[00:17] things are looking quite a bit better on the automated tests with this one - mir is still pretty rough of course [08:36] lool, https://code.launchpad.net/~saviq/unity8/revert-r376/+merge/189455 [09:11] cjwatson: gnutls26 >> sure; thanks for the heads up [09:11] Saviq: looking [09:12] rsalveti: ok thanks [09:12] lool, I've a package coming up [09:12] lool, like in 2 mins [09:13] lool: uploaded, thanks [09:13] http://people.canonical.com/~ubuntu-archive/proposed-migration/saucy_uninst.txt is actually looking almost acceptable now, with the ignorable exception of arm64 [09:14] Saviq: cool [09:15] Saviq: should we sync a bit for plan over the week-end? [09:15] I do have some other planned outside activities too :-) [09:15] Saviq: so right now, we have the reverts in archive, I built an image with them; it's not the highest pass rate ever, but nothing to be too ashamed off and it generally works; most AP fails are in apps, a brightness thing in unity8 AP tests [09:16] generally speaking, I would be happy to promote is and call SF done [09:16] ogra_: ^ your toughts [09:16] *thought [09:16] lool, i'm just flashing [09:16] (morning etc) [09:16] ogra_: Morgen ! :-) [09:17] lool, ogra_ the biggest difference from my PoV is WiFi password input, which is in the indicator in the rolled back unity8, unity-notifications, indicator-network [09:17] Saviq: Now Mirv actually did the testing of the CPU pegging unity8 and was happy with it, so what I'd like is either getting us all these revs into the image with the fix you've identfied [09:17] the failures look like many of them are toolbar related [09:17] Saviq: Yes, I rolled back the 3 [09:17] lool, yeah, I'd rather do that [09:17] (from the apps) [09:18] lool, http://people.canonical.com/~msawicz/unity8/ here is unity8 trunk + the fix/revert [09:18] Saviq: so we do that on top of what was tested yesterday? I think you have landed other things in unity8 trunk already [09:19] lool, we only have a revert and a test-drop in trunk over what's released currently [09:19] lool, so with the "new revert" we'd have two reverts and the test drop, so rather safe [09:19] lool, and that's what you have in the packages I put up [09:20] bug 1235480 [09:20] bug 1235480 in upstart "upstart-local-bridge clobbers $PATH with its events" [Undecided,Fix committed] https://launchpad.net/bugs/1235480 [09:21] Saviq: looking [09:22] Saviq: right, trunk looks good [09:22] Saviq: I'll take the previous unity-notifications debs to test your unity8 build [09:22] lool, so if you confirm the fix/revert helps with CPU - let's merge, release [09:23] lool, yeah, it needs the "new" unity-notifications and indicator-client [09:23] s/-client/-network/ [09:23] lool, so full dist-upgrade + my packages should work good [09:25] Saviq: well I just took image 82 + your pcakages + old unity-notifications (qtdeclarative5-unity-notifications-plugin) [09:26] just trying to unmount now [09:27] lool, old unity-notifications and indicator-network will conflict with my packages - you need all three upgraded to their latest releases from trunks [09:27] lool: yeah I didn't notice the CPU hoggage, it didn't seem bad performance wise as such [09:28] maybe the autopilot tests themselves should notice it [09:28] "idle, check cpu usage" [09:30] lool, that is to say "sure, it'll work, but expect issues with WiFi password entry" [09:30] Saviq: "old" means the one before my revert [09:30] lool, ok then - yes :) [09:30] Saviq: yesterday I reverted indicator-network, unity8, and unity-notifications in archive [09:31] so I might get FS corruption, hopefully not [09:31] lool, yeah, saw that now [09:32] Saviq: so I have 7.82+13.10.20131004.1-0ubuntu2~sbuild3 installed, and top if quiet \o/ [09:33] *is quiet [09:33] screen of or screen on [09:33] Saviq: so that looks good [09:34] Saviq: let's upload all of this [09:34] lool, yay [09:34] lool, FYI some kind of SDK weirdness happened there - still need to investigate [09:35] lool, the spinner 'caused the CPU usage even when it was, in theory, off and invisible [09:35] -' [09:36] Saviq: you mean in the code that you reverted, it should not be using CPU but is due to some SDK bug? [09:37] Saviq: to clarify: are we good with just the revert