[02:10:03] hey everyone, can an admin please run "sudo chmod 777 /home/vacation9/VoxelBot/logininfo.pyc" for me [02:10:07] on bots-4 [02:10:15] since we don't have root any more [02:10:23] actually, just chown'ing it would work [02:11:06] but i'm obviously not a labs admin :) [02:11:33] true [02:12:00] sudo chown vacation9 /home/vacation9/VoxelBot/logininfo.pyc on bots-4 if any labs admins are here [02:12:33] Gimme a sec. [02:12:44] You're lucky I was walking in front of the screen right that moment. :-) [02:13:51] Coren|Sleep: heh... thanks [02:14:12] ... if bots-4 lets me in. It was slow. [02:14:19] {{done}} [02:14:37] thanks you :) [02:14:45] * Coren|Sleep goes back to idle. [05:35:51] * legoktm hugs Coren|Sleep [06:06:23] Coren|Sleep pong [06:16:44] Ryan_Lane: hi [08:08:26] !log account-creation-assistance updated packages on all instances [08:08:29] Logged the message, Master [09:13:12] * addshore pangs Coren|Sleep [09:16:37] addshore ping [09:16:44] I created qdeploy tol [09:16:46] tool [09:16:54] it install packages across cluster [09:17:03] so you can use it to install some package on all servers [09:17:18] @notify ceradon [09:17:18] I will notify you, when I see ceradon around here [09:17:41] @notify Coren [09:17:41] I will notify you, when I see Coren around here [09:20:05] oh cool [09:49:16] !botsgrif [09:49:18] !botsgrid [09:49:28] !gridb [09:49:28] http://helms-deep.cable.nu/~rwh/blog/?p=159 [09:54:39] !log bots petrb: created new queue for irc bots [09:54:43] Logged the message, Master [09:59:22] @notify addshore [09:59:22] This user is now online in #huggle so I will let you know when they show some activity (talk etc) [10:22:16] @notify andrewbogott_afk [10:22:16] This user is now online in #wikimedia-labs so I will let you know when they show some activity (talk etc) [10:25:35] wtf addshore o.O [10:28:37] ? [10:28:54] mhm it seems that qsub randomly submit jobs to any queue [10:29:01] I thought that there is default queue [10:29:04] main.q [10:29:22] some of addshore jobs were submitted to minimalload [10:29:33] What's the difference between the queues? [10:29:51] minimalload is for irc bots - it contains server where should be almost no load [10:30:06] it's for interactive bots [10:30:17] which need to respond quickly and not be affected by system load at all [10:31:41] This should be documented somewhere (is it already ?) [10:31:55] Ah, lunchtime. [10:32:10] I am setting that queue up now [10:32:13] will document it once it works [10:36:01] !log bots petrb: changed addshore's cron so that it does not affect minimal server [10:36:04] Logged the message, Master [10:37:40] petan i know how to fix that [10:37:52] there is a setting somehwere ;p [10:38:22] but qsub by itself should be fine... you can select the queues for qsub to use as default [10:38:31] currently it just uses the whole list of queues [12:10:27] addaway how? [12:10:30] @notify addaway [12:10:30] This user is now online in #huggle so I will let you know when they show some activity (talk etc) [12:10:39] not sure :P [12:10:56] i saw it when i was looking through everything the other day [12:10:57] ok [12:11:03] I couldn't find it [12:31:43] !log bots petrb: qdeploy python-irclib [12:31:45] Logged the message, Master [13:07:23] legoktm ping [13:07:28] pong [13:07:36] there is something of you [13:07:39] running on bots-3 [13:07:44] yes [13:07:47] my cron is on there [13:07:48] is it possible to move that to grid? [13:07:51] ergh [13:07:55] i need to do that [13:08:00] ok, no rush [13:10:14] @seenrx [Mm]ad[mM]an [13:10:14] petan: Last time I saw RogueMadman they were quitting the network with reason: no reason was given at 11/17/2012 6:21:49 AM (118.06:48:25.1352900 ago) (multiple results were found: RogueMadman|2, AMadman, AMadman[food], AMadman[nap], AMadman[cranky|c and 3 more results) [13:10:37] @seenrx madman [13:10:37] petan: I have never seen madman [13:10:46] @seenrx ^[Mm]ad[mM]an$ [13:10:47] petan: I have never seen ^[Mm]ad[mM]an$ [13:10:56] 118 days... that's a while [13:11:03] He's usually AMadman [13:11:09] @seenrx ^A[Mm]ad[mM]an$ [13:11:09] petan: Last time I saw AMadman they were quitting