser... [19:18:17] ...yes [19:18:44] Is the data backing ganglia.wikimedia.org publicly available in a processable format? [19:19:51] I'd like to do some general analysis on the impact on the services of adding/removing a machine [19:21:17] Dispenser: What do you think of https://www.mediawiki.org/wiki/Extension:Disambiguator [19:21:27] Dispenser: I was thinking about turning it on for en.wiki soon [19:21:44] and deprecating the old disambiguation special page/api [19:22:29] Dispenser: Disambiguator has 2 API and should give you more data than the old API did [19:22:52] but it's probably not completely backwards compatible [19:23:01] There was a disambiguation API!? [19:23:12] yeah, but it sucked [19:23:42] Dispenser: if you weren't using it, I guess I should worry about killing it :) [19:23:49] er shouldn't [19:24:58] It really depends how it's implemented and if it can stack against the Dab Challenge https://toolserver.org/~dpl/ch/dab_challenge.php [19:25:32] wow, I have a challenge ;) [19:26:09] Also disambiguation is screwed up on enwiki with "Set index pages" [19:26:26] "set index pages"? [19:27:21] Dispenser: nevermind about the old API, it looks like Special:Disambiguations wasn't a queryable special page after all. [19:27:46] They're like a cross between an article and a disambiguation page [19:27:55] example? [19:28:55] http://enwp.org/WP:SETINDEX [19:29:55] Basically we can't tell if the link is intentional or accidently [19:30:44] What would help is if the editor told users what kind of article they were linking to [19:31:26] Dispenser: set indexes seem to be the same thing as a list. I don't understand the difference. [19:32:17] I guess they're supposed to be in-between a list and a disambiguation page [19:33:14] I think they should just call them 'lists' and not make it so confusing though [19:34:09] Most editors link blindly, that's the real problem [19:34:33] Dispenser: Anyway, let me know if you endorse turning on the Disambiguator extension and if you think it needs any extra features. [19:35:27] since you are one of the main people involved in disambiguation work [19:35:50] It looks like its using the page_props table, so it offers no performance improvements [19:37:01] If it were in the page table, like page_is_redirect, then you could include it in the DOM like mw-redirect [19:40:50] That would make disambiguation highlighting gadgets far faster. Currently its a question of using the slow API or unstable Toolserver. [19:44:46] Dispenser: we have a dablink highlighting gadget at pl.wp, if you're looking for one [19:45:09] Dispenser: it's surprisingly well-written ;) you'd just have to change some constants and it should work everywhere [19:45:12] Thank you, but I've written several myself [19:46:19] Best is Anomie's linkclassifier which highlights things besides disambiguation pages [19:49:15] We use categorylinks table since it's smaller than templatelinks [22:57:16] anyone here? I am getting this error while saving a page [be6e4cdb] 2013-05-21 22:56:38: Fatal exception of type MWException [22:58:22] hello? [22:58:47] i'm sure there are people arround. That error message is much too generic to mean anything to me personally unfortunatly [22:59:06] Url for that message was http://commons.wikimedia.org/w/index.php?title=Commons:Community_portal&action=submit [22:59:18] It doesn't mean anything more to anyone else except shell users who can look into the log files to find the stack trace. [22:59:37] hmm [22:59:53] RoanKattouw, could you do that please? ^ [23:01:35] Looking [23:02:54] Mono: http://pastebin.com/qyP3nfZU [23:03:16] That's the full error message and back trace [23:03:37] lol, parser [23:03:42] yay, parser stuff -.- [23:03:46] Yeah, good luck with that one ^_^ [23:03:53] And with that, I'm outta here. It's 1am, I should get some sleep [23:04:02] heh. good night RoanKattouw :P [23:04:03] this is gon' suck. :P [23:04:23] does anyone know what that means? [23:04:43] Mono: that means bad things happened [23:04:53] Mono: did you try saving the page again? [23:05:06] ebernhardson, such exceptions are investigated by looking up their token (be6e4cdb in this case) in /a/mw-log/exception.log on fluorine [23:05:25] MaxSem: good to know, slowly i'll figure out what all is going on around here :) [23:05:26] Mono: i'm no expert here, but this looks like a failed sanity check :) [23:05:31] MatmaRex: yep [23:05:47] same error, different code and timestamp [23:05:49] two browsers [23:06:03] huh. [23:06:06] ebernhardson, though nobody really understands what's inside the parser anyway;) [23:06:08] let's file a bug, then [23:06:16] would other people have this problem? [23:06:19] MaxSem: yea that seems to be the concencus :) [23:06:19] Mono: can you pastebin your text somewhere? [23:06:22] sure [23:06:51] * MatmaRex goes to file a bug [23:09:23] http://pastebin.com/TydSKcsK [23:09:25] https://bugzilla.wikimedia.org/show_bug.cgi?id=48697 [23:10:20] Mono: yukon: are you the same person? [23:10:22] MatmaRex: thanks [23:10:32] Yep :) [23:10:38] also Mono[Away] [23:10:59] MatmaRex, the cloak is a clue :) [23:11:26] Just copied your text into the edit box and looked at the diff, seems you're adding lots of translate tags [23:12:03] Yep [23:12:07] yukon: hmm, i attempted to sve that [23:12:17] MatmaRex: did it work? [23:12:19] yukon: and got an error mentionin unmatched tags [23:12:32] or , in fact [23:12:39] let me look through it [23:12:58] It shouldn't be a permission/spam filter problem, I'm an admin and translation admin on Commons [23:14:12] I found an error in my code, let me try again [23:14:13] Mono: ;Over 15 million files uploaded [23:14:17] this line looks funky [23:14:32] * mono fixes it [23:15:17] that's the error I found [23:17:19] it saved [23:20:34] mono: https://bugzilla.wikimedia.org/show_bug.cgi?id=48697#c2