=== elopio_ is now known as elopio === duflu_ is now known as duflu [08:16] mzanetti: ping [08:22] mzanetti: goint to need a qmluitests VM detached when you have time to try to repro that test that is now suddenly blocking us from out of the blue [08:42] tsdgeos: hey [08:42] tsdgeos: I'll get you the vm now [08:42] tx [08:46] tsdgeos: ps-trusty-server-amd64-1 it is [08:46] kgunn: o/ [08:46] tx [08:47] mzanetti: \o yo! [08:47] kgunn: fixed the last glitches [08:47] cool...i'm all set up...i'll resync this afternoon [09:29] * tsdgeos does the evil eyes at Saviq [09:29] :D [09:30] * tsdgeos does the evil eyes to himself [09:35] mzanetti: https://code.launchpad.net/~aacid/unity8/fix_test_dash_shown/+merge/201899 [09:35] :D [09:36] mzanetti: you can bring back the vm to the pool [09:36] * mzanetti wonders why tsdgeos doesn't do the evil eyes to him [09:36] vm is back in pool [09:36] mzanetti: you didn't do nor review the change, no? [09:37] ah right... no, I didn't do that change [09:37] you're saved from evil eye-ing [09:37] tsdgeos, uh oh, sorry [09:37] Saviq: there's a few more checks against undefined that should be probably null [09:37] but they are == or != and that still works [09:38] it's === or !== that doesn't [09:38] tsdgeos: hmm... but how could this have gotten merged? I mean, shouldn't those tests have failed already with the bad commit? [09:38] mzanetti: because it's a timing thing [09:38] ah [09:38] that function there [09:38] checks that everything we're going to need is loaded [09:38] oh, I got it [09:38] by making sure they are there [09:38] i guess we were unlucky and the merge ran on a fast machine :D [09:39] yeah, the vm's performance is quite variable depending on how many jobs are running [10:29] MacSlow, hey, could you review https://code.launchpad.net/~sam92/notify-osd/focusfollowdefault/+merge/198541 as well? [10:31] greyback: hi [10:31] greyback: just merged trunk and as I feared,everything broke badly [10:31] TaskController::startApplication appId='dialer-app' FAILED [10:31] Asking Upstart to start application 'dialer-app' failed [10:31] any idea? ^ [10:32] greyback: if I try for a second time it crashes: (null):dbus_error.c:69: Unhandled error from nih_dbus_error_raise: Name "com.ubuntu.Upstart" does not exist [10:33] seb128, ok [10:33] MacSlow, thanks [10:34] seb128, but mpt also approved it already... but it's good to go from my side too [10:34] seb128, I can top-approve it [10:34] seb128, done [10:34] MacSlow, thanks [10:36] greyback: forget about the above... no idea what went wrong, now it works - sorry about the noise === _salem is now known as salem_ === pete-woods is now known as pete-woods|posti [11:25] mzanetti, hey, you happy to show what you have for right edge? [11:26] Saviq: yip yip [11:26] mzanetti, what do I do? [11:26] Saviq: when did you last flash your phone? [11:26] Saviq: do you already have the sidestage code from ricmm on your device? [11:26] mzanetti, I don't think so, checking [11:27] I'm at image 127 [11:27] Saviq: hmm.. no idea which one that is [11:27] Saviq: better flash it with a fresh one to be on the save side [11:28] mzanetti, 0.2+14.04.20140108.1-0ubuntu1 [11:28] mzanetti, will do, thought it was "you should have an old one"-type of a problem [11:28] Saviq: I just merged everything with trunk [11:29] Saviq: if your image is an older one you could use my branches before the "merge trunk" commit [11:29] mzanetti, upgrading now [11:29] ack [11:29] Saviq: so, you need this branch: lp:~mzanetti/unity-api/new-screenshot-and-focusing-api unity-api [11:30] it's enough if you copy the Application*.h files to /us/include/... [11:30] then you need this one: bzr branch lp:~mzanetti/unity-mir/screenshotting-focusing-api unity-mir [11:31] need to compile + install [11:31] mzanetti: hey, sorry was away, that error sounds like upstart broke a bit, as that message says the dbus message to upstart failed to send [11:31] and then this one: bzr branch lp:~mzanetti/unity8/appmanager-rework [11:31] you can run_on_device this one ^ [11:32] greyback: no worries. I still don't know what happened, at some point it was working again... the first 2 reboots I always got this error [11:32] greyback: but can't reproduce