the network with reason: Ping timeout: 276 seconds at 3/7/2013 1:16:51 AM (8.11:54:17.4018660 ago) [13:11:10] @seen amadman [13:11:10] legoktm: Last time I saw amadman they were quitting the network with reason: Ping timeout: 276 seconds N/A at 3/7/2013 1:16:51 AM (8.11:54:18.1109390 ago) [13:11:21] @notify amadman [13:11:21] I will notify you, when I see amadman around here [13:12:29] you two guys are each having 1 job on bots-2 and bots-3 :P [13:12:45] which is only reason why we still have them :D [13:14:28] o.O [13:14:34] well it shouldnt be too hard to move them [13:15:00] addshore I created queue for irc [13:15:02] can OG access stuff in my homedir? [13:15:08] but I didn't find a single bot to move there :D [13:15:27] legoktm OG can't but YOUR user which is running the task through OG can [13:15:39] so if you qsub a job, that job will have access there [13:15:42] so…. [13:15:42] ok [13:15:45] er wait [13:15:59] should be the same if my crontab qsub's it right? [13:16:12] and did the "qcronsub" thing ever get set up? [13:16:15] your crontab has access to your home as well [13:16:21] I am working on it [13:16:36] problem is that the script in mail isn't complete and I wasn't told what is purpose of that [13:16:51] have you talked to Merlissimo? [13:17:04] I already wrote a script which check if a job is already running, it's a 2 line script, so... [13:17:10] no I didn't, I don't know who he is [13:17:19] neither where he is [13:18:40] well he's in this channel :P [13:18:48] Merlissimo ping [13:18:57] @notify Merlissimo [13:18:57] This user is now online in #wikimedia-tech so I will let you know when they show some activity (talk etc) [13:19:00] :P [13:33:58] :O [13:34:41] !log deployment-prep -squid shows a ton of stalled `redirector` processes. Killed them all. [13:34:43] Logged the message, Master [13:38:14] !log deployment-prep -squid ran puppet manually which deployed the new redirector from https://gerrit.wikimedia.org/r/53935 [13:38:16] Logged the message, Master [13:39:11] !log deployment-prep -squid started puppet service [13:39:13] Logged the message, Master [13:39:18] !deployment-prep -squid restarting squid [13:39:18] deployment-prep is a project to test mediawiki at beta.wmflabs.org before putting it to prod [13:41:03] !log deployment-prep -squid killed nrpe and restarted it (just to be sure) [13:41:07] Logged the message, Master [13:45:53] Coren hey [13:45:55] you wanted me [13:46:07] I will again, in about 15. :-) [13:46:11] k [13:46:42] !log wikidata-dev wikidata-dev-9: Installed Scribunto extension on dev client [13:46:46] Logged the message, Master [13:46:50] addshore ping [13:47:06] we have 3 irc bots and I failed to move all of them [13:47:18] they are too complex for grid :/ [13:47:36] too complex? [13:47:39] yes [13:47:53] one of them needs to listen on static IP address [13:48:05] second one is debianized as system service using data in local etc [13:48:11] third is just too complex :D [14:03:03] hey andrewbogott [14:03:17] do you think it's possible to rewrite labs-morebots so that it doesn't run as a service? [14:03:26] Sorry for the delay; some bozo was attempting to DOS my SIP. [14:03:27] so that it can be run from grid [14:03:41] people DoSing my boxes every week :P [14:04:04] btw there was a good blacklist somewhere [14:04:12] but it banned like half of planet [14:04:22] at least whole asia [14:04:29] There are several, but eagle830.startdedicated.com. seems like some bozo got a vps. [14:04:37] :/ [14:10:12] petan: hi. [14:10:40] Yes, it would be possible. I don't think we've decided where the official wikitech bot should live yet... [14:10:49] making that decision will largely dictate how it should be designed. [14:11:09] ok [14:11:20] I wanted to move current morebots but it doesn't seem to be possible :/ [14:11:50] Yeah, it's pretty entrenched as it is now. [14:12:12] Hm… I presume there's a way to run bots on the grid engine while still keeping things like passwords moderately secure? [14:12:34] andrewbogott: Yes to both grid architectures. :-) [14:12:48] (Through different means) [14:13:14] OK. petan, I'll make a note to review the morebot setup