we tested, or do we need another change for that spinner thing? [09:39] lool, yes [09:39] lool, we're good [09:40] lool, I reverted the change that was broken - and was due to a SDK bug [09:40] lool, with my MP unity8 is no-CPU-hogging anymore [09:40] lool, but we still need to merge / release it [09:40] Saviq: I've happroved the mp [09:40] lool, thanks [09:41] Saviq: I'm preparing the other uploads [09:41] Saviq: will send MPs to resync bzr trunks with changelogs of uploaded reverts [09:41] lool, yup, let me know if you need me to approve [09:41] (since nothing sits in trunk, can't ask cu2d to resync somehow) [09:42] * Saviq hates nautilus + gvfs recently [09:43] it was working so well last release, but with the new nautilus all kind of flakiness happens [09:45] that reminds me, I've noticed now that mtp coming up kills my adb shell connection for a sec [09:45] then it comes back [09:45] like the connection gets closed, I have to reopen it [09:50] uploaded indicator-network [09:50] it will be blocked in unapproved, then in proposed [09:51] Saviq: https://code.launchpad.net/~lool/indicator-network/resync-indicator-network-changelog/+merge/189458 [09:52] lool, +1 [09:52] lool, actually +2 - mtp is flaky for me, too - but that's with android as well [09:54] setting a property resets the gadget [09:55] when mtp comes up it adds a property to the gadget settings [09:56] Saviq: https://code.launchpad.net/~lool/unity-notifications/resync-indicator-network-changelog/+merge/189459 [09:56] ogra_: aha [09:56] ogra_: so it's "by design" [09:56] yeah [09:57] lool, +1 [09:57] Saviq: sadly, indicator-network needs someone from release team to ack [09:57] and setting the property persistent would make your gvfs on the desktop freak out [09:57] since there is an mtp device but no server attached [09:57] Saviq: and we're saturday, so might take some time [09:57] lool, but we can merge the changelog regardless, right? [09:57] I think so [09:57] (so gvfs goes into a loop of mount attempts spilling errors) [09:57] Saviq: I've put a block hint in place so that we can upload all 3 and let them in -proposed [09:58] so that we don't get half in unapproved half in archive [09:58] lool, oh cool [09:58] lool, and then just flip the switch to let through to saucy, got it [09:58] lool, thanks! [09:58] Saviq: right, so I can upload unity8 when it's there [09:59] lool, granted, the deps should prevent going into an inconsistent state [09:59] lool, actually there's one question you might be able to answer [09:59] Saviq: the provides are there, but I didn't update deps for indicator-network [10:00] lool, ah right [10:00] lool, we had a situation last week when unity-notifcations Provides was already unity-notifications-impl-2 (in the upstream merger's local repo) [10:00] lool, but unity8 Depends was still at unity-notifications-impl-1, which was available in saucy [10:00] lool, and apt crapped out when trying to install that unity8 version [10:01] which was the goal? [10:01] lool, is it not smart enough to install a non-highest version to fulfil the deps? [10:01] ah no [10:01] it will select the highest version [10:01] it's basically a complex scoring mechanism [10:01] which I don't know all the details of [10:01] but a higher version will get lots of point [10:02] lool, mhm [10:02] a downgrade wont happen unless your force it [10:02] lool, yeah, but it wasn't about a downgrade [10:02] lool, it was about installing (new package) a non-highest version [10:02] which I was hoping would work [10:02] that I would expect to work [10:02] but then maybe not :-) [10:03] Saviq: but if you remember the specifics, I guess some APT maintainers could possibly look into it [10:03] lool, fortunately we even have a log - 'cause it was on mediumtests [10:03] Saviq: Ok, so we're waiting on a bunch of things right now [10:04] lool, yup, all things to merge [10:04] to sum up: a) upstream merger picks up your MP, merges it b) I press the cu2d buttons and get that uploaded c) release team reviews + approves new indicator-network d) we unblock all 3 e) we build animage [10:04] ogra_: what did you think of latest image? [10:05] ogra_: should I promote it? [10:05] ogra_: we might get another SF one EOD [10:05] lool, well, just