=== maclin1 is now known as maclin [08:41] Saviq, hey, how are unity8 landings going? i've 1 silo ready in the pipeline and 1 that will be ready soon (filters) [09:23] pstolowski: it's all complicated [09:24] new sdk needs to land first afaik [09:24] and that's always complex because it's a big beast [09:31] tsdgeos, hmm silo 50? it failed in qa afaict [09:32] pstolowski: i think so yes, theywe're working on make it go thought qa again [09:32] i may be wrong in the current status though [09:32] it's all second or third hand info [09:33] tsdgeos, Saviq can we try to land silo 76 first then? it shouldn't cause any hiccups [09:33] that's for Saviq to decide, i abstain :D === JMulholland_ is now known as JMulholland [10:23] pstolowski, hey, getting some info from Zoltan [10:24] pstolowski, unity8 not built in there? [10:25] Saviq, may need rebuilding again [10:27] Saviq, rebuilding unity8 there [10:29] pstolowski, so, when you're ready, press Lander: Approved, let's see where we are with SDK at that point? [10:30] Saviq, okay [12:46] Saviq, I'm still puzzled. can I run something on britney manually? [12:46] Saviq, the inputMethodSurface is even created in the SurfaceManager's ctor... so it can't be a timing issue. it's like it would use the wrong mock or something. [12:46] mzanetti, not really, let me find out what I can do (but adding debugs is your best bet) [12:47] kk, will add some [12:48] it's all in cloud these days, so should be possible to spawn a similar one... [12:49] ok... I might have found the issue [12:51] yep, confident I found it [12:51] if only I could reproduce it locally [12:52] but it has to do with garbage collection [12:52] mzanetti, actually pitti can give us ssh access to the node where it fails [12:53] ah, that's good to know... let's have one normal run before (preparing my changes for pushing), if that doesn't help, I'd need that [12:56] Saviq, pushed, please give it another try [12:57] mzanetti, going [13:05] mzanetti, pitti gave me an idea for a local reproducer (qemu), trying that too === alan_g is now known as alan_g|lunch [13:53] ltinkl, if you just could update the comment from "simulate the model reset" to "simulate a changing icon name" in order to have it matching with what the code does [13:53] ltinkl, apart from that, it looks good to me, will test it [13:53] mzanetti, kk [13:53] ltinkl, is this really needed? + cache: false [13:54] mzanetti, yup, didn't work without [13:54] mzanetti, it would still show the old icon [13:54] ok... might want to add a comment there too on why we disable caching then [13:55] mzanetti, I even looked at the implementation, it's too smart - if the source path doesn't change, it won't go on and try to reload the image (QQuickImage) [13:55] mzanetti, hmmm but... we do change the path now right [13:56] yeah... that's what I mean [13:56] mzanetti, to empty and then back to the real one [13:56] ltinkl, please check if this is still really needed with the new change [13:56] mzanetti, aha! will fix then, this was valid for the model reset case [13:56] kk [13:59] mzanetti, MP updated [13:59] cool. I'll give it a try [14:00] * ltinkl bbl food [14:10] Saviq, any news from britney? [14:10] mzanetti, oh it will take a while, they only just built [14:11] mzanetti, trying in qemu locally now [14:12] mzanetti, http://packaging.ubuntu.com/html/auto-pkg-test.html for reference, a --qemu-options='-cpu SandyBridge' will take you as close to britney nodes as possible [14:13] Saviq, can we run QQuickView in qemu now? [14:13] mzanetti, I think you're mixing up qemu emulated (armhf) [14:14] this is just virt [14:14] ah, right... true [14:19] pstolowski, I've put silo 76 to Lander: Approved so we get britney feedback [14:20] Saviq, ah, I'be jest tested it and found prolly unrelated issue - reflashing without the silo to check [14:20] * just [14:20] kk === alan_g|lunch is now known as alan_g === dandrader is now known as dandrader|afk [14:45] mzanetti, http://pastebin.ubuntu.com/15259322/ [14:45] 'setState' of object function() { [code] } [14:45] dafuw [14:45] dafuq [14:47] how is that even possible.... [14:51] mzanetti, that happened for me on qemu [14:51] qemu is a liar! [14:51] it is not a function [14:52] oh well, I guess I gotta set up stuff here then [14:52] mzanetti, wait, that was with previous packages I think [14:52] * Saviq tries again [14:52] hmm [14:52] right... the previous one would have a dangling pointer [14:52] so