[09:49:21] hello, maybe someone has some seconds to help me. I need to remove the \n (newlines) inside tags in a wiki page with python. Is there any way to do this beyond regex? is it possible to do it with regex? (I haven't succeeded yet) [09:52:00] Arnaugir: use beautifulSoup lib [09:52:40] gry: I've just read about it, will try. [10:38:51] gry: do you klnow if with beautifulSoup I could remove newlines inside wiki templates? ie from {{cite web\n |title=TITLE\n \url=URL}} . Or it does only work for like HTML tags? [10:39:03] no I don't know but #python could [10:39:15] I'm trying but not succeeding :P [10:42:47] ok just discovered https://www.mediawiki.org/wiki/Alternative_parsers [15:49:04] Cyberpower678: ping [15:49:28] hedonil, I'm watching the Hunger Games. Better be important. :p [15:49:47] Cyberpower678: you know, you have multiple resource issues [15:50:01] Cyberpower678: https://tools.wmflabs.org/wikiviewstats/screenshots/Screenshot_xtools_1.png [15:50:09] Cyberpower678: https://tools.wmflabs.org/wikiviewstats/screenshots/Screenshot_xtools_2.png [15:50:24] Cyberpower678: takes ages to load [15:50:44] hedonil, the tools are pending a redesign. [15:51:02] Those things calling from tool server will be resolved. [15:51:17] Cyberpower678: I know, but some users are suffering from low performance [15:51:53] Cyberpower678: for labs-logo, you may want to try this from commons https://upload.wikimedia.org/wikipedia/commons/b/bf/Powered-by-tool-labs.png [15:52:31] hedonil, okay, I'll look for and disable those bad resources. [15:52:43] Sometime during this week however. [15:52:57] I'm enjoying my Christmas presents now. [15:53:01] :p [15:53:37] And a slice of Pumpkin Cheesecake. [15:54:09] And a crescent rolls, Cinnamin Buns, scrambled eggs and Cider Bacon. :p [15:54:20] And raw milk [15:54:41] * Cyberpower678 eats too much [15:55:20] Cyberpower678: especially this 'vcss-blue' resource from screenshot no.2 - ages... [15:56:03] Cyberpower678: ... and your current eating habit: you're getting fat :P [15:56:23] Actually, I'm losing weight. [15:56:27] :D [15:56:34] Which is quite funny. [15:56:49] * Cyberpower678 weighs 190 lbs. at 5'11" [15:57:48] * Cyberpower678 is now eating chocolate chip cookies. [16:00:04] Cyberpower678: but there are some advatages http://www.flickr.com/photos/110698835@N04/11227399814/ :P [16:01:41] Cyberpower678: 190 lbs 5'11 ... [16:02:05] !log wikimania-support Updated scholarships-alpha to f61cb21 [16:02:06] * hedonil is a metric guy [16:02:06] Logged the message, Master [16:03:54] hedonil, that's normal weight for a guy [17:05:25] hey guys, i habe a newbie problem. after i hav uploaded files via WinSCP to my data/project-dir i have to change ownership of the files to my project. however, sometimes using "take ./" doesn't work as i get "You need to share your group with the file.". What am I missing, how can I make my tool become owner of the file? Maybe a rights problem? [17:24:18] hi Yellowcard :) [17:24:32] oh, hi Giftpflanze :) [17:25:25] is there a group mismatch? what is the groups of your files? [17:25:34] *are [17:29:26] how can i find out the group? the owner is 'yellowcard'. However, after deleting the whole dir and re-uploading it, there was no problem using the take command. But the problem tends to show up all day once in a while, but i cannot really reproduce it. [17:30:45] hm [17:31:02] the group shows up with "ls -l" [17:32:28] i think it also depends on whether you are your user or the tool user [17:34:43] on WinSCP I'm always my user (yellowcard), I cannot become my tool user with that program. I use git bash for the become and take commands. [17:35:34] oh yeah, i was talking about the shell stuff [17:36:01] if you run become you become the tool user obviously [17:36:30] exactly, on shell i'm the tool user and run take to make the files owned by the tool user as well [17:37:48] alright, currently the problem shows up as well. the file is in the wikidev group. [17:38:35] mh, i see [17:40:39] do