[00:06:08] Cyberpower678: I figured out the db issue [00:06:37] Betacommand, send the details to me in memoserv. My work break is over now. [00:06:46] Gotta go. [00:07:04] Cyberpower678: sendt [08:03:43] Coren: ping [08:09:42] Coren: ping [09:16:40] Coren: ping [09:27:03] Coren: Where are you? [09:31:38] zhuyifei1999: i hav eating ^^ [10:18:47] zhuyifei1999: given timezones probably sleeping [10:19:20] Betacommand: ok [11:26:28] Coren: ping [11:39:13] petan: when is Coren going to wake up? [12:01:47] petan: [[en:WT:WikiProject Articles for creation#Category:AfC submissions with missing AfC template]] is looking for you with an AFCbot question. [12:01:54] @link [12:01:54] http://enwp.org/WT:WikiProject_Articles_for_creation#Category:AfC_submissions_with_missing_AfC_template [12:05:37] T13|needsCoffee: what is needed from me [12:06:12] zhuyifei1999: unfortunatelly I don't share room with Coren, so I have no idea when he wakes up :P [12:07:10] hasteur was paging you because apparently AFCbot hasn't done anything since 28 Aug 2013. [12:15:22] petan: That's correct [12:17:34] aha [12:17:40] might have died? [12:18:01] or got blocked :/ [12:19:22] LOL [12:19:37] someone enforced https on all wikimedia sites without some kind of preparation I guess [12:19:55] well afcbot doesn't support https [12:20:05] so it can't work in this moment [12:20:12] Whell POO! [12:20:40] that's trully funny to enforce https on wikipedia [12:20:45] what a silly idea [12:21:08] Don't blame the sysops, blame the herd mentality [12:21:31] I don't really blame anyone I just think it's silly [12:22:15] especially given that 90% of traffic comes from people who don't even use any account and don't give a fuck about security [12:24:46] petan: How do you populate en:User:Petrb/Wierd pages? I might slap together a quick script to traverse the list and finish out the current set. Do you have any objection as a temporary workaround? [12:25:08] that is NOT a current set [12:25:13] it's a very obsolete set in fact [12:25:18] current set is on labs [12:27:51] Hrm... I'll see if I can get with you in ~8 hrs to see if I can put together a temporary script to populate up the list. [12:37:22] @link [[en:User:Petrb/Wierd pages]] [12:37:22] http://enwp.org/User:Petrb/Wierd_pages [12:48:27] Coren: ping [13:12:02] !ping [13:12:02] !pong [13:12:35] !ping [13:12:35] !pong [13:19:35] Coren, ping [13:22:09] Coren: ping [13:22:25] poor C oren [13:22:40] petan: :P [13:23:10] you guys know I can help you if it's not DB related? [13:23:35] petan: I know [13:23:47] but it's DB related [13:26:04] petan: see PM [13:29:15] petan, the cyberbot node doesn't seem to be working right. [13:29:48] node? [13:29:51] what do u mean [13:30:01] don't tell me cyberbot has own node XD [13:30:22] Coren created a new node for me with a lot more available job slots. [13:32:22] aha [13:32:24] name? [13:33:20] I got it [13:33:38] Coren: ping [13:33:58] petan, do you know what's worng? [13:34:01] zhuyifei1999 maybe create a ticket for that instead of pinging him? [13:34:08] Cyberpower678: you didn't describe the problem [13:34:37] I submit jobs to it and it deosn't show up in qstat. Those that do can't seem to access the scripts. [13:35:03] petan: I'd rather ping [13:35:30] hmm [13:36:00] Cyberpower678 maybe because you have many jobs currently running on other nodes? [13:36:12] makes no sense. [13:36:18] how you start them? [13:36:36] same as I start normal jubs but with -q cyberbot [13:37:01] does it say anything as error? [13:37:15] Can't access source file. [13:37:21] aha [13:37:23] that would be it [13:37:37] is that error from qsub or jsub? [13:37:43] jsub [13:39:16] :/ [13:39:33] maybe try with qsub then? [13:39:41] jsub is just a wrapper for qsub [14:04:11] [bz] (8ASSIGNED - created by: 2DrTrigon, priority: 4Unprioritized - 6enhancement) [Bug 53625] Install python module opencv (v1 and v2) - https://bugzilla.wikimedia.org/show_bug.cgi?id=53625 [14:07:59] Coren: ping [14:10:26] zhuyifei1999: what do you need? [14:11:11] Betacommand: the question is: where's the table "text" [14:11:44] zhuyifei1999: not sure its available [14:12:16] so I have to ask coren [14:13:10] zhuyifei1999: Im actually fairly sure that due to how we use external text storage we dont have access to the text field [14:13:31] Betacommand: ??? [14:14:31] Coren, ping [14:14:57] /grep Coren.