[00:27:54] 220200960 bytes exhausted (tried to allocate 189006053 bytes) [00:28:16] Wants nearly 86% more memory!? [05:35:12] [[Tech]]; Auntof6; /* Where/how to report issues with new release of software? */ new section; https://meta.wikimedia.org/w/index.php?diff=6336277&oldid=6333521&rcid=4660840 [08:11:53] How does a Wikipedia page resize itself as to fit to the size (widht, heigth) of a webview on a mobile device ? Using what mechanism? [08:43:30] CSS, JS, HTML viewport? [08:49:12] OMG, search engines were invented https://www.google.com/cse?cx=010768530259486146519%3Atwowe4zclqy&q=responsive+mobilefrontend#gsc.tab=0&gsc.q=responsive%20mobilefrontend&gsc.page=1 [08:49:24] proposal for the future: https://www.mediawiki.org/wiki/Requests_for_comment/Grid_system [09:01:20] Nemo_bis: sorry, I'm totally ignorant of frontend development [09:03:09] well, not totally, but pretty noobish [09:03:16] in the mobile space [09:05:00] * Nemo_bis even more [09:24:21] anyone else having issues with enWP search engine? [09:24:23] An error has occurred while searching: The search backend returned an error: [09:24:39] 3 searches in the past minute [09:25:10] 5 searches in 2 minutes [09:26:03] nobody uses enwp) [09:26:58] yep aha [09:28:19] how strange, "An error has occurred while searching: The search backend returned an error:" returned a highlight from my script. My regex, she is sad. [09:31:33] oh well, seems like a bugzilla is required [09:34:53] For anyone who can see or resolve the issue, I have entered into bugzilla ... Bug 56936 - [enwiki] search engine failing [14:28:00] [[Tech]]; Ruslik0; /* Where/how to report issues with new release of software? */; https://meta.wikimedia.org/w/index.php?diff=6340561&oldid=6336277&rcid=4661318 [14:33:49] [[Tech]]; Ruslik0; /* Where/how to report issues with new release of software? */; https://meta.wikimedia.org/w/index.php?diff=6340630&oldid=6340561&rcid=4661320 [15:36:55] Hi, I got this message "Our servers are currently experiencing a technical problem. This is probably temporary and should be fixed soon. Please try again in a few minutes." [15:37:23] for some time. After saving the page @ml.wikipedia. [15:37:54] While checking recent changes it can be found that edit is saved [16:18:02] Praveen is right, we're still having a few order of magnitudes more 5xx errors than usual https://gerrit.wikimedia.org/r/94918 [16:18:06] * https://gdash.wikimedia.org/dashboards/reqerror/ [16:25:49] Nemo_bis: you will want to repeat that in wikimedia-operations in a few minutes [16:39:03] apergos: yeah, I checked the channel but I saw it busy with something :) [16:39:08] yeah [18:29:14] greg-g: Poke - I'll likely need a deploy window for VectorBeta on Thursday (CC jdlrobson) [18:41:27] marktraceur: can you send an email, I don't want to forget (and can't do it now) [18:41:39] 'kay [18:45:46] Enwiki's API seems incredibly slow. [18:46:18] I've been timing it, and I'm at 2:37 seconds, to push an edit and receive a response. [18:49:50] HELLOOOOOOO? [18:50:01] Guess no ones home. [18:50:52] ... [18:50:58] That's totally not how IRC works [18:52:29] he didn't even ask any questions [19:02:32] WE'RE HERE! [19:02:49] greg-g, can you fix the API? [19:03:08] Cyberpower678: can you tell me more specifics about the issue? [19:05:18] greg-g, http://pastebin.com/6sr9cLSD [19:05:36] The API seems to be causing trouble since the PHP update. [19:05:51] Bad serialization, I'm guessing. [19:06:23] Seems to be happening through POST requests. [19:07:42] Cyberpower678: will you commit to not running away from this channel? [19:08:04] No one responded so I left. [19:08:12] !quiet [19:08:23] what, no bangcode for that? we used to have one. [19:08:28] Cyberpower678: Coren and I have both told you to log the raw data you're getting back from the API. Without that, we can't help you. So stop forum shopping and just do it. [19:08:28] As long as I know my comments are being responded to, I stay. [19:08:31] people don't sit on irc full-time. [19:08:35] ... in 5 minutes? we're all a bit busy, watching IRC and immediately answering isn't always possible [19:08:54] and, what anomie said, I suppose :) [19:09:28] yeah, that pastebin log is approximately useless [19:09:43] doesn't say what you're POSTing [19:09:50] nor what the response was [19:09:57] nor how long it waited [19:09:59] * YuviPanda hopes we can kill all non JSON formats at some point in the future [19:10:20] * Cyberpower678 says no to YuviPanda [19:10:32] I wasn't asking, Cyberpower678 [19:10:35] just hoping [19:10:42] anomie, you didn't tell me anything. [19:11:09] I think he just did "provide raw data you're getting back from the API" [19:11:18] Cyberpower678: http://lists.wikimedia.org/pipermail/labs-l/2013-October/001772.html, and http://lists.wikimedia.org/pipermail/labs-l/2013-October/001776.html [19:12:50] alright, given anomie's links, I'll just wait until you reply on list, Cyberpower678 [19:13:14] replying on list also helps others who might potential hit the same issue [19:13:25] auto-documentation ftw [19:13:31] anomie, I don't remember that email. [19:13:40] Oh well. [19:13:55] anomie, whoops. [19:14:00] anomie, sorry. [19:14:34] * YuviPanda proposes we never upgrade PHP ever on the server [19:14:38] I'm installing