[01:12:27] Technical_13: fyi, i'm also opposed to renaming #mediawiki-api and expanding scope to include bugzilla and friends [01:15:21] jeremyb: ssshhh, just don't talk about it, T13 will soon find something else to tinker with :) [15:00:17] parent5446: hello [15:00:46] svick: hey, sorry I missed Friday. Accidentally overslept. [15:01:02] no problem [15:01:11] So what's today? [15:02:34] so, since last time, i have extended README to describe how to use the application (including diff dumps) [15:03:09] and i have also written a file format spec for diff dumps: http://www.mediawiki.org/wiki/User:Svick/Incremental_dumps/File_format/Diff_specification [15:03:35] Nice, and I just saw the commit for sorting revisions. [15:04:18] yeah, that was actually a tiny change (it didn't require many changes) [15:04:23] now i think it's time to start working on delta compression [15:05:01] Ah it's finally that time. [15:05:16] so today, i'm going to read some papers, but it looks like the best choice will be zdelta [15:05:38] OK, hopefully it gives us some good compression results. [15:06:07] though it's last version is from 2004, so i'm considering contacting some people (authors of zdelta, authors of similar papers) to ask whether there is something newer and better [15:06:56] Mhm, it looks to be a bit like a research project. Do you know how it's compression compares to other methods? [15:08:13] Ah I remember, the research paper itself has some benchmarks. [15:08:14] that was what the delta.pdf paper was about; it compared xdelta, vcdiff and zdelta; and zdelta won [15:08:30] Yeah, almost forgot they did their own. [15:08:52] Well if you're gonna contact somebody you should get in touch with Nasir Memon. [15:09:08] He's still a professor at NYU-Poly and is in charge of the CySec department I believe. [15:13:06] ok, i guess i'll contact him; i thought about writing to the other author of the paper Torsten Suel, who also seems to be still teaching, but it probably doesn't matter much [15:13:17] Ah well that works as well [15:13:31] I just wasn't sure if the other authors were still active [15:13:43] Also the README looks pretty good now [15:15:26] good; if you think there's anything unclear or something like that in there, let me know [15:16:42] OK, well that's all I have. Do you have anything else? [15:17:14] i can't think of anything; see you tomorrow [15:17:23] Yep see you tomorrow. Good luck with delta encoding. [15:21:10] parent5446: btw. i'd love to see your securesessions extension enabled, has there been any discussion about such a thing in any of the fora ? [15:21:41] Not yet. [15:22:13] Unfortunately I'd have to disable some features before we deployed it (because it uses HSTS even though WMF can't use that yet). [15:22:45] k, after the https deploy, we should open a discussion on this, i'd love to have this (possibly even fully integrated into core). [15:23:47] Yeah. A lot of the features I took from LastPass, e.g., banning login from certain countries. It provides enough security for the paranoia to hang themselves. :P [15:23:53] session management like that should be a default feature for users. not necessarily an "in your face" feature, but there when neede. [15:24:40] Although I'm beginning to not like the idea of having so many cookies with identifying information in it. [15:25:02] Maybe my own paranoia is geting to me. [15:41:29] anomie: heya, the tabs vs spaces bug in CE that Alex/Ori patched, should we update CE soon to get that out there or just wait until Thursday? [15:42:42] greg-g: Personally, I want to touch spaces vs tabs as much as I want to touch vi vs emacs. ;) But it probably should get some testing before the initial deploy everywhere. [15:44:03] anomie: ditto :) [15:44:36] anomie: can I have you deploy the change sometime soonish? [15:45:15] anomie: like, this afternoon would be fine [15:45:35] post 1pm Pacific, is what I mean by that [15:45:54] greg-g: Soonish as in "before 21:00 UTC", sure. After that, I can't today. [15:46:06] though I guess