you might end up on function() { [code] } indeed === shookees is now known as Guest64232 [15:25] Saviq, so, we really want this? https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1540502 [15:25] Launchpad bug 1540502 in Canonical System Image "Alt+Tab doesn't work when in staged mode" [Critical,Triaged] [15:25] still seems quite useless to me tbh... [15:26] mzanetti, actually staged mode is quite nice with kbd (on a tablet) [15:26] I mean, sure the target is to unify the code, at which point it would just work, but blocking an ota because a phone doesn't do alt+tab, not sure :D [15:26] hmm [15:26] makes defintely more sense on a tablet, yes [15:27] mzanetti, so yeah, we really need to unify the switchers ;P [15:27] Saviq, yes, about that, we need to talk at some point this week [15:27] mzanetti, ack [15:27] had a meeting with vesa today, it's really not simple === dandrader|afk is now known as dandrader [15:28] I'm not saying it is [15:28] not only from an engineering point of view [15:28] it's 3 different things from a design point too... and changing the phone expierince is very likely to make it worse [15:29] while using the phone stuff on the desktop isn't good either [15:29] at which point we'll have 3 different things from a design point of view... if they are different for design, it doesn't really make having it the same from a code point of view easier either [15:30] anyways... let's hangout the next days [16:03] tsdgeos, hola [16:03] hola hola [16:03] tsdgeos, can you please recap me on the two reviews for the filters? [16:04] tsdgeos, I forgot if we were blocked on some pawel work, and now is fine or not yet [16:04] or was design [16:05] cimi: should be good to review as far as i remember [16:06] tsdgeos, yeah I dont remember :D is it? :D [16:06] pstolowski: what is the silo filters are on? and can you point cimi to your book scope? [16:06] thanks guys [16:07] tsdgeos, cimi silo 54, it's currently rebuilding. ~stolowski/+junk/scope-filters2 is the scope (click/ dir has armhf package) [16:07] thanks pawel [16:17] mzanetti, got the same "is not a function" error with the newest code, got the testbed set up so I can change things [16:25] Saviq, can you print SurfaceManager and SurfaceManager.inputSurface [16:25] (in a OOBE hangout atm) === dandrader is now known as dandrader|afk [17:00] cimi: https://code.launchpad.net/~cimi/unity8/use-PageHeader/+merge/284942 === dandrader|afk is now known as dandrader === alan_g is now known as alan_g|EOD === dpm is now known as dpm-afk === dpm-afk is now known as dpm === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader [20:22] mterry, do you know (or knows who might know) what displays those permission dialogs (like "do you allow camera-app to access the camera")? [20:23] dandrader, no I don't know [20:23] nor do I know who would :) [20:23] mterry, ok [20:25] mterry, camera-service [20:25] dandrader, ↑ [20:26] or location-service, or *-service [20:26] Saviq, I wanna reset permissions so that I can trigger those permission dialogs again. any idea where they're stored? [20:26] dandrader, find ~ -name trust.db [20:29] Saviq, didn't find anything. but thanks for the clues [20:30] dandrader, /home/phablet/.local/share/PulseAudio/trust.db [20:30] /home/phablet/.local/share/CameraService/trust.db [20:30] /home/phablet/.local/share/UbuntuLocationService/trust.db [20:30] if those are not there, the prompts should pop up (assuming the app is confined() [20:30] ) [20:31] Saviq, I don't even have those files. wiped out my .config .local and .cache hoping to get the prompts back to no avail [20:31] dandrader, you need to restart [20:31] dandrader, at least those services [20:34] Saviq, another thing: scopes in untiy8-dash are completely empty. what do I have to kick to get it working? I recall I had it before and tsdgeos helped me out.... [20:35] dandrader, restart scope-registry would be one thing [20:35] dandrader, ah and you might've lost the favorites [20:37] Saviq, excellent. got the permission dialog back again. thanks [20:37] gsettings reset com.canonical.Unity.Dash favorite-scopes [20:37] dandrader, ↑ [20:38] Saviq, I did have Apps, Music and Videos already. they remain empty though... [20:41] dandrader, ok so you do have scopes but they're empty... I assume pulling down for refresh does not yield anything? at least it should yield some logging, so clear ~/.cache/upstart/* before === ljp is now known as lpotter