[00:11:30] Coren! [00:11:34] can you take a look at https://bugzilla.wikimedia.org/49046 [00:33:33] Coren, well can I at least connect to it? [00:34:27] Krenair: become your tool, you'll be able to connect to it from there. [01:10:35] Coren, honestly so far I've just been messing around with the database replicas, I haven't got around to creating a tool yet :p [01:49:48] Krenair: Do like many do, create a dev tool for playing around with. :-) [01:50:00] That's basically what I'm doing [02:21:09] Coren: Take a look at p50380g50577__wmf_wiki_info on tools-db :) [02:39:17] Haven't figured out how some of the 'family' values, is_meta, is_sensitive, is_multilang, or root_category are supposed to be generated yet, and the private wikis are missing some data of course [02:39:34] Also need to work out how to make it readable by all (assuming it's not already) [02:48:04] MariaDB [p50380g50577__wmf_wiki_info]> GRANT SELECT ON p50380g50577__wmf_wiki_info.* TO '%'@'%'; [02:48:04] ERROR 1044 (42000): Access denied for user 'p50380g50577'@'%' to database 'p50380g50577__wmf_wiki_info' [02:48:06] huh [02:55:38] O.o That should have worked. [02:58:36] Ah. That be my fault; the privileges on the local DB does not have with grant option. Poop. [03:01:35] Krenair: Try again? [03:02:27] MariaDB [p50380g50577__wmf_wiki_info]> GRANT SELECT ON p50380g50577__wmf_wiki_info.* TO '%'@'%'; [03:02:27] ERROR 1410 (42000): You are not allowed to create a user with GRANT [03:03:38] Silly mysql. [03:03:57] It needs to already know a user '%'@'%'. :-) [03:04:15] It does now. [03:05:42] Looks like it worked that time [03:06:38] Can you see the data? [03:26:45] How do I get a page from the db? [03:33:24] a930913: "Get a page"? [03:37:10] Coren: As in the current revision I guess. [03:49:32] a930913: Well, you can't get the text from the DB, you'd use the API for that -- and since you're on the API anyways getting the latest revision is the default. [04:28:34] ohai legoktm [04:28:41] omg [04:28:43] i literally [04:28:46] just got home [04:29:24] omg [04:29:25] i literally [04:29:25] just [04:29:26] woke up [04:29:27] :P [04:29:33] nice saturday night? [04:30:56] yes :D [04:31:07] trying to find a movie on netflix to watch [04:31:10] :) [04:31:10] ah [04:31:18] FDR: American Badass? [04:31:19] :P [04:31:52] did you get started on editcountitis at all, legoktm? [04:31:55] or shall I just start? [04:32:09] go for it [04:32:12] i didnt get a chance [04:32:15] sweet [05:20:42] @notify anomie [05:20:42] I'll let you know when I see anomie around here [07:13:14] YuviPanda: whats Werkzeug? it got installed when i installed flask [07:13:25] it's the underlying wsgi wrapper [07:13:26] > The Python WSGI Utility Library [07:13:28] o.O [07:13:34] or rather [07:13:35] implementation [07:13:42] wsgi is sortof like a protocol (like CGI) [07:13:47] werkzeug wraps around it [07:13:50] a little [07:13:56] flask just sits on top of it [07:14:02] werkzeug adds error reporting [07:14:04] debugging [07:14:07] and all that stuff [07:14:13] think of it like cgitb [07:14:46] legoktm: it also I think can be extended as 'middleware' if necessary [07:15:54] so you can do things like auth, gzip, etc without needing to touch the app [07:16:00] ah, ok [07:50:49] YuviPanda: its not working :( [07:50:57] where is it? [07:51:02] which tool? [07:51:22] hold on a sec, pastebinning [07:52:38] data.py https://dpaste.de/HP7OH/raw/ [07:52:58] api.py