[02:37] unity-scope-tool used to properly show cards with overlay and title. it *just* broke after I dist-upgraded. [02:37] Saviq: FYI ↑ [02:38] my art aspect ratio is 2.5 [02:38] card-size is large [02:40] as if the title has been moved downwards [07:02] karni, are you by any chance using "fill-mode": "fit"? [08:51] greyback: howdy [08:51] dednick: morning [08:52] greyback: so just looking quickly at the unitymir::session before goingt through review comments. looks like the only real use is in the surfacemanager now. [08:53] where i changed the surface to have a ref to session instead of application [08:53] what's the reason for that? Application was just a wrapper around mir::Session [08:57] greyback: hm. it was because sessions for prompt sessions may not have an application set when the surface is created. [08:57] greyback: if they have not yet been added to the prompt session. [09:00] greyback: so it's basically doing the app lookup lazily. [09:05] greyback: and the surface managment code is not threadsafe. If we don't keep reference to all the sessions, then they may have been deleted once the queued slot gets called. [09:06] which is possibly why i did the list in the first place. [09:06] Saviq, is it desired that the shadows appear before the actual icons in carousel? [09:06] Saviq, or should i open a bug about that? [09:06] greyback: queued slot = sessionCreatedSurface [09:07] dednick: it should be thread safe, if it's not then it's something that needs fixing [09:07] dednick: keeping copies of all Mir's internal data is not pleasant [09:07] greyback: it's not threadsafe. [09:07] void SessionListener::surface_created(ms::Session& session, std::shared_ptr const& surface) [09:07] mhr3, dunno [09:08] session is passed by reference [09:08] Saviq, opening bug then [09:08] mhr3, against ubuntu-ux too, please [09:08] dednick, re urls, Qt.openUrlExternally is official API that we support, it needs to work [09:09] dednick, so no point in duping that effort [09:09] Saviq: yeah. i just used that. [09:10] dednick: you have a point there. I'd forgotton about that. That needs fixing somehow [09:10] greyback: yeah, unity-mir needs to be threadsafe! [09:10] :) [09:11] Saviq, mikenagle, https://bugs.launchpad.net/ubuntu-ux/+bug/1333604 [09:11] Ubuntu bug 1333604 in Unity 8 "Shadows in carousel appear before the icons" [Undecided,New] [09:14] greyback: but yeah, not sure if it's even practical/possible to assume direct connections over all slots coming across threads. maybe we need all listeners to pass ownership, although don't know if that's going to fly with mir peeps. [09:16] dednick: no we can't, hence using queued connections as much as poss. Since the events do pass ownership of the object created/destroyed, I had hoped it was enough. [09:16] greyback: mir does pass by ownership for the create/destroy of objects, but all operations on those objects seems to be pass by reference. So I think that if you want to use cross thread, you need to store. [09:17] dednick: yeah. I might get them to change that to a weak reference [09:17] greyback: ok. so for now, can I leave the session storage? [09:17] dednick: something to be done later [09:18] greyback: although just using shared_ptr<>, not the wrapper [09:19] dednick: yeah. Will need to be careful we always release those shared pointers [09:19] greyback: yup. [09:59] mzanetti, ping? [10:05] mzanetti, and unping, rm-ing builddir sometimes helps [10:13] mhr3: sorry, was in a meeting. [10:13] mhr3: the QGSettings thing? :) [10:13] oh, i wasn't the only one? :) [10:13] nope :) [10:14] mzanetti, for your amusement http://paste.ubuntu.com/7694378/ [10:15] huh? [10:15] haven't seen that before [10:15] Saviq: yes, fill-mode: fit [10:15] looked a bit like finnish at a first glance :D [10:15] karni, don't use that [10:15] karni, like *ever* [10:15] karni, it's only meant for apps icons atm [10:16] Saviq: what's the default behavior? [10:16] karni, crop, but that's also the designed