and

bits [13:54:55] I will note that the gains are negligible practically, and I'm not sure how hard it is to implement [13:55:01] but it'd be nice-to-have [14:29:27] eh, WP app reloads page on pressing back? [15:38:41] MaxSem: yes [15:42:02] well, that sucks;) [15:42:22] browsers stopped doing it more than 15 years ago:) [15:42:32] MaxSem: what do they do instead? [15:42:41] load cached version [15:42:58] MaxSem: so, we do that if we've no netwokr [15:42:59] *network [15:43:52] MaxSem: cache first then network was our initial behavior. Then we realized we had no control over cache [15:44:02] MaxSem: so main pages started showing up several days old [15:44:35] cache expiry strategy would have fixed that [15:44:59] MaxSem: something which Android doesn't let you control [15:45:10] MaxSem: so the alternative was to do our own cache [15:46:24] since I have an Android phone now, I might be interested in doing that [15:47:35] MaxSem: aha [15:47:37] awesoem :D [15:47:49] MaxSem: it'd be super easily doable once the API work is complete. [15:47:59] because then we'll have a 'Page' object that can be thrown around and serialized [15:48:03] I'm already regretting this decision - that thing is huge [15:48:14] instead of the stateless ui-centric design we have [15:48:16] MaxSem: what phone? [15:48:33] SGS2. thanks crap not Galaxy Nexus [15:48:52] MaxSem: Tried the Galaxy Note? :P [15:49:36] why not windows-based 15" tabletm then? [15:50:29] MaxSem: because you're not out of your mind :) [15:51:17] really? I'm not that sure [15:51:57] you don't have a 15" Windows Tablet [15:52:02] hence you're not out of your mind [15:52:27] that's quite a narrow definition of sanity:P [15:53:12] MaxSem: :P [15:53:17] it is one that you pass [15:55:44] MaxSem: thoughts on mobileformat=html5 [15:55:50] and just adding divs to mobileformat=html [15:55:51] ? [15:56:58] not convinced that HTML5 is needed atm [15:57:07] me neither [15:57:10] but something on a backburner [15:57:13] adding divs? [15:57:26] wanted to discuss this with Jon [15:58:26] MaxSem: out till Monday then :( [15:58:34] well, I could do the other parts of thist ill then. [16:01:00] what's the problem with dynamic section loading? Jon already did this for MF [16:02:50] MaxSem: MF gives you divs [16:03:02] MaxSem: action=parse&mobileformat=html does not [16:11:26] action=mobileview even gives you the sections separately [16:53:01] maplebed: Message from VCC-compiler: [16:53:02] DNS lookup(195.189.142.0/23): Name or service not known [16:53:03] ('input' Line 27 Pos 9) [16:53:04] "195.189.142.0/23"; [16:53:05] --------##################- [16:54:47] preilly: this should be: [16:54:49] acl digi_ips { [16:54:49] "203.92.128.185"; [16:54:51] "115.164.0.0"/24; [16:54:52] "116.197.0.0"/24; [16:54:54] "195.189.142.0/23"; [16:54:55] "91.203.96.0/22"; [16:54:57] "80.239.242.0/23"; [16:54:58] "217.212.230.0/23"; [16:55:00] "141.0.8.0/21"; [16:55:00] "82.145.208.0/20"; [16:55:01] } [17:01:48] maplebed: https://gerrit.wikimedia.org/r/4004 [17:01:51] greetings all [17:02:03] we'll get started at 10:05 [17:02:10] preilly is just finishing something up [17:04:20] is wmf_office "carrier" for testing? [17:09:00] https://gerrit.wikimedia.org/r/#change,3930 is what i was referring to [17:13:12] ummm, was it terminated prematurely? [17:19:50] * YuviPanda waves. [17:20:11] sorry everyone, portable internet got shot and had to travel elsewhere with internet [17:20:41] preilly: load looks fine on 1041. Continuing to the rest. [17:20:59] tfinc: ^^ [17:21:07] YuviPanda: ahh .. that explains it [17:21:37] * maplebed imagines