*ping for this channel looks funny [14:15:18] zhuyifei1999: the WMF doesnt use the text table, they store it in a external format [14:15:46] Betacommand: so only api can access? [14:16:11] zhuyifei1999: that or using a dump [14:16:26] let me grab the docs on text storagd [14:16:44] Betacommand, thanks. I'm such an idiot. [14:17:02] I mixed up the db connection settings. [14:17:21] I attempted to connect to enwiki_p via tools-db [14:17:24] :/ [14:18:15] zhuyifei1999: https://wikitech.wikimedia.org/wiki/External_storage [14:20:38] petan, the cyberbot node doesn't seem to have access to the files. [14:20:43] zhuyifei1999: so yeah, you need to use the API [14:23:07] grabbing ~70,000 pages from the api? [14:24:51] if you need to grab 70,000 pages, use the dump then [14:26:14] Cyberpower678 what files? [14:27:17] petan, brb [14:34:43] YuviPanda, I'm running into trouble generating gpg keys because the VM is entropy-bound. That said, I think that things should be mostly working now… can you install the package without complaint? [14:34:58] testing now [14:35:17] also haha, I just was reading about entropy sources and the problems they can cause to VMs today, and here it is... :D [14:35:19] * YuviPanda sshs [14:36:39] YuviPanda: is it possible to read within a script? [14:36:53] I mean the dump [14:37:10] yeah, but they're kinda big and you'll have to write your code to take that into account [14:37:21] andrewbogott: no luck, still 'cannot be authenticated!' failures on puppet run [14:37:27] crap [14:37:36] andrewbogott: hmm, perhaps I need to run puppet twice [14:37:43] andrewbogott: once for labsdebrepo to do its thing [14:37:48] and then for it to take effect [14:37:52] let me try that [14:38:25] YuviPanda: use which one of http://dumps.wikimedia.org/commonswiki/20130827/ ? [14:38:36] I tried with a normal apt-get install; does that not complain about unsigned repos? [14:38:37] zhuyifei1999: they're already there in toollabs [14:38:42] andrewbogott: that also did [14:39:01] for all category, currrent version only? [14:39:11] YuviPanda: where? [14:39:21] OK, short of duplicating your puppet run, how can I test this? [14:39:37] zhuyifei1999: https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Help#Public_dataset_dumps [14:39:49] andrewbogott: ssh to proxy-dammit and run puppet? [14:39:50] :D [14:40:01] andrewbogott: it seems 'stuck' on labsdebrepo [14:40:05] andrewbogott: on the puppet run [14:40:09] andrewbogott: probably looking for entropy? [14:40:24] Um… wait, which problem are you getting? A minute ago you said 'cannot be authenticated' [14:40:52] andrewbogott: so I ran puppetd -tv [14:40:57] andrewbogott: and it complained 'cannot be authenticated' [14:41:01] andrewbogott: and failed installing [14:41:06] andrewbogott: however, puppet itself continued to run [14:41:18] andrewbogott: and it is currently at the labsdebrepo part [14:41:20] let me do a paste [14:41:30] I'm running it too, and I get different behavior... [14:41:37] everything runs except for the timeout [14:41:38] andrewbogott: https://dpaste.de/XYytG/ [14:42:06] what the heck? Totally different from what I see on the same VM [14:42:06] andrewbogott: aaha, and now I get a timeout [14:42:08] for keygen [14:42:15] err: /Stage[main]/Misc::Labsdebrepo/Exec[keygen]/returns: change from notrun to 0 failed: Command exceeded timeout at /etc/puppet/manifests/misc/labsdebrepo.pp:49 [14:42:18] notice: Finished catalog run in 315.99 seconds [14:42:19] Anyway… is there a standalone command that will test for the status of that package? [14:42:34] YuviPanda: which to use for all category, currrent version only? [14:42:44] andrewbogott: sudo apt-get install nginx-extras? [14:42:50] zhuyifei1999: I think it is current version only? [14:43:17] YuviPanda: revision [14:43:45] zhuyifei1999: I think so, yeah. Look at the files? [14:43:54] zhuyifei1999: addshore worked on them a while ago, he might have more details [14:44:08] @notify addshore [14:44:08] This user is now online in #huggle. I'll let you know when they show some activity (talk, etc.) [14:45:02] YuviPanda, are you in another puppet run? I can't apt at all, it's locked. [14:45:10] andrewbogott: nope [14:45:11] YuviPanda: thanks [14:45:24] andrewbogott: ah, try again? [14:45:33] yep, better [14:45:38] andrewbogott: I had tried an apt-get install nginx-extras and let it hang. Sorry about that [14:45:41] killed it now [14:45:46] should be unlocked [14:47:22] petan, I'm back. [14:47:30] zhuyifei1999: if you are looking for category based information, you should also look at https://wikitech.wikimedia.org/wiki/Nova_Resource:Catgraph [14:47:39] So I execute the same commnd and add -q cyberbot [14:47:59] The cyberbot node can't seem to access the scripts in the cyberbot project. [14:48:16] Cyberpower678: I couldn't send you with memoserv: I don't know your freenode registration username. :-) [14:48:39] Cyberpower678: Your problem was that your commandline had -mem 8m. 8m isn't even enough to start the shell that will start your command. :-) [14:49:16] Oh? [14:49:30] Coren, err. My username is Cyberpower678. :p [14:49:57] Huh. I tried that and it said it didn't know you. Maybe I had a typo. [14:50:06] Coren, how much memory do I need? [14:50:28] It depends on what you are trying to run. As a rule, PHP is happy with ~350m [14:51:00] That means, I will eat up my 2gb in no time. [14:51:17] Cyberpower678: No, no, don't confuse vmem and actual RSS [14:51:30] dammit, commandline apt-get doesn't care about repo signing. So all of my testing to this point was bogus [14:52:02] If you are running all PHP scripts, you can probably fire dozens of those before you actually run afoul of 2G of ram. [14:52:04] andrewbogott: oh? it does to me [14:52:13] WARNING: The following packages cannot be authenticated! [14:52:13] nginx-common nginx-extras [14:52:14] Install these packages without verification [y/N]? [14:52:59] Mine just says [14:52:59] After this operation, 1,688 kB of additional disk space will be used. [14:53:00] Do you want to continue [Y/n]? [14:53:05] andrewbogott: say Y [14:53:10] andrewbogott: it will give you the warning after that [14:53:15] Oh, dang. OK. [14:55:50] andrewbogott: There is a workaround. [14:56:44] andrewbogott: Add [ trusted=yes ] to the *.list pointing at the repo. [14:57:19] Which, in a case of a repo we control, is true. :-) [14:58:31] okay, brb [14:59:31] Well, dammit, that sounds a lot easier [15:00:07] Coren, why does it need that much memory to start? [15:00:28] Also can anybody access the cyberbot queue? [15:00:45] Cyberpower678: That's the sum of php, the memory it allocates, and all the dso its linked to. [15:00:52] Cyberpower678: No, you alone. [15:01:49] Tasks successfully started. [15:07:52] Coren, I've transferred all, but 5 tasks to the cyberbot node. [15:08:06] Why not the other five? [15:08:07] !logs [15:08:08] http://bots.wmflabs.org/~wm-bot/logs/%23wikimedia-labs/ [15:08:09] !logs del [15:08:10] Successfully removed logs [15:08:23] !logs is raw text: http://bots.wmflabs.org/~wm-bot/logs/%23wikimedia-labs/ cute html: http://tools.wmflabs.org/wm-bot/logs/index.php?display=%23wikimedia-labs [15:08:24] Key was added [15:08:37] Coren, 2 of them are in the middle of doing something, and the other 3 are python scripts. [15:09:02] Coren, what happens when I hit the memory limit? [15:09:06] Cyberpower678: Python also works on your queue; my comment about "all PHP" was re the actual footprint. [15:09:39] Python, eats a lot of memory. [15:10:00] Cyberpower678: The box will get fairly slow as it starts trashing to swap. If you go *way* over, the OOM killer will start murdering your bots. But you're using, like, 12% of your memory atm so it's not panicky. :-) [15:14:20] Coren, according to Ganglia, where my node is also visible, I'm using 50% [15:18:16] Cyberpower678: You're not looking at the right numbers; your current usage is (in MB) 247 / 2003. When you look at the ganglia stats; you have to ignore the green regions (buffers, cache). [15:20:13] Cyberpower678: You get your own node, but only if you stop using the general ones. :-) [15:23:27] I will move them in due time. [15:24:35] Well, the time is due now really -- but I don't mind leaving the general queues open for you over the weekend so that you can do so when you have time to yourself. [15:25:21] the spamscan task won't finish until 9/13 and the spambot taks won't finish until tomorrow. [15:26:51] Oh, no worries, I'll not kill any of your jobs. [15:27:22] Just make sure their next runs, at least, or on the right queue. [15:27:43] Coren, trusted=yes on the source line? "deb file:///data/project/repo/ / trusted=yes" doesn't parse [15:28:09] andrewbogott: "deb [ trusted=yes ] file:///data/project/repo/ /" [15:28:18] ah, ok, thank you. [15:29:50] Coren, I like how my own exec node shows up with all the other exec nodes. :D [15:30:36] ... [15:30:48] did we really need whole exec node? [15:36:50] petan, it was Coren's alternative to raising the job slot's limit. [15:36:54] I accepted. [15:39:30] petan: It's a more efficient use of resources for edge cases like Cyberpower. [15:40:17] I just don't think that cyber bot is edge case :P IMHO it's just another bot that does a lot of tasks... [15:40:37] 16 processes may eat less resources than 1 heavy process [15:40:50] he could as well create 1 task which runs in 16 threads [15:41:25] I mean, he does run lot of processes, but that doesn't really mean it's resource expensive bot [15:42:10] petan: He could have, but this we he specifically gets 1 cpu and 2G of ram to manage as he sees fit, and can neither consume slots nor resources from the other tools. Win-win. [15:42:20] ok [15:42:20] s/this we/this way/ [15:43:07] * Cyberpower678 is extremely happy, so another win. [15:44:36] drugs can accomplish that too [15:44:44] -.- [15:44:57] That wouldn't solve my bot problem. [15:45:03] who knows :P [15:49:30] YuviPanda|brb: I think it's happy now. [15:56:14] andrewbogott: swell! checking [15:56:31] * YuviPanda puppets [15:57:25] !logs [15:57:25] raw text: http://bots.wmflabs.org/~wm-bot/logs/%23wikimedia-labs/ cute html: http://tools.wmflabs.org/wm-bot/logs/index.php?display=%23wikimedia-labs [16:00:53] andrewbogott: works! [16:00:54] yay, ty! [16:08:56] Coren, I've updated the crontab. The only scripts that are still running in the normal nodes are those 5 I mention earlier. [16:10:01] Cyberpower678: Cool beans. [16:16:36] Coren: btw, IIRC we don't have snapshots anymore, so perhaps update https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Help#Time_travel? [16:17:10] They're coming back soon, almost certainly during the two weeks I'll be in SF. [16:30:37] Coren, if the grid fails, does that mean my node will keep running on it's own? [16:30:54] Leaving my bot operational? [16:31:15] That depends what you mean by "grid