[07:01:37] YuviPanda: https://github.com/legoktm/webwdapi/blob/master/api.py#L25 [07:01:54] mmm [07:02:16] :D [07:02:31] legoktm: maybe I could, but not for a couple of days [07:02:35] doing Redis stuff [07:02:39] for SuchABot [07:03:17] ok :) [07:03:25] what are you planning on using redis for? [07:04:34] legoktm: persistant job queue [07:04:48] legoktm: i've a continuous job that keeps a connection open on ssh to gerrit stream-events [07:04:52] so I get realtime notifications for all events [07:04:55] ah [07:04:57] and then it puts it into redis queues [07:05:12] wait, why do you need to listen to gerrit though? isnt it just github-->gerrit? [07:05:21] legoktm: no, it's supposed to be two way sync [07:05:30] what goes the other way? [07:05:35] legoktm: so if you comment on Gerrit it sohuld sync to GitHub [07:05:41] line comments -> back to GitHub [07:05:47] ah [07:05:47] same thing the other way [07:05:56] also need to figure out a way to get patchset updates [07:05:58] back to GitHub [07:06:13] legoktm: mediawiki.org/wiki/User:Yuvipanda/G2G has complete info on what to do [07:06:39] legoktm: Redis also has an embedded Lua interpreter [07:07:04] legoktm: https://github.com/yuvipanda/SuchABot/blob/master/suchabot/stream-receiver.py [07:07:07] prettty sweet [07:07:32] ooh [07:07:32] nice [07:08:50] legoktm: so if this goes well, we can actually use this to implement it for the IRC RC Feed [07:08:55] legoktm: perhaps with zeromq [07:09:05] legoktm: so then writing a new consumer of the RC feeds would be much ismpler [07:09:13] :D [07:09:20] plus this has reliable delivery, so you won't lose messages if your bot crashes [07:11:57] What, no Atom+PubSubHubbub? [07:12:40] hmm? [07:12:43] for IRC RC Feeds? [07:12:52] ;) Yeah [07:13:25] DanielFriesen: nobody actually still uses it in production do they? [07:13:32] idk if it survived the flickr redesign [07:13:53] Huh, what does IRC RC have to do with Flickr [07:14:05] DanielFriesen: no I mean PubSubHubbub [07:14:19] flickr and wordpress were, IIRC the highest profile implementations [07:14:46] GitHub is using it (well a bit of a less featured modified version) [07:15:11] hmm, that. [07:16:17] ;) I plan to both make use of that in Gareth for watching repos and implement near-full PubSubHubbubthe Gareth API. [07:16:52] nice [07:17:01] DanielFriesen: does pubsubhubhub do reliable delivery? [07:17:05] in case of client crashes? [07:17:31] I can't remember the exact specifics of the standard or GitHub's impl. [07:18:02] But there should be repeat pokes if the subscriber times out doesn't return an OK [07:18:40] ah, hmm [07:18:41] that'll be nice [07:23:33] Using it will be nice in Gareth. Instead of forcing users to manually add Web hooks for GitHub a plugin will just add a PubSubHubbub subscription automatically. [07:24:14] nice [07:24:41] if the volume on SuchABot gets big enough I should go the pubsubhubbub route too perhaps [07:28:11] legoktm: We should probably deprecate memcached on tools-mc for redis [07:28:16] legoktm: redis is secure-er [07:28:20] Er, [07:28:22] if we use a cryptographically strong prefix [07:28:24] But memcache is so simple. [07:28:31] redis too :P [07:28:32] None of my data needs to be protected [07:28:38] hmm th at [07:28:40] *that [07:28:47] in that case someone needs to upstream the changes [07:28:57] * YuviPanda puts it on his TODO list [09:52:16] addsleeep addsleeep addsleeep [11:41:23] hey petan ! [11:59:11] addshore [11:59:14] :D [11:59:29] we need Coren for that [11:59:35] @notify Coren [11:59:35] You've already asked me to watch this user [11:59:37] you clearly didnt ping me enough :P I slept through it :> [11:59:47] I was sleeping as well [11:59:53] i returned home like 3 am [12:00:17] :O i got home at 6:30am :) [12:00:38] ok [12:43:43] legoktm: https://github.com/wikimedia/apps-android-commons/pull/6 [12:43:45] :) :D <3 [13:14:30] hi [13:36:39] do we have folks who start a process (such as a bot) at the labs server ? [13:38:19] gry: Sure. What's your problem? [13:39:38] I got access to bastion and to tools-login, how do i proceed if i would like to run a bot for a wikimedia project, what do i login to ? [14:06:44] gry: Have you