Krenair or ori-l should be around too [15:46:43] that's 2pm Pacific, sure thing. so 1pm Pacific should be fine (20 UTC) [15:47:51] You want me to be around for a spaces vs. tabs patch? :) [15:50:22] Krenair: you wrote it right? :) [15:51:28] well, yeah [15:52:20] Krenair: so, are you going to be online around 20 UTC? [15:52:33] I guess so, yes. [15:53:42] Krenair: cool, just in case things blow up when that is deployed [15:55:02] Krenair: anomie I guess the obvious question though is: Is that hook enough for that religious issue for now? [15:55:51] greg-g, Krenair: I wonder what happens if someone who sets "use spaces" edits a page that is currently using tabs, or vice versa. [15:56:11] Do we get diffs filled with whitespace changes? [15:56:44] When I tested it, it looked like it only applied when you created a new line. But feel free to prove me wrong [16:02:44] Krenair, greg-g: Even if it's only for new lines, we end up with a file that is mixed tabs and spaces. [16:03:28] that's assuming the file starts with spaces [16:03:46] Without this patch, the opposite would happen which would have the same problem [16:05:22] Where was that interwiki table again? [16:08:30] T13: Interwiki table? [16:08:46] T13: https://meta.wikimedia.org/wiki/Interwiki_map ? [16:09:01] T13: https://meta.wikimedia.org/wiki/Special:SiteMatrix perhaps? [16:09:50] Yeah, that will work. [16:10:11] Trying to see which ones are missing from wm-bot's db [16:10:58] reading http://www.mediawiki.org/wiki/Manual:Maxlag_parameter there is no distinction between reading and writing actions. Is that correct, or should it be taken to mean just not to write during higher lag? [16:16:06] anomie: Krenair votes on whether or not to deploy the new code? [16:20:37] Well we have 4 hours. I'll test it again [16:21:16] greg-g: If we're going to keep it merged, deploy it. I'd personally have questioned merging it in the first place, but not to the extent of -2ing it. Sure, people can paste text with tabs into a wiki page now, but when people can make gadgets to set tabs-vs-spaces per account is it going to be a mess? [16:21:40] yeah [16:21:51] it's per account, not per script? [16:21:54] I missed that part [16:23:48] is Elsie around? [16:24:07] He was the one who poked at this recently, tool. [16:24:10] s/tool/too. [16:24:22] s|.|/| [16:24:38] greg-g: Unless I missed something, there's a JS hook in there to allow for people to adjust the tabs-vs-spaces setting for their account. [16:25:25] gotcha, I misunderstood how the hook was called [16:25:54] thanks. Ugh, so edit wars on scripts. :/ [16:27:32] anomie, greg-g: it does only appear to affect new lines inserted, not existing ones [16:31:13] greg-g, Krenair: Personally I prefer tabs, but in this case I'd say just stick with spaces (and WONTFIX the bug) because it's the status quo and because if you toggle the editor back to plain-text mode it's difficult to enter tab characters in common browsers because the tab key will move to the next form element instead. But, meh. [16:33:38] touche on that last point [17:53:34] brion's not here at moment is he? [17:57:33] russavia: i guess not... what's up? [17:57:49] 26 17:25:19 -!- brion_ [~brion@c-24-6-11-12.hsd1.ca.comcast.net] has quit [Quit: Leaving] [17:57:53] 26 17:53:34 < russavia> brion's not here at moment is he? [18:00:11] apergos: i was sure there was a bug/ticket on getting image rsync/mirror/tarball working in eqiad. can't find it. help? :) [18:00:33] found https://bugzilla.wikimedia.org/51001 [18:00:41] but that's not quite what i had in mind [18:39:13] anomie: Krenair alright, for now, given the mixed use of spaces/tabs if this goes live, I say we hold off on the CE update (and just stick to the original plan). [18:39:17] good? [18:39:46] greg-g: So do what exactly? [18:40:29] nothing today, re-engage that bug report and see what a compromise solution would be that doesn't cause mixed use in a file [18:40:48] anomie: will you be at the MW Core meeting today? [18:41:01] greg-g: Yes [18:41:41] anomie: i won't