[08:15] mzanetti, these probably fall on your lap... https://code.launchpad.net/~saviq/unity8/refactor-carousel-activation/+merge/225743 https://code.launchpad.net/~saviq/unity8/refactor-scopeitem/+merge/225745 [08:15] mzanetti, and I had to re-do most of the scope customizations branch... [08:18] Saviq, i can do the first [08:18] Cimi, it requires new UITK though [08:18] there was a silo no? [08:19] Cimi, there is, yeah, 16 [08:19] Cimi, claim it in that case [08:20] Saviq, ok first thing, why it was necessary to move the signals to the delegates? [08:20] Cimi, because it's the right thing to do ;) [08:20] really? [08:20] this way the carousel is a widget with no signals [08:20] Cimi, everywhere else you have those on the delegate [08:21] Cimi, and why is that bad? [08:21] I see [08:21] Cimi, in any case, the real need was that we needed to pass result out to scope activation (like all the other delegates do) [08:21] result as in model.result [08:21] we deal with it as a container, like listview [08:21] yeah [08:22] that need (temporarily) lifted, but the change is valid still [08:22] and cleans things up [08:25] dednick, re: layouts, when we want something to be constant size, it should just have explicit width/height, the layout system picks that up and tries to adapt [08:25] Saviq: i c [08:25] dednick, preferred/maximum/minimum only really make sense for cells that you want to resize [08:27] Saviq: ok. i'll change the preferred to explicit [08:29] Saviq: hm. no, that doesnt work [08:29] i thought i had tried.... [08:29] dednick, orly? how about implicitWidth/Height? [08:30] Saviq: yeah, that does [08:30] * Saviq doesn't really get all of how layouts work, just reading docs [08:30] Saviq: indeed. i always had a bit of trouble when using in qt. [08:31] dednick, in that case let's be explicit and leave the preferred in [08:33] Saviq: CI on unity8 trafer branch fails because the MenuItemFactory is dependent on the TransferMenu type, which is in the USC transfer branch [08:33] dednick, oh of course [08:33] Saviq: i think thostr_ created a silo for it all [08:34] Saviq: https://launchpad.net/~ci-train-ppa-service/+archive/landing-004 [08:34] dednick, ktx [08:35] hm. unity8 failed to build there... dep on libunity-mir-dev [08:35] Saviq: Do you know what's the network cache size that's been recently implemented? [08:36] Saviq: although succeeded on armhf/amd64&i386 [08:40] karni, the default is 50MB [08:41] Saviq: is there a manual way to increase it on demo devices? [08:41] karni, http://qt-project.org/doc/qt-5/qnetworkdiskcache.html#setMaximumCacheSize [08:42] karni, you'd have to recompile [08:42] karni, we could add support for an env var or something, but are you sure that will help you? [08:42] karni, can you see that your .cache/unity8/network is around that size? [08:43] Saviq: ok thanks. we'll look into that and follow up, if there's need. [09:01] Saviq: http://paste.ubuntu.com/7759274/ [09:01] looks like it's locale dependent [09:01] dednick, LC_ALL=C [09:02] Saviq: i just built and ran. should be in the env vars for the test? [09:02] Saviq: unless my cmake didn't update [09:02] dednick, yeah it should, but isn't... [09:02] dednick, I need to rework/fix the QmlTest cmake module to make it work properly [09:02] Saviq: ok [09:03] Saviq, eh, i actually meant to increase the size, but forgot about that, think ~200 would work better [09:03] mhr3, mhm [09:03] mhr3: 50MB for 'global' unity8 cache does sound somewhat little [09:04] mhr3: I just didn't want to force you guys update it before I confirm we actually need more (but I think might, eventually) [09:05] mhr3, on that note, I can't see anything in http://qt-project.org/doc/qt-5/qnetworkdiskcache.html that says it will actually auto-expire [09:05] would probably work betterer with unity-api cache, since it does lru [09:06] Honestlyl, I was wondering how exactly QNetworkDiskCache works - the documentation is not great. If I set ::PreferCache, when does it read from network? Does it do HEAD requests first anyway to the resource? [09:07] karni, basically, never, it's nice enums that aren't really implemented :) [09:07] Saviq, where do you see that exactly? [09:07] mhr3: so once it caches the resource, it never