looked at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Help yet? [14:38:40] scfc_de: I read down to the "Getting access to the Tool Labs" section, which I did. Past it, I don't know what to do. [14:49:39] gry: Just read the /next/ section: "Your tool account" :-), i. e. create a tool account for your bot. [14:50:18] scfc_de: I don't know the toolname. Can I make it up or do I have to request approval for it first? [15:03:29] gry: You *must* make it up :-). [15:04:26] ok [15:14:37] Need Coren for what? [15:14:41] petan: ^^ [15:17:34] hi Coren [15:17:50] what if we create a new project for testing of tools and delete bots entirely? [15:18:04] what do you think? [15:18:56] petan: That's probably the easiest way forward, but it does mean that anyone on bots- have to start anew. Depending on how many people are using it and how complex their stuff is this may or may not be too hard. [15:19:57] petan: On the plus side, configuring it /should/ be almost as easy as just copying the puppet classes. (Except for the gridengine config and web server config, which can still be just copied over) [15:20:53] Are we short on resources? Otherwise we could have tools-testing and bots at the same time, until the last user has migrated from Bots to Tools. [15:21:26] Coren: I think all people on bots should just move to tools [15:21:41] until they do it, we can keep the servers they use [15:21:42] petan: Probably, for the most part. [15:22:16] scfc_de yes that is what I mean [15:22:39] Coren can you create the new project then? I think tools-beta or tools-test is short and clear [15:23:23] How about betatools instead? Keeps naming simpler (no extra dashes) [15:23:50] Doesn't show up in lists next to tools. [15:23:53] I dont really care :o [15:23:57] true [15:24:02] toolsbeta [15:24:19] tbh I dont see anything wrong on dashes [15:24:51] Me neither. [15:25:18] It'd just end up confusing if tools needed a -beta or -test server someday. :-) [15:25:22] !log bots removing most of instances we dont use anymore (keeping only -login -gs -*sql* -bnr* [15:25:24] Logged the message, Master [15:26:04] if we create a project called tools-beta we will never create server in tools called tools-beta :P [15:26:11] that makes sense [15:27:26] petan: people have a few things running on bots-4 ;p [15:27:33] I know [15:27:35] petan: {{done}} [15:27:41] Coren thanks [15:27:57] heh 3701 cdpark 20 0 313m 272m 3488 R 99.4 13.5 3619:42 python [15:28:06] Coren is there any restriction for having sysadmin / root on that project or is it free for any volunteers? [15:28:06] 'toolsbeta' Dashes in project names bug /me/ if noone else. :-) [15:28:50] petan: Provided that everyone understand that the policy privacy doesn't cover tools running on that project then I see no reason to not let volunteers be root there. [15:29:00] ok [15:29:07] i will put it to motd and help page [15:29:15] addshore, scfc_de want sysadmin there? [15:29:19] petan: The web server will need the big disclaimer too, so would any tool's landing page. [15:29:24] ok [15:30:10] !log toolsbeta importing security groups of tools [15:30:10] toolsbeta is not a valid project. [15:30:21] * petan slaps labs-morebots [15:30:29] petan: Sure. [15:31:17] scfc_de dont forget to prepare security group before creating an instance [15:31:22] !toolsadmin | scfc_de [15:31:22] scfc_de: https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Documentation/Admin [15:31:32] this may be useful too :P [15:32:06] scfc_de reming me your labs name? [15:32:09] * remind [15:32:22] scfc [15:36:48] !log toolsbeta blah [15:36:50] Logged the message, Master [15:37:01] !log toolsbeta importing security groups of tools [15:37:01] Logged the message, Master [15:37:07] !log toolsbeta importing sudo policies of tools [15:37:09] Logged the message, Master [15:39:29] * scfc_de will be back later. CY! [16:23:35] is wikitech db available from labs? [16:25:59] Coren: you know? [16:26:54] AzaToth: ... actually, funnily enough, I'm not sure. I *think* it should be, but I never tried. :-) [16:27:09] heh [16:27:30] was looking into search for bad external links to old gitweb [16:28:14] !s7 [16:28:39] * Coren glares at wm-bot [16:28:52] offcourse I could use Special:LinkSearch... [16:29:15] I think it's a DB on S7, AzaToth, which means "will be there as soon as Asher is done with it" :-) [16:29:21] hehe [16:29:44] I notice /usr/bin/sql doens't exists to any package, is it manually placed there? [16:30:25] AzaToth: From puppet. It should probably be packaged with misctools like the others. [16:30:32] ok [16:30:40] https://wikitech.wikimedia.org/w/index.php?title=Special:LinkSearch&limit=500&offset=0&target=https%3A%2F%2Fgerrit.wikimedia.org%2Fr%2Fgitweb [16:30:49] too many to change manually :-P [16:31:38] is labs puppet same as ops puppet? [16:37:40] * Coren nods. [19:22:22] ? :)Cyberpower678: hows peachy [19:49:44] addshore, unstable as usual. Why do you ask? [20:00:57] Coren: a question, https://gerrit.wikimedia.org/r/#/c/62350 you +2+2 but never merged [20:01:09] any reasons? [20:02:02] which seem to be same as https://gerrit.wikimedia.org/r/#/c/62236 [20:02:25] and https://gerrit.wikimedia.org/r/#/c/66294 [20:02:57] petan: ping [20:04:50] New review: AzaToth; "duplicate of I6d46145ba184b5ec07aa8fa49913c2fb51b36b68" [labs/toollabs] (master) C: -1; - https://gerrit.wikimedia.org/r/66294 [20:05:58] New review: AzaToth; "Duplicate of I6d46145ba184b5ec07aa8fa49913c2fb51b36b68" [labs/toollabs] (master) C: -1; - https://gerrit.wikimedia.org/r/62350 [20:08:56] New review: AzaToth; "Please rebase onto master and place the file under jobutils/bin" [labs/toollabs] (master) C: -1; - https://gerrit.wikimedia.org/r/62236 [20:22:09] Change merged: coren; [labs/toollabs] (master) - https://gerrit.wikimedia.org/r/62350 [20:22:46] Coren: erm [20:22:59] I think that was a bad merge [20:23:19] unless you magically changed the path in the commit before you merged it [20:23:39] Coren: as the commit placed it in usr/local/sbin/rmtool [20:24:09] AzaToth: Oh, didn't notice that. I thought it'd have been rebased, but I note that this wasn't dependent on the reshuffle dance. [20:24:30] yea becuaue it's a new file [20:24:58] you close the other two as well? [20:25:16] Not yet, I was looking a them. [20:25:38] Change abandoned: coren; "Duplicate, will add new changeset to move it to the right spot." [labs/toollabs] (master) - https://gerrit.wikimedia.org/r/62236 [20:25:49] that are "petan failed with gerrit" changes [20:26:04] Change abandoned: coren; "Duplicate; will move to the right spot with a new changeset." [labs/toollabs] (master) - https://gerrit.wikimedia.org/r/66294 [20:26:38] New review: coren; "More tests == More goodness" [labs/toollabs] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/68330 [20:26:39] Change merged: coren; [labs/toollabs] (master) - https://gerrit.wikimedia.org/r/68330 [20:26:46] hehe [20:27:41] uh? [20:28:09] ah [20:28:27] New review: coren; "(1 comment)" [labs/toollabs] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/68331 [20:28:27] Change merged: coren; [labs/toollabs] (master) - https://gerrit.wikimedia.org/r/68331 [20:33:34] New patchset: coren; "Tool Labs: move rmtool to the proper spot" [labs/toollabs] (master) - https://gerrit.wikimedia.org/r/68960 [20:39:24] New review: AzaToth; "nice move, but need to add it to debian/tools.install as well" [labs/toollabs] (master) C: -1; - https://gerrit.wikimedia.org/r/68960 [20:39:42] Coren: I see in "sql" it links to a /usr/share/dblist [20:42:51] New patchset: coren; "Tool Labs: move rmtool to the proper spot" [labs/toollabs] (master) - https://gerrit.wikimedia.org/r/68960 [20:57:45] !ping [20:57:46] pong [20:57:50] !s7 [20:57:59] Coren there is no such key I think [20:58:28] petan: I though I put one in listing the databases on S7 [20:58:32] * YuviPanda looks around [20:58:48] @searchrx [sS]7 [20:58:48] No results were found, remember, the bot is searching through content of keys and their names [20:58:57] !S7 [20:59:00] nope [21:06:07] petan: where is /usr/share/dblist? [21:06:38] AzaToth ? [21:06:41] I have no idea what is that [21:06:52] petan: it's referenced from /usr/bin/sql [21:06:57] but I can't find it [21:06:58] huh? [21:07:03] it doesnt exist [21:07:15] and I dont remember I ever pushed the version that reference it [21:07:39] fb3156fc (Petr Bena 