[00:00:15] Or perhaps Microsoft might sponsor some research? [00:00:22] Fae: you are welcome to set up a project and get it reviewed by the Research Committee [00:01:06] Okay I get the message. I'll go away and the trustees can take the WMF position that you don't believe this is a problem until we get legal advice. [00:01:13] and you might be delighted to know that WMDE is running a research project in partnership with Google and Telefonica [00:01:29] I look forward to seeing those free adverts. [00:01:47] I cannot respond for legal matters, but only as far as the RCom support is involved [00:03:10] Okay, I have put my case forcefully enough, there seems not point in repeating it. [00:03:53] I'll work on it tomorrow with our Chair and/or CEO. [00:04:52] Thanks Fae, I hope we can reach an understanding of the legal implications of this project for your chapter [00:04:56] Krinkle, are you responsible of the HTTP-only logos in the banner? :) [00:05:10] I checked everythign for // [00:05:32] Krinkle, reedy had to fix it https://meta.wikimedia.org/w/index.php?title=MediaWiki:Centralnotice-template-HSP_final&diff=3135155&oldid=3134785 [00:05:54] DarTar: Sure, shame we resort to it being a legal issue rather than resolving it as a community and a common sense interpretation of our values. [00:06:01] Fae: I am really sorry you take this as an ad, especially given the huge community compliance we tried to have throughout this long process... [00:06:12] but, Krinkle, next time this filter will disallow it! :-p https://meta.wikimedia.org/wiki/Special:AbuseFilter/42 [00:06:24] (advanced wheel wars) [00:06:28] Nemo_bis: I doubt CentralNotice will abide that [00:06:37] since it saves in the background [00:06:40] it does [00:06:50] hmm, I didn't test from the special page [00:06:51] I am happy to discuss the community and research implications, but you seem to raise legal issues and I think our Legal team is in the best position to address them [00:06:55] okay, whatever. It was all protocol-relative, 100% sure [00:07:01] Nemo_bis: Hi. [00:07:06] they probably made changes and reverted that unintentionally [00:07:16] Joan, ;-) [00:07:17] Jerome_: Email me a link to the community consensus on the banner, it will help my discussion tomorrow. [00:07:25] Mmmmm [00:07:27] Duplication [00:07:28] Nemo_bis: The abuse filter only catches CSS URLs currently. [00:07:34] It needs to be trained to do tags too. [00:07:41] Joan, I changed it a bit [00:07:44] \o/ [00:07:50] but just a bit [00:07:53] I don't know HTML [00:07:56] Fae: I invite you to read the project page on Meta if you haven't done so [00:08:08] note url() can have '' or "" as wlel [00:08:19] hence them doing the 2 seperate parts [00:08:25] Jerome_: please make sure all the discussion links are prominently placed so people can find them [00:08:43] DarTar: on the Research page? [00:08:49] Reedy: yes they cover ' and " for src= not for url() [00:08:57] If this banner is being kept, the logos are horribly distorted-looking. [00:09:01] Nemo_bis: You don't know HTML? You can work in Wikimedia's fundraising department! [00:09:06] Jerome_: yes, or at least at the top of the discussion page [00:09:21] Joan, hm, thank you for the suggestion! I'll consider it. :-) [00:09:25] It is already there, and Nemo put even more highlight on it [00:09:26] Nemo_bis: Can you add url(' and url(" support in addition to url( ? [00:09:45] Krinkle, sure [00:10:00] but I don't know when regex become more efficient than all those checks [00:13:46] Joan, Krinkle was right, sigh https://meta.wikimedia.org/w/index.php?title=MediaWiki:Centralnotice-template-HSP_final&diff=3135276&oldid=3135155 [00:16:49] Nemo_bis: yesterday when I synced changes back from Meta to testwiki protocol fix was reverted [00:16:49] https://test.wikipedia.org/w/index.php?title=MediaWiki:Centralnotice-template-Harvard_prototype&diff=prev&oldid=120440 [00:16:57] before they were relative, after not anymore [00:17:21] Krinkle, you didn't need to provide the diff... [00:17:34] no problem. i'm waiting for code to compile [00:17:47] heh [00:18:04] I'm so disappointed by this failure of advanced wheel war hopes. [00:18:21] AbuseFilter should overrule CentralNotice, not the opposite! [00:19:10] Shell > * [00:19:34] umpf [00:19:44] Reedy is a turtle. [00:20:32] I like turtles. [02:07:57] This page [http://grebdioz.sciences-po.fr/limesurvey/services/login.php] does not work in Safari [02:08:12] /limesurvey/scripts/jquery/jquery.js:19 -- TypeError: Result of expression 'document.body' [null] is not an