http://www.nytimes.com/2012/03/13/technology/homeless-as-wi-fi-transmitters-creates-a-stir-in-austin.html in combination with YuviPanda's recent comment... poor internet. [17:21:55] YuviPanda: for next time see if you can get that figured out beforehand . these check ins are better over voice then irc [17:22:14] but for now .. lets us know what you've been doing since our last check in, what your doing today, and any blockers [17:22:19] tfinc: will make sure, I think i've missed two in a row now. :( [17:22:25] yes. [17:23:10] yesterday was spent fixing some bugs in the code that produces 'Wikipedia Experimental' releases, and started work on moving to action=parse API [17:23:36] found a few issues that I've discussed with MaxSem, and I'm going to continue side stepping those issues for now until we could discuss with jon on monday. [17:24:17] today i'll start with investigating a few complaints fo regressions in the market (saved pages disapperaing, intent filter not working) [17:24:20] and then get back to the API [17:25:07] * tfinc pulls upgrade stats [17:25:14] I've seen a complaint about text becoming very small on market [17:25:37] preferences got reset again? [17:25:41] YuviPanda: https://support.google.com/googleplay/android-developer/support/bin/answer.py?hl=en&answer=2528691 !?! [17:25:51] finally [17:26:02] tfinc: :) [17:26:09] we don't have to do a stupid shared account [17:26:09] MaxSem: that technically shouldn't happen. [17:26:13] tfinc: \o/ [17:26:52] YuviPanda: i just sent you an invite [17:26:55] let me know what it looks like [17:26:56] tfinc: do you have more devices with 1.0.x market release around? [17:27:01] YuviPanda: yes [17:27:11] tfinc: test out a thing for me? [17:27:20] sure. what do you need tested ? [17:27:21] 1. Save a few pages [17:27:26] 2. Change font setting [17:27:28] 3. Upgrade [17:27:37] check if it stays [17:28:24] tfinc: i accepted, looks like you need to confirm the account [17:28:30] YuviPanda: we've had 500,000 updates in the last 24hrs [17:28:35] woohoo [17:28:44] 610,678 V1.1 users now [17:29:04] we'll overtake V1.03 tomorrow [17:29:08] :D [17:29:27] which means it takes less then a week to update all of our current users to a new version [17:29:36] thats pretty insane [17:29:42] with how quickly we can do it [17:29:50] YuviPanda: did you get the invite ? [17:29:52] maplebed: i'm technically homeless (been couch surfing for a few months now), so that makes sense :) [17:30:01] tfinc: i accepted, it told me that you'll need to 'confirm' [17:30:12] tfinc: I wish Android versions upgraded that fast [17:30:16] thats dumb as i sent out the invite [17:30:22] our daily uninstall rate is growing [17:30:25] thats a bit troubling [17:31:01] tfinc: do you know how it is compared to other apps? [17:31:08] YuviPanda: i'll ask around [17:31:45] YuviPanda: access granted [17:32:57] YuviPanda: i'm seeing lots of reports about lost pages [17:33:01] why is that happening ? [17:33:09] preilly: the change is live on all mobile varnish hosts. [17:33:15] tfinc: shouldn't, that's why I asked you to check. [17:33:32] i'm testing it right now as well [17:34:43] maplebed: okay, thanks! [17:40:20] YuviPanda: we've really bifurcated our reviewers its mostly 5 star with a trickle of 1 star [17:41:05] YuviPanda: update our new features section on the market and put a note that the update *might* wipe your saved pages [17:41:10] tfinc: I see the issue. [17:41:16] hmm ? [17:41:25] tfinc: upgrading doesn't show you your saved pages [17:41:41] however, they aren't gone - going back to v1.0.3 shows them back. Looks like a path issue [17:41:46] hehe [17:41:47] nice [17:41:58] we didn't really