fails". If you mean the scheduler, then yes. If you mean your node specifically, no -- but the jobs will get restarted when it comes back up. [16:42:55] * valhallasw wonders why Eclipse always makes him want to smack his head into a wall [16:43:25] valhallasw: why are you using Eclipse? [16:43:54] YuviPanda: I'm trying to see if it makes using Gerrit from windows less painful [16:44:05] That was before I remembered it's Eclipse. [16:44:07] hahahahahahaha [16:45:01] Internal error validating repository [16:45:02] java.lang.reflect.InvocationTargetException [16:45:03] \o/ [16:45:30] \o/ [16:46:36] basically, the screenshots always look really nice: oooh, integration with both gerrit and bugzilla! let me try that! [16:46:49] but it is Eclipse. Never Forget! [16:47:46] ok, so it doesn't want to login to gerrit [16:48:41] and now it's in an infinite loop! [16:49:28] and of course, clicking the 'cancel' button does nothing except graying out the 'cancel' button. [16:50:38] * valhallasw cries in a corner [16:50:59] * YuviPanda pats valhallasw [16:51:01] there there [16:51:08] this is why you don't use eclipse [16:53:47] ok, let's try the nightly build of the connector [16:54:43] Gerrit 2.5 is supported. Support for Gerrit 2.6 is still work in progress [16:54:47] Well, that explains. [16:55:51] I'd almost consider paying for pycharm, which is also supposed to have gerrit integration [16:56:08] they have open source licenses, after all... [16:56:58] http://i.imgur.com/TS77f.jpg :D [16:57:08] valhallasw: Writing a stupid-simple plugin for our stripped down workflow might not be too hard [16:58:35] I'd almost consider learning COM programming to interface with TortoiseGit ;-) [16:59:04] Oh, right, Windows. Never mind. [17:12:49] heh [17:12:58] * YuviPanda 's workflow doesn't involve git review anymore [17:13:47] Coren: you around? [17:20:30] how do you take ownership of all files and sub directories of a folder? [17:25:51] Betacommand: use 'take'? [17:25:59] Betacommand: 'take '? [17:26:31] YuviPanda: that will work for sub files too? [17:26:49] Betacommand: yes, it is recursive by default, IIRC [17:27:47] YuviPanda: what do you do instead of git review? [17:28:04] chrismcmahon: git push gerrit HEAD:refs/for/master [17:28:55] YuviPanda: huh. istr at one point that didn't work. guess it changed. [17:29:08] chrismcmahon: really? that's always been the official way to get changesets into gerrit [17:29:20] git review isn't really endoresed by the gerrit folks [17:29:30] I might be wrong then [17:29:54] <^d> At its core that's what git-review does. [17:29:59] <^d> Plus some other rebasing black magic. [17:30:15] yeah, and I guess now that I prefer doing the black magic myself, if needed... [17:30:19] <^d> YuviPanda: And aliases, man. `git push-review` [17:31:01] ^d: hmm, for some reason I never have gotten big into aliases. [17:31:37] * Betacommand is finally finishing up his labs move [17:32:22] <^d> [alias] [17:32:22] <^d> repack-everything = repack -a -d -f --depth=250 --window=250 [17:32:23] <^d> push-review = push origin HEAD:refs/for/master [17:32:24] <^d> up = pull --ff-only [17:32:26] <^d> log-graph = log --graph --oneline --branches --all [17:32:28] <^d> rollback = reset --hard HEAD~1 [17:32:30] <^d> amend = commit -a --amend [17:32:32] <^d> crap = reset --hard origin/master [17:32:34] <^d> sub-up = submodule update --init [17:32:36] <^d> YuviPanda: ^ [17:32:41] haha@ crap [17:33:05] <^d> rollback is nice too. [17:33:07] <^d> I do that one alot. [17:33:56] you repack a lot? [17:36:22] ^d are other people besides me getting the "unpacker error" when trying to push? [17:37:28] chrismcmahon: I got it y'day, a couple of other people did too, IIRC [17:37:51] YuviPanda: just lucky I guess [17:45:48] <^d> chrismcmahon: Yes, it's known, across multiple