expires it (unless it goes over the cache size) and we always read from cache?! [09:08] :( [09:08] mhr3, I said I "can't see" [09:08] Saviq, eh, sorry [09:09] karni, it's read from cache if the header said it doesn't need to be validated [09:09] mhr3, hmm ::expire is only there for reimpl purposes [09:09] mhr3: oh ok [09:09] mhr3, so it should probably be fine [09:10] Saviq, yes [09:10] but yes, it means that most of the time the cache will be full, and inserting new entries just deletes the oldest ones [09:12] FWIW if it's over max cache size, it'll purge until 10% is freed [09:12] starting with oldest, yes [09:53] Saviq: is this wanted? http://i.imgur.com/ccYYPlC.jpg [09:54] mzanetti, you mean the button? [09:54] yes [09:54] mzanetti, that's how it comes from the scope, so yes [09:54] ok [09:55] mzanetti, https://sites.google.com/a/canonical.com/unity8dash/scopes/app-scope-home/app-scope-landing-page [09:55] ok... surprises me a bit though [10:14] Saviq, i suppose i can turn the "See more | See less" widget into a simple "See more" version, right? categories have just that, and previews shouldn't be using it anyway [10:14] mhr3, yeah ok [10:15] Saviq, i'll need your help with some things that broke though :) [10:15] or maybe i can annoy dednick [10:20] Saviq, why did you change the test number of items? [10:21] 58 - tryCompare(carousel, "createdDelegates", 10); [10:21] 159 + tryCompare(carousel, "createdDelegates", 9); [10:21] Cimi, because the reinit of carousel caused the loaded items to be reduced [10:21] ok [10:21] number fo [10:21] of [10:21] mhr3, lemme know if you need me [10:23] Saviq, where is this reinit? [10:23] Cimi, between cleanup() and init() [10:23] ah in tests, ok [10:23] Cimi, yeah, it's only in tests [10:27] i'll try asking again... [10:27] sonne, hey, I saw your question [10:27] i'm having a strange problem with unity: a QT program, that works fine under osx / windows / gnome, doesn't seem to behave properly under unity, as in once you switch focus to another program you can't switch back to it [10:27] clicking on the program on the unity bar has no effect, and it doesn't show on alt-tab. this only happens under unity.. any ideas? [10:27] sonne, and yeah, it's a bug between unity and bamf, some Qt apps behave weird [10:27] ah cheers.. didn't read your reply if you did one :) [10:28] sonne, but I don't think there's been a real reproducible test case yet [10:28] i see.. so isn't there anything i could look into? [10:29] sonne, no, I can't find a reported bug, please file one [10:29] sonne: check how bamf handles that via dbus [10:29] Trevinho, you suggest the problem only relies on the underlying system, and not in the app itself? because things like amarok have no problems at all [10:30] Saviq, where should i file it? on unity? [10:30] sonne: it might be the app indeed, but to understand why unity doesn't work it would be nice to check what it's trying to do [10:31] Trevinho, unfortunately i have no ideas on how to inspect what you're talking about :( [10:32] sonne: give me 1 minute [10:32] sure [10:33] (on an offtopic side note, are you the guy i met at fosdem outside the station, trying to find public transportation stops?) [10:33] (if not, you have very similar names :P) [10:38] sonne: I was at the FOSDEM, but I think I was not that guy :) [10:39] right :) [10:39] sonne: btw if you call this: http://pastebin.ubuntu.com/7759576/ [10:39] sonne: and if you select your app [10:39] you get the dbus path of your app in BAMF... [10:39] at that point [10:39] hold on, firing up unity.. [10:40] Saviq: any idea? https://jenkins.qa.ubuntu.com/job/unity-phablet-qmluitests-utopic/353/console [10:41] dednick, it can't install qtdeclarative-ubuntu-telephony0.1 [10:42] Setting up ofono (1.12.bzr6868+14.10.20140625-0ubuntu1) ... [10:42] invoke-rc.d: unknown initscript, /etc/init.d/ofono not found. [10:42] dpkg: error processing package ofono (--configure): [10:42] subprocess installed post-installation script returned error exit status 100 [10:42] dednick, ↑ [10:42] Saviq: hum [10:43] Saviq, we can probably merge click and press and hold tests [10:43] Saviq, so we don't dup code [10:43] Cimi, mhm, lemme [10:44] Trevinho, ok [10:44] i have /org/ayatana/bamf/application/311805604 [10:45] sonne: so if you now use d-feet with that object path you can verify that bamfs gets all the infos for that correctly [10:45] d-feet? [10:45] sonne: or just use gdbus call --session --dest org.ayatana.bamf --object-path $YOUR_PATH --method [10:45] Cimi, done [10:45] sec, installing d-feet [10:46] Saviq, did u push? [10:46] Cimi, yes, refresh [10:46] Saviq, launchpad doesn't show [10:47] Cimi, LP has some latency after pushing, you need to wait (or pull yourself) [10:47] sonne: if you want to check the app has a desktop file correctly matched with it you can just use [10:47] gdbus call --session --dest org.ayatana.bamf --object-path /org/ayatana/bamf/application/311805604 --method org.ayatana.bamf.application.DesktopFile [10:48] Saviq: happy with https://code.launchpad.net/~nick-dedekind/ubuntu-settings-components/transfer-menu/+merge/224672 ? [10:49] Trevinho, oh my, it returns gnome-terminal... [10:49] i guess your earlier script screwed up :P [10:51] Trevinho, ok, now i have the correct .desktop file [10:51] and it's the right one [10:51] ...apparently it's working now(!) [10:52] heisenbug! [10:53] dednick, yup, requested a silo for the two [10:54] Saviq: should be @ https://launchpad.net/~ci-train-ppa-service/+archive/landing-004 [10:54] dednick, ah both already? [10:54] thostr_: ^ is that the transfer indicator silo? [10:54] dednick, I can see now, will tweak it [10:54] dednick: yes [10:56] thostr_, had to add the prerequisite branch, doing recon and build [10:57] Saviq: ok, thanks for heads up [10:59] dednick, the .pot file has "..." still? === MacSlow is now known as MacSlow|lunch [11:00] dednick, the .pot file has "..." still? [11:01] Saviq: eh. forgot to update [11:02] Saviq: ./qml/Launcher/LauncherDelegate.qml:203: warning: unterminated string literal ? [11:02] dednick, just a warning [11:02] dednick, it's complaining about QML's multi-line strings [11:02] dednick, which don't require \ [11:03] Saviq: ah. wrong line. [11:03] xgettext: Non-ASCII string at ./qml/Panel/Indicators/MenuItemFactory.qml:583 [11:03] dednick, ugh? [11:03] dednick, gimme a mo, let's see [11:04] dednick, one more thing - commented on the MP [11:04] doens't like the … [11:04] dednick, yeah, stoopid [11:06] Saviq, I don't think you pushed [11:06] Saviq, bzr doesn't have the revision [11:06] Cimi, pushed to the wrong branch ;| [11:06] Cimi, will fix in a sec [11:10] dednick, http://paste.ubuntu.com/7759677/ [11:10] dednick, just include in your MP [11:11] Cimi, fixed, sorry [11:11] np [11:15] Saviq: thanks. pushed [11:18] dednick, so last thing... should we bump changelog and Depends in unity8? [11:18] (changelog in u-s-c) [11:19] Hola [11:19] \o [11:26] Saviq: um. probably [11:26] dednick, please do [11:33] Saviq: done. [11:34] i think [11:34] Saviq, why in CardCarousel function clicked and not onClicked ? [11:35] Saviq, Carousel calls clicked signal [11:35] Cimi, calling foo.signal() equals calling foo.function() [11:35] yes I know [11:35] but we have already the signal [11:35] Cimi, I'll have a look in a mo [11:36] isn't more appropriate to use onSignal? [11:36] sure [11:36] rest is fine [11:36] dednick, ffr, it's enough to just do 0.3-0ubuntu1 [11:36] tests pass [11:36] dednick, train will do the rest [11:38] Saviq: ah. thanks [11:53] mzanetti, on https://code.launchpad.net/~mzanetti/unity8/launcher-update-home-button-design/+merge/224102 [11:53] thostr_, all transfer menu MPs ACKed from our side, kicked the last rebuild of unity8 and ubuntu-settings-components [11:53] mzanetti, was only the mousearea meant to be on top with z:1? [11:54] Cimi: yes [11:54] or you want to keep that z: 1 inside rectangle? [11:54] Cimi: ah... hmm, let me check [11:55] Cimi, re: CardCarousel, I'd need to introduce the signal in the delegate, felt unnecessary [11:55] Cimi, 'cause in CardCarousel we're not using BaseCardDelegate [11:57] pete-woods, welcome back! [11:57] pete-woods, we have conflicts in silo 9, could you have a look please? [11:58] Saviq: sure, will handle it [12:01] /food [12:01] Cimi: its ok as it is I'd say [12:02] Saviq: okay, I've done the camera-app one now [12:02] Cimi: actually you're right... this doesn't make too much sense [12:10] Saviq: thanks. will test it as soon it's built === MacSlow|lunch is now known as MacSlow === _salem is now known as salem_ [12:29] pstolowski, Saviq, anybody, may you know why I'm getting "SSQueryObject::run(): unity::InvalidArgumentException: Result::set_uri(): Invalid empty uri string" here? http://pastebin.ubuntu.com/7759946/ [12:29] that is the smartscopeproxy output for a query I did to the 7digital scope I'm building [12:30] mmm... the first item has no uri, that could be === Trevinho_ is now known as Trevinho [12:31] * facundobatista tries changing that [12:32] Saviq, cardCreator should add those signals for all cards no? [12:32] yay, that was it [12:33] Saviq, it is AbstractButton [12:33] but not the loader, the loaded item [12:33] lunch#! [13:01] Cimi, yes exactly, the direct delegate is the Loader [13:15] facundobatista, this can only happen if the the sss result doesn't have uri attribute or it's empty [13:15] pstolowski, thanks! [13:21] Saviq, I'm lil confused, if you can follow me and confirm all is fine :) [13:22] Saviq, in Carousel.qml we call delegateItem.clicked() for example [13:22] Saviq, this is relative to the Loader, but also the Loader item has clicked [13:22] are we sure those never clash? [13:30] Cimi, in CardCarousel [13:30] Cimi, we have another loader [13:30] Cimi, to load the component from cardcreator [13:30] yeah exactly [13:31] Saviq, but this has as well clicked and pressandhold [13:31] Cimi, so Carousel calls clicked and pressAndHold on that Loader [13:31] indeed we use those signals in cardfiltergrid [13:32] Cimi, no, the *card* has clicked and pressAndHold [13:32] Cimi, and in CardFilterGrid we can use those [13:32] that is inside the loader [13:32] Cimi, because it's the MouseAreas of the Card itself [13:32] so if all the calls from carousel just go to the loader, right? [13:32] Cimi, that send those signals [13:32] without if [13:32] Cimi, instead, in Carousel, it's the Carousel that calls clicked and pressAndHold on its delegates [13:33] Cimi, FilterGrid doesn't do that, the delegates themselves get the mouse events [13:33] Cimi, in Carousel you hijack them [13:33] Cimi, and calculate the item number yourself [13:34] Cimi, so the clicked and pressAndHold signals from the card are never triggered [13:34] Cimi, Carousel would need to know that there's a Loader in its delegate and call the signals like delegate.item.clicked() [13:35] Cimi, which is not great [13:35] nope [13:35] fine [13:35] I just wanted to make sure those signals were never arriving to this CardCreator [13:36] Cimi, you wrote the Carousel, you should know ;) [13:36] mhr3: If I do this: result["uri"] = cannedQuery.to_uri(); -- when I touch the result, it should issue the CannedQuery, shouldn't it? [13:36] Saviq, you modified it :) [13:36] karni, well... it will preview [13:37] karni, and if the preview action doesn't handle it, it will activate it [13:37] == issue the query [13:37] Saviq: Cimi: on that note, I think overlays should be visible on all cards of a carousel. It's quite loseless to see images and not know what they refer to, if you ask me. Just my 2cents. [13:37] mhr3: so I have to implement the onActivate thingy in the scope? [13:37] karni, useless you mean? [13:37] mhr3: can you define "review action doesn't handle it"? [13:38] karni, problem is the text would get scaled down a lot [13:38] karni, or we'd need to reflow [13:38] karni, or scaled up a lot [13:38] karni, for the current item [13:38] Saviq: Say, you have a bunch of articles with titles. if there's random pictures on them, I have to first scroll to any item to know what it relates to. [13:38] karni, probably shouldn't use carousel then ;) [13:38] Saviq: yes, I recognize the text would be much smaller [13:38] karni, http://developer.ubuntu.com/api/scopes/sdk-14.10/unity.scopes.ActivationResponse/#a726cf6ec58d4e035e2cdc14cda152d02aa3c4b9a3c5fcb54e5cce7d473cecfde3 [13:38] Saviq: fair enough ;) [13:39] karni, (which is the default) so you don't need to implement anything really [13:40] mhr3: reading and trying to understand how this works. I want my result to trigger search. I get a preview. [13:41] karni, right, you can't do that right know, we asked design to figure out a way to visually distinguish