test upgrade paths :( [17:42:07] YuviPanda: i asked you to test upgrade paths ;) [17:42:25] tfinc: I hope you put a note on the iOS instructions about wiping bookmarks [17:42:32] nope, but i will [17:42:42] there is nothing that we can easily do on that one [17:42:57] tfinc: I was testing MFE upgrade paths. Not this. I figured that since that is unchanged urlcache code shouldn't need changes. [17:42:59] looking into it now [17:43:16] tfinc: me clicking 'my apps' tells me 'sorry there was a problem' [17:43:23] am unable to access it [17:46:24] my apps ? [17:46:45] 'my android apps'? [17:46:49] i guess that's the management interface? [17:47:00] takes me to https://play.google.com/apps which error [17:47:02] *errors [17:47:09] ohh [17:47:17] you mean the management interface [17:47:25] yes [17:47:37] YuviPanda: what does this show you https://play.google.com/apps/publish/Home [17:47:41] thats the link that i use [17:47:51] well, that loads [17:48:36] YuviPanda: so you can see stats, binaries, crash reports, etc ? [17:49:32] tfinc: yes i do [17:49:38] excellent [17:49:43] i'm no longer a bottle neck then! [17:49:54] YuviPanda: take a look at the upgrade stats [17:50:04] yes, moving fast! [17:50:30] tfinc: updating description [17:50:59] woot! [17:51:14] we should issue a 1.1.1 that fixes the saved pages issue [17:51:22] yes. i'm working on it [17:51:57] modified [17:55:54] hello [17:56:24] Hello, is this air boats? Yeah, hi, it's me space bot. [17:56:46] wow [17:57:04] My tongue says sangria by my heart says michelada. [17:57:27] preilly is this what happens when you delegate puppetization to jerith and company ;) [17:57:30] we're getting popular [17:57:38] :) [17:57:57] oh, I thought it was a run fo the mill spam bot [17:57:59] nevermind [17:58:21] /mode +q is ur frayend. [17:58:26] ArcherBot: karma YuviPanda [17:58:41] preilly: can we get an Eliza bot too? i'd love for archer to talk to her [17:58:44] so much karma that it crashed [17:58:44] hilarious ... [18:00:07] terrychay [18:00:09] tychay [18:00:10] Fact: The Internet doesn't go down, Terry Chay's laptop just goes into sleep mode [18:00:20] tychay [18:00:21] Fact: Terry Chay does not walk, or run, or drive, or take an airplane to a place. He simply uses a Keynote transition. [18:01:16] Announcement: Terry Chay joins WMF as Director of Features Engineering [18:01:17] Fact: Terry Chay doesn't write code, he just thinks it, and minions do it for him [18:02:59] thats too perfect for him just joining [18:04:06] tfinc: looks like a PhoneGap issue. [18:04:08] tfinc: You're getting our new improved deployment infrastructure before we are. ;-) [18:04:24] jerith: mwhaha … woo! [18:04:27] jerith: ha ha [18:04:29] Preferences persist in same app version when reinstalling, but not across app versions [18:04:46] Although you're getting a different one, probably. [18:06:02] How do you puppet-manage services that make persistent connections? [18:06:37] jerith: what do you mean? [18:07:19] Let me actually stop making assumptions and start again. [18:07:27] How does your deployment process work? [18:07:42] YuviPanda: when the app gets move to an sd card .. do our saved pages move as well ? [18:10:13] tfinc: yes, android takes care of that [18:10:36] I investigated that when we made the move [18:10:45] let me try again just to confirm. [18:10:58] jerith, I'm surprised with not hearing "wow, your code sucks" from you:P [18:11:11] cause I know it sucks:) [18:11:28] MaxSem: I try to be diplomatic. ;-) [18:11:38] YuviPanda: i'm tweeting about the saved pages issue [18:11:42] But actually, I haven't had a chance for a proper review