you directly copy your files to the tool directory? [17:42:47] yes. [17:43:15] maybe first copy it to your home directory, then use 'chgrp local- ' 'mv ~local-' 'take ' [17:44:08] respectively 'chown local-:local- ' and then the 'mv' [17:44:14] then you don't need to take it [17:44:57] ok the latter doesn't work [17:46:38] mhm, seems pretty complicated to me. usually, directly copying and then using take works. but when i upload single files into existing directories, the error seems to come up [17:47:28] what is this where you don't 'upload single files'? [17:48:03] just use unix ;) [17:48:13] then you don't have any problems ;) [17:49:35] can you use scp with 'git bash'? that should also work [17:52:31] haha yes i'm a windows user but i see i should setup a unix system too ;) i upload my files with winscp to my tool dir (logged in as yellowcard). then i use git bash with become PROJECT take ./ that my project becomes owner of the files [17:52:51] when i upload directories it works properly, when uploading single files into those directories it doesnt [17:52:57] ah [17:53:53] maybe that could help: https://wikitech.wikimedia.org/wiki/Help:Putty#Copying_files_from_your_computer_to_Labs_via_the_command_line [17:55:31] this is almost one year old but proved to work ;) [17:55:48] i can't believe that i'm here this long [17:57:28] 1 year and 5 month plus cookie crumbs [17:57:32] *months [17:59:30] i do not want to upload via command line if possible but use winscp as gui to upload files. windows user :D i know and have configured winscp that way. i do not have putty running but git bash, might that be the problem? [17:59:53] sorry for these noob questions, i'm really not familiar with all that stuff. :( [17:59:59] no, i think winscp is the problem here [18:00:15] and i'm no longer familiar with windows :) [18:01:14] hi Yellowcard , I am also a Windows user [18:01:20] maybe I can help you? [18:01:24] oh, hi benestar :D [18:01:33] hi giftpflanze . Whats up? [18:01:37] hey benestar, that would be great! :) [18:01:49] do you also use winscp to upload files? [18:01:55] benestar: this is just so coincidential [18:02:00] and thanks a lot giftpflanze [18:02:06] Yellowcard: yes [18:02:27] Yellowcard: what is your problem? [18:02:37] * benestar is too lazy to read the whole discussion :P [18:02:51] -i [18:03:13] i could tell you but i think yellowcard will soon tell you [18:03:15] haha that's okay :) i upload files via winscp and then use git bash to change ownership of the uploaded files to my project using become and take [18:03:34] hmm, I always use putty for that purpose [18:03:54] when uploading directories this works all fine but when uploading single files into that dir i get a strange error [18:04:11] Yellowcard: why do you need to change ownership? [18:04:12] then just upload directories ;) [18:04:17] giftpflanze: :P [18:04:38] this is a typical windows solution, what? ;) [18:04:51] probably right [18:05:25] shut down your computer and start it again, maybe it works then, yellowcard ;--) [18:06:03] * giftpflanze is very impudent today [18:06:16] giftpflanze: are you joking? :P [18:06:16] unfortunately i've had the problem for several days :/ [18:06:30] :O not shut down for several days... [18:06:31] indeed i am :) [18:06:32] benestar: so you use putty instead of git bash? [18:06:46] Yellowcard: right [18:06:57] or only partially right, for git I use git bash :P [18:06:59] do you think there could be any difference? [18:07:18] in using shell commands [18:07:54] I don't know, but I unfortunately can't help you with git bash in that case [18:08:34] as i said before: winscp seems to do something wrong, can you tweak configuration, yellowcard? [18:09:31] i configured winscp as described in [18:09:57] .oO(this is brought to you by a scriptkiddie) [18:10:10] *cough* [18:11:03] mhm. i will try the 'take' command using putty [18:11:24] but for what i understand it must output the same error, but we'll see [18:14:35] benestar: exactly same