had to reboot trying to send an SMS [10:05] ogra_: but I'd rather we promote this one if you feel it's good enough [10:05] ogra_: what happened? [10:05] it hung [10:05] unity8? [10:05] not sure [10:06] any crash files? [10:06] i made a call, afterwards saved the contact ... (which used some http formatting for the plus sign in +49 ... which in turn couldnt send to that number full of special chars ...) [10:07] Saviq: I dont think we can fix the current SF images much more over the week-end; the original plan was to fix a couple of critical issues and switch to Mir; the critical issues were input in the webbrowser (might be related to volume up / down or not), and input focus on snap decisions [10:07] then i edited the contact to have a proper +49 ... got an error that contact saving failed and after that no input was possible anymore [10:07] Saviq: I can't fix the input issues myself, but if they get fixed I can land the "switch the default to Mir" over the WE [10:07] ogra_: fun [10:07] lool, nah, let's get back to it on Monday [10:07] ogra_: but is it a _regression_? :-) [10:08] lool, there's no one working on those [10:08] lool, let's all have our WE now [10:08] or after we fix SF [10:08] ogra_: is this something you experienced successfully in the past? [10:08] ogra_: basically, would this image be more or less substantially better than the previous stable one? [10:08] lool, the mangled number definitely is ... [10:09] lool, but i'm not sure thats a blocker [10:09] I fear we're running out of runway [10:09] right, lets promote it [10:09] after reboot there dont seem any issues [10:10] * lool reads ~ogra/README.mark-current [10:10] should i ? [10:10] ogra_: so mark-current is for the cdimage part [10:10] (if you feel safer with that) [10:10] ogra_: and the second is for system-image? [10:10] right, that needs to happen first [10:10] yep [10:10] ogra_: Actually, I like to do it because you're around :-) [10:10] ogra_: build is is 82? [10:10] just call both commands in succession with the right numbers [10:11] 82 for system image [10:11] ah and for cdimage I need to use the ubuntu version? [10:11] so 20131004.1? [10:11] 20131004.1 [10:11] right [10:12] the second one churns [10:12] I guess deltas [10:12] done [10:12] yeah, the second one takes a bit [10:13] ogra_, popey: 82 promoted! [10:13] the first one just sets a link [10:13] yay [10:13] ogra_, popey: Bulding an 83 tonight with updated 83 and SF, but otherwise leaving it there [10:13] popey, thats 82/20131004.1 [10:13] sweet! [10:13] might want to promote 83 [10:13] * popey mails [10:13] if you guys feel like it [10:13] ok, I'll be around [10:13] * ogra_ cant promise but will help if he is around [10:14] it's mostly what was in 81, with no CPU hogging, so 82 + new unity8 [10:14] and indicator-network [10:14] ogra_: yeah if you can't then so be it [10:14] dont you need a FFe for that one ? [10:14] we can always promote it monday [10:14] indicator-network is a desktop package [10:14] ogra_: it says ubuntu-desktop, but I don't actually see it pulled by ubuntu-desktop [10:14] an dseeded-in-ubuntu also doesn't [10:15] intresting [10:15] ogra_: well, I don't understand why it's still in the desktop packageset [10:15] i always thought we have it installed [10:15] missing meta update ? [10:15] or a dep ... [10:16] Saviq: So I'll step away for a bit, possibly only back after lunch + some family activities outside, near 5pm UTC+2 [10:16] lool, yeah, enjoy [10:16] Saviq: if upstream merger doesn't run or fails, would you merge by hand? [10:16] lool, will do [10:16] I might have commit access, not sure [10:17] Saviq: thanks for the fix, will be back in a bit [10:17] lool, I'll force-hand generic-land [10:17] lool, he has access for sure :) [10:17] ogra_: i dont see changes at http://people.canonical.com/~j-lallement/touch/changes/ [10:17] Saviq: http://www.chiark.greenend.org.uk/ucgi/~cjwatson/blosxom/ubuntu/2012-01-29-apt-resolver-bugs.html has some random wittering about how to investigate apt resolver bugs, if it happens to be useful [10:17] cjwatson, thanks - I'm not really sure it's a bug, and rather a feature ;) [10:18] cjwatson, but will be a useful read anyway [10:18] well, how to investigate the