again… meanwhile probably best to just ignore those instances if that's OK [14:13:28] petan: pm check? :-) [14:18:49] !log wikidata-dev wikidata-dev-9: I had the same non-responsiveness again that we saw on Wednesday during demo time. Restarting apache helped again, though this time the server was idling around while last time apache was causing high load. Hm.... [14:18:51] Logged the message, Master [14:25:34] petan: going for lunch, also what does the i stand for in ibnr? [14:27:07] addshore interactive [14:27:20] that mean - interactive bots run there :P [14:27:40] (bots that needs to interact fast) [14:27:55] :) [14:28:08] I couldn't think of anything better :D [14:28:33] i would have just gone for nr1 ;p as its smaller also ;p [14:28:55] yes I was thinking the same, but some people might confuse it with previous nr1 which was for something else [14:29:14] "ib" stands for interactive bots while "b" is just big :D [14:29:25] haha [14:29:46] ahh xD [14:42:43] btw addshore what irc client are you using [14:42:46] :D [16:46:36] LOL [16:46:46] addshore once you figure out how to set up a default queue let me know [16:47:04] otherwise minimal become maximal soon [16:49:27] queue instance "minimalload@bots-ibnr1.pmtpa.wmflabs" dropped because it is overloaded: np_load_avg=1.010000 [16:49:31] :D [17:24:46] Coren: Are there different queues for SGE on Tools Labs? [17:25:07] Because the Toolserver has a queue for long time jobs and for short jobs [17:25:24] Jan_Luca: There are two at the moment, a "normal" one 'task' and another for long-running and continuous jobs 'continuous' [17:31:39] haha petan we will fix it :P [17:33:46] Hi guys! Can be installed git-review on bastion1? [17:35:28] Or which instance is better for submit patches on gerrit [17:35:43] beta16: For Mediawiki? [17:36:09] for an extension [17:36:22] Scribunto [17:36:50] beta16: Why you need a Labs instance and not use your own PC? [17:37:23] I think it is easier when you have not only a command line [17:39:37] beta16, um... you have a git repository on bastion? [17:40:21] I read Help:Git on wikimedia labs wiki, so I think that I can use labs [17:40:43] yes [17:40:49] but you shouldn't be doing stuff like that on bastion I don't think [17:41:33] In fact, I asked if another instance is better [17:42:01] beta16: You can request a project for this extension or use maybe mediawiki-dev project [17:43:35] ok, thanks Jan_Luca [17:43:47] for now i do in my pc. If continue I ask for a dedicated project [17:45:21] [bz] (NEW - created by: Chris McMahon, priority: Unprioritized - normal) [Bug 46166] Move automated browser tests from test2wiki to beta labs (tracking) - https://bugzilla.wikimedia.org/show_bug.cgi?id=46166 [17:46:21] [bz] (ASSIGNED - created by: Antoine "hashar" Musso, priority: Immediate - normal) [Bug 45084] autoupdate the databases! - https://bugzilla.wikimedia.org/show_bug.cgi?id=45084 [17:57:33] Coren: An idea: Could you add a wrapper around sudo -iu with the as parameter and not local-, e.g. "become "? [17:58:08] Jan_Luca: That... is surprisingly simple. :-) [17:58:31] Coren: But it reduce the confusion of new users :-) [18:06:42] Jan_Luca: {{done}} [18:06:59] Coren: The name is ? [18:07:05] Jan_Luca: 'become' [18:07:10] ok [18:07:16] Jan_Luca: Why mess with a good idea when it's a good idea? :-) [18:07:39] Coren: The idea is from Toolserver [18:07:52] Jan_Luca: Still a good idea. :_0 [18:07:52] for the MMP [18:08:14] Coren: That's right :-) [18:09:25] become also (a) checks that it's actually the name of a tool and (b) doesn't confusingly ask for a password if you're not allowed. [18:11:24] Coren: Can you install php5-cli on login and the submit hosts? [18:12:02] huh. Didn't even notice it wasn't there. [18:12:17] "The program 'php' is currently not installed. To run 'php' please ask your administrator to install the package 'php5-cli'" [18:12:52] No, I believe you, I'm just confused it's not already there. That's, like, one of the first things I install usually. :-) [18:13:26] Coren: There should be a puppet class for this ;-) [18:14:26] Yes, there should. I'm working on them, but the current cycle to get the +2 on