[00:02:24] binasher is there a downside to that approach? i'm trying to think if we'd need to change anything else in software [00:02:38] MaxSem: passing by reference isn't needed in this case but, I added it in a new patch set anyways [00:02:58] +1'd [00:03:08] * MaxSem kicks gerrit-wm [00:03:42] awjr: can you approve https://gerrit.wikimedia.org/r/#change,4379 [00:03:54] awjr: actually, i don't think it would work [00:04:03] :( [00:04:10] whynot? [00:04:33] since the desktop view link takes you to non-m wikipedia with &useformat=desktop [00:04:55] but you'd get redirected to m since the cookie hasn't been changed yet [00:05:29] this is obnoxiously challenging for such a seemingly trivial thing to do... [00:06:24] tfinc: i've got those MS folks coming in tomorrow, around 4:30 if anybody other than me's interested in seeing their metro goodies :) [00:06:42] binasher is it possible to make an acl that looks at the cookie AND the request query string? [00:06:45] brion: ask lindesy, heather, or brandon [00:06:51] jorm: --^ [00:07:02] mrh? [00:07:07] cannae. [00:07:14] well, maybe. i'll have to see. [00:07:14] np [00:07:15] jorm: Microsoft metro designers are coming in [00:07:30] or at least some of the people who worked on it [00:07:36] jorm: brion can tell you more [00:07:48] or at least folks who want us to make our app for windows 8 and are showing us the style :) [00:08:36] preilly approved. what happened to the bot? [00:08:42] preilly merged, also [00:08:58] * awjr pokes gerrit-wm [00:08:59] awjr: that is a zero thing bot isn't checking that yet [00:09:03] ohho [00:11:44] YuviPanda: http://www.geonames.org/img/charts/ws-credits8163-NhBKEPcu6P86DGxC7SYivA.png <-- Geo traffic [00:11:47] MaxSem: --^ [00:12:02] nice spike [00:12:06] tfinc: wow, [00:12:10] binasher it should be possible to make a squid acl that checks cookie an url query string, right? [00:12:17] we've used up 78% of our queries [00:12:19] but also quickly dropping [00:12:21] * tfinc goes to buy some more [00:12:43] awjr: i don't think its actually a good idea, since m. isn't intended for all url types [00:12:59] MaxSem: looks like 'oooh, let me play with this!' type initial look [00:13:02] binasher: what do you mean? [00:13:07] YuviPanda: exactly [00:13:11] i like to see things! brion [00:15:05] edit links, etc. if i have two wikipedia windows open, and get the mobile cookie, then click an edit link in the other window, i don't want it to open the edit page in the m domain [00:16:24] binasher i see what you mean but that usecase could also be acl'd, right? i mean, as it is now, if you try to edit a page on your mobile device, you go to the desktop version on .m to edit it [00:17:16] i guess my point is, that problem already exists in production [00:17:40] having to come up with a different acl for every case where we wouldn't want this new redirect to trigger doesn't seem like a good path to follow [00:17:57] tfinc: http://www.applestocks.net/wikipedia-drops-google-maps-in-favor-of-openstreetmap-for-use-in-its-mobile-apps/ [00:18:01] again we're a 'company' [00:18:01] sigh [00:18:08] things are already way more complex on the php side [00:18:21] binasher i agree, but that problem *Already* exists in production [00:18:42] tfinc: disregard last link, that was a splog with real source http://www.cultofandroid.com/8092/wikipedia-drops-google-maps-in-favor-of-openstreetmap-for-use-in-its-mobile-apps/ [00:19:22] i think the majority of people who are clicking on 'mobile view' are doing it on mobile devices that we're not detecting. i wish i had numbers to cite, but i would imagine that very few people on desktops are switching to mobile view, unless they're testing. [00:19:27] in my opinion desktop should only ever be visible on en.wikipedia.org and mobile should only ever be visible on en.m.wikipedia.org [00:19:54] +1 [00:19:58] preilly, binasher, in that case we'd need an acl to detect useformat= in the query string and redirect the request appropriately [00:20:12] it might also make sense to record to UA of the user when they click on mobile view [00:20:13] YuviPanda: now imagine being Jay and having to mange that for all of our communications ;) [00:20:20] preilly agreed [00:20:47] tfinc and i had chatted about that last week but im not sure of a sensible way to do that [00:20:49] tfinc: I'm guessing a lot of people think 'Wikipedia' is a company which writes this huge encyclopedia [00:20:53] would it make sense to step back a bit [00:21:04] like remove all of this logic [00:21:16] and attack it again with a clean slate [00:21:32] tfinc: do we try to get these people be right or let it go? [00:21:55] awjr: what is the context of, "tfinc and i had chatted about that last week but im not sure of a sensible way to do that"? [00:22:02] HeadAsplode: we focus on what were good at … dev .. not managing comm [00:22:03] preilly perhaps, but we still face the same useformat problem if we go back to how things were before we started down the stickycookie road [00:22:13] preilly: sorry - context was in regards to recording UAs [00:22:18] tfinc: by 'get these people be right' I mean do we poke jay about it or let it go [00:22:22] not send emails ourselves [00:23:18] awjr: what's the useformat problem? [00:23:51] lets not spend our time cleaning up the internet [00:23:52] let it go [00:24:03] ok [00:24:28] binasher using "useformat=mobile" ni the query string to toggle the view, so you'll see the mobile view on non- .m site if you do en.wikipedia.org/wiki/article?useformat=mobile [00:25:02] so en.wikipedia.org/wiki/article?useformat=mobile should redirect to en.m.wikipedia.org/wiki/article [00:25:37] just handle it like a canonical redirect [00:25:45] preilly redirect in php? [00:25:57] yeah [00:26:07] yes [00:26:10] that would work [00:26:10] that should be easy? [00:26:15] yeah it is [00:26:36] should we do the same thing for the action=edit cases as well? [00:26:46] like action=edit, action=history … not sure if there are others [00:27:58] it looks like useformat=mobile is ignored in those cases? [00:28:14] binasher correct, in those cases you will see the desktop version [00:30:23] so from the php side, don't redirect to .m. if there's a useformat=mobile and an action defined [00:30:45] i'm not sure what the correct behavior should be on mobile site [00:31:31] preilly would there be a problem if someone had useformat=mobile but shouldDisplayMobileView() === false and then got redirected to desktop site? [00:31:36] or vice versa? [00:32:16] binasher you mean like if someone's on the mobile site but then tries to edit an article? [00:39:00] awjr: how would that happen? [00:39:27] er wow that made no sense [00:39:35] ok if someone had useformat=mobile but action=edit [00:41:34] YuviPanda: can you pass me a link to 1.1.1 apk .. i want to send it to upload [00:42:01] tfinc: signed one? [00:42:04] yes [00:42:15] the exact same one you sent to the market [00:42:21] ok [00:43:33] tfinc: http://dl.dropbox.com/u/8768784/Wikipedia-v1.1.1.apk should be it [00:43:35] double checking [00:43:39] yeah, so useformat=mobile would trigger a redirect to the canonical mobile version of the url, unless there isn't such a thing (i.e. there's an action defined) [00:43:53] yup [00:43:55] that's it [00:43:57] tfinc: ^ [00:49:37] thanks. http://dumps.wikimedia.org/android/ is up to date now [00:49:46] cool :) [00:52:57] this release, saved pages is going to drive me nuts :D [00:53:11] YuviPanda: 1.1.1 or 1.2 ? [00:53:15] 1/2 [00:53:17] err [00:53:18] 1.2 [00:53:31] i'm trying to figure out the way I can get it to work with the least amount of code changes. [00:54:23] tfinc: ideally, i'd like to 'relegate' offline save pages functionality to someone else (Read It Later, etc). Thoughts? This would be on a much future release, wtih lots of advanced warning. [00:54:38] YuviPanda: thats a no go [00:54:44] but why [00:54:49] we can't build external dependancies [00:54:51] tfinc: ha ha [00:54:58] there are other