object. [02:08:43] This is because it is just two script tags, no body element :S [02:09:06] lol [02:12:32] ... [02:16:33] Is it supposed to be a(n) (X)HTML document? [02:20:37] *PiRSquared17 waits for a response [02:24:30] PiRSquared17: probably have better luck at whenever normal working hours is in the US [02:27:32] PiRSquared17: yeah, they seem to have disappeared, sorry [02:28:03] you can always email wikitech-l also to make sure someone sees it at some point [02:29:34] although, I don't really think the survey is being handled by wmf folk (Although i did see krinkle working on the banner earlier) [04:00:34] umm [04:00:43] Anyone now? [04:41:04] *PiRSquared17 is tired [04:41:09] Nobody? [04:43:35] PiRSquared17: file a bug (arguably its not in scope, but the person closing it will probably redirect it to the proper people) put someone involed with wmf reasearch in the cc field ( Dario Taraborelli would probably be a good choice) [04:44:01] I don't have bugzilla, or e-mail [04:44:22] *I do have e-mail* [04:44:29] just don't want it public [04:44:53] leave a message on http://en.wikipedia.org/wiki/User_talk:DarTar [04:45:07] (I'm guessing DarTar would probably know who to tell) [09:31:37] hello [09:36:11] hi [15:33:45] <^demon|away> brion: Mind taking a look at http://svn.wikimedia.org/core-repack.git.gz? I'd like a sanity check before I push to gerrit for general testing. [15:34:24] .git.gz? is it a tarred directory? [15:34:31] *brion is dl'ing [15:34:33] <^demon|away> Yep [15:34:37] spiff [15:35:24] will look it over in a few [15:35:28] <^demon|away> It's just the repo so you won't see a working copy when you untar. You can do stuff like git log and git branch on the repo, and clone from it to get a working copy. [15:35:39] <^demon|away> Thanks a bunch, I've got class for the next few hours so there's no rush :) [15:38:51] how big is it? if it's gbs then... it would take me a few hours to download :-/ [15:46:16] <^demon|away> apergos: Managed to get it down to 100M. [15:46:23] ah much better [15:46:31] <^demon|away> Yeah, shaved off 4.6G ;-) [15:46:47] holy crap, what did you purge? [15:47:53] <^demon|away> I repacked it. The raw import is really really inefficient. [15:48:23] <^demon|away> So basically I told git to rebuild all the deltas it could. [15:48:37] good little git :-) [15:49:03] <^demon|away> http://metalinguist.wordpress.com/2007/12/06/the-woes-of-git-gc-aggressive-and-how-git-deltas-work/ is a pretty good blog post explaining it. [15:49:18] checking [15:50:01] I lie. I opened the tab and I need a break. I didn't eat lunch and ti's almost 6pm [16:06:48] hexmode, how common is your surname? [16:07:47] Platonides: meh... my spelling? not too common [16:09:21] I recently saw a Katy Hershberger in the 'Thanks' section of a book [16:09:39] I immediatly wondered if you would be family [16:09:45] ^demon|away: taking a quick peek at the git thing... [16:09:47] http://www.ancestry.com/facts/Hershberger-family-history.ashx [16:09:51] seems to clone ok, branches visible [16:10:10] ^demon|away: no release tags? [16:10:13] Platonides: probably not, but maybe distant [16:11:07] installer seems to be running ok on master checkout [16:11:13] would probably have been too much coincidence [16:11:15] there were no census in the US after 1920? [16:11:49] Platonides: I know of one guy who shares my name... he used to be the top hit on google [16:12:08] Platonides: full census data is released after something like 75 years [16:12:09] <^demon|away> brion: Not yet, I didn't know how to do those with the tool I was using. [16:12:18] <^demon|away> I think we could fairly easily find them and tag them on the branches [16:13:10] question i guess is 'use tags for tags, or branches for tags' :) [16:13:30] <^demon|away> We should probably use tags for tags. That's what they're for :) [16:13:37] in theory yes :DD [16:13:46] <^demon|away> And using branches for tags is just being svn-esque all over again ;-) [16:13:51] *hexmode goes to look at the 20% to see who he can bug [16:14:16] ^demon|away: ok i can install from a clone of that running master. yay! [16:14:34] special:version may need updating to show the current git rev [16:15:10] if git elseif svn [16:15:16] switching to REL1_18 and REL1_17 branches seems to work nice [16:15:47] and git blame works \o/ [16:15:53] Reedy: could I get you to look at an API bug? [16:15:55] heh [16:15:55] 2cb4eea9 (Erik Moeller 2004-02-28 03:04:02 +0000 1) 9f146cee (Brion Vibber 2007-03-12 15:57:17 +0000 2) /** [16:16:05] lol [16:16:09] hexmode, ? [16:16:50] Reedy: https://bugzilla.wikimedia.org/show_bug.cgi?id=32895 [16:20:22] <^demon|away> Other people feel free to test that repo too :) So far it sounds good and I'll try to get it posted on gerrit this afternoon so we can start playing with it as a group. [16:22:03] what's "posting it on gerrit"? [16:22:50] (I had downloaded it, now was searching for some file with tricky history) [16:23:43] how can you know the svn rev in that repo? [16:24:28] <^demon|away> I meant pushing the repo to gerrit so we can start playing with the repo like we would for real. [16:24:44] <^demon|away> So we can get used to it, tweak settings, and get more people playing with it. [16:24:56] <^demon|away> As far as "which svn rev"...hrm. [16:25:00] <^demon|away> That's a difficult question. [16:25:31] <^demon|away> Ah hrm, there's some options I could've tried. [16:25:40] <^demon|away> --add-metadata if passed, each git commit will have svn commit info [16:25:40] <^demon|away> --add-metadata-notes if passed, each git commit will have notes with svn commit info [16:25:49] some conversions leave it in the commit summary, but doesn't seem to be the case [16:26:19] <^demon|away> I can try it with those options and see what I get as a result. Sorry for overlooking that before. [16:27:16] it's a tricky thing to do right [16:27:31] I was getting empty diffs with gitk, yet restarting it they are show now :S [16:28:03] ^demon|away: ok here's a problem [16:28:12] we're missing things that were checked out via svn:externals [16:28:34] <^demon|away> Did we ever have any externals in phase3? [16:28:44] wait ???.. nevermind [16:28:50] <^demon|away> :) [16:28:55] i'm on crack or something, that was a) an extension and b) not the git checkout [16:29:01] FALSE ALARM [16:29:01] xD [16:29:10] *brion needs more coffee? [16:29:15] <^demon|away> Platonides, brion: This is the tool I ended up using, by the way https://gitorious.org/~marcguenther/svn2git/marcguenther-svn2git [16:29:36] does it currently allow to go from rXYZ to the git commit? [16:29:48] we have plenty of references in our commit logs [16:29:53] <^demon|away> No clue. [16:30:39] <^demon|away> I'm a little less concerned about having that sort of patchset metadata. [16:30:46] <^demon|away> That's really hard to get right. [16:30:46] they can be used with git svn [16:31:06] (if you directly clone the repo, never make things work cloning from github) [16:32:13] bookmarkin' that ... [16:32:28] ^demon|away: it would be handy if we had a revision map that we could query from mwbot :D [16:32:44] or use to shuffling things from old codereview pages [16:32:51] <^demon|away> That's probably doable, yes. [16:33:12] <^demon|away> Especially if I can get the svn commit info into the git commits, which I will try later. [16:33:30] they are complementary features [16:34:21] when using git-svn, the mapping seems to be left at .git/svn/refs/remotes/git-svn/.rev_map.3e45a915-b2e0-c146-c98e-857251466b57 [16:34:29] (that trailing is probably the repo uuid) [16:34:47] then you can do git svn find-rev r50123 [16:36:50] <^demon|away> I'll start a fresh dump of this with that info. It'll take awhile anyway [16:38:02] don't worry, we will probably need several iterations [16:39:44] <^demon|away> Dumping in a screen, hopefully it'll be done by the time I get back and I'll repack it and we can give that a shot. [16:39:52] *^demon|away goes |away for real now [16:40:26] ok :) [18:25:21] brion? [18:26:50] Jerome_, just for consistency, I think that the discussion has been on #wikimedia-rcom lately [18:27:29] there you are [18:27:39] Nemo_bis: I'm there too. Happy to discuss :) [18:27:59] why do you have so many different usernames? :) [18:29:23] hi jeremyb [18:29:58] er Jerome_ [18:30:04] tab completion is awful in this version of xchat :P [18:30:13] Well, my tail grows each and every time I log in ;) [18:40:33] ^demon|away, hey you about? we're trying to see what it'll take to run automatic nightly builds of the android app on the jenkins system [18:40:45] if we can't figure that out, maybe we'll do it on labs ;) :P [20:07:13] <^demon|away> brion: If you can script the build process in ant, you can do it in jenkins. [20:08:27] ^demon|away, 'ant debug' to create a debug build [20:08:58] needs android sdk available [20:09:29] <^demon|away> Yeah, that's the only gotcha I could think of. [20:09:46] any impediments to that? [20:09:48] <^demon|away> But as long as there's a package for it we should be fine. [20:10:35] packages... may or may not be standardish [20:10:39] it's food time, bbiab