it any more. so everything fine I guess [11:32] mzanetti: I had it once before too [11:39] ricmm_, greyback https://bugs.launchpad.net/ubuntu/+source/unity-mir/+bug/1269414 you saw that? [11:39] Ubuntu bug 1269414 in unity-mir (Ubuntu) "Sound is cut when another application is launched" [High,Triaged] [11:55] tsdgeos, mzanetti, can you please do https://code.launchpad.net/~mterry/unity8/hide-greeter-on-focus-request/+merge/201817 ? [11:55] we should add an ap test for that... [11:57] Saviq: ack [11:59] greyback: so i was trying to build unity-mir on my device...which i did yesterday, i just updated [12:00] the branch....but now, when i hit qmake...it spews the help [12:00] seemingly [12:00] ever seen this [12:00] kgunn: migrated to cmake [12:00] build with dpkg [12:00] thanks [12:00] yeah, we switched to cmake [12:01] need to see if I can X-compile it now [12:01] greyback: if you manage to, let me know [12:01] last time I tried click chroot didn't like my system at all [12:01] mzanetti: will do [12:04] mzanetti, FOOK 6k diff? think we could split it up in smaller chunks? [12:04] Saviq: no [12:04] :/ [12:04] Saviq: and it'll get about double the size [12:04] Saviq: because we're not supposed to break anything [12:04] Saviq: so the new sidestage has to be in this too [12:05] Saviq: but most of it is dropping old FIXME's :) [12:05] mzanetti, we're hammering out the side stage case this week, too, so hopefully there will be direction (or there might be an interruption anyway) [12:06] Saviq: yeah, I heard. vesar is busy prototyping it. looking forward to it === MacSlow is now known as MacSlow|lunch [12:06] mzanetti: you doing mterry's review, then, right? [12:06] tsdgeos: yes, can do === pete-woods|posti is now known as pete-woods [12:20] Saviq: where did you see a 6k diff btw? [12:20] mzanetti, unity8 [12:20] Saviq: https://code.launchpad.net/~mzanetti/unity8/appmanager-rework/+merge/199815 [12:20] 1.5, no? [12:20] mzanetti, against your branch [12:21] mzanetti, hmm wait [12:21] mzanetti, merged on wrong branch [12:22] mzanetti, 1.5k is fine [12:22] Saviq: yeah. it's ok. so this is only the Phone stuff... the tablet will come in with another 500 - 1000 lines approx [12:26] Saviq: does it work? [12:26] mzanetti, not yet, couldn't cross-build :/ === alan_g is now known as alan_g|afk [12:30] noOoOOoOoO [12:30] whitespace [12:30] :D [12:31] Mirv: if you have not started a qtdeclarative rebuild [12:31] you want to include this one [12:31] Mirv: https://codereview.qt-project.org/#change,75689 [12:32] tsdgeos: ok. well I "started" but KDE is eating all the builders so it has not started yet. reuploading with that added. [12:32] Mirv: awesome, it should fix that crash in i386 [12:34] tsdgeos: do you have qtdeclarative release branch git checked out? could you check if I should update from the cb78684ae or if the 10 commits on top of that are not crucial? like by looking at the commit descriptions. [12:34] xnox, hey, something changed and Qt cmake modules are not found any more when cross-building :/ any idea? [12:34] Mirv: sure, let me check [12:35] to me maybe thiago's d580411ca looks important, but not much else [12:35] on the other hand if it doesn't look too important, it shouldn't hurt either... [12:36] Mirv: 2f9099443d9acd6583e92785afbb38b2e4dcbfd5 looks like something we want to have (no clue what it does but sounds complicated, i.e. could break/fix stuff :D) [12:36] Mirv: if updating to current tip of release is not hard i'd do it [12:37] ok then :) [12:40] mzanetti, ApplicationImage.qml: File not found? [12:40] damn ^W [12:40] salem_: where do you get this? [12:40] Saviq: ^ (sorry salem_) [12:41] mzanetti, hmm it didn't build, let me [12:41] mzanetti, ah that's "real" u8 [12:41] Saviq: ? [12:41] mzanetti, distro u8 with new libunity-mir1 [12:41] mzanetti, so makes sense [12:42] ah ok, yeah [12:42] I dropped AplicationImage.qml [12:42] mzanetti, yup [12:43] Saviq: btw... u8 trunk won't start with my libunity-mir, which causes libhud to go to 100% cpu [12:43] mzanetti, mhm [12:43] Saviq: so make sure to reinstall unity-mir from the repo before leaving the device unattended for a while [12:43] or it'll get hooot [12:43] mzanetti, :) [12:44] actually