behaviour [10:16] ok [10:17] karni, it wasn't really thought out TBH, we just slapped the fill-mode: fit to be able to do app icons [10:17] ack [10:23] Saviq: same effect. should I explicitly use fill-mode: crop? http://paste.ubuntu.com/7694461/ [10:23] karni, no [10:23] karni, trying [10:23] thanks [10:24] karni, weird, this works fine here... [10:24] Saviq: unity-scope-tool on utopic? [10:25] * karni dist-upgrades once more [10:26] karni, apt-cache policy unity-scope-tool? [10:26] karni, also, any chance I could have a look at the scope? [10:26] Installed: 7.89+14.10.20140623.1-0ubuntu1 Candidate: 7.89+14.10.20140623.1-0ubuntu1 [10:26] let me arrange that [10:28] karni, ok, reproduced [10:28] Saviq: just sent you the source. ignore if you want then :) [10:47] karni, fix: http://paste.ubuntu.com/7694526/ [10:48] Saviq: thank you! [10:48] karni, sorry about taht [10:48] Saviq: np [10:51] Saviq: this doesn't belong to unity8, does it? where should I patch? [10:52] found it ./usr/lib/x86_64-linux-gnu/unity8/qml/Dash/CardCreator.js [10:52] karni, yes, unity8 [10:54] Saviq: sweet, thanks [10:56] Saviq: last question - when will categories by default not-collapse if they're a sole category in the result set? do you guys have this bug on your radar? [10:56] Saviq: no rush, I just want to know what to tell people (if other than "bug is filed" :) ) [10:57] karni, they won't default, but you'll be able to set collapsed-rows: 0 [10:57] karni, and yes, it's on our radar [10:57] works for me! :) thanks [11:02] * greyback has dentist visit, bbiab === MacSlow is now known as MacSlow|lunch [11:10] Cimi, rather small one for you https://code.launchpad.net/~saviq/unity8/fix-header-height/+merge/224269 [11:31] seb128, silo 8 looks good, although I was unable to test on my laptop, unity8 isn't started for some reason (neither distro or silo packages) [11:31] seb128, could you (did you) test the functionality of the session indicator? [11:32] Saviq, ok, sure, let me upgrade/install the ppa on my test config [11:35] Saviq, how do I test it? [11:36] Cimi, with the test tool, go to amazon scope for example, and override with this: http://paste.ubuntu.com/7694461/ [11:36] Cimi, only s/image/art/ [11:37] upgrading utopic [11:37] see you or not :) === pete-woods is now known as pete-woods-lunch [11:53] Saviq, /home/cimi/Development/unity8/card_touchdown/plugins/Unity/Launcher/backend/launcherbackend.cpp:27:22: fatal error: QGSettings: No such file or directory [11:53] #include [11:53] Saviq, maybe something missing in the deps [11:54] Cimi, ./build.sh -c [11:57] Saviq, what service to run prior unity scope tool? [11:57] Cimi, smart-scopes-proxy === MacSlow|lunch is now known as MacSlow [12:04] Saviq, I have quite different result on amazon scope [12:04] for category === alan_g is now known as alan_g|lunch === _salem is now known as salem_ [12:08] * Cimi lunch [12:11] still here :D [12:13] Cimi, not sure what you mean? yes, you should override it, disregard what amazon sends itself [12:13] ah ok [12:13] Saviq, results? [12:14] Cimi, what about results? [12:14] Saviq, override category for results [12:14] Cimi, just override the category template with the pastebin [12:14] or toys etc [12:14] Cimi, whatever [12:14] Cimi, doesn't matter, as long as it has a title and art [12:14] it's empty for me [12:15] Cimi, any other scope will do [12:15] ok [12:15] Cimi, that has an image and title [12:15] weird [12:16] Saviq, hey, getting an error with unity8 desktop (after upgrading). Seems to be when the switch was made to use ubuntu-app-launch [12:16] “WARNING: QApplication was not created in the main() thread” [12:17] if you don't know anything about that error message no worries, ill just poke around [12:17] bschaefer, that warning was always there [12:17] o really? [12:17] soo... i wonder why unity8 isn't working...with no error message [12:17] just black screen [12:17] bschaefer, it's because Mir is the one starting the shell, and not in the main stage [12:17] Saviq, that makes sense, it was just the only log message :( [12:17] bschaefer, doesn't run here either, but it actually never starts (as in I have no unity8-mir log file from upstart) [12:18] bschaefer, hmm [12:18] bschaefer, ps aux | grep unity8 ? [12:18] Saviq, you're right, no process [12:18] Saviq, http://i.imgur.com/FYzSbDt.png [12:18] Saviq, weird the overlay on digitally imported radio... [12:19] Cimi, transparent icons → broken overlay [12:19] Saviq, seems like different compositing [12:19] Cimi, the overlay is only drawn where the icon isn't transparent [12:19] we should have OVER [12:19] Cimi, really, don't try to explain this [12:20] Cimi, I know what the reason is [12:20] Cimi, and it's because there's no new shape [12:20] Saviq, thanks! I should be able to figure out where it should be getting started [12:20] bschaefer, there's a unity8-mir job that should start it [12:20] bschaefer, .crash in /var/crash in any chance [12:20] ? [12:20] Saviq, no new shape? [12:20] Saviq, the overlay fills just the icon? [12:21] Cimi, just look at the code [12:21] if you really need to know [12:21] but basically, [12:21] Saviq, our code or ubuntu shape? wondering if we can fix it [12:21] haven't looked yet [12:21] Cimi, no we can't [12:21] ok [12:21] if (alpha != 0) { draw overlay } [12:22] Saviq, so it's a shader basically [12:22] but that obviously breaks if the image has transparent pixels [12:22] Cimi, yes of course [12:22] Saviq, theres only a crash about the keyboard indicator service [12:22] bschaefer, so it indeed looks like it never starts [12:22] strange indeed [12:23] Saviq, thanks! Forgot to check for a crash file :) [12:23] soo that should be u-s-c that should start it right? [12:23] bschaefer, upstart actually [12:23] Saviq, cool, ill dig through there. Thanks! [12:23] bschaefer, unity8-mir.conf has a "start on unity8-mir session" or so [12:23] bschaefer, bregma knows more obviously [12:25] bschaefer, there's always an error message about the keyboard service, are you sure it's a crash? [12:25] bregma, its in /var/crash after starting unity8 [12:26] bschaefer, bregma, Saviq: talking about unity8 in utopic? [12:26] seb128, yes [12:26] seb128, yeah [12:26] it segfaults in libqubuntumirserver.so/qtsensor [12:26] bregma, could the keyboard indicator keyboard service cause unity8 to not start? [12:26] I guess an issue with the recent platform-api/qtubuntu-sensors changes [12:26] seb128, no fun :( [12:26] rsalveti, ^ [12:26] I was just looking at that [12:26] at a sprint, and attempting things with it atm :) [12:27] Saviq, well looks fine [12:27] Cimi, compare to trunk, where it doesn't ;) [12:27] Cimi, the header was 2gu smaller, so labels were cut off at the bottom [12:29] still fine to me [12:29] weird [12:29] but yes code looks fine [12:29] acxtually didn't test with trunk [12:29] but with my card emblem branch [12:30] maybe I fixed it :D [12:30] Cimi, no, maybe you don't have trunk ;) [12:30] Cimi, it got broken by Albert's branch that was merged yesterday [12:30] bschaefer, bregma, Saviq: back to working after installing https://launchpad.net/~ci-train-ppa-service/+archive/landing-005/+build/6112664/+files/qtubuntu-sensors_0.6%2B14.10.20140619-0ubuntu1_amd64.deb [12:30] ok [12:30] rsalveti, ^ [12:30] seb128, awesome, thanks! [12:30] yw [12:31] I moved to -touch, ricmm is not on -unity [12:31] * bschaefer isn't in -touch [12:32] bschaefer, well, just downgrade that binary and let us fix the issue in distro [12:32] seb128, yeah, i've downgraded it on my machine, attempting to downgrade on other machines has been somewhat of an issue [12:32] but ill try again [12:32] we really need to start testing that those landing don't make the desktop session stop working [12:33] seb128, +1 [12:33] seb128, that was one of the requirements before 14.04, I guess they "forgot" about it again [12:34] bregma, requirement for who? anyone doing landings? because I think the unity8 team is good about it, but the platform-api