apps that do this as their core competence [00:55:04] negative ghost rider [00:55:26] tfinc: we can let people pick whoever they want. [00:55:37] we're not having external dependencies. [00:55:53] if people want offline functionality, they will have an app that provides it for them. Why not integrate with those? [00:58:03] other services like evernote actually charge you to save things offline [00:58:08] http://www.evernote.com/about/premium/ [00:58:33] tfinc: there will be plenty of free ones. Android makes sure the choice is with the user, and not with the app developer [00:59:03] see https://readitlaterlist.com/ for a free one [01:00:12] YuviPanda: i know that list and i'm happy to integrate with more of them [01:00:21] tfinc: 'make the best app to consume (and eventually contribute to) wikipedia' is what we should be building. Saved pages is something we should leave to other apps. [01:00:32] YuviPanda: but unless we have a compelling reason i'm not going to remove a feature that people are actually using [01:00:53] tfinc: hmm, what would actually convince you to remove it? [01:01:04] YuviPanda: wouldn't you like to know ;) [01:01:09] indeed [01:01:21] ha ha ha [01:01:34] it's our last major platform specific code [01:01:57] and we're doing crappily what other apps are doing awesomely *and letting us use them* [01:02:14] why not integrate and free up time to work on awesomer stuff? [01:04:13] tfinc: our iOS app has 'save page' and 'save to read it later' right next to each other. I bet people would use RIL (or similar service) more than the actual save page. If they need to be reading stuff offline, they will already have an app. [01:04:25] if i had stats that showed a minority of users were using our native saving in favor of external services then we could have a conversation about it [01:04:58] i'll be happy to set up instrumentation to see how many people use 'save page'. [01:05:07] and ofcourse, that'll take about 9 months to set up servers for [01:05:24] given that a) we don't have that and b) people use the current functionality easily without any account level features. i'm not going to take it away unless I'm convinced that they don't need it [01:05:46] in summary - show me the numbers :) [01:06:05] how do you suggest I do that? [01:06:28] I can look into third party services that let me do analytics, and then watch as the community comes to my home with pitchforks [01:06:55] hmm, is there a way of using our pre-existing analytics stuff via an API? [01:07:08] YuviPanda: ask diedrik .. he'd know [01:07:30] time for me to head out [01:08:14] sigh [01:15:17] aaaah [01:15:28] i think i know a way to collect numbers on this particular thing [01:15:40] but first I need to make it work :D [01:36:38] YuviPanda: you might be able to abuse the clicktracking extension for that sort of means [01:36:46] Reedy: link? [01:37:04] Reedy: my other thought was that since i'm switching to the API, I can pass an extra parameter that goes unused, but is logged [01:37:14] that assumes that all API GETs are logged, which might not be true [01:37:20] but am guessing that'll be quite easy to measure too [01:37:22] https://en.wikipedia.org/w/api.php "Track user clicks on JavaScript items." [01:37:42] wooo, that's pretty much *exactly* what I need :) [01:37:49] err [01:37:56] https://www.mediawiki.org/wiki/Extension:ClickTracking [01:38:00] or [01:38:01] maybe not [01:38:09] it requires an edit token, which means I need to have a logged in user :( [01:38:35] Err [01:39:47] Edit token for anon is +\ [01:39:54] ah [01:40:02] sweet [01:40:42] Anons can edit with the api if they can edit via the gui [01:41:06] that never hit me. And I don't remember reading this in API:Editing [01:41:07] :( [01:41:21] Reedy: is there documentation on 'proper' usage of this? [01:41:30] Does Roan count? [01:41:30] ;) [01:41:49] I suspect users of the api are just even more likely to be logged in [01:42:08] i'm sure that someone can kidnap Roan and ransom