=== chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [09:41] alan_g: Is there a rule of thumb for what to do in place of "static virtual"? [09:41] It's been years since I needed one [09:41] What do you want it to do? [09:42] alan_g: Return a constant value. But that is configurable based on the class you ask [09:42] Hmm [09:42] Not static [09:43] Determined at runtime or compile time? [09:44] alan_g: Compile time [09:45] But the implementation class isn't known till run time [09:45] I know it's theoretically reasonable but maybe not possible in C++ [09:45] It's just like querying RTTI in a way [09:46] virtual Value value() const = 0; === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader === ara is now known as Guest11257 [11:35] didrocks: is there a CI problem? Jenkins doesn't seem to have done anything for 10 hours or so: https://code.launchpad.net/mir/+activereviews [11:36] alan_g: you should ping the vanguard from the CI team for that I guess [11:36] on #ubuntu-ci-eng [11:36] (we are only doing the landings) === alan_g is now known as alan_g|lunch === dandrader is now known as dandrader|afk === alan_g|lunch is now known as alan_g [13:22] alf__: hey, so josharenson is working on getting glmark2 running as part of our ci...however, glmark2 that's in archive for arm is only for x11 [13:22] my understanding is that there is a glmark2_es2_mir [13:22] that can be built, but its not in the archive [13:23] so i was hoping to add it, do you see any problem there ? [13:23] mir is in universe as well, so should be fine...right? [13:23] (need to do this in order to have the ability for the jenkins job to apt-get install glmark2-es2-mir as part of the ci run) [13:26] kgunn: when I released 2014.03 I also updated a packaging branch so that it produces packages for all variants https://code.launchpad.net/~glmark2-dev/glmark2/glmark2-package [13:28] hello guys [13:28] I import Unity.Application 0.1 in my qml [13:28] I start this when unity8 is not running [13:28] but still, my app crashes with this error [13:29] [libprotobuf ERROR google/protobuf/descriptor_database.cc:57] File already exists in database: unityrpc.proto [13:29] [libprotobuf FATAL google/protobuf/descriptor.cc:954] CHECK failed: generated_database_->Add(encoded_file_descriptor, size): [13:29] terminate called after throwing an instance of 'google::protobuf::FatalException' [13:29] what(): CHECK failed: generated_database_->Add(encoded_file_descriptor, size): [13:40] Cimi: so you start before unity8...but are you using unity-mir ? or...rather...are you trying to launch the mir server ? or going through u-s-c's instance of the mir server ? [13:41] kgunn, just running the app with upstart... [13:41] kgunn, don't know the internals [13:41] kgunn, so probably mir is already running [13:42] Cimi: mmm...do you see the greeter ? [13:42] kgunn, I'm starting on xsession-init [13:42] nope [13:42] Cimi: yeah...something needs to get you a mir server fired up... [13:42] Cimi: are you just doing this for testing ? or is this actually how you're planning on doing it ? [13:43] e.g. does the user see the greeter first? then dismiss into the welcome wizard ? [13:43] kgunn, I suppose the welcome wizard needs to be before unity [13:43] Cimi: sure... [13:43] before everything [13:43] it can be.. [13:43] Cimi: well....there's some effort there tho... [13:43] wonder if you can just glom onto greeter ? [13:43] mterry: ^ [13:44] seems Cimi needs a mir server....but wants to be "before everything else" [13:44] should he start his own mir server ? [13:45] kgunn, Cimi: my plan was to insert it right before greeter [13:45] Cimi: you're suffering since split hasn't landed yet [13:46] Cimi: today there is only 1 mir server....shared by unity8 (session) & greeter (system) [13:46] kgunn, Cimi: well, with split greeter, you'd have same problem [13:46] kgunn, mterry, Cimi, he doesn't need a mir server, thouhg [13:46] Why do we need a mir server? [13:46] it could be just a client [13:46] This is for notifications? [13:46] mterry, no, no need for a server [13:46] mterry, we can be just a client [13:47] Saviq: client to u-s-c's server ? or a 3 rd instance? [13:47] kgunn, client should be good [13:47] the current issue is that Cimi loads the Unity.Application API , which probably requires to be a mir server [13:47] or well, wait [13:47] I think he needs to be a server for the OSK [13:48] we don't want the OSK to go directly to u-s-c do we [13:48] mmm [13:48] how is it handled in greeter for pin code before unity8 is started ? [13:48] same case isn't it? [13:48] yeah [13:48] mterry, ↑ ? [13:48] in split greeter, that is [13:49] kgunn, Saviq: in split greeter, maliit talks to split greeter, which is a server [13:50] The unity.Application plugin can only be used by an