yet. [18:11:43] okay [18:11:50] tfinc: move to SD card doesn't affect thing [18:11:51] s [18:11:51] s [18:13:07] YuviPanda: "Were aware of some users upgrading to V1.1 of the Wikipedia Android app who are not seeing saved pages. Were investigating now" [18:13:18] *We're [18:13:25] sure [18:13:46] sent [18:17:19] preilly: i've found some bugs in piano bar .. it doesn't properly share audio on osx. for instance I've had a number of cases where its not playing and no other app has audio until i kill piano bar [18:19:33] tfinc: lame [18:19:35] archer [18:19:35] : Karate? The Dane Cook of martial arts? No, ISIS agents use Krav Maga. [18:20:59] do it have a lana trigger ? [18:20:59] wow [18:21:37] we need a preilly bot. [18:25:45] archer [18:25:45] : Karate? The Dane Cook of martial arts? No, ISIS agents use Krav Maga. [18:25:48] archer [18:25:48] : Karate? The Dane Cook of martial arts? No, ISIS agents use Krav Maga. [18:25:52] archer [18:25:52] : If you want breakfast you should try the diner; you're obviously into greek. Get it? [18:26:18] tfinc: mobile-feeds:/var/www/archer/index.php [18:26:18] : I have to go, but if I find one dog hair when I come back I'll rub sand into your dead little eyes. I also need you to buy sand. I don't know if they grade it, but ... coarse. [18:34:20] tfinc: found the issue. [18:34:41] YuviPanda: nice. whats was it? [18:35:01] tfinc: PhoneGap. They changed the path in which the database is stored between 1.3 and .14 [18:35:03] *1.4 [18:35:14] so the data is still all there. [18:35:42] I'll write a small piece of code that's run on upgrade that'll migrate data from old path to new one [18:37:45] spiffy [18:39:44] very cool http://dillinger.io/ [18:40:33] tfinc: on it. But sortof pissed with the PhoneGap people. [18:40:43] this is a breaking change. And was not mentioned in the CHANGELOG [18:40:44] YuviPanda: was it in the release emotes ? [18:40:46] or anywhere. [18:40:47] no [18:40:51] http://phonegap.com/2012/01/31/phonegap-1-4-released/ [18:40:56] make sure to flame them appropriately [18:41:05] indeed. [18:41:16] their CHANGELOGS are simply commit messages [18:41:21] and they write shitty commit messages [18:44:18] * tfinc wonders what archer would say about RagingPanda [18:44:32] archer [18:44:33] : Hey Cyril. Cyril. CYRIL! I'm saving Lana, as usual! [18:44:59] archer [18:45:24] : You better call Kenny Loggins. 'Cause you're in the danger zone. [18:45:35] archer [18:45:35] : How'd you get life insurance, Lana? Don't they know you're in the danger zone? [18:45:43] archer is all well and good to have but it would be better to have mobile fronted check in bot be in here as well [18:45:43] : See that? He was putting on his pants, and I stopped him. So you just watch your step, mister ... dammit. [18:46:34] ArcherBot! [18:46:34] : How'd you get life insurance, Lana? Don't they know you're in the danger zone? [18:46:40] how boring [18:46:58] MaxSem: do you want me to remove it? [18:47:24] depends how much it will annoy us in the future:) [18:47:34] long long time ago, when I was a GSoC student, I started writing a brionbot. [18:47:36] hmmm... [18:47:37] later [18:48:22] archer [18:48:23] : Hm? Sorry I was picturing Whore Island. [18:49:22] archer [18:49:22] : So, thanks for nothing Carol. Now, I have to break into ISIS headquarters—in a $900 turtleneck. And if it gets ruined, I'm going to make you drink heavy cream, you ... Carol. [18:49:47] tfinc: the commit message that made this breaking change is... [18:50:00] Proved generating sqlite database path to open database without permission error [18:50:09] how very descriptive and accurate [18:50:37] useless [18:51:06] let me fix it first, can rage later [18:52:09] archer [18:52:19] ArcherBot has left the building [18:52:42] grrr, I've just vandalised it [18:53:18] preilly: pulse check on eta of work on project-specific js? [18:53:30] MaxSem: Why did you change it to GAME OVER? [18:53:41] Amgine: still no progress [18:53:59] because "MAX WAS THERE" would have been too obvious:P [18:55:09] maplebed: can you look at this file: /var/lib/gerrit2/review_site/logs/mobile.log: [18:55:17] maplebed: on that host [18:55:59] sure. [18:56:22] FixingPanda: this is the bug that i noticed last week but couldn't reproduce https://bugzilla.wikimedia.org/show_bug.cgi?id=34661 [18:56:52] tfinc: yup, but I think that can wait for 1.21 [18:56:53] *1.2 [18:57:11] sure, if its not a quick fix then it can wait [18:57:21] preilly: there is no file called 'mobile.log' in that directory. [18:57:23] sorry. [18:57:23] let me get this one out first. [18:57:41] FixingPanda: why don't you want to use "Beta" in the android market ? [18:57:54] tfinc: because my idea is [18:57:57] 1. finish one feature [18:57:58] 2. put it up [18:58:04] beta sounds a bit too 'late cycle' for that [18:58:12] This is in between a beta and a nightly [18:58:18] so i'd like to ideally call it 'experimental' [18:58:46] i don't know. people seem to have understood beta just fine over the last handful of years [18:59:13] what i don't want to happen though is [18:59:21] that we refer to it as a beta internally and something else externally [18:59:27] it has to be the same thign [18:59:35] hmmm [18:59:42] okay. Beta then. [18:59:49] Actually beta makes sense [18:59:55] since I can keep doing 'beta1', 'beta2' [19:00:00] 'experimental 1' doesn't quite work [19:00:03] nope [19:00:04] :D [19:00:05] easier for bug reporting [19:00:29] tfinc: alright, 'beta' then. Waiting on Icons from heather *and* an actual new feature before we push it out :) [19:01:00] woot. beta channel here we come [19:01:11] *nod FixingPanda [19:01:29] tfinc: we still haven't figured out 'feedback mechanism' for it [19:01:38] tfinc: the market reviews are one (they show us version too) [19:04:54] maplebed: I think it's templates/gerrit/hookconfig.py.erb [19:06:28] maplebed: so, do you think: "operations/*": "operations.log", [19:06:28] + "mobilefrontend/*": "mobile.log", [19:06:29] [19:06:32] maplebed: would work? [19:06:40] all irc bots suck [19:07:48] maplebed: https://gerrit.wikimedia.org/r/4015 [19:07:54] echo "all irc bots suck" >> mobile.log [19:08:06] maplebed: okay, cool [19:17:30] maplebed: mediawiki/extensions/MobileFrontend [19:19:33] maplebed: https://gerrit.wikimedia.org/r/4021 [19:20:48] ok, committing. [19:22:32] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/3960 [19:22:34] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/3960 [19:22:44] tfinc: ^^ [19:23:05] maplebed: thanks, for your help... [19:23:18] you're welcome! lunch time. [19:25:08] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (history); V: 1 C: 2; - https://gerrit.wikimedia.org/r/3894 [19:25:10] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (history) - https://gerrit.wikimedia.org/r/3894 [19:54:02] preilly: thanks for getting gerrit in here [19:54:08] so were getting mw->mf [19:54:18] are we also getting wikimedia mobile -> * ? [19:54:31] wait [19:54:35] i'm thinking about this wrong [19:54:50] what would it take to have a bugzilla bot in here [19:55:18] tfinc: like wikibugs? [19:55:24] yes [19:55:38] we would need it to monitor at least three areas [20:03:11] tfinc: okay, I'll look into it [20:11:04] lovely .. another budget meeting [20:11:06] * tfinc dies