the knowledge in his brain [01:42:10] There aren't going to be many casual users using the APi [01:42:21] agreed [01:42:55] AFTv5 uses ClickTracking (among others) [01:44:05] Reedy: where does data collected by this go to? [01:44:18] a database table [01:44:24] is that publicly visible? [01:44:36] nope [01:44:41] might be replicated to toolserver [01:44:44] ah [01:44:46] ok [01:45:19] Reedy: I think this should help me find the 'numbers' to (hopefully) kill Saved Pages in a few releases :) [01:45:21] thanks! [01:45:53] Reedy: is there anything you don't know? :) [01:46:11] Haha. Pi to a million places? ;) [01:47:00] i'm sure you could find a way around that :P [01:47:02] If you poke Dario/Howie they might be able to advise a bit better how they get access/use the data [01:47:14] 1415926535897932384626433832795028841971693993751058209 7494459230781640628620899862803482534211706798214808651 3282306647093844609550582231725359408128481117450284102 7019385211055596446229489549303819644288109756659334461 2847564823378678316527120190914564856692346034861045432 6648213393607260249141273724587006606315588174881520920 9628292540917153643678925903600113305305488204665213841 4695194151160943305727036575959195309218611738193261179 [01:47:14] 3105118548074462379962749567351885752724891227938183011 9491298336733624406566430860213949463952247371907021798 6094370277053921717629317675238467481846766940513200056 8127145263560827785771342757789609173637178721468440901 2249534301465495853710507922796892589235420199561121290 2196086403441815981362977477130996051870721134999999837 2978049951059731732816096318595024459455346908302642522 3082533446850352619311881710100031378387528865875332083 [01:47:16] 1415926535897932384626433832795028841971693993751058209 7494459230781640628620899862803482534211706798214808651 3282306647093844609550582231725359408128481117450284102 7019385211055596446229489549303819644288109756659334461 2847564823378678316527120190914564856692346034861045432 6648213393607260249141273724587006606315588174881520920 9628292540917153643678925903600113305305488204665213841 4695194151160943305727036575959195309218611738193261179 [01:47:16] 3105118548074462379962749567351885752724891227938183011 9491298336733624406566430860213949463952247371907021798 6094370277053921717629317675238467481846766940513200056 8127145263560827785771342757789609173637178721468440901 2249534301465495853710507922796892589235420199561121290 2196086403441815981362977477130996051870721134999999837 2978049951059731732816096318595024459455346908302642522 3082533446850352619311881710100031378387528865875332083 [01:47:17] 1415926535897932384626433832795028841971693993751058209 7494459230781640628620899862803482534211706798214808651 3282306647093844609550582231725359408128481117450284102 7019385211055596446229489549303819644288109756659334461 2847564823378678316527120190914564856692346034861045432 6648213393607260249141273724587006606315588174881520920 9628292540917153643678925903600113305305488204665213841 4695194151160943305727036575959195309218611738193261179 [01:47:17] 3105118548074462379962749567351885752724891227938183011 9491298336733624406566430860213949463952247371907021798 6094370277053921717629317675238467481846766940513200056 8127145263560827785771342757789609173637178721468440901 2249534301465495853710507922796892589235420199561121290 2196086403441815981362977477130996051870721134999999837 2978049951059731732816096318595024459455346908302642522 3082533446850352619311881710100031378387528865875332083 [01:47:20] 1415926535897932384626433832795028841971693993751058209 7494459230781640628620899862803482534211706798214808651 3282306647093844609550582231725359408128481117450284102 7019385211055596446229489549303819644288109756659334461 2847564823378678316527120190914564856692346034861045432 6648213393607260249141273724587006606315588174881520920 9628292540917153643678925903600113305305488204665213841 4695194151160943305727036575959195309218611738193261179 [01:47:20] 3105118548074462379962749567351885752724891227938183011 [01:47:22] ... [01:47:24] that's not a million [01:47:47] Reedy: i emailed diederik [01:47:55] as tfinc suggested [01:48:08] preilly: You missed the 3. ;) [01:48:18] and the rest of the million digits [01:48:19] YuviPanda: http://pastebin.mozilla.org/1555565 [01:48:34] still not a million [01:48:36] prepend the 3. [01:48:49] still not a million [01:49:04] damn it [01:49:04] http://www.piday.org/million.php [01:49:31] :D [01:49:36] "Computation timed out" on WA [01:49:45] :D [01:51:19] WA? [01:51:39] Reedy: ^^ [01:51:58] WolframAlpha [01:52:46] Reedy: ah [02:33:44] YuviPanda: so, what's the status on WCCCE? making reservations? [02:33:53] Amgine: i… don't know [02:33:55] i'd love to come [02:34:04] but i'd have to wait for response from tfinc or sumanah [02:34:41] Okay. The registration for the conference apparently costs money, and it will increase in price on April 12. [02:35:13] Only $25 more [02:36:22] I can't even find the dang registration page. [02:37:05] Amgine: getting into the conference is going to be easy. However, since all the teleportaiton devices in the world no longer work, getting me to Canada is going to be difficult [02:37:13] https://apps.cs.ubc.ca/WCCCE2012/frontpage.jsp [02:37:40] let's all chant: REEDY KNOWS EVERYTHING [02:37:49] 3rd result on google [02:37:50] srsly [02:38:03] chanting needs no reason [02:38:46] that actually dies for me, Reedy. [02:38:58] Weird [02:39:04] Are you on their network? [02:39:37] no, but I do use noscript. [02:39:52] apps.cs.ubc.ca seems a complete fail for this browser. [02:41:00] * Reedy blames Canada [02:41:10] Amgine: anyone of the students applying for GSoC? [02:42:25] I *hope* so. I have been pushing them to apply. I would love to see a GSoC to build the iOS/RIM or /other [02:43:01] RIM? Pffft [02:43:09] [02:43:11] * Mono[Away] was going to say that [02:43:45] No one I know has a Blackberry. Hopefully that will soon be true about Android. Some things should just die. [02:43:57] There's interesting stuff happening at RIM atm: executives flying out windows, but developers being bought up. [02:44:10] Doesn't sound like a CEO who is going to take apart the company. [02:46:04] Amgine: today is the last day, I guess [02:46:15] Android is adding millions of new users a day, it's not going anywhere for a while [02:46:18] O RLY? [02:46:19] Mono[Away]: hopefully that will soon be true of iOS. [02:46:20] * YuviPanda runs [02:46:31] Mono[Away]: 4.0 is *waaay* more polished than 2.x [02:50:18] YuviPanda: but most users are still on 2.3 [02:50:24] which makes the platform suck [02:50:27] for now [02:50:33] things get better :) [02:50:45] it's like the IE6 transition period [02:50:45] I read an article speculating that even Google would abandon it [02:50:48] :P [02:50:53] I hope not. [02:50:59] Except upgrading isn't even supported. [02:51:08] people buy new phones [02:51:11] Supposedly, to compete with the iPad and iOS platform, [02:51:36] they will launch 'Google Play' - a polished, Google-branded experience [02:51:52] 'and Android will become the cheap and open source option for the new generation of dumbphones [02:51:53] they have, and it is mostly slightly better crap than Android market [02:51:57] but sitll crappier [02:52:04] while also keeping their new glasses around [02:52:11] ah [02:52:20] I did not realize today is bash-the-google day [02:52:25] :) [02:53:01] At least I don't bash them for their privacy policy [02:53:16] YuviPanda: the article is at http://articles.businessinsider.com/2012-04-02/tech/31272944_1_the-android-android-marketplace-android-brand - it's an interesting read regardless [02:54:14] Just did. [02:54:17] speculation. [02:54:43] I, for one, prefer reading and writing code over speculation :) [02:55:50] of course. but it's a big deal for the industy [02:55:53] *industry [02:56:03] btw, who decided to go with MapQuest for tiles? [02:57:55] YuviPanda: ^ [02:58:48] Mono[Away]: I think