stuff keeps screwing us recently [12:34] Saviq, on that note, confirmed that the silo for logout works [12:34] it was one of the items in the QA list -- a manual desktop sanity check [12:34] seb128, \o/, ACKing it [12:35] Saviq, ;-) [12:35] we had a documented checklist and a status spreadsheet [12:35] and people assigned and given devices just for that purpose [12:35] wwwhhhhyyyy dafuck am I being logged out all the time :| [12:35] Saviq, yeah, happens for me every time I close firefox :/ [12:36] hate google some days [12:36] seb128, must be fx 30 :| [12:36] bregma, right, but I guess nobody is going through the whole QA list for landing individual bits [12:36] seb128, I also noticed that it suddenly offers to remember passwords for my banking account, for example - not great [12:36] Saviq, that has been happening for a while for me, other mentioned it as well, seems rather on the google side [12:37] seb128, hmm only happened recently to me [12:37] Saviq, they might opt-in more people for "increased security" over time, dunno [12:56] also touch screens have stoped working in unity8 desktop [12:56] not a huge deal, just a regression [12:57] bschaefer, did they? [12:58] when? [12:58] seb128, not sure, i don't have a touch screen, but i've one in front of me right now and its not working [12:58] bregma, possibly could know [12:59] bschaefer, on my inspiron 11 I had to create a file with the input device id to get Mir to work with the touch screen [12:59] bregma knows the magic [12:59] bschaefer, you may beed to add a custom /usr/idc file to get the Android code to recognize the touchscreen as a touchscreen [12:59] seb128, that sounds resonable [12:59] reasonable [12:59] bregma, can you share yours? I don't remember the name/content (I wipe my install since budapest to test the iso install mode) [13:00] * bregma needs to put something on the wiki [13:00] bregma, that would be nice [13:00] I was going to do that [13:00] but I lost the required knowledge :p [13:01] step 1: do an 'lsusb' and look for the touchscreen device (you may need to guess here) [13:01] I get something like 'Bus 002 Device 007: ID 04f3:016f Elan Microelectronics Corp.' [13:02] bregma, yeah that one [13:02] Elan Microeslel [13:02] step 2: create a file '/usr/idc/Vendor_04f3_Product_016f.idc' (where the hex codes match the ones in the lsusb line) [13:02] give it these two lines as content: [13:03] device.internal = 1 [13:03] touch.deviceType = touchScreen [13:03] step 3: ??? [13:03] step 4: PROFIT!!!! [13:03] er, um, restart the Mir server [13:03] bregma, cool thanks! [13:04] ill test that out === alan_g|lunch is now known as alan_g === salem_ is now known as _salem [13:17] bregma, yay it works === _salem is now known as salem_ [13:29] mterry, the spinning ubuntu logo is done by unity-system-compositor, right? [13:30] dandrader, yeah, by a tiny helper in the USC source tree [13:30] dandrader, separate process though [13:30] ok [13:31] dednick: hey, i was trying to build your trusted demo...first tried on the device, ran out of mem... [13:32] kgunn: uuuh. that's weird. give me a minute [13:32] dednick: its likely me...take your time [13:39] kgunn: just built on newly flashed device. didn't have an issue. did it seem to be taking a ridiculous amount of time? [13:40] kgunn: although i did have to get required packages manually to build. so not sure if you had something missing... [13:40] dednick: so i think i might have been loading too many deps [13:41] kgunn: i just used build-dep unity-mir to source package deps. [13:41] dednick: for the demo app source ? [13:42] kgunn: lp:~nick-dedekind/+junk/trusted_sessions_app [13:42] dednick: yes thats it...ok...will do [13:43] $ apt-get build-dep unity-mir [13:43] $ mkdir builddir && cd builddir [13:43] $ cmake .. -DCMAKE_INSTALL_PREFIX=/usr [13:43] $ make -j4 [13:43] kgunn: ^ [13:48] dednick: ta [13:48] dednick: ...i was trying to load like the whole sdk :) [13:49] kgunn: ah :) [14:04] Saviq, branch approved [14:05] Saviq, how about some of mines? :) [14:06] Saviq, hm, looks like temp dash pages blow up when you lock+unlock the screen while in one [14:07] Cimi, I was almost there with the emblem, but then bug #1284233 happened, I'd like that to land first so we don't have to do it [14:07] bug 1284233 in Ubuntu UI Toolkit "Can't use uris for Icon source" [Undecided,In progress] https://launchpad.net/bugs/1284233 [14:07] Cimi, will also want to fix bug #1284235 [14:07] bug 1284235 in Ubuntu UI Toolkit "Icon has to be sized explicitly" [Undecided,Incomplete] https://launchpad.net/bugs/1284235 [14:08] Cimi, but it'll probably end up with "add TODO" for once UITK lands [14:09] Cimi, did you see email I forwarded? looks like we need/want a bigger background than the card itself [14:09] for the touch highlight [14:09] Saviq, I will see === pete-woods-lunch is now known as pete-woods [14:09] Cimi, but really, you just need to go to the office [14:09] Saviq, I can work on the latter bug as well [14:09] and show them what works, what doesn't [14:09] Saviq, ok [14:09] Cimi, since with what Josh just sent, cards that *do* have background will look crap [14:10] mhr3, explode how? [14:10] mhr3, everything seems to work here (except that with current design it *should* reset back to dash) [14:10] Saviq, having an effect outside of the region looks weird [14:11] Saviq, because the target click area will be smaller [14:11] Cimi, well, why would it be? [14:11] Saviq, now we are almost anchoring to root [14:11] so we might want to create something with negative margins [14:11] Cimi, yeah, we'll have to [14:12] Cimi, or well, maybe not, we have 2gu spacing between medium cards [14:12] Saviq, outside of the clickable mouse area [14:12] Cimi, we might not need to [14:12] yeah 1gu [14:12] Cimi, 2gu [14:13] which is too much, actually... [14:13] it should be 1gu [14:13] Cimi, which means the highlight would clash, but maybe that's fine [14:13] Saviq, it cannot clash [14:13] Saviq, since you touch only once at a time :) [14:14] Cimi, right, mousearea can't clash [14:16] Cimi, btw, apparently the sizes changed, and indeed there's 2gu spacing [14:16] Saviq, let me try adding negative margin [14:16] Cimi, look at the same mail [14:16] Cimi, I think it'll still look weird for cards with background... [14:18] Saviq, ah ok, it's related to preview, if you search for something in a scope, then preview it, lock+unlock, it disappears [14:18] Saviq, related to the search string being reset i guess [14:19] mhr3, yeah, that's the designed behaviour - dash was *meant* to reset when you focus on something else [14:19] mhr3, although I fully agree that's wrong [14:19] mhr3, and we'll yank it out with going to dash-as-app for sure [14:19] Saviq, http://paste.ubuntu.com/7695330/ [14:19] * Cimi tries [14:19] Saviq, i'd argue that lock+unlock isn't necesarrily focus change [14:19] mhr3, you and your arguments [14:20] :) [14:21] Saviq, weird [14:25] Saviq: hi. For LP:1330957 I can have the margins http://goo.gl/ZMUrfv [14:25] Saviq: Is there some other cases that I missed? [14:25] Saviq, I think our approach is best [14:26] Saviq, woule be ideal to have bigger target without background [14:26] but will not be consistent [14:27] paulliu, hmm interesting, didn't work for me [14:27] paulliu, maybe I didn't build [14:28] Saviq: yes. have to run build.sh everytime when changing the CardCreator.js [14:28] paulliu, correct, I take that comment back [14:31] paulliu, on the test fixes, testCard passes fine in trunk, and on your branch without the changes to tst_Card.qml, (why) are those needed [14:35] pete-woods, how do I test camera app and infographics? [14:36] Cimi: well I hope you can now use the debian version of the camera app [14:36] i.e. upgrade everything from silo 009 [14:37] pete-woods, but camera app is installed already [14:37] pete-woods, how do I overwrite it? [14:37] Saviq: I don't know. Trunk testCard always failed on my machine. I have to have those