the git review would destroy any semblance of responsiveness on my part while I'm still early in the setup. [18:14:38] Hence my taking notes whenever I make a change like this. :-) [18:14:40] Coren: I have created one... [18:16:21] Oh right... It was never merged... (https://gerrit.wikimedia.org/r/#/c/29975) [18:19:06] Coren: Do you already have a change for your puppet classes? When yes, I could review them [18:19:53] Jan_Luca: There is, but just a skeleton while I repackage some pips. [18:21:22] Coren: Now php5-cli is installed but php-mysql not... ;-) [18:21:51] Coren: Repackage = to .debs? [18:22:02] scfc_de: Aye. [18:22:38] Jan_Luca: I don't know what you're talking about. php5-mysql appears to be installed. Mysteriously. :-) [18:23:11] Coren: Now it is ... [18:23:44] Jan_Luca: How... mysterious. :-) [18:24:28] Coren: If your assembly line is running smoothly, you could add oursql and wikitools (abandoned ITP: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=594130) to your queue :-). [18:25:18] (Debian Policy is tedious compared to Fedora.) [18:25:29] oursql is already on my list of pips to debianize, so is a more recent version of requests [18:26:07] Change on 12mediawiki a page Wikimedia Labs/Tool Labs/Needed Toolserver features was modified, changed by Beta16 link https://www.mediawiki.org/w/index.php?diff=660225 edit summary: [+128] /* Database */ [18:26:26] Jan_Luca: Speaking of user-friendly, I now have a better way to start grid jobs (that includes qcronsub-like functionality): jsub [18:28:45] Coren: When you have some debs done, please drop a note (duplication of efforts :-)). [18:28:58] scfc_de: Ayup. [18:29:12] scfc_de: I'll have the pushed to the labs repo too. [18:29:43] Coren: Past, present, future, what? :-) [18:29:54] scfc_de: future. :-) [18:30:12] scfc_de: I.e.: as soon as they are tested. [18:30:55] Coren: The labs repo is enabled on all projects? [18:31:52] scfc_de: I think it is by default. [18:31:57] Coren: libgcc_s.so.1 must be installed for pthread_cancel to work [18:32:09] The first try of jsub :-( [18:32:16] o_O? [18:32:30] jsub works for me. That error message is... strange. [18:32:52] Wait, did you get that from your job in the .err file? [18:32:58] yes [18:33:17] Ah. Jsub worked, but your actual job didn't. [18:33:35] When call the script directly it works [18:33:55] o_O [18:34:35] -rw-r--r-- 1 root root 88384 Apr 15 2012 /lib/x86_64-linux-gnu/libgcc_s.so.1 [18:34:37] mlitn: https://github.com/MrMEEE/bumblebee-Old-and-abbandoned/commit/a047be85247755cdbe0acce6f1dafc8beb84f2ac#diff-1 [18:34:48] Probably the best single-character bug ever [18:34:59] ouch [18:35:05] wow that's harsh [18:35:34] That... ow! [18:35:57] Jan_Luca: Oh, wait, how much memory does that code of yours eat up? :-) [18:36:00] I of course meant to send that to -tech [18:36:02] Coren: do you think you can "install" the software described in https://wikitech.wikimedia.org/wiki/GiftBot#Dependencies on tools in a reasonable timeframe? [18:36:04] But irssi had other ideas :P [18:36:38] giftpflanze: 'fcron'? Whyfor? [18:36:44] Coren: I don't know, I have to show [18:36:58] Coren: i need timezone enabled cron, that's all [18:37:11] FastLizard4: I think that would have warranted to bump the major version :-). [18:37:22] scfc_de: Hehe, ikr [18:37:23] i haven't found it in the used distributions [18:37:30] giftpflanze: That's a breaking change, possibly, so I have to look into it before I can answer you. Can you open a bz for it, please? [18:37:38] Bugfix: Installing won't destroy operating system [18:37:40] :P [18:37:52] mh [18:38:03] Jan_Luca: Try to jsub with, say, '-m 1g' [18:38:39] Coren: is it better to backport cron or to install fcron? [18:38:52] BTW, fcron was (at least for quite some time) rather unmaintained. I think making it work with cronie should be much better. [18:39:44] giftpflanze: I don't know yet, that's why I need to look into it. Open a bz with your actual requirement and the fcron suggestion, and I'll try out and examine alternatives on a throwaway VM to find the best solution. [18:39:59] ok, thx [18:40:11] what about the tcl8.6 branch? [18:40:23] giftpflanze: That