2013-05-26 17:23:56 +0200 45) list=/usr/share/dblist [21:07:59] ah this [21:08:07] that variable isnt used anywhere [21:08:13] the line should have been removed [21:08:17] I forgot to do that [21:08:30] oh I see [21:12:20] New review: Tim Landscheidt; "(1 comment)" [labs/toollabs] (master) - https://gerrit.wikimedia.org/r/68331 [21:15:25] New review: coren; "(1 comment)" [labs/toollabs] (master) - https://gerrit.wikimedia.org/r/68331 [21:18:34] New review: coren; "Yeay new version." [labs/toollabs] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/68960 [21:18:34] Change merged: coren; [labs/toollabs] (master) - https://gerrit.wikimedia.org/r/68960 [21:22:15] @searchrc centralauth [21:22:15] No results were found, remember, the bot is searching through content of keys and their names [21:22:19] Hm. [21:22:21] New review: Tim Landscheidt; "(1 comment)" [labs/toollabs] (master) - https://gerrit.wikimedia.org/r/68331 [21:22:22] I could have sworn. [21:23:37] !log tools 1.0.3 deployed (jobutils, misctools) [21:23:39] Logged the message, Master [21:28:39] petan: I'm writing a patch to memcached that lets us specify a command line param that disables STATS and FLUSH_ALL [21:28:48] and then I'm going to send them it and see if they will accept it :) [21:29:06] * YuviPanda compiles libevent [21:29:09] YuviPanda no point in doing that. sharp memcache can already do that [21:30:51] petan: tch tch. [21:31:05] well, let's just think of it as a way to re-touch my C skills :) [21:33:33] petan: plus memcached has millions of users [21:33:42] plus... I honestly doubt C# can beat epoll, etc [21:33:55] I bet it can + much more [21:35:21] well, okay :) Once you get sharp-memcache working with any of the existing test suites, would be simple to compare them :) [21:35:23] good luck! [21:42:11] Coren: "Huh. -w FTW. I've been using that syntax for ages without even knowing I was a silly cow. :-)" - hehe [21:43:30] knowing the difference between my $foo, $bar and my ($foo, $bar) is what separates perl programmers to the rest of the pack [21:43:47] (those who actually knows how to use pack is an other story) [21:47:34] AzaToth: I use pack, and I my ($foo, $bar) = (somelist) idomatically; I just never heeded the "use -w" mantra. Too old school. :-) [21:55:34] hehe [22:12:51] petan: I started sending in patches, but memcached work has sortof mostly died down for a while now [22:12:57] i'm going to try kicking the tires but :) [22:18:53] YuviPanda it is not hard to make my memcahe working with test suites, it is hard to get the test suites working by themselve :P [22:19:06] I wasnt able to compile the official one [22:19:12] petan: not really, I just managed to submit a trivial patch after passing the test suite on my local machine :P [22:19:15] it seems to be too outdated for modern compilers [22:19:18] in what, 20 minutes? [22:19:25] (https://github.com/memcached/memcached/pull/41) [22:19:51] petan: in fact it is possible that my patch might have fixed the issue you had :P [22:19:54] which test suite you used? [22:19:58] though idk why these guys just don't use autotools [22:20:00] the official one [22:20:05] I found only the one for binary protocol [22:20:05] make test on the official repo [22:20:13] binary protocol is undocumented evil [22:20:21] https://github.com/memcached/memcached/tree/master/t [22:20:22] specifically [22:28:10] YuviPanda: autohell? [22:28:35] AzaToth: well, considering that memcached seems to be some of the auto* tools but not all... [22:29:08] that sentence doesn't make any sense to me ヾ [22:29:55] AzaToth: https://github.com/memcached/memcached/blob/master/autogen.sh [22:29:59] felt a little... weird to me [22:31:05] YuviPanda: classic autohell.sh [22:31:09] heh [22:31:47] though "$AUTOMAKE --foreign --add-missing || $AUTOMAKE --gnu --add-missing || exit 1" I've no idea why [22:32:02] that just fells stupid [22:32:51] feels* [22:34:36] AzaToth: indeed. [22:34:40] also the perl at the top feels stupid to me too [22:34:55] despite the fact that I haven't touched auto* since my GNOME dev days from a long time ago [23:50:15] > It's not just you! http://tools.wmflabs.org looks down from here. [23:50:18] Coren: ^