such items, and they you will be able to... but for now users would have no idea whether tapping something is going to do an action or open a preview, so it always opens a preview [13:42] s/they will be able to/then you will be able to/ [13:43] mhr3: :/ is there a workaround? I'm totally fine for the user not knowing what to expect (for now). I need to show a list of results when a "department card" is pressed (list of results has 1-1 mapping with departments. long story, if you really want to understand the reason behind this. short story: *subdepartments* *are* the results that I want to surface.) [13:43] karni, no, it needs to go through a preview for now [13:44] that's pretty bad :( [13:44] mhr3: you're saying I should put a button on the preview to trigger the search under the result["uri"] basically? === larsu_ is now known as larsu [13:45] karni, until design gets back to us and we a global solution for allowing preview-less results, yes [13:47] mhr3: I know we're dependent on Design here, but it puzzles me how we need spec for such a natural thing. "Best things to do this week" - most simple example of a result that naturally would return a list of results, instead of a preview. [13:48] dednick, hey, any update on https://code.launchpad.net/~nick-dedekind/unity8/indicator.call-hint/+merge/218627 ? [13:48] mhr3: What about this activation, could I use that to force search upon result click on specific results? If so, I would really want that (or Joe would, in fact) [13:49] karni, nope, that only applies when you're inside a preview [13:50] Saviq, app [13:50] Cimi, what app? (j/k) [13:50] mhr3: I thought Activation can tell whether we should show a preview, or issue search, or [other things], no? [13:50] approved [13:50] Cimi, yeah, j/k [13:50] karni, yes, activation in reply to a preview action [13:51] I don't know what j/k means [13:51] Cimi, just kidding [13:51] hah [13:51] mhr3: then we're talking about the same thing, yes [13:51] Saviq: um. no. not sure why it's not installing ofono... [13:51] mhr3: click on a result -> don't preview, search instead [13:51] karni, preview action == button in a preview [13:51] uhh [13:51] ok, I guess I have to give up [13:52] dednick, I was asking about Daniel's review comments rather [13:53] dednick, why do we need to install ofono in qmluitests though? [13:53] Saviq: ah. havent got to it yet. [13:54] dednick, we shouldn't install it Build-Depends ideally [13:54] dednick, we should mock it instead [13:54] (this applies to other plugins, too, probably) [13:54] that we already have in B-D [13:55] Saviq: right. i have mocked telephony plugin. maybe just shouldnt be in build-deps [13:55] dednick, yeah, add it to build.sh, too [13:56] dednick, the whole dep situation should get better once we move to autopkgtests, where we can declare test dependencies [13:56] Saviq: ok [14:05] Saviq: i'm not going to make standup today. I'll add my notes. [14:05] dednick, k [14:13] mhr3: I get this when I press the preview button with a canned query in it's uri field: http://paste.ubuntu.com/7760352/ [14:14] "Scope instance 0x94e68020 doesn't have Scopes as a parent" [14:14] Nothing else happens, doesn't launch the search. am I doing something wrong? [14:14] karni, right, no bug [14:15] mhr3: "no bug", or "no, it's a bug"? :D [14:15] karni, https://bugs.launchpad.net/unity-scopes-shell/+bug/1329890 [14:15] Ubuntu bug 1329890 in unity-scopes-shell "Canned queries don't work if they come from a scope that doesn't have a page in the dash" [High,Triaged] [14:15] phew, so there's an easy workaround [14:15] mzanetti: https://code.launchpad.net/~paulliu/unity8/reboot/+merge/221159 [14:16] karni, and yea, sorry about the missing comma :) [14:16] np [14:16] and thanks, mhr3 [14:24] tedg, poing [14:27] dednick, hmm shouldn't the new settings components conflict with -assets? [14:27] Saviq: yeah. i was thinking something like that... [14:28] Saviq: just had some issues reverting from new to old. [14:29] Saviq: is the infographics silo good to go yet? I think I resolved the conflicts.. [14:31] bregma: hello! [14:32] pete-woods, we need to wait for new UTIK [14:32] UITK [14:32] uh-oh [14:32] pete-woods, after that it seems it'd be good to go in [14:32] sil2100, how can I help? [14:32] bregma: so, is the compiz SRU that's currently in -proposed verified and works OK? I was wondering if we could maybe merge it in already [14:32] Saviq: okay cool, just wanted to check I had done everything you needed [14:32] pete-woods, yeah, I wanted to have it prepped for when new UITK is there, will rebuild now [14:32] (we're low on silos, so if the fixes are all verified then we could merge it in already) [14:33] sil2100, yes, except it seems to be stuck in the queue pending verification of a bug not in the SRU [14:33] bregma: I see one bug not verification-done, but I'm not sure if it's still valid [14:33] Ok, as I thought then [14:33] hmm [14:34] it seems the automated SRU report generator doesn't take into account bugs that were there and then removed [14:34] I suspect it needs a manual intervention by someone who knows what they're doing [14:39] mzanetti, sorry, pressure from above ;) [14:39] mzanetti, but hopefully it's an easy one [14:39] Saviq: no worries... I thought I did that one already [14:40] mzanetti, yeah, I just redone it over the weekend to be more future-proof [14:40] ah [14:40] Saviq: hey, about josh doing those things with autopilot [14:40] do you think that's a good idea? [14:40] Is there a way to clear suggestions/history from recent changes? I'd like to test something. [14:42] karni, recent searches you mean? [14:42] yes [14:42] karni, just restart unity8, they're not persistent [14:43] mzanetti, that'd be the highest level of those tests, I think it's fine with AP [14:43] mzanetti, closest to real life [14:44] Saviq: bad news is, it seems I *need* a search (even empty space) to force the requery when using CannedSearch. I'd like the CannedSearch to only contain department, but that, funny enough, only changes the department, without touching/affecting the results visible underneath. If I use any search term (and specifically ignore that search term in the scope, like a space), I get proper department *AND* results shown. [14:45] karni, this'd be a mhr3 thing [14:45] Saviq: ack [15:26] paulliu, can you please update the status on (and assign yourself to) bug #1234062 [15:27] bug 1234062 in Ubuntu UX "no power off dialog" [High,Triaged] https://launchpad.net/bugs/1234062 [15:36] Saviq, ideas why my theme doesn't have "chevron"? [15:36] Saviq: ok [15:37] mhr3, it doesn't exist in suru, does exist in ubuntu-mobile [15:37] mhr3, somehow on desktop the fact that ubuntu-mobile is parent to suru [15:37] Saviq: only unity8 right? [15:37] mhr3, doesn't work [15:37] paulliu, yes [15:37] Saviq, i see... thx [15:37] mhr3, there isn't a bug yet I think, worth filing one [15:48] Saviq, http://paste.ubuntu.com/7760732/ fixes it for me, know who deals with that project? === alan_g is now known as alan_g|EOD === josharenson1 is now known as josharenson === salem_ is now known as _salem === _salem is now known as salem_ [21:26] hi, i am trying to run unity8 from 14.10 daily (pulled and built from bzr) but ./run.sh just says "Unity8 is already running, please stop it first" [22:10] mterry: hey, could you help me debug an issue on my desktop where USC fails to come up? [22:10] greyback, k [22:10] mterry: this looks bad, right? http://pastebin.ubuntu.com/7762308/ [22:11] mterry: USC is running anyway [22:11] but I don't see a greeter, I get just a blank screen [22:11] no mouse-cursor either [22:13] there is an X server running too: /usr/bin/X -core :0 -seat seat0 -auth /var/run/lightdm/root/:0 -mir x-0 -mirSocket /run/mir_socket -nolisten tcp [22:13] greyback, that's not great [22:14] greyback, so USC is running and presumably didn't crash [22:14] I don't see any errors in Xorg.0.log [22:14] greyback, the proximate cause of the crashes seems to be connection to X [22:15] mterry: suspect it's a (X)Mir problem? Will I ping the mir team? [22:16] greyback, is there anything suspicious in /var/log/lightdm/unity-system-compositor.log? [22:16] I can kill USC and start a plain-jane mir server and client just fine. So not driver problem [22:17] mterry: looks ok to me http://pastebin.ubuntu.com/7762332/ [22:17] greyback, yeah... confirms that it never sees a client session though [22:17] greyback, yeah, I'd poke Mir team [22:17]