repos. [17:57:35] OK, Eclipse, it was nice seeing you again. Farewell. [17:57:40] rm -rf eclipse [17:58:03] :D [17:58:14] valhallasw: try IntelliJ and friends, if you have to [17:58:25] YuviPanda: yeah, that's pycharm [17:58:35] right [18:09:07] ^d: fwiw, my push just went through. not sure what you did, but thanks! [18:12:01] Betacommand: Am now (just done with lunch). What can I do to you? [18:12:28] Coren: fixed it already [18:12:31] Heh. Should have been "ohcrap = reset --hard origin/master" [18:13:06] Coren: do you know how to get git-review working in windows? [18:15:20] Betacommand: Except for my gaming box, I haven't used Windows since well before git existed. I know we have a number of Windows users on labs-l though, so asking there will probably find at least one who has some git-fu. [18:15:32] Betacommand: I'm not sure anyone uses git review on windows [18:15:48] Betacommand: ask valhallasw, or Reedy, perhaps? [18:15:58] I don't [18:16:04] well, valhallasw then [18:16:11] I do all git stuff (well, gerrit based) stuff in a linux VM [18:16:15] YuviPanda: Ive been talking with valhallasw and no luck [18:16:19] I think siebrand does git-review from Windows [18:16:33] IIRC, petan also does some Windows; and I know he does git reviews, so perhaps he does them from there. [18:16:43] * Betacommand wonders why the hell this has to be such a pain in the ass [18:17:08] Betacommand: Your question mentionned "Windows"; that in itself is the answer. :-) [18:17:09] Does it play ball under cygwin? [18:17:24] Reedy: I'd expect it would. Cygwin is posixy enough for most things. [18:17:25] Coren: svn works like a charm [18:17:27] well... [18:17:32] Betacommand: xqt had mixed results with tortoisegit, but it's not great [18:17:41] Betacommand: you don't need to use git review for gerrit [18:18:01] Ryan_Lane: then how does one submit patches? [18:18:03] also, if you install python for windows, installing git review should be relatively straightforward [18:18:16] Ryan_Lane: ive installed it fine [18:18:17] Ryan_Lane: no, it's not, because it needs both command line git and command line ssh [18:18:19] git push origin HEAD:refs/for/master [18:19:00] you also need to add the git hook for gerrit [18:19:06] Betacommand: one more reason to have github integration, really - github has a nice windows client. [18:19:23] valhallasw: or just say fuck git and use svn [18:19:28] :P [18:19:33] From my experience today, I can say pycharm's gerrit plugin and eclipses gerrit plugin are both crap. [18:19:39] Betacommand: you can do that with github. [18:23:38] trust me. use git for a while and you'll never understood why you liked svn [18:24:56] Ryan_Lane: Ive been trying and just cursing out git the whole time [18:25:02] Ryan_Lane: compare tortoisesvn and tortoisegit, and you'll understand. [18:25:14] I use tortoisesvn at work [18:25:15] Ryan_Lane: try running gerrit integration on windows, and you'll understand [18:25:27] easy. don't use windows [18:25:35] in 10 years windows will probably not exist anymore [18:26:15] windows is a piss-poor environment for OSS devs [18:26:45] that said, git-review is just there to make things easier [18:26:49] it's not needed at all [18:26:54] ^d, for instance, doesn't use it [18:27:21] <^d> s/doesn't use it/refuses to use it/ [18:27:21] as long as you have the necessary git hook installed, and you push origin HEAD:refs/for/master you're fine [18:29:15] I really don't get the purpose of tortoisesvn/git they just hide things from you [18:29:24] Ryan_Lane: the point is there are people who would like to contribute who are also running windows. We need to help these people instead of sending them away with a 'well, better switch to linux!' [18:29:40] yeah, we do [18:29:42] Not everyone is a power-user. Some people just have a