shouldn't pose a major issue. [18:40:32] ok :) [18:40:53] what about time deadline? [18:40:56] I think I could use the one from Raring. [18:41:12] what's Raring? [18:41:16] giftpflanze: Depending on the cron issue, I can probably have this all up for you sometime next week. [18:41:35] giftpflanze: the 13.04 release of Ubuntu. [18:41:44] oh, i forgot … [18:41:48] Coren: Now it works [18:42:01] What's the problem [18:42:23] Jan_Luca: The default maximum memory usage is a bit conservative. :-) [18:42:39] Jan_Luca: Once your job is done, you can use qacct to see how much memory it actually needed. [18:43:25] ... if accounting worked, which I haven't yet tested. [18:43:27] :-/ [18:44:01] Ah. Doesn't work from random submit hosts, only from the queue master. I'll write a wrapper for it. [18:44:05] Coren: /var/lib/gridengine/default/common/accounting: No such file or directory [18:44:26] Jan_Luca: What was your job number? [18:44:41] 102 [18:44:47] or name "linky" [18:46:22] giftpflanze: If the requirement for fcron is only the time zone bit, I think it's much easier to schedule your jobs for hours 22,23 and then test in your script if it is hour 0 in :Europe/Berlin. [18:46:59] scfc_de: true, but i want to avoid that [18:47:03] Jan_Luca: maxvmem 140.641M [18:47:03] Jan_Luca: the default max is 128M :-) [18:47:35] i'm not responsible for ubuntu being so super-stable ;) [18:47:39] Coren: The most of this is php [18:47:43] itself [18:48:11] Jan_Luca: Hm. good point, php has a fairly large footprint. I'll bump the default up to 256M [18:48:21] Since there'll be a lot of small php stuff [18:48:57] … and not keen on reinventing the wheel [18:48:58] Jan_Luca: Should work now with no need to -m [18:49:07] Ok I test it [18:50:20] * Coren is not entirely surprised that php doesn't deal well with ulimits [18:50:45] For some stupid reason, error checking has never been a forte of the php devs. [18:51:24] Coren: Waiting for immediate job to be scheduled. [18:51:24] Your qsub request could not be scheduled, try again later. [18:51:41] Jan_Luca: Did you try a continuous job? [18:51:43] giftpflanze: Last time I looked, fcron didn't work with SELinux enabled, and for a system daemon that's usually not a good sign. [18:51:48] yes [18:52:01] Jan_Luca: Hm. From your /tool/ account? [18:52:05] yes [18:52:40] Damn. I know why it did that. Silly -now semantics. Gimme a sec [18:53:02] Jan_Luca: You should be able to do it now. [18:53:09] A moment please [18:53:29] Jan_Luca: I tried something to make it less strange for users, but it doesn't quite work like I would have hoped. [18:54:12] Jan_Luca: Undocumented feature: non-tool users cannot run continuous tasks. There is no message, your job will just sit forever quietly in queue and never run. [18:54:22] scfc_de: i don't require it, it just was the only tz one easily available on bots before it was broken (although i like some details of fcrontab(5) :)) [18:54:28] Jan_Luca: I'll find a workaround to tell the user why it's not working. [18:55:06] Coren: Now it works but when using -once there is no error when tring to queue it a second time [18:55:24] ... really? [18:55:36] yes [18:55:36] Does it actually start? [18:56:05] Because, by default, the error will have been put in the .err file, not on the console. :-) [18:56:06] Coren: Uups, there error is in the .err file [18:56:51] Jan_Luca: Yeah, that's because the typical use of -once is from cron, and you (a) don't want email spam and (b) want to have a trace that it failed. :-) [18:57:09] (You can add -stderr if you want to see the fail on console) [18:57:30] Coren: But why is the output of "echo ...;" in the .err file and not in the .out one? [18:57:48] Do you use "-j y"? [18:58:36] Jan_Luca: Not by default - but if you're talking about the echo /in the script/, notice the >&2 at the end. :-) [18:58:57] It's an error message, so I send it to stderr. :-) [18:59:40] Wait, no it isn't. [18:59:41] Hm [18:59:55] Coren: PHP's echo is a error message? [19:00:02] Oh! php's [19:00:14] Oh. No. I see the bug now. [19:00:50] Coren: In jsub there is the line: '-a' => 1, '-b' => 1, '-cwd' => 0, '-e' => 1, '-hard' => 0, '-i' => 