a little [20:50:04] http://hacks.mozilla.org/2012/03/people-of-html5-andrew-betts-on-building-the-ft-com-html5-app/ [21:19:05] tfinc: you sister is here [21:19:15] preilly: thanks. she just messaged me [21:46:08] oh, fun — localStorage to cache your JavaScript libraries [21:46:57] http://www.joezimjs.com/news/localstorage-script-caching-with-basket-js/ [21:48:27] http://addyosmani.github.com/basket.js/ [21:49:26] preilly: nice [22:02:14] tfinc: is tychay with you? [22:07:29] * FixingPanda slams self on wall [22:09:44] FixingPanda, take it easy [22:10:04] MaxSem: i've been spending the last 3 hours trying to get a fucking sqlite insert to work [22:10:09] still nothing [22:10:38] problem with sqlite or interface? [22:10:41] people who 'design' Java APIs must hate their users [22:10:57] Java is very interprisey [22:11:04] MaxSem: not sure. It tells me 'constraint failed', which is baffling because SQLite doesn't really enforce any constraints except Unique ID [22:11:05] even enterprisey [22:11:11] and the *one* row I am inserting doesn't violate that [22:11:22] nooo [22:11:27] Java is evil. [22:11:46] it also enforces NOT NULL - even stricter than that joke called MySQL [22:11:59] yes, but i did a .schema and there are no NOT NULLs [22:12:07] besides, i'm feeding it data for all three columns [22:12:15] show me your schema and query [22:12:25] String[] params = { oldData.getString(0).substring(0, 30), oldData.getString(1), oldData.getString(2) }; [22:12:26] v10Db.execSQL("INSERT INTO savedPagesDB VALUES (?, ?, ?)", params); [22:12:26] query [22:12:37] CREATE TABLE savedPagesDB (id NVARCHAR(32) UNIQUE PRIMARY KEY, value TEXT, timestamp REAL); [22:12:38] schema [22:12:57] MaxSem: even предприятие-y [22:13:13] ololo [22:13:33] Russian words aren't built this way;) [22:13:34] Java pushes old blind women down stairs while wearing an expensive suit. [22:13:53] MaxSem: I know [22:14:09] MaxSem: [https://en.m.wikipedia.org/wik, {"key":"https://en.m.wikipedia.org/wiki/Groundbreaking","title":"Groundbreaking"}, 2012-03-30T18:21:59.283Z] [22:14:19] is data for it [22:14:23] I'm assuming it doesn't enforce REAL [22:14:35] because this data is being read from another database with the exact same schema [22:14:39] (I am merging two databases) [22:16:54] SQLite has dynamic type system [22:18:50] http://techcrunch.com/2012/03/30/wikipedias-next-big-thing-wikidata-a-machine-readable-user-editable-database-funded-by-google-paul-allen-and-others/ [22:19:06] go wiki data [22:19:21] are you trying to insert "https://en.m.wikipedia.org/wik" exactly? [22:20:21] MaxSem: nope, I was trying to insert the full string, then *thought* maybe the NVARCHAR(32) constraint was being respected and throwing me the error and hence did a substr(0,30) [22:20:47] well, that's how it failed [22:21:02] first 30 chars are always the same [22:21:06] sqlite> select * from savedPagesDB; [22:21:07] https://en.m.wikipedia.org/wiki/Groundbreaking|{"key":"https://en.m.wikipedia.org/wiki/Groundbreaking","title":"Groundbreaking"}|2012-03-30T18:21:59.283Z [22:21:16] you'd notice that the first column is way more than 30 chars [22:21:20] also that's the only row in the table [22:21:21] yup [22:21:31] its type is TEXT [22:21:52] FixingPanda: when do you want to push 1.1.1 fix? [22:21:58] there's no nvarchar or other strongly typed heresy [22:22:13] MaxSem: indeed, so that create statement seems to be the work of some poor misguided soul [22:22:39] tfinc: in a few hours. That's what I'm bitching to MaxSem about, once that bit is done this should be ready to go [22:22:41] luckily, sqlite eats unknown types easily [22:22:59]