algorithm anyway. it's a useful though tricky thing to understand [10:18] popey: http://people.canonical.com/~ogra/touch-image-stats/20131004.1.changes has some [10:18] yeah, i have that [10:19] i usually link to the pretty version jibel has [10:19] nvm [10:19] ah [10:19] will link and say it's pending [10:19] given we can guess the url ☻ [10:19] I guess it's running from jibel's desk! :-) [10:19] popey: ah you probably want the diff between stbale images [10:19] I forgot where ogra has it [10:24] lool: no, I deal with that [10:24] lool: I manually merge the individual diffs [10:27] ok [10:31] mail sent [10:34] popey: thanks! [10:35] ugh [10:35] ok, so upstream merger is broken - the old IP was hardcoded in the jobs ;( [10:36] gosh, where is it hardcoded? [10:37] I love it when we move stuff to a new IP. Great way to shake out these kinds of issues. #glasshalffull [10:40] popey, yeah, and finally *name it* somehow [10:40] ☻ [10:41] popey, and whoa! here's an idea - let's provide a DNS server behind the QA VPN! [10:41] Steady on there! ☻ [10:41] that would be like The Enterprise [10:41] * popey emails /etc/hosts files around to everyone [10:41] ghetto dns [10:41] rotfl [10:42] bugs against ps-qa-tools? [10:42] pass [10:43] actually some other jenkins dep had the great idea to use the public facing jenkins.qa.ubuntu.com to get a stable name :-) [10:43] I don't remember which cron it was that got pasted here, but because it used jenkins.qa.ubuntu.com instead of 10.x.x.x it was immune to the change :-) [10:44] Saviq: where's this hardcoded IP you say? [10:44] https://bugs.launchpad.net/ps-qa-tools/+bug/1235621 [10:44] Error: ubuntu bug 1235621 not found [10:45] lool, so, first of all the links to jobs from history don't work (that, hopefully, is a configuration issue - otherwise only new jobs will have correct jobs cross-links) [10:45] lool, but http://10.97.0.26:8080/job/generic-mediumtests-saucy/4632/console [10:46] downloads artifacts directly from the old IP [10:47] indeed [10:47] popey, my DSL died after a storm and have limited connectivity. I'll relocate the job that publish changes during the week end [10:48] jibel: ouch! [10:48] jibel: thanks. [10:49] Saviq: I've just checked and don't have permission to fix it [10:49] lool, yeah, me neither [10:49] mailing Francis + Larry [10:49] fginther, https://bugs.launchpad.net/ps-qa-tools/+bug/1235621 https://bugs.launchpad.net/ps-qa-tools/+bug/1235622 [10:49] Error: ubuntu bug 1235621 not found [10:49] Error: ubuntu bug 1235622 not found [10:51] Saviq: ming merging unity8 by hand? [10:51] Saviq: (I've mailed them now, listing bugs and Cc:ing you) [10:51] lool, already there [10:51] lool, everything's is merged by now [10:51] let's see if cu2d is still working [10:52] lool, it's not been moved (yet), so hopefully yeah [10:52] lool, but it'd make sense to solve the above bugs for them *before* they move, yes [10:52] A version (7.82+13.10.20131004.2-0ubuntu1) is available at the destination for that component but is not in trunk which is still at 7.82+13.10.20131004.1-0ubuntu1. Ignoring that component for source: unity8, branch: lp:unity8, series: saucy. [10:53] guess we need to merge the changelog there too [10:54] Saviq: https://code.launchpad.net/~lool/unity8/add-reverts-in-unity8-changelog/+merge/189463 [10:55] Saviq: could you hand merge this one if you're ok with it, and then I'll run cu2d again? [10:56] off for lunch [10:57] lool, going [10:57] lool, it's going to be merged in 30s [10:57] lool, if you can kick off cu2d again [10:58] lool, ready [11:02] ok [11:03] running [11:22] * ogra_ notes that the session upstart goes wild after some time on 82 [11:24] * ogra_ sees a load of 17.08 [11:27] (using Mir btw) [11:27] Saviq, the IP has been updated [11:28] ogra_, oh, I wonder if it'd be bug #1235190 [11:28] bug 1235190 in mir (Ubuntu Saucy) "[mako] Unity8 on Mir got slow" [Critical,Confirmed] https://launchpad.net/bugs/1235190 [11:29] Saviq, qwll, i suspect it is also some issue with upstart [11:30] i see init consuming a lot of CPU [11:30] ogra_, although CPU was idle for us mostly - same for IO and RAM - everything looked fine resource-wise, only display slowed down to a crawl [11:30] ogra_, so yeah - not the