patch they would like to submit. [18:29:46] agreed [18:29:53] valhallasw: well, you were missing when I said you don't need git-review [18:29:58] as long as you have the necessary git hook installed, and you push origin HEAD:refs/for/master you're fine [18:30:40] I have git on windows :\ [18:31:09] valhallasw: it's not a matter of switching to linux. installing a linux vm is simple [18:31:38] For most people, it's not. [18:31:44] I understand where you're coming from, but Windows is a relatively hostile environment for OSS devs [18:32:06] I'd rather call it an 'impedance mismatch'. [18:32:35] the windows way of doing things - with an explorer plugin et cetera - does not match well with a command-line based workflow [18:32:45] microsoft encourages its own platform, and does absolutely nothing to enable OSS dev [18:34:00] * Ryan_Lane shrugs [18:34:19] there's CoApp [18:34:29] either way, it's just OS wars. the problem is that most devs don't use it, and it isn't supported well. [18:34:46] Ryan_Lane: if you want the counterexamples to how MS actually does try to encourage OSS dev (albeit not through Windows changes), I can give 'em [18:35:27] sumanah did you see http://tools.wmflabs.org/wm-bot/beta/ ? I think you might like it... [18:35:38] thank you petan! [18:35:41] once I implement some search etc [18:35:45] I had not taken a look yet! [18:35:55] it even display part / joins if you request it [18:35:55] actually I do quite a bit of open source python programming in windows that I port to other OSes without any issues [18:36:09] cool! [18:36:36] Ive had quite a few pywiki patches over the years [18:36:44] I've had ... 1 :) [18:36:51] I started working on pywikipedia in windows, and I know at least two pywikipedia developers who like to develop from windows. [18:36:53] basically fixing the copyright year on some files to include 2013 [18:37:19] valhallasw: Im a third [18:38:02] valhallasw: i'll see if I can get g2g back up at the tech days [18:38:07] valhallasw: or I can give you access to the tool account... [18:38:16] yeah, Windows support is good. YuviPanda I didn't know GithubToGerrit was down! D: [18:38:18] Ive got a gerrit account and was trying to submit patches per the documentation and getting pissed off [18:39:01] sumanah: it's been for a week, yeah. I rewrote it to make it a lot more robust, and then haven't had time to swtich it on [18:39:05] nod [18:39:14] I ended up saying fuck it, and making valhallasw commit the, [18:39:17] *them [18:39:28] Betacommand: in the meanwhile, feel free to mail me patches at any time. [18:39:41] at least Git makes it easy to distinguish author & committer [18:39:43] I might take a few days to actually submit, though :-) [18:39:55] valhallasw: Ill look at my code base and see what I should push upstream [18:39:59] valhallasw: in the meantime, also feel free to just send pull requests. I can manually sync fast enough :) [18:40:27] * Betacommand has a shitton of custom code [18:41:28] petan: I found a bug. http://tools.wmflabs.org/wm-bot/beta/index.php?start=07%2F01%2F2013&end=09%2F06%2F2013&display=%23wikimedia-office is only showing me stuff from Sept 5 [18:41:44] YuviPanda: not sure what I can send requests on, though, as I don't know what does what ;-) [18:42:19] sumanah: yes I didn't convert text logs to mysql, so older data are available in txt logs only [18:42:26] valhallasw: :) i get notified on any pull requests to github.com/wikimedia/* [18:42:29] oh, sorry for calling it a bug petan [18:42:38] no problem :D [18:42:41] YuviPanda: oh! I see :-) [18:42:44] Betacommand: ^ [18:42:53] valhallasw: does https://www.mediawiki.org/wiki/Developers/Maintainers help at all? :) [18:42:56] oh, *what* does what [18:43:16] Betacommand: I'll try to see how well the github windows client works. [18:44:07]