I should file a bug about it [12:59] Saviq: ping: are we eventually planning on having the unity automatically detect the screen resolution settings? or will we always need to drop a file in /etc/ubuntu-touch-session.d/? [13:00] cwayne, you mean DPR / GRID_UNIT_PX? [13:00] yeah, specifically GRID_UNIT_PX [13:00] cwayne, those are arbitrarily set, so AFAICT there will always be a per-device file there [13:00] cwayne, 'cause it's not really DPI [13:01] right, i was worried about that :) [13:01] Saviq: where are those files actually read? [13:01] i.e. is it just unity8 that actually reads the contents of those files? [13:01] cwayne, no, it's being exported in the session I think [13:02] mzanetti, can't launch apps, and run_on_device complains about pkilling processes that're not running? [13:02] mzanetti, ah wait, got dialer [13:03] mzanetti, what's the chopping at the top? [13:06] mzanetti, is the "I can swipe enough to go back to the app when reached the end of the stack? [13:06] +designed behaviour? === MacSlow|lunch is now known as MacSlow [13:14] Saviq: no. that's todo still [13:14] Saviq: design is not exactly sure yet where the point of no return shall be [13:14] mzanetti, you mean todo to fix? [13:15] mzanetti, ok, how do we fix the cropping? [13:15] Saviq: what cropping? [13:16] mzanetti, - yOffset is not needed any more [13:16] ? [13:16] mzanetti, application.cpp in libunity-mir [13:16] mzanetti, the screenshots coming from mir are now the actual size of the surface [13:17] Saviq: re launching of apps: currently you need to launch apps with the launcher as we didn't merge the fix to unity-scopes yet [13:17] mzanetti, 'stood [13:17] Saviq: I don't see any cropping [13:17] mzanetti, merge trunk libunity-mir [13:18] mzanetti, you'll see it [13:18] Saviq: hi === ricmm_ is now known as ricmm [13:18] ricmm_, FUCK OFF [13:18] again? I did that today already [13:18] damn. conflicting again [13:19] so yea yOffset is not needed as app buffers are now the size of the render area requested [13:19] in accordance to the available geometry in the selected stgage [13:20] ricmm, in the selected what?\ [13:20] Saviq: sthfgage [13:20] stg007Fage [13:20] mzanetti, you're being missed here [13:21] am I? why is that? [13:21] mzanetti, just because ;D [13:21] mzanetti, it's late, I didn't sleep much, hungover, don't listen to me ;D [13:21] is the beer any good down there? [13:22] lol. do you have the drunken voice again? [13:22] mzanetti, it's *CHEAP* to start with [13:22] not any more ;) that was Monday :D [13:22] :D [13:22] mzanetti, a pound a beer, 3-course dinner with a bottle a wine a head, 16 pounds [13:23] Saviq: hi [13:23] and they give you half a cow [13:23] ricmm, FUCK OFF [13:23] ↑↑ he's sitting →→ [13:23] o/ [13:23] :D [13:25] * mzanetti just realizes that the app suspend whitelist will mess with the screenshot stuff we're doing [13:25] i.e. the screenshots are most likely always outdated [13:26] oh not at all [13:26] it still takes screenshots of suspended app's buffers [13:26] and technically the app wont have changed if its suspended [13:26] yeah, but the app is not suspended so it will change after the screenshot has been taken [13:26] that's what I mean... the whitelist [13:26] oh the whitelist [13:26] sorry [13:26] np [13:27] mzanetti, the two last (bottom ones) screenshots get swapped sometimes [13:27] mzanetti, you know about that? [13:27] so the whitelisting is a hack [13:27] Saviq: yep... that will implicitly be fixed with the point of no return [13:27] lets not write special code to support that outdate screenshot of music app [13:27] ultimately it wont use whitelisting [13:27] mzanetti, ok [13:27] but the media hub [13:27] ricmm: ack [13:39] mzanetti, lookin' good! === alan_g|afk is now known as alan_g [13:40] Saviq: nice :) [13:40] mzanetti, noticed that fullscreen apps might need to be handled differently [13:40] mzanetti, or are we not meant to put the toolbar on them? [13:40] Saviq: yeah, that's one of the questions I had for design [13:40] they don't know a solution yet [13:40] the prototype and mocks only dealt with non-fullscreenones [13:41] mzanetti, no I mean that weren't we supposed to put the panel on the non-fullscreen