same - our issue was with idle CPU [11:32] root@ubuntu-phablet:/# ls /var/crash/|wc -l [11:32] 6 [11:32] bah [11:32] i had cleared it for the last reboot [11:33] its only up since a few minutes [11:34] seems /usr/lib/arm-linux-gnueabihf/upstart-app-launch/zg-report-app is one of the crashers [11:36] hwee [11:36] whee even ... [11:37] and the session upstart is leaking ram like crazy [11:37] seems RES rises by 2M every second while the screen is on [11:38] it is at 141M now [11:38] * ogra_ makes the image writable and installs last nights upstart upload [11:39] i wonder if that fixes it [11:41] i have one unity crash and one _usr_lib_arm-linux-gnueabihf_upstart-app-launch_zg-report-app.32011.crash [11:41] but don't see the memory leak [11:41] using Mir ? [11:43] ogra_: it might be when dealing with crash files [11:43] ogra_: there's a file watch on this [11:43] lool, does upstart deal with them in the session ? [11:43] hmm no, you're right [11:43] i thought thats the system upstart [11:43] ogra_: it could logging [11:43] ogra_: ls -ltr .cache/upstart/ perhaps [11:44] well, i installed the ubuntu6 package from the archive [11:44] lets see how this behaves [11:46] no wilddly growing files in the logdir [11:47] the session init is at 27M already [11:47] and pumped itself to 32M while i typed the above [11:47] 39 [11:47] 41 [11:47] 44 [11:48] it only rises while the screen is on [11:48] seems to stay where it is with the screen off [11:48] 53M btw [11:48] check initctl list as phablet? [11:49] http://paste.ubuntu.com/6196152/ [11:50] nothing obvious [11:50] gtg [11:50] unity8 is building in cu2d [11:50] didn't pick up bzr the first time, but did the second time [11:50] ping once you build a new image [11:52] ok [11:54] ogra_: no, sf [12:09] Saviq, lool, the jenkins IP problem is resolved [12:15] so after ~30min it starts swapping like crazy [12:15] http://paste.ubuntu.com/6196223/ [12:15] which is what i saw initially as UI hangs [12:16] * ogra_ reboots into SF [12:51] popey, bug 1235649 [12:51] bug 1235649 in upstart (Ubuntu) "session upstart leaks massive amounts of memory on Ubuntu Touch" [Critical,New] https://launchpad.net/bugs/1235649 [12:52] how do I enable mir? [12:52] touch /homa/phablet/.display-mir [12:52] and reboot [12:53] k [12:53] apport as soon as I boot [12:53] *home (indeed) [12:59] confirmed [12:59] thx [14:53] fginther: awesome, thanks [14:54] == Publishing unity8 == [14:56] in Unapproved [14:57] in -proposed [14:57] I'll lift the blocks in an hour or so, once it's all built [15:19] lool: you mentioned lifting blocks in an hour. is that why I can't upgrade to 82 right now? [15:46] jdstrand: nope [15:47] jdstrand: what's your issue upgrading? [15:47] jdstrand: I'm just about to lift some block hints on 3 packages we want to let out of proposed together [15:48] (lifted) [15:53] lool, did you see the above bug ? [15:53] pertty much a Mir blocker i think [16:12] ogra_: yeah [16:13] ogra_: it would also explain why it crashes after a while [16:13] right [16:13] the device seems to start swapping once i hit ~250M ... from then on i can only use the UI every few mins [16:13] ogra_: I'll raise this to jodh [16:13] but I guess only Monday [16:14] what i find really odd is that this hasnt occured before to me ... and i actually ran Mir for a day or more beofre [16:14] it's also weird that upstart would have loads of memory [16:14] *before [16:14] ogra_: did you see anything pile up in initctl list as ~phablet? [16:14] like, many apps starting [16:15] cause I mean upstart shouldn't have much tracking to do [16:16] the amount of lines initrctl list returns doesnt change [16:17] the amount of running apps neither [16:17] it is somehow related to screen activity ... [16:31] ogra_: I was actually thinking, you might be able to strace this upstart from the user session [16:31] it's not great to see where memory goes, but you might be able to tell what it's doing [16:31] like starting processes, opening fds, writing to files [17:07] ok, pkgs are in release pocket now [17:07] uh no [17:09] looks like if I ever block them once, removing the block isn't enough, I need to unblock them explicitly [17:13] ah no, britney stopped running [18:11] popey, ogra_: No -proposed