it was a collective decision. Openstreetmaps.org didn't want the traffic, and MapQuest was okay with it, so we're with mapquest for now [02:59:49] Mono[Away]: aude plans on helping us set up our own tile server soon. [03:29:08] Which reminds me, aude, what kind of server would be needed for that? [03:47:28] Amgine: IIRC we were waiting on some gluster storage to come online with labs [03:47:32] and then git happened [03:47:41] woohoo, saved pages now work [03:47:43] sortof [03:47:49] \o/ [03:48:19] Amgine: we're in the process of decoupling the *data* of the mobile site and the app [03:48:24] while sharing more of the code [03:48:25] YuviPanda: Storage of the maptiles is different from generating maptiles. [03:48:38] Amgine: yes, but they are dependent on each other [03:48:52] i am setting up a small personal server of mine with tilestache, just for serving mbtiles [03:48:57] am sure that won't work for the wikipedia app :) [03:49:27] Yah, but I work with a host which has no throttle on bandwidth, but the storage on the server is limited. [03:50:16] We have some third party measures of 250GB transfer rates. [03:50:25] (rare peaks, but still...) [03:54:17] Amgine: :D [03:54:28] Amgine: the entire size of my tileset is about 5MB so I guess i'm good [03:55:50] [08:53:06] jdlrobson, do you see something on https://www.mediawiki.org/wiki/Special:Code/MediaWiki/114240 ? :P [08:53:16] * jdlrobson looks [08:54:26] looks like i'm closing a section before opening it ? :) [08:59:22] this function confuses me a lot MaxSem [08:59:47] yeah [09:00:14] gotta make a final push today to make it last a few weeks more:) [09:02:29] mmm so I'm not seeing where that div should be closed [09:55:59] MaxSem: I can't work out what's going on with those sections [09:56:12] I'll fix it later [10:00:54] much appreciated! giving me headaches.. [10:46:10] New patchset: Siebrand; "Fix some breaks." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4392 [13:49:52] New patchset: Jdlrobson; "hide clear search button when no text in input" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4401 [13:49:53] New patchset: Jdlrobson; "hide references on touchstart or clicking outside the reference box" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4402 [13:49:53] New patchset: Jdlrobson; "prevent jump when toggle" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4403 [13:49:54] New patchset: Jdlrobson; "improve smoothness of expansion transition" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4404 [13:49:55] New patchset: Jdlrobson; "on touchstart hide references instantly" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4405 [13:49:56] New patchset: Jdlrobson; "prevent automatic text resizing for landscape mode" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4406 [13:49:56] New patchset: Jdlrobson; "prevent page zooming when rotating to landscape" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4407 [14:23:58] MaxSem: I've built new debs in a new PPA with much nicer packaging magic and less insane versioning. [14:24:04] The new PPA is https://launchpad.net/~jerith/+archive/vumi-snapshots [14:37:57] whee [14:39:26] MaxSem: The copyright stuff in vumi-wikipedia says "Praekelt Foundation and individual contributors". Do you know what WikiMedia does about that kind of thing? [15:00:56] ajerith, https://gerrit.wikimedia.org/r/gitweb?p=mediawiki/core.git;a=blob;f=CREDITS;h=d0c4623d1a2ab63073b5a5ee6a0cd515a101db5e;hb=refs/heads/master [15:01:21] also https://en.wikipedia.org/wiki/Special:Version [15:02:59] MaxSem: Cool, thanks. [15:04:07] MaxSem: If and when you see preilly, can you point him at the new PPA? [15:04:23] sure [15:05:05] Everything installs fine on my VM, but that's the VM I built the debs on. [15:05:50] The vumi and vumi-wikipedia package names have changed, since they're more app-written-in-python than python-library. [15:06:21] Other than that, it's just version numbers and metadata and stuff that's changed. [15:07:00] That's