error message using putty. :( [18:14:50] that was to be expected [18:14:55] yeah [18:15:08] maybe they meant that they use pscp? [18:15:17] or psftp [18:15:38] that was what i taught them iirc [18:15:54] instead of winscp? [18:15:58] yes [18:17:34] i don't know, i'll ask benestar when he's back ;) [18:31:10] Coren: I'm making changes on the betaweb instance and they aren't reflecting on page load. Is there some funky cache thing I need to reset or summat? [18:31:43] ... there really shouldn't be, unless it's browser-side. [18:32:37] no, i'm making php-level changes. [18:33:03] First time I hear of someone having that issue. [18:33:14] * Coren tests things. [18:33:31] * Cyberpower678 wishes Coren works at Nintendo. [18:33:44] o_O Why Nintendo? [18:34:00] Coren, so you can fix the 407 error. [18:34:22] And make millions of ads users happy, and allow everyone to use internet services again. [18:34:43] ihaven't tested it this morning; doing so now. it was behaving that way last night around midnight. [18:34:47] ads>3DS [18:35:03] Ah, I don't have a 3DS; my WiiU seems okay. [18:35:03] yeah, still happening. [18:35:29] Coren, try accessing the shop. [18:35:29] *eShop [18:35:30] the main index view should now have "ARGH" in front of every tool name, but it doesn't. [18:35:48] Coren, you'll notice that 99% of the time, you may not be able to access it. [18:36:23] jorm: I just added a comment in index.php and it reflects properly for me when I go look at view-source:https://tools.wmflabs.org/betaweb/ [18:36:42] anomie, thank you. [18:37:09] wtf crazypants! [18:37:35] jorm: Also I don't see ARGH anywhere in the source. [18:37:40] this was inside data/project/betaweb/www/index.php ? [18:37:48] jorm: Yes. [18:37:54] yeah, i put the ARGH inside of data/project/betaweb/www/content/list.php [18:38:07] Wait, hang on. [18:38:17] !log wikimania-support Updated scholarships-alpha to 643d4e3 [18:38:18] Logged the message, Master [18:38:20] Coren, or never mind. The Wii U is functional again. [18:38:28] but the 3DS is still down. [18:38:29] johang: You realize that the live directory is public_html [18:38:38] jorm: ^^ [18:38:44] not www [18:38:45] :-) [18:38:50] grrr. [18:39:08] yes, i'm an idiot. [18:39:22] I can symlink the former to the latter and save you the trouble though. :-) [18:39:59] Or not; I see you just removed www [18:40:00] :-) [18:42:19] yeah, i realized just now that www was the untar directory. [18:42:34] * jorm trying to fix a smallish bug before diving into other things. [18:43:10] for +2 purposes, do you want me to do incremental changes and keep bugfixes separate from features/refactoring? or would just prefer everything at once? [18:54:23] !log deployment-prep performing in place index rebuild for wikis in beta after recent cirrus update [18:54:25] Logged the message, Master [20:26:23] andrewbogott: howdy [20:26:31] hello! [20:26:37] andrewbogott: any idea where we're at with the migration? [20:26:44] have you been checking in with mhoover? [20:26:52] I haven't because I thought you were :( [20:27:00] heh [20:27:01] I've just been trying to get hardware up and ready. [20:27:02] no worries [20:27:13] he and I both had a set of tasks [20:27:27] if he finished, I'm sure he would have mailed me. I haven't had a chance to check mail lately [20:28:12] * andrewbogott nods [20:28:29] I broke nova-precise2/3 last week, I hope no one needed it [20:28:58] Hm… I guess I didn't :) [20:29:07] they are set up as different connected regions [20:29:20] with a replicated database for keystone [20:29:23] and replicated LDAP [20:31:42] Are you hoping to have eqiad and pmtpa work at the same time? Or are you just planning for a new as-yet-unnamed region? [20:31:58] we need both to work at the same time, don't we? [20:32:09] so that tools and deployment-prep can migrate themselves [20:32:27] and in the future we'll have a region in the new datacenter [20:33:33] Hm… yeah, it would certainly be /good/ for both to work at the same time. But I was never clear on whether that was part of our ambition. [20:33:41] Like, that will require us to upgrade pmtpa won't it? [20:40:02] andrewbogott: nope [20:40:13] because we're using versioned APIs [20:42:52] paravoid: http://lists.openstack.org/pipermail/openstack/2013-December/004221.html [20:53:52] andrewbogott: oh, so we're going to need to write a simple keystone token driver [20:53:56] https://github.com/openstack/keystone/blob/stable/havana/keystone/token/core.py#L194 [20:53:59] for havana [20:54:09] because it doesn't actually have redis support [20:54:14] just caching support for redis [20:54:33] all of the driver functions that have exception.NotImplemented() need to be implemented [20:54:44] except flush_expired_tokens [20:54:52] which can just do nothing [20:55:16] once I finish the work on openstackmanager I was planning on doing that [20:56:05] Ryan_Lane: Where does redis fit into the picture? [20:56:23] token replication [20:56:36] for keystone [20:57:32] replication between regions, you mean? [21:00:49] * andrewbogott feels like he came in in the middle [21:04:35] TS is going down by 6.1.2014 evben though tool-labs is still missing features? Is that correct? [21:06:22] andrewbogott: right, so that authentication only has to occur once [21:06:32] sure, makes sense [21:06:36] the keystone tokens will be usable in either region [21:07:08] Hm… somehow I would've imagined that we would only have one keystone [21:07:16] which managed services for both regions [21:07:30] I guess that might not work so well with different API versions [21:09:14] that doesn't work so wel [21:09:15] *wrll [21:09:18] *well [21:09:30] if you lose network for one region then both regions would be down [21:09:54] the way I'm setting it up is that the web interface will be set to use one keystone for writes [21:09:58] and it will replicate the tokens [21:10:05] the other region will read from its own keystone [21:10:31] I take it the havana keystone implemetation has backwards API compatibility? [21:10:44] yeah [21:11:21] hope it works :) [21:12:01] Hello Admins! Some time for me? [21:12:13] well, I'm doing something stupid in the code [21:12:22] not totally sure what yet [21:12:36] the service catalog is getting stuck doubled into memcache somehow [21:13:38] DrTrigon: I can help you with labs stuff… tools less so [21:13:42] Boxing day, I suspect Coren is off. [21:13:50] Being canadian and all. [21:13:52] TS is going down by 6.1.2014 even though tool-labs is still missing features? Is that correct? [21:14:08] coren was here earlier … [21:14:21] Oh, um… not sure. I think silke would be the one to ask about that. [21:14:28] But, six more months doesn't seem crazy to me. [21:14:58] 6.1. translates to january 6th [21:15:49] um… [21:15:55] it might, depending on who's saying it? [21:16:09] I am missing mail support on labs - and it does not seam very cleaver to shut down TS befor labs has all features... [21:16:36] Well, as I said, I'm not really the right guy to complain to about this. [21:16:59] I wrote a mail to TS maillist also... ;) [21:17:07] I believe that account expiry is not the same thing as toolserve shutdown. [21:17:19] It's something about purging old accounts I think? Anyway, probably someone on the list will explain. [21:17:39] It says "You can renew it with acctrenew as usual" [21:17:46] That doesn't sound very final to me [21:18:03] DrTrigon: in the ml, silke said that accounts expire on 6/1/14 and can be renewed, ts runs until mid 2014 [21:18:18] oooo ... GOOD! [21:18:22] that sounds better.... [21:19:17] I will wait for answers on the maillist also in order to sort this out... thanks! [22:17:38] nova client is such a huge piece of shit [22:17:44] at least for the folsom release [22:19:01] well, I guess it's at least working. it returns the incorrect set of endpoints, though [22:19:08] since it only lists one region [22:20:11] well, good news is that setting the region correctly does indeed send the correct commands to the correct endpoints