1, !!!! '-j' => 1, !!!! [19:00:51] Fixed. Future jobs won't mix output with input. I used the wrong default if you didn't use an explicit -o option. [19:01:23] Jan_Luca: Yes, that line is there. Why? :-0 [19:01:44] I mean the "'-j' => 1" [19:01:58] this seems to set -j to true [19:02:12] Jan_Luca: Ah. You misunderstand; that is the list of /allowed/ qsub options. :-) [19:02:24] The 1 means 'takes argument' [19:02:40] OK [19:03:04] There was a bug where I accidentally set -o to the error file if you didn't specify one manually. That's been fixed now. [19:03:29] So future jobs won't mix output and error. [19:06:35] Coren: Now it's goog [19:06:54] good [19:08:03] Yeay! [19:08:28] BTW: Other shortcut: [19:08:35] job [19:08:56] Give the job number if that job is running [19:09:05] job -v is the human readable version [19:09:36] Coren: On Toolserver, a major hassle was PATHs being different in interactive and cron environments. cron sets PATH to /usr/bin:/bin, so it would be nice if your scripts - properly packaged :-) - would be installed to /usr/bin. [19:10:43] scfc_de: That breaks the FHS [19:11:07] Coren: Why? [19:13:02] scfc_de: Machine-local stuff is supposed to always be under /usr/local. Though I admit I'm probably more anal than most about the FHS. :-) [19:14:18] * Coren is the kind of dude that makes the careful distinction between /var/spool, /var/run and /var/cache :-) [19:16:37] Coren: Well, the FHS just says "The /usr/local hierarchy is for use by the system administrator when installing software locally." That fits *all* installed software :-), and for example SGE is installed in /usr/bin (and if you can live with log files in $HOME - *brrr* - I don't think you're that anal :-)). [19:19:47] scfc_de: What makes you think that I don't abhor the stupid SGE log file stupidity? :-) [19:20:12] * Coren needs to afk for a bit. BBL [19:21:09] scfc_de: But, honestly, you should /never/ rely on $PATH in a crontab. [19:32:42] Coren: The location is configurable in SGE, and *you* put the httpd logs there as well :-). The problem on Toolserver was (and is) that the interactive PATH changed from time to time and on Solaris included for example version numbers of various packages. So I think it is very useful to have all programs in /usr/bin as that PATH will be in crontab and will not need to be updated anytime (or update cronie to include /usr/local/bin in [19:32:42] PATH). [19:34:11] scfc_de: Yeah, I'm not sure I don't prefer making sure that cronie knows about /usr/local/bin instead, but I'll think about it and see if I can do something elegant yet simple. :-) [19:34:22] And now, I leave for real [19:35:11] Okay :-). [19:43:14] * Damianz reads 'it would actually be nice if we had a cPanel or Webmin setup ' and stabs a kitten [19:57:12] [bz] (NEW - created by: m.p.roppelt, priority: Unprioritized - normal) [Bug 46170] Evaluate installing timezone aware cron - https://bugzilla.wikimedia.org/show_bug.cgi?id=46170 [19:57:13] [bz] (NEW - created by: Chris McMahon, priority: High - normal) [Bug 46166] Add automated browser tests to beta labs (tracking) - https://bugzilla.wikimedia.org/show_bug.cgi?id=46166 [20:22:39] Coren|AFK: drdee: did you hear where the next pycon is? hint: very short commute :) [20:24:04] wtf is hashar doing in topic [20:24:24] that was a while ago [20:24:36] I need to make wm-bot log topic [20:24:37] and if you read the topic so much then you should have seen it earlier!! [20:24:38] :D [20:24:47] what was there [20:25:05] (petan was saying he reads it often) [20:25:22] I didn't say that [20:25:24] but I don't have backlog [20:25:27] I restarted bnc [20:25:29] :/ [20:25:37] because of lmao bug [20:25:43] 15 19:55:38 < petan> at least I am reading them when i join :P [20:25:54] I created a perfect constructor :D [20:26:09] * jeremyb_ runs away [20:26:13] jeremyb_ yes but usually I am not joining [20:26:18] because I am idle for days [20:26:39] 15 15:21:38 -!- petan [~pidgeon@wikimedia/Petrb] has quit [Remote host closed the connection] [20:26:45] indeed [20:26:47] 15 15:21:38 -!- petan [~pidgeon@wikimedia/Petrb] has quit [Remote host closed the connection] [20:26:51]