changes. [14:38] Saviq: but I think those changes it quite reasonable. Isn't it? [14:38] paulliu, Saviq to me is not... if we have no art why would we quit the test? [14:38] we should report the test is broken [14:38] Cimi, paulliu, yeah, I don't like the "return" part there [14:38] Cimi: I'm not really sure, just try installing the debs too? [14:38] paulliu, I wonder if it's your screen size that's the problem [14:39] pete-woods, can you test it? :) [14:39] paulliu, try `make -C builddir xvfbtestCard` [14:39] paulliu, it will run in a virtual X [14:39] Saviq: ok [14:39] pete-woods, if I use your commands to generate a fake one, it appears on unity [14:39] Cimi: is there a working version of unity8 in the silo now? [14:40] pete-woods, yes [14:42] Saviq: Cimi, I'll remove those lines after I tried the xvfbtestCard [14:42] on trunk [14:42] Cimi: you can hack in the click support by running touch touch ~/.local/share/libusermetrics/sources/com.ubuntu.camera_camera_2.9.1.json [14:42] paulliu, thanks [14:53] mterry, hey, could you please review https://code.launchpad.net/~albaguirre/unity8/use-new-display-power-state-interface/+merge/219552 ? [14:53] mterry, it's in silo 20 for testing pleasure [14:54] Saviq, OK === dandrader is now known as dandrader|afk [15:11] mzanetti, yeah, until it's merged, it's INPROGRESS [15:11] that might be the wrong approach, though... [15:11] sure yeah. that's why I didn't change it [15:12] but still didn't want someone to wait for it [15:21] dednick: ok, lots of interrupts...just tried, but i'm hitting same spot where i started to install sdk :P [15:21] dednick: https://pastebin.canonical.com/112272/ [15:22] dednick: maybe a better way to ask, how are you building? ...on device i assume ? (not cross) [15:22] kgunn: on device [15:22] kgunn: which device? [15:23] kgunn: mine is N4 [15:23] kgunn: did you apt-get build-dep unity-mir ? [15:23] kgunn: and 'apt-get update' first [15:26] dednick: not sure i did apt-get update first...lemme try that [15:27] dednick: i think that was it [15:28] too many interrupts [15:28] actually i updated...but forgot to get the build-deps === dandrader|afk is now known as dandrader [15:55] kgunn: looks like we don't have an armhf version of unity-system-compositor based on trust sessions. [15:55] kgunn: in staging i mean [15:56] dednick: ah...i need to tweak a recipe [15:56] kgunn: looks like the gcc issues we were having. https://launchpad.net/~mir-team/+archive/staging/+build/6110724 [15:56] dednick: just a sec [15:57] mzanetti, I think John didn't get the bug https://bugs.launchpad.net/unity8/+bug/1316512 [15:57] Ubuntu bug 1316512 in Unity 8 "Greeter can be flicked away in the wrong direction" [Undecided,Won't fix] [15:58] as in he probably didn't understand the problem [16:14] dednick: ok...first alf is going to have to do all the abi bumping to get staging back in shape [16:14] incl creating a u-s-c branch [16:16] kgunn: ok. thanks. [16:17] dednick: are you broken...do you need this asap ? [16:17] kgunn: i can work on something else for now. [16:17] kgunn: and can always build it locally. [16:19] ack [16:28] alberto_, I tested silo 020, and the greeter didn't appear after locking screen. Is that known? [16:31] mterry: no [16:31] mterry: it's because there's a new unity8 package [16:31] mterry: which apt-get selects over the silo [16:31] alberto_, oh crud [16:31] mterry: so I need to rebuild in silo [16:33] mterry: startup wizard notifies me that my N7 has no SIM card installed. There a bug about that? [16:34] greyback, yes... let me see [16:34] greyback, bug 1332583 [16:34] bug 1332583 in ubuntu-system-settings (Ubuntu) "Wizard suggests I can "insert a SIM card later" on a SIM-less tablet" [Medium,Triaged] https://launchpad.net/bugs/1332583 [16:34] mterry: also in the wizard, I've 4 blobs indicating pages. First page was Language. But next "page" was that SIM card message. I hit next, and now page 2 is "wifi" [16:34] greyback, that's actually by design [16:35]