ones? [13:41] Saviq: hmm... dunno. I'd think that looks weird, but who knows. maybe that'll be designs answer [13:42] mzanetti, we will probably have to "composite" the OSK on the apps anyway [13:42] well, it solve some more issues tho [13:42] mzanetti, yeah, and I thought videos were doing that (panel) [13:42] yeah, martins videos did [13:43] Saviq: vesar's prototype doesn't [13:43] mzanetti, I think the video is more of what we should be looking at, but yeah a confirmation is needed [13:44] hmm... no that was the mistake I did at first... I implemented the video while the prototype is the one where the changes from the reviews are in already [13:45] but we'll see... I have another design feedback call either today or tomorrow - depending on how busy things are on the southern hemisphere [13:45] mzanetti, right, visual vs. behavioral [13:46] mzanetti, great [13:46] but overall it's quite cool isn't it? [13:46] makes the right edge so much more useful [13:50] Saviq: I merged with trunk again, still nothing clipped off [13:51] mzanetti, there's been changes to platform api and qtubuntu that changed that [13:51] ah... [13:51] mzanetti, so apt-get update/upgrade should show you that [13:52] * mzanetti wishes he waited for another hour with the whole updating/merging thing :D [13:52] mzanetti, conflicts? [13:52] yeah, lots of them [13:53] :/ [13:53] but merged already... just seems I'm not up to date again [13:56] still good after an upgrade... /me tries dist-upgrade [13:56] nope, nothing in there [14:01] Saviq: did you manually install something on top? ^ [14:02] Hey guys. Is the Y axis in Qt layouts pointed downwards from upper left corner, or upwards from lower left? I seem to be calculating correct values for vertical journal row spacing, but with the opposite sign. second.y - (first.y + first.height) [14:02] karni: 0, 0 -> upper left [14:02] mzanetti: Thank you [14:03] mzanetti, no, I flashed latest trusty-proposed and installed your unity-mir, built unity8 [14:03] strange... looks fine here [14:11] Saviq / tsdgeos: https://code.launchpad.net/~unity-team/unity8/new-scopes-vj-integration/+merge/201932 [14:11] tsdgeos: Thank you for your e-mail :) [14:11] karni: no worries [14:11] * karni already notices copyright header wrong year heh [14:12] :D [14:13] karni, measuredTwoLinesHeight → out [14:15] Saviq: ow WOW sorry, that wasn't supposed to land there. yes, we talked about it [14:15] * karni removes [14:15] karni, no worries [14:16] karni, I'd also s/columns/maxColumns/ probably [14:17] Saviq: The thing is, that's useful for ResponsiveGridView. In this case, we never make the number of columns less than "columns" set, because of large column spacing or other factors. [14:17] Saviq: That's why I thought naming it just columns would be more appropriate, as it always matches maxColumns. Thoughts? [14:18] karni, not sure I followed [14:18] Saviq: ResponsiveGridView - when you set maxColumns = 1000, it may happen number of columns will end up 23, because of delegate item with and column spacing [14:19] karni, yup, why is here different? [14:19] Saviq: in ResponsiveVerticalJournal - whatever you set the columns (i.e. maxColumns) value, that WILL always be the value of columns [14:19] karni, ah [14:19] karni, I don't think that should be the case [14:19] That's where my question came from if we shouldn't limit the number of columns from arbitrary values [14:19] karni, the number of columns should be dynamic [14:19] in which case I'd agree we should name it maxColumns instead [14:20] karni, we *know* the size of the cards (so column width) [14:20] karni, what we don't know is spacing or column number [14:20] hrm [14:21] I ended with conclusion we don't know delegate size, but what you're saying makes sense. [14:21] karni, we do [14:21] That'll need fixing, and with that, columns -> maxColumns makes sense [14:21] karni, it's small, medium or large [14:21] correct [14:21] my bad! [14:21] karni, no worries [14:21] Saviq: -> Needs fixing :) [14:22] karni: https://code.launchpad.net/~aacid/unity8/journal_misc_fixes ? [14:22] karni, done [14:22] * karni looks [14:22] since daniel is back pained [14:22] i'll let you review it [14:23] who reviews