a logger into Peachy to store it's activity and log errors in a seperate file. [19:15:02] * Cyberpower678 sides with YuviPanda and proposes we downgrade PHP to it's previous version. [19:15:44] yeah, maintaining the status quo is the most important part of being a programmer [19:15:50] * YuviPanda goes to do something more useful [19:15:58] evening devs, are you guys aware of an issue where the user and user talk pages for usernames ending in .exe (and possibly other file extensions) is causing some strange behaviour for some Chinese users ? [19:16:54] anomie, I'll let you and Coren know when Peachy is generating log data and error data regarding this issue. [19:17:51] anomie, oh and to clear it up, I wasn't attempting to forum shop. Sorry. [19:18:40] Well I have to go. See ya. [19:18:48] Got a Wikibreak to finish. :p [19:19:03] csteipp: evening devs, are you guys aware of an issue where the user and user talk pages for usernames ending in .exe (and possibly other file extensions) is causing some strange behaviour for some Chinese users ? [19:19:22] thanks Cyberpower678 [19:19:38] NotASpy: define "strange behavior" [19:19:58] got an admin who is being directed to http://60.13.74.166:83/1Q2W3E4R5T6Y7U8I9O0P1Z2X3C4V5B/en.wikipedia.org/wiki/User_talk:HTMLCODER.exe when they try to edit en.wikipedia.org/wiki/User_talk:HTMLCODER.exe [19:21:04] uhhhh [19:21:28] that's not our IP [19:21:32] NotASpy: we redirect urls ending in something that looks like a file extension to &*, to prevent IE6 from setting the mime type based off of the "extension"... [19:21:49] there's something proxying here [19:21:58] * greg-g wonders what, snarkily [19:22:02] But, this looks like something entirely different. [19:22:48] Probably either some kind of anti malware extension/ proxy/ whatever going crazy or malware [19:23:24] greg-g: from what the user has said, it looks like it's a Chinese firewall issue, which is why I raised it. [19:23:28] yeah [19:23:36] well, we can't patch the firewalll [19:23:38] -l [19:23:39] Yeah, I agree. Sounds like a "security" feature doing that in the browser or proxy [19:24:21] that ip is owned by the hainan province, so, the question is, why are they getting in the middle there :) [19:25:07] we've tested it with en.wikipedia.org/wiki/User:TESTPAGE.exe [19:25:18] someone probably wants to filter .exes before delivering them to users, so they are proxied... but that doesn't seem to be implemented terribly well [19:25:18] from a non-China computer? [19:25:27] hoo: oh man, probably [19:25:29] wow [19:25:36] if that's it.... :) [19:26:15] load them, check the header against $whatever and then deliver them to the user [19:27:06] greg-g: we checked en.wikipedia.org/wiki/User:TESTPAGE.exe with the user and it's not causing any difficulties, no proxying or any trouble editing. [19:27:31] yeah, I just did https://zh.wikipedia.org/wiki/User:TESTPAGE.exe too [19:27:39] (that's zhwiki) [19:28:07] so, I'm leaning towards hoo's unsubstantiated theory :) [19:28:14] but if it's working on specific regex, then that makes sense. [19:28:20] NotASpy: Almost impossible to figure out what they are using to filter things... maybe a stupid regex [19:28:28] * greg-g sighs [19:28:52] If you got access to the network yourself, you can figure that out rather easily (with a lot of testing) [19:28:58] time eating, though [19:28:59] A) chinese user can edit en.wiki Something.exe fine [19:29:09] B) chinese user can't edit zh.wiki Something.exe [19:29:19] C) us user can do en.wiki something.exe [19:29:27] D) us user can do zh.wiki something.exe [19:30:06] NotASpy: You can have someone in country try out https://en.wikipedia.org/w/index.php?title=User_talk:HTMLCODER.exe&action=edit, and see if they're just filtering .exe on the end, or if .exe anywhere in the url is enough to trigger it [19:30:35] But yeah, not a lot we can do if their request doesn't get to our servers, unfortunately. [19:30:42] Oh, http that link ^ [19:30:46] heh [19:31:15] Beacuse, who really wants integrity, when the mitm is protecting you.... [19:31:16] https would be solution, but that's only working in very few networks in China [19:31:30] csteipp: :) [19:31:36] mostly probably because those are badly configured [19:32:23] I've asked the user, if they're able to spend a few minutes testing, I'll let you know. [19:34:47] thanks NotASpy [22:27:31] Hi [22:28:23] ? [22:28:47] Does someone knows if the administrator of a mailing list is technically able to delete a thread/mail? [22:29:01] which list? [22:29:09] i.e. is it @lists.wikimedia.org ? [22:29:34] In general [22:29:50] 12 22:29:09 < jeremyb> i.e. is it @lists.wikimedia.org ? [22:29:50] That kind of list, of course [22:30:18] no, there is no deleting, not even for root :P [22:30:25] mutante++ [22:30:50] Okay, thanks :) [22:31:11] so, first of all it needs root. second it risks breaking links to the archives. third, you're pretty much guaranteed that the info you're deleting will stay out there forever [22:31:45] (on other archives besides the lists.wikimedia.org copies) [22:31:59] LlamaAl: you should assume undeletable unless you're willing to give an example of something you want to delete [22:32:05] and even then it's unlikely [22:34:57] jeremyb: Thanks for your help, I got it