application that instantiates a QMirServerApplication. Here's a simple example: lp:~gerboland/+junk/qml-demo-shell/ [13:50] Saviq, out of curiosity, what is wrong with OSK talking to USC? [13:50] mterry, you tell me :) [13:51] different users could have different OSKs... [13:51] I don't know enough about differences between mirservers, but I'd guess nothing [13:51] mterry, I think at least one thing is that shell, greeter, wizard needs to control the input routing [13:51] greyback, yeah but this is a very special case (first boot wizard) [13:51] mterry: ah I was missing context, sorry [13:52] mterry, greeter is not as special, though [13:52] greyback, that would include the input areas and such, right? [13:52] Saviq, sure. But greeter acts like unity8 [13:52] Saviq: for USC, yes === ara_ is now known as ara [13:52] mterry, in that sense welcome wizard does, too, 'cause it's the only "shell" running at the time, even if the only client would be the OSK [13:53] greyback, I mean if the welcome wizard wants to use import areas (which it needs for OSKController), it needs to be a QMirServerApplication? [13:53] Saviq, yeah... but it's not a shell. It's just a client itself [13:54] I mean, we can change it to be a server too. [13:54] mterry, yeah, as long as it can control the input area of the OSK... [13:54] but wait...isn't Cimi finding out he alreaday needs more clients (following the shell argument...like notifications, settings app) [13:54] Saviq: yes [13:54] kgunn, I don't think notifications are clients... I think shell renders those [13:54] kgunn, they are within the app [13:54] (Or whomever loads the Notifications plugin, which maybe could be wizard) [13:54] notifications is in-surface [13:55] settings? why? [13:55] anyway, gtg for a bit [13:55] actually, bringing the laptop with me... === dandrader|afk is now known as dandrader [13:55] Cimi: do you launch settings ? or just pirate pieces of it ? [13:55] kgunn, I import the plugins I need [13:55] kgunn, it's standalone app [13:57] So looks like only client needed is OSK so far [13:57] ok...so sounds like we're settled.... mir client ? mterry ? [13:57] ;) [13:58] But sounds like it needs to be a server to manage the OSK [13:58] Cimi, we night need to make the wizard a little server... greyback, do we have any experience with OSK and servers that aren't unity-mir? Or should we just use unity-mir here too [13:59] the server needs logic to show/hide the OSK surface, and do a little input filtering (top of OSK surface transparent, so should not steal events) [13:59] greyback, is that stuff we get magically for being a server, or do we have to reimplement that logic? [14:00] mterry: you'll have to re-implement, if your server is not using unity-mir [14:00] which it isn't, it's using mostly default mir stuff, isn't it? [14:00] greyback, do you remember how closely tied unity-mir is to the unity8 shell? Last time I remember looking, it was checking surface names and such [14:01] greyback, right now the wizard isn't a server at all [14:01] greyback, we were hoping to get away with just being a client [14:01] mterry: but the wizard is a client, and should stay that way [14:01] mterry: what is the server here? [14:02] unity-mir needs one thing from a shell, that it names its surface something particular. That's all. [14:02] greyback, well this is all up for debate. But the plan I was envisioning was having wizard run under USC as a client with maliit server running beside it [14:02] mterry: that's fine. So USC needs logic to handle the OSK surface [14:02] Hrm [14:03] greyback, and that logic is stealable from unity-mir? [14:03] I don't know if I like USC adding that feature just for this one time wizard [14:03] mterry: it's mostly implemented in QML, but it's pretty easy [14:03] Ugh or gaining QML [14:03] 1) try to detect the OSK surface by its name [14:04] right now USC doesn't need to bother with the UI parts of Qt [14:04] 2) listen for maximize/minimize evnets from that surface, and when they happen, show/hide the surface [14:04] 3) set an input area on the OSK surface so the transparent part does not take events [14:06] greyback, out of curiosity, why doesn't OSK just show/hide itself and such? [14:06] And set it's size itself [14:07] mterry: because surfaces cannot do that. That's a shell's decision [14:07] Instead of needing to have a transparent part [14:07] ah the transparent part for hte top row of keys, when you press them, the letter pops up higher [14:07] Ah OK [14:07] Details details everywhere [14:08] indeed [14:08] some keyboard too, a long press pops up a accent selector, which needs to receive events [14:08] greyback, if we add those features to USC, I'd want to avoid QML. How easy is it to do that input filtering without a nice Qml input area? [