https://code.launchpad.net/~aacid/qtubuntu-camera/no_priv_headers/+merge/201933 now that gusch is gone? [14:23] Mirv: ↑↑↑ [14:23] sil2100: ↑↑↑ ? [14:23] rsalveti: ↑↑↑↑ ? [14:23] tsdgeos, start with bfiller [14:24] tsdgeos: Jenkins failure is expected on that branch, right? [14:24] karni, one interesting thing to know would actually be whether cards should be resized or spaced out [14:24] karni: i retriggered the build [14:24] we have a pretty bumpy CI [14:25] tsdgeos: approved [14:25] Saviq: Correct, I came through that, gave it all much thought. Should I consult Katie? [14:25] anyone know if it's possible to detect a click on a qml area, but not consume the events for it? ie. it gets passed still gets passed on. [14:27] dednick: MouseArea has a property for that afair [14:27] dednick: perhaps http://qt-project.org/doc/qt-5.0/qtquick/qml-qtquick2-mousearea.html#propagateComposedEvents-prop [14:27] tsdgeos: right, bfiller is the right person, but I guess I can take a look at this as well [14:28] i've tried, but seems to get messed when there are mouse areas at a different/deeper parent level... [14:29] Saviq: added the "point of no return" as I think it makes sense. fixes the weird flipping issue you had [14:31] dednick: standup? [14:52] tsdgeos, still no feedback from design regarding half-filled organic grids/ [14:52] ? [14:52] dandrader: nope :_/ [14:52] Saviq: can you get katie to answer the email we sent? [14:53] mzanetti: you're preview man? you doing https://code.launchpad.net/~diegosarmentero/unity8/purchase-service/+merge/201807 ? [14:53] tsdgeos, or I would exercise common sense and do what gallery does [14:53] tsdgeos: ack [14:54] dandrader: ok, let me finish what i'm doing and i'll have a look at what the galllery does [14:56] good morning everybody. [14:57] I now need help with the cmake task to install the fake scopes and previews. Can somebody help me? [15:03] mterry: hey, would you agree that an AP test makes sense here? https://code.launchpad.net/~mterry/unity8/hide-greeter-on-focus-request/+merge/201817 [15:04] mzanetti, uh, sure. The bug I fixed I don't think needs it (because the bug is in how we respond to events from upstart), but that code in general could use it (for testing integration boundary with upstart) [15:04] I'll look at it [15:04] yeah [15:05] mterry: the functionality and the unit test look good to me. I'll approve once we have an AP test. Just drop it into the other app_lifecycle ap tests === alan_g is now known as alan_g|afk === alan_g|afk is now known as alan_g [15:49] mzanetti, do the autopilot tests not have an example of locking the screen? ... I'm not sure how to induce a fake power button press [15:49] mterry: hmm... good question [15:51] mterry: how does it actually work? [15:51] mterry: I remember we were using Qt.KeyPower at some point, but that seems gone [15:52] mzanetti, how does locking actually work? we get a signal from powerd that the display setting is changing [15:52] mterry: well, I guess in theory we'd need to change the display setting then [15:53] mzanetti, seems like something we should mock [15:53] mterry: not really in autopilot tests [15:53] ideally autopilot would not use any mocks at all [15:54] mzanetti, I get that they are integration tests, but I feel like actually changing hardware settings might be too much? [15:54] mzanetti, cameras watching the devices and all might not like the screen going out. Though I suppose it would only be for a moment [15:54] and that's what we're striving for. the only reason we're using some mocks now is because we don't have an applicationmanager for the desktop yet [15:56] mterry: the more I think about it, the more I think we should do that... its the proper thing to do [15:56] mzanetti, actually change display brightness? [15:56] mterry: yes. [15:56] mterry: anyways, that's not really related to your tests here [15:57] mterry: so why don't you just not unlock it after starting up unity and unlock it directly by starting some other app? [15:57] mzanetti, well. That's one test. That test always worked though. [15:57] mzanetti, the thing that broke was focusing an already open app while the greeter was up [15:57] ah, I see [15:57] mzanetti, but... [15:58] hmm... so yes. my