migration for now until britney/firewall are unstuck; will ping if there is an update [18:14] k [18:24] lool: ok [19:36] ogra_, popey: Unstuck, building an image [19:37] great [20:01] ok [20:41] popey, ogra_: Image was up; I can't install Clicks from the store anymore; could you do this with 82? can't remember whether I tested it [20:42] either it was already in #82, or it's an unity8 thing given the list in http://people.canonical.com/~ogra/touch-image-stats/20131005.changes [20:43] lool: pretty sure I installed from store [20:43] let me try now [20:43] with #83, I press Install and nothing happens [20:43] lool: yes, i have ureadit on my 82 image, which i installed from the store [20:43] cool [20:44] want me to try? [20:44] popey: sure [20:44] ok [20:44] just tried again on 82 to make sure it's not a backend issue, and installed something fkine [20:44] popey: important thing for me was confirm #82 wasn't broken in this respect [20:44] *fine [20:44] * popey flashes [20:45] if #83 works for you, then perhaps there's something wrong here I need to figure out [20:45] * popey flashes [20:45] and it's bad for you too, then I guess I can open the unity8 bug early and we know we wont promote it [20:45] ya [20:45] (I actually upgraded, didn't flash) [20:46] [unity-scope-click] - DEBUG: click-scope.vala:253: got details: Akari [20:46] [unity-scope-click] - DEBUG: click-scope.vala:255: getting creds [20:46] [unity-scope-click] - DEBUG: ubuntuone-credentials.vala:35: Using account id: 1 [20:46] that's where .cache/unity-scope-click.log stops [20:46] it might be a server side issue too [20:46] in which case you would see this with #82 too [20:46] i happened to install that same app [20:46] just now, on 82 [20:46] they were changing stuff recently to cope with authenticated downloads for free apps failing [20:46] popey: and that worked? [20:46] yes [20:46] cool [20:47] and the game runs [20:47] :-) [20:59] lool: installed fine on 83 [20:59] popey: apps? [20:59] yes [20:59] ok, in a way it's good news [20:59] hmm [21:00] in another, well [21:00] well. they installed [21:00] I suspect something related to the oauth stuff happening behind the scenes with my U1 account [21:00] but i dont see them in the dash [21:00] I also played for the first time with the download-manager settings in OS updates [21:00] I told it to never automatically download [21:00] but it should not affect downloads of click-scope [21:00] but if i search, i find it and it runs [21:00] popey: oh yeah, that's a known bug, you have to reboot to see them [21:00] ah [21:00] or search forthem [21:01] this has been for weeks [21:01] no, just locking is sufficient [21:01] just locked / unlocked and both apps show up [21:01] ah that's an improvement [21:12] filed https://bugs.launchpad.net/ubuntu/+source/unity-scope-click/+bug/1235784 for the click thing [21:12] Ubuntu bug 1235784 in unity-scope-click (Ubuntu) "Click install stalls; nothing happens when pressing Install button" [Undecided,New] [21:12] if it's only me, then it shouldn't get in the way of promotion; it's probably unrelated to the recent upgrades, just a bad oauth state somewhere I'd guess [21:12] lool, popey: just discovered the lock/unlock in 82, may have work previously I'm not sure, I can't reproduce reliably it may just be that some process needs time so lock/unlock late enough and it seems to reveal the apps ? [21:12] anyway, tests will be running, we can assess promotion on Monday [21:13] vila: what do you mean? [21:13] vila: there's sometimes a window where the lock screen doesn't appear immediately; is that what you meant? [21:14] lool: the lock/unlock trick to reveal click apps already installed [21:14] oh ok [21:14] lool: I was rebooting before more or less randomly to get them back, found them back after a lock/unlock, came here, saw popey mentioning it. Synchronicity ;) [21:15] :-) [21:15] Ok, going off, I don't really expect we can do much more this week-end anyway [21:15] ok [21:15] popey: Thanks a lot for the testing [21:15] np [21:15] popey: if you're playing with #83 and you find it's good, let me know and we can promote early on Monday with ogra's ack [21:15] I'll keep an eye on the tests [21:15] will do [21:16] they just started