[00:09:44] Will do that tomorrow [14:06:45] RoanKattouw_away: let's talk about where you think the code for collapsing the TOC should go when you get back [14:38:38] adam_miller: I see you just wanted it in SVN to take it home; that's cool, but I don't think it's the right place in the long run [14:39:06] yeah, i had it in jquery.wikiEditor.toc.js initially [14:39:18] but thought i'd get heckled if i committed it in there :) [14:40:06] That makes more sense to me (haven't asked Trevor), as long as there's a var to disable it [14:40:12] AND with this svn multiple file example you just posted - wouldn't that also commit any changed file inside the directory where i ran commit? [14:40:42] a JS var, or something that we can stick in LocalSettings.php? [14:42:45] adam_miller: It would, unless you do svn commit file1 file2 dir1 [14:43:08] Both; you'd have a LocalSettings.php setting that's exported to JS [14:43:55] how do I hook that up? [14:43:58] *RoanKattouw goes away again; feel free to fire more questions at me, I'll answer them at some point [14:44:22] adam_miller: The UsabilityInitiativeHooks class has an addVariable() static function or something, ClickTracking also uses it [14:44:37] cool, thanks i'll check that out [16:59:32] RoanKattouw_away: please ping me when you are back [17:04:25] nkomura: Pin [17:04:27] g [17:04:38] one sec [17:07:18] hey [17:07:51] wanted to check in with you where you stand on the survey data dump [17:07:57] it's holding up howie's work [17:08:08] Yeah I just got an e-mail from Howie [17:08:20] I said I'd get on it rightaway [17:08:25] k [17:08:30] He wants a new dump that goes right up to today [17:08:40] right [17:09:01] would you do a complete dump, meaning survey data, pref stat data, and total? [17:09:16] so that we can have a even start on all three front? [17:09:47] OK [17:09:52] The optin one is quick [17:12:18] another thing i want to coordinate with you is to schedule a time to make the enhanced toolbar default for en.wikinews and remove beta [17:12:31] what is your availability tomorrow? [17:12:43] I'll be available all day [17:13:28] nkomura: Optin done [17:13:41] are you 8 hrs ahead of us right now? [17:13:46] Yes [17:14:08] That'll go back to 9 after Nov 1 [17:14:10] so let's say 11amPDT/7pmCET? [17:14:25] Should be fine [17:14:40] what time would that be in UTC? [17:14:47] 6pm [17:14:48] 18:00 [17:15:02] k [17:15:08] and are you comfortable with config change in production? [17:15:27] I am [17:15:35] Strategy and usability already have it [17:15:40] And I did one config change before [17:16:06] right [17:16:27] no more disabling the enhanced toolbar, i hope ;) [17:24:18] nkomura: I discussed anonymous preferences earlier today in #wikimedia-tech [17:24:39] how did it go? [17:24:52] There are some technical objections [17:25:34] All three ops people I talked to (and one developer) thought it was a bad idea; one called it "Pandora's box" (cf. opening the floodgates to all kinds of crazy cache-defying uses) [17:25:51] hmm [17:25:54] The only scenario in which it would reasonably work is when anonymous opted-in users are rare [17:26:27] But it's kind of crazy to enable a feature and say "let's hope it will hardly be used, or it'll break" [17:26:50] the objective is to promote the opt-in [17:27:30] so hoping that not too many people would use it defeats the purpsoe [17:27:36] Exactly [17:28:59] and your position was that this anon opt-in was just for beta right? [17:29:11] Yes [17:29:46] But even that is pretty unpopular [17:29:56] i was afraid of that [17:30:30] will you post a discussion to wikitech-l? [17:30:36] I could [17:30:56] I don't expect any significantly different responses, though [17:31:06] i suppose that the response would be similar, but it is good to get the feedback from wider participants [17:32:38] what do you think it would happen if we proceed without buy-in? [17:33:10] adam_miller_lunc: FYI, you don't need to raise $wgNavigableTOCStyleVersion if you haven't changed any files in the NavigableTOC/ directory (similar for other extension style version vars) [17:34:18] RoanKattouw: right on, i had made a change, but reverted it before committing [17:34:54] The only realistic scenario in my opinion is you convince Erik to pull his rank and force the ops folks to do it, which will result in vocal disagreement, hard feelings and more of that bad stuff, and probably result in an I-told-you-so scenario where they end up being right [17:35:08] (assuming disagreement is pretty much unanimous) [17:35:46] so you felt the majority folks are against it [17:36:02] k, let me discuss it with howie [17:36:22] we are trying to find a way to reach out to more people for tying the beta [17:37:19] I'll still do the wikitech post for wider input if you want me to, but I don't think we'll get a more diverse response from that [17:37:56] how many folks are in #wikitech room? [17:38:37] RoanKattouw: is the issue mostly that squid caching would break? Because I don't see how it would...it should still be able to cache the vector version of the skin just as much as any other [17:38:40] Plenty [17:38:51] nimish_g: Yes, but the required cache size would double [17:38:57] nkomura: 126 people right now [17:39:07] Yeah but most of those were inactive at the time [17:39:12] that's a big crowd [17:39:15] It was still early morning in PT [17:39:21] I only got response from like 3 people [17:39:37] it would only double if half the people used vector and half used monobook [17:39:42] nimish_g: Additionally, the number of backend requests (= the amount of initial misses for a cacheable resource) would double [17:39:46] No it wouldn't [17:39:54] As long as the crowd using Vector visits every page [17:39:58] right [17:40:23] hm [17:40:26] Even a relatively small crowd could already visit like 75% of all cached pages [17:50:56] nkomura, howiefung: I gotta go now, will be back around lunchtime to finish the dumps [17:51:37] thanks roan [20:33:30] howiefung, nkomura: Dumps are all done @ http://toolserver.org/~catrope/survey/ [21:16:54] parutron: About the "wipe dialog fields after clicking Insert" thing we discussed yesterday: do we want this for all three dialogs (link, table, s&r)? [21:17:59] *RoanKattouw assumes so [21:18:02] Tell me if not [21:45:57] parutron: Done, and live on prototypes. I didn't do it for the search&replace dialog, since there's only one way to close it. Also, it's not gonna be in of our upcoming deployment, as it's part of the link/table dialog revamp [21:47:16] Oh wait those /are/ part of our upcoming deployment. Crap. [21:48:33] nkomura: I see the link/table dialog revamp /is/ part of our upcoming deployment now; I previously assumed it wasn't, which is why I didn't include this in the deployment staging area (making life harder on myself). I assume we still want it in there? [21:52:01] Oh wait, it's not [21:52:06] I'm an idiot, ignore me [22:00:13] adam_miller_away: I just noticed you added a new config var ($wgNavigableTOCCollapseEnable) without setting it to a default value anywhere (the appropriate place for this one being NavigableTOC.php); this is not only unclear and undocumented, but also a security vulnerability if PHP's register_globals setting is on [22:01:14] adam_miller_away: conflicting info right? "RoanKattouw> I'm an idiot, ignore me" :D [22:01:37] *thedj pets RoanKattouw [22:02:09] Woof [22:02:13] Or did you mean to pat me? [22:02:24] nah [22:02:53] aai poesje, rot kat ! :D [22:41:50] parutron__: RoanKattouw> parutron: Done, and live on prototypes. I didn't do it for the search&replace dialog, since there's only one way to close it. Also, it's not gonna be in of our upcoming deployment, as it's part of the link/table dialog revamp [22:42:11] (it = wipe form values post insert)