a snapshots PPA, so we can pull updates from the dev versions of whatever to chuck in there as needed. [15:07:20] Once we have a proper vumi 0.4 release I'll build a releases PPA. [15:07:51] But now it's hometime. :-) [15:24:45] New patchset: Jdlrobson; "when initialising beta opensearch check whether search is already focused" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4424 [15:24:46] New patchset: Jdlrobson; "prevent page zooming when rotating to landscape" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4407 [16:27:49] New patchset: MaxSem; "Yet another fix for div fun. Looks like everything started with r114240" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4432 [16:57:11] * YuviPanda looks around [16:58:12] preilly, https://launchpad.net/~jerith/+archive/vumi-snapshots [16:59:19] MaxSem: thanks! [16:59:33] MaxSem: do you know where the redis PPA is? [16:59:44] no idea [17:00:00] preilly, http://dpaste.com/727848/ [17:00:23] greetings all [17:00:30] lets do our standup [17:00:36] * tfinc notes the jon and arthur will be missing [17:00:45] MaxSem: ah, okay — thanks, for relying the message [17:01:03] * YuviPand_ checks logs [17:02:39] preilly: Ping? [17:04:01] i heard sonar again [17:04:09] jeroid: he's in a call, give him a few minutes [17:04:47] Cool, thanks. [17:06:06] jeroid, I've passed everything to him [17:06:26] interesting. connection works fine whenever tfinc or brion are speaking and kills itself when preilly is talking. [17:06:27] hmmm [17:09:36] MexSem, preilly: you want rwky/redis for the redis PPA. [17:10:12] MaxSem: Thanks muchly. [17:10:58] YuviPanda: what? [17:11:08] preilly: your voice kills my internet [17:11:11] jeroid: okay [17:11:11] :) [17:11:17] YuviPanda: NOT OKAY [17:11:56] It is probably scared and/or intimidated. [17:12:03] I'll calm it down, should be fine [17:12:19] I'm pretty much out for the evening, but I may check in later if there's a suitable gap in the proceedings. [17:12:27] YuviPanda: i saw diedriks response [17:12:39] tfinc: and Reedy found me a way to actually measure things :D [17:12:43] woot! [17:12:46] do tell [17:13:10] tfinc: api has a action=clicktracking [17:13:16] yes! [17:13:24] and since were a mw extension .. we can use that [17:13:33] plus since the feature requires internet [17:13:41] and save pages also requires internet [17:13:45] we can definitely use that [17:13:47] lolz [17:13:48] i like this [17:13:53] i'll be slipping it into the beta [17:14:03] don [17:14:04] t [17:14:08] we have to check with legal [17:14:21] Www [17:14:22] i'll slip it into a branch then. [17:14:27] jeroid: okay [17:14:29] actually .. let me talk with howie to see what they did for the other projects [17:14:30] after other beta stuff is done [17:14:31] Oops. [17:14:40] so that we can see what we have to tell our users [17:14:44] but assuming thats fine [17:15:11] I'd *definitely* want to get this out for 1.2, so either I can convince myself save page is useful or convince you that it is not :) [17:15:33] we can 1) add the analytics to see current usage … useful in its own right [17:15:52] 2) add more ways to save [17:15:54] 3) measure [17:15:57] 4) re-asses [17:16:17] yes, this would indeed be useful data. [17:16:23] reassess* [17:16:27] reassess [17:16:46] YuviPanda: that way we give users more options and can migrate them over [17:17:14] yes. Once we have the data we could figure out what we could do. [17:17:19] Rather than speculating [17:17:24] OMG WIKIPEDIA IS SPYING ON US ONEONEONE [17:18:01] just name the js file da_man_at_da_nsa.js [17:18:09] tfinc: I also have another way to measure saved pages, even without using clicktracking [17:18:22] it's definitely a hack though, but it'll work. But only for saved pages. [17:19:45] since i've moved to using the API for saving, I can just pass an extra parameter that'll be ignored by the API, but can be used for analytics [17:23:51] YuviPanda: I've taken your vimrc lead and disabled the arrow keys :D [17:24:05]