opinion is that we shouldn't mock in autopilot tests [15:58] powerd-cli should be able to futz with screen for us [15:58] yeah, that sounds like the proper thing to do [15:59] not that anyone was asking, but /me agrees :) [15:59] mzanetti: can you help me with the installation of the fake scopes ? [15:59] elopio: hi [15:59] elopio: what up? :) [16:00] what exactly do you want to install? [16:00] mzanetti: I need the scopes to be copied to builddir/install/lib/{arch}/unity8/qml/Unity/ [16:00] but I don't know anything about cmake. [16:01] elopio: afaik they should already be packaged up into unity8-fake-env [16:01] need to reboot, something broke in here [16:01] mzanetti: some of them, but not the ones I need. [16:02] builddir/tests/mocks/Unity has more files than [16:02] builddir/install/lib/{arch}/unity8/qml/mocks/Unity/ [16:04] so I suppose I'm missing the install function in tests/mocks/Unity/CMakeLists.txt [16:04] but I don't know what to write in it [16:04] elopio: oh boy... that's going to be tricky now :/ [16:04] elopio: so the thing is this: [16:04] mzanetti, hm. powerd-cli doesn't let you turn off display, only on. Do you happen to know a better way? [16:05] elopio: ideally we wouldn't use any mocks at all but we need to for now. that's why we install some of them [16:05] mzanetti, sorry if you replied, I had connection issue [16:05] elopio: the problem now is, if you just install your mock along with the others it will always be used for all tests [16:05] elopio: so we'd need a way in autopilot to tell which mocks to use and which not [16:06] mterry: no, I didn't reply yet. still thinking about something [16:06] mzanetti: in this case, I think that's a must. Otherwise, all the tests will access the u1 production servers. [16:06] mzanetti: we will cover the tests for the real scopes and servers in a different branch. [16:06] Saviq: help me :) [16:07] mzanetti, no rush, sorry [16:10] mzanetti: oh, but btw, what you said is doable. We can tell the tests to use just some of the fakes, but for that we would need them to be in a separate folder. [16:10] I suppose cmake can do that. [16:12] elopio: yeah, cmake can do that [16:13] mterry: really weird... powerd-cli help says and the descroption of that says "Off, On or Don't Care" [16:13] mterry: does that seem like a bug to you too? [16:13] mzanetti, maybe. It has always been like that. There is a documentation bug for certain. Maybe also a missing functionality bug [16:14] mterry: hmm... interesting... if I manually suspend the screen and then call powerd-cli display on it turns on [16:14] mterry: but as soon as I kill powerd-cli again, it goes off again [16:15] mterry: so off wouldn't work anyways as unity still holds the "on" state afaiu [16:15] mzanetti, a powerd-cli command is valid just as long as it's running [16:15] mzanetti, yeah. powerd-cli will install an 'on-hold' that goes away when it does [16:15] mzanetti, so you have to keep it running in a separate tab/ssh session [16:15] mzanetti, fair... [16:15] so I guess that architecture doesn't allow an off if someone else holds a on-lock [16:16] mzanetti, and that's why it doesn't accept off as an argument, sure [16:16] still a doc bug there I guess [16:17] And yet, the user can press a button and the android drivers deliver a signal to powerd that a screen turn off should happen [16:17] mzanetti, we could emulate/create that driver event? [16:17] mterry: yeah... either that, or there is some admin mode to powerd where you can kill other locks [16:18] mterry: I guess you need to ask someone from phonedations [16:19] elopio: check out tests/mocks/Unity/Application/CMakeListsts.txt [16:19] elopio: there are to install() statements at the end [16:20] elopio: add the same to tests/mocks/Unity/CMakeLists.txt for the target FakeUnityQml and put some other directory in there [16:20] mzanetti: let me give it a try. [16:22] elopio: along with that, please group the add_subdirectory() from the top and the bottom of the file [16:23] np [16:27] mzanetti: it doesn't work. [16:27] the libFakeUnityQml.so gets copied to the install lib, but I suppose we are still missing some files. [16:28] elopio: yeah, the qmldir file [16:28] ah, lets see [16:32]