be, unfortunately, but could you bring it up? I have it on the agenda. [18:42:07] greg-g: ok [18:42:13] anomie: thanks [18:42:48] greg-g, this causes mixed use in a file now? [18:42:55] hang on [18:43:02] Do we want this to cause mixed use in a file? Or dirty diffs? [18:43:53] well, I think that question is unanswered [18:44:24] but, I have to run, will be online a bit later. [19:37:06] greg-g: Hi. [19:37:27] CodeEditor is just a wrapper for ACE (or is it Ace...). [19:37:39] It has a bunch of options, but I don't think most are exposed in the MediaWiki extension. [19:38:37] https://en.wikipedia.org/wiki/Ace_(editor) [19:38:43] https://en.wikipedia.org/wiki/ACE_(editor) [19:39:27] Not to be confused with "Ace (software)". [19:39:27] "ACE (software)", rather. [19:39:27] Rage. [20:01:32] greg-g: Ping me when you're around re: CodeEditor. [20:09:07] hrmmmm, i wonder if watchmouse tests ipv6? i guess they could have locations that are all v6, and all v4 and dual stack [20:29:40] jeremyb: I think it can, not sure if it actually does [20:31:12] well i still wonder! [20:31:21] prompted by the wikidata AAAA bug [20:31:54] MissGayle: hi! http://thecommandline.net/2013/07/08/chris_miller_management/ [20:31:58] fyi [20:33:42] Thank you, jeremyb! Taking a look [20:35:15] Whoah. I couldn't believe Gerrit really does not allow to download a patch as plaintext (available in 2.8, we run 2.6) and I am way too lazy to start my other machine but there seems to be no other option. [20:36:06] <^d> andre__: We don't really run 2.6 [20:36:30] <^d> 2.7-rc2-505-g7502a46 [20:36:40] <^d> Which is master, roughly resembling 2.7rc2. [20:36:48] ah, nice to know [20:48:48] YuviPanda: ping [20:48:53] dMaggot: pong [20:49:47] YuviPanda: should the UploadWizard be displaying a select widget by now? [20:51:14] dMaggot: not sure, let me check the deployment calendar [20:54:16] Reedy: hmm, how do I find out what version of an extension was deployed as part of wmf14? [20:54:21] Reedy: in the mediawiki/extensions repo? [20:55:28] hmm, not there [20:56:23] greg-g: ^ [20:57:32] checkout the wmf14 branch [20:58:18] reedy@tin:/a/common/php-1.22wmf14$ git submodule status extensions/CodeReview/ [20:58:18] 65b3228e6f145807a03a48ba6e0e1d097eb5d93b extensions/CodeReview (heads/master) [21:00:17] Reedy: of mediawiki/extensions? ok [21:00:26] github tells me there are no branches, but let me clone [21:10:24] YuviPanda: No, 1.22wmf14 of core [21:10:39] oh [21:14:27] thanks Reedy [21:14:34] dMaggot: i checked, it isn't deployed with wmf14, I'm afraid [21:14:45] dMaggot: will be deployed with wmf15, which should be next tuesday. [21:14:55] dMaggot: it is deployed on betalabs, though. [21:15:05] dMaggot: http://commons.wikimedia.beta.wmflabs.org/ [21:15:11] and I can give you appropriate rights there if you want [21:18:19] YuviPanda: yes please, my nich there is DNarvaez [21:19:01] dMaggot: moment [21:19:44] dMaggot: you should have sysop now :) [21:20:00] YuviPanda: thx [21:20:30] dMaggot: also, any other patch that needs to go in along with this should be merged before Thursday. [21:21:23] YuviPanda: ok [21:21:32] dMaggot: also, thanks for the patch :) [21:24:34] what's a good resource to learn about handling cookies on the client side manually? [21:24:44] is the en.wiki article basically sufficient? [21:24:50] [21:24:52] MartijnH: in the context of MediaWiki gadgets / extensions? [21:24:55] or generally? [21:25:06] in the context of a MediaWiki bot [21:25:18] with a wonky web-client library that doesn't do cookies [21:25:37] what is it implemented in? and did it recently break? [21:25:42] so I'll be parsing and generating the headers myself [21:25:53] no, it didn't recently break, it's not working yet [21:26:52] ah. well, the enwiki article should be sufficient. you may find the various /cookies endpoints provided by http://httpbin.org/ useful for testing. [21:27:09] ah, that's good, thanks [21:27:44] that's a really handy resource by the way [21:27:48] will simplify testing a lot [21:28:44] :) [22:55:15] gn8 folks