https://dpaste.de/EJScj/raw/ [07:53:11] and test url is http://tools.wmflabs.org/legobot/cgi-bin/av/api.py/aiv/ [07:54:51] legoktm: what error are you getting? [07:55:11] http://tools.wmflabs.org/legobot/cgi-bin/av/api.py/aiv/ : "A server error occurred. Please contact the administrator." [07:55:41] for some reason i'm giggling at 'import gf ' :P [07:55:44] okay getting serious now [07:55:45] * YuviPanda reads [07:56:15] legoktm: https://dpaste.de/EJScj/raw/ has two hashbangs [07:56:28] errrr [07:56:31] also what is gf? [07:56:34] is it importable? [07:56:40] yeah [07:56:43] give me a sec [07:56:59] can I haz access to the repo? [07:57:02] even temperoarily? [07:57:13] removed the second hash bang, no difference [07:57:30] um, its in "legobot" which has my bot password so i cant give that out [07:57:37] i can create another tool? [07:57:44] right [07:57:45] well [07:57:49] i can just copy paste this [07:57:50] so [07:57:50] moment [07:59:03] legoktm: added you to tool editcountitis [07:59:06] copy there? [08:00:07] legoktm: just copy the entire folder, including virtualenv, there? [08:02:21] ok sure, give me a sec [08:06:55] YuviPanda: its in cgi-bin/av/ [08:08:01] ok checking it out [08:15:37] legoktm: hmm, the cgi script isn't being executed at all [08:15:48] o.O why not? [08:15:58] i set it to +x [08:16:40] legoktm: even if i set it to using /usr/bin/python [08:16:46] legoktm: see ~/cgi-bin/avi/av/ [08:16:50] (moved it to a new virtualenv) [08:17:24] ok [08:17:59] legoktm: hmm, they're not beigne executed. I removed the .pyc files, hit it again, no execution [08:18:09] this is failing much before it reaches wsgi [08:18:53] petan: can you grep apache error logs to see if there's anything? [08:19:14] do we need to set up the rewrite rules or name it .cgi like the example at http://flask.pocoo.org/docs/deploying/cgi/#creating-a-cgi-file ? [08:19:55] legoktm: i put a file write based log on the first line of api.py, before it even tries to import wsgiref [08:20:00] hm [08:20:03] so this should run as a normal python cgi file [08:20:04] not happening [08:21:41] legoktm: oh i'm an idiot [08:21:41] nevermind [08:21:53] didn't set 'w' mode on file [08:21:56] o.O [08:21:58] cgi runs [08:22:08] need to figure out the issue with wsgi [08:22:08] debugging [08:28:10] legoktm: ah, so now it is 404 [08:28:13] found a chunk of the problem [08:28:20] legoktm: it was missing memcached and simplejson packages [08:28:24] oh [08:28:25] :P [08:28:39] package is python-memcached btw [08:28:49] yeah, installed it [08:28:53] and renamed simplejson to json [08:28:59] so now i'm getting 404 [08:29:02] which is... more appropritae [08:31:17] legoktm: tada http://tools.wmflabs.org/editcountitis/cgi-bin/avi/av/api.py [08:31:18] :) [08:31:29] yayay :D [08:32:06] legoktm: :) [08:32:10] legoktm: you need to 'return' not 'print' [08:32:15] ohhhhhhhhhhhhhhhhh [08:32:20] legoktm: however, there's some weird rewrite rule for http://tools.wmflabs.org/editcountitis/cgi-bin/avi/av/api.py [08:32:25] that's redirecting it to an internal url [08:32:49] ok but wait, why doesnt http://tools.wmflabs.org/editcountitis/cgi-bin/avi/av/api.py/aiv/ work? [08:33:26] good question :P [08:33:29] i'm trying to find out now [08:34:17] legoktm: http://tools.wmflabs.org/editcountitis/cgi-bin/avi/av/api.py/aivdd/ works [08:34:20] also it seems app.debug=True doesnt work [08:34:27] hm [08:34:31] legoktm: so that seems to be an error being thrown up by mc_get [08:34:37] ok [08:34:57] legoktm: you should cd to the virtualenv directory and use 'source bin/active' [08:35:49] >>> import gf [08:35:50] >>> gf.mc_get('aiv') [08:35:50] [] [08:35:52] works fine? [08:38:00] hmm [08:38:38] YuviPanda sure [08:38:49] petan: also can you install 'ack-grep' package? :D [08:38:59] ye [08:39:09] thank you :) [08:39:13] everywhere? [08:39:16] yes [08:39:25] it's small, and very useful [08:39:38] !log tools petrb: installing ack-grep everywhere per yuvipanda and irc [08:39:40] Logged the message, Master [08:39:45] thank you [08:40:45] it was already everywhere [08:41:18] YuviPanda what you want to grep in log [08:41:51] hmm, we're looking for python tracebacks from cgi-bin from the tools project editcountitis [08:42:35] legoktm: http://flask.pocoo.org/docs/quickstart/#debug-mode i think we might not be able to use debugging mode on cgi-bin, I suppose it is a forking environment... [08:42:46] :/ [08:43:53] legoktm: nevermind :) [08:43:57] legoktm: http://tools.wmflabs.org/editcountitis/cgi-bin/avi/av/api.py/aiv/ tracebacks ahoy :) [08:44:01] legoktm: unicode issue :P [08:44:13] legoktm: werkzeug's debug mode works [08:44:14] ahhhhhhhhh [08:44:22] ok thats easy [08:44:23] petan: nevermind, we found it! :) [08:44:24] YuviPanda http:/pastebin.com/S0LxBw83 [08:44:27] omg [08:44:36] * http://pastebin.com/S0LxBw83 [08:45:06] :) [08:45:07] same error [08:45:15] http://tools.wmflabs.org/editcountitis/cgi-bin/avi/av/api.py/aiv/ :DDDDDD [08:45:48] legoktm: :D [08:46:12] dammit I wanted to play with the debugger more but now there are no errors :( [08:46:20] legoktm: is this in a git repository somewhere? [08:46:27] also welcome to WSGI :) [08:46:30] and flask :P [08:46:37] also do you really have to name your methods with one variables? :P [08:46:42] and call them 'thing'? [08:46:44] heheheheh [08:46:58] i was being lazy [08:47:10] tch tch :P [08:48:37] im working on a WP:AIV parser and so the next step is to have a JSON api that will be used [08:48:50] NIICEE! [08:48:52] :) [08:49:01] petan: memcached being used ^ :) [08:50:45] so i have an IRC bot updating that memcached key :) [08:51:07] ... [08:51:12] why?! [08:51:48] is this all in a git repo? [08:53:59] well the bot does a lot more. https://github.com/legoktm/adminbots and ##legoktm-bots [08:54:15] nice :) [08:56:19] live running feed of AIV, UAA, RfPP and then tracks the users/articles listed, provides info about previous blocks, edit counts, etc [08:56:48] niiice! :) [08:56:51] all running on toolslabs? [08:56:53] soon im going to add in a CSD module that tracks when articles get tagged [08:56:54] was in on ts before? [08:56:55] yup [08:56:59] no, i wrote it last week [08:58:57] nice [08:58:59] :) [08:59:34] it's like how I write tools for fun but you write tools that people actually use :P [09:00:19] heh, well right now i've just been using it [09:00:34] i've blocked over 150 users in the past 2 days [09:00:48] wah [09:00:51] ips? [09:02:16] and users too [09:52:40] legoktm: i think the tracebacks are interactive [09:52:46] you can execute arbitrary code [09:52:48] i think [09:52:49] o.O? [09:52:53] thats not a good idea. [09:52:54] that's why they should be off [09:52:56] yes [09:52:59] they have a REPL [09:53:02] we can turn it off [09:53:08] we should.... [09:53:13] by passing False for the second parameter [09:53:15] in api,py [09:53:17] api.py [09:53:20] ok [09:53:23] however it is still a cool feature :D [09:53:29] yeah [09:53:43] nothing's as cool as unauthenticated remote execution of untrusted code :P [09:53:48] im gonna sleep in a bit, but tmrw ill try and transition it to its own tool [09:53:53] lolol [09:54:09] legoktm: i'll turn off debug execution [09:54:14] should leave tracebacks intact [09:54:14] thanks [09:54:35] done [09:54:36] :) [09:54:45] yes tracebacks, no code exec [09:54:50] legoktm: good night :) [09:55:02] legoktm: are you on Facebook / Twitter? [09:55:13] twitter.com/legoktm [09:55:41] :) [09:55:43] thanks [09:58:09] :D [10:14:05] YuviPanda I update the memcache I made, which libraries are you using for it? [10:14:17] I will try to create some that works with my version as well [10:14:18] currently it is legoktm's tool [10:14:22] python-memcached [10:14:25] ok [10:17:44] YuviPanda this version support stat command and it restrict flush_all to user [10:17:53] nice! [10:17:56] if anyone connect to shared mc and do flush_all they will remove all keys [10:18:07] you really should add one of the test suites but [10:18:10] ah [10:18:15] we should comment that out too? :P [10:18:17] what test suite you mean? [10:18:18] how? :P [10:18:18] !cyberpower678 [10:18:32] !Cyberpower678 [10:18:33] addshore, how do you rollback? with the rollback button? :D [10:18:35] https://code.google.com/p/memcached/wiki/TestSuites [10:18:36] "{ [10:18:47] aha [10:18:49] ok [10:18:52] I will [10:20:14] Warning: There is 1 user waiting for shell: AleXXw (waiting 0 minutes) [10:21:24] petan: :) [10:26:56] YuviPanda https://github.com/benapetr/memcached [10:27:04] mmm [10:30:49] https://github.com/benapetr/memcached#descriptive-errors [10:30:56] :P [10:32:37] YuviPanda I dont know python :/ [10:32:47] it is too hard for me to update the library but modification is very easy [10:32:53] learn! [10:32:55] :D [10:33:01] it is just needed to create new function Authenticate [10:33:16] hm [10:33:26] which only send 1 stanza to server [10:33:28] depends. [10:33:37] does it use persistant connections? [10:33:39] Warning: There is 1 user waiting for shell: AleXXw (waiting 13 minutes) [10:33:42] once per set? [10:33:45] if it was in language I know it would be matter of 20 minutes for me to implement it [10:33:51] YuviPanda yes I know... [10:33:54] https://github.com/linsomniac/python-memcached/blob/master/memcache.py [10:33:59] thats the file that needs to be modified [10:34:20] if it doesnt it means the stanza needs to be added everywhere after connection [10:34:21] maybe https://github.com/linsomniac/python-memcached/blob/master/memcache.py#L1110 ? [10:34:42] legoktm how are you using it? [10:34:47] brb food [10:34:50] it connect only when you need to access it [10:34:55] or it connect once and then you access it [10:35:06] connect only when you need access i think [10:35:10] mhm [10:35:16] i never explicitly call a .connect() or anything [10:35:20] i just create the object [10:35:26] then call a get() or set() on a key [10:35:29] that would be easy anyway [10:35:34] the library knows if it is connected or not [10:35:43] you would just create a constructor with name and password [10:35:54] and in every connect it would automatically login [10:36:17] you just need to send to server these bytes [10:36:30] authenticate user:psw\r\n [10:36:34] user is username [10:36:37] psw is password [10:37:06] ok [10:37:09] shouldnt be that hard [10:37:21] I think it is matter of few minutes for someone who knows the language [10:37:35] yeah [10:47:08] Warning: There is 1 user waiting for shell: AleXXw (waiting 27 minutes) [10:54:04] YuviPanda I made a cool stuff :D [10:54:10] I started my memcache in mixed mode [10:54:21] that means it allows authentication but it doesnt require it [10:54:30] users who dont login share the same realm [10:54:41] so it should be compatible with regular memcache now [10:55:05] legoktm ^ [10:55:15] can you try some python script with this memcache server? [11:00:32] Warning: There is 1 user waiting for shell: AleXXw (waiting 40 minutes) [11:01:44] sure [11:01:50] whats the hostname/port? [11:01:58] tools-mc 11212 [11:03:54] Setting a key didn't work [11:04:03] what it did? [11:04:03] It also took much longer for it to return a failed error code [11:04:08] hmm [11:04:13] i>>> import memcache [11:04:13] >>> mc=memcache.Client(['tools-mc:11212']) [11:04:13] >>> mc.set('test', '1234') [11:04:13] 0 [11:04:40] I would need to see the traffic hold on [11:06:33] ok try now [11:06:40] I am watching the traffic [11:07:13] but LOL [11:07:15] just did it [11:07:15] it worked [11:07:22] set test 0 0 4 [11:07:23] 1234 [11:07:26] response: [11:07:27] STORED [11:07:34] wtf [11:07:39] but [11:07:41] >>> mc.get('test') [11:07:41] >>> [11:07:49] it sent literally same response as regular mc o.O [11:07:58] waaaaa [11:08:14] now it didnt send any data [11:08:39] no data whatsoever :/ [11:08:50] idk ://///// [11:09:00] are you sure you are trying get on same mc server? [11:09:06] because it didnt try to ask it [11:09:49] yes, it is [11:10:05] but it didnt even sent 1 byte to this port [11:10:18] did you just see something? [11:10:20] yes [11:10:22] now I did [11:10:24] get test [11:10:32] ok, but i didnt get any data back [11:10:34] response was VALUE test 0 4 [11:10:36] 1234 [11:10:40] hm [11:10:45] just as it should be o.O [11:13:53] Warning: There is 1 user waiting for shell: AleXXw (waiting 53 minutes) [11:15:47] back [11:15:53] legoktm: i think we should document the wsgi setup [11:16:17] petan: do you know where would be a nice place to document 'How to Run your WSGI scripts with VirtualEnv on Labs'? [11:16:28] wikitech [11:16:37] yeah but where on wikitech/ [11:16:37] ? [11:16:45] no idea XD [11:16:50] Nova_Resource:something/something? [11:17:07] User:YuviPanda/wsgi python something [11:17:14] what userspace [11:17:15] noooo [11:17:37] legoktm: petan also I was reading through some of the threads on toolserve-l [11:17:41] that.... looks painfully dramatic :( [11:17:49] link [11:17:58] oh stuff like http://lists.wikimedia.org/pipermail/toolserver-l/2013-April/005978.html [11:18:14] and this thread http://lists.wikimedia.org/pipermail/toolserver-l/2013-May/006147.html [11:19:22] oh yeah YuviPanda, out of curiosity, how hard do you think it would be to set up a SVN mirror of a git repo? [11:19:23] i read this [11:19:25] long time ago [11:19:39] legoktm: so... The answer to that is... sadly... [11:19:40] not very hard [11:19:58] assuming readonly mirror [11:20:03] Yes, read only. [11:20:10] legoktm: so, *easiest* way.... is https://github.com/blog/626-announcing-svn-support [11:20:21] provided that it is already on github [11:20:23] if not [11:20:38] (or https://github.com/blog/966-improved-subversion-client-support) [11:20:42] OH SRSRLY [11:20:46] YA SRSLY [11:20:55] (it wasn't an april fool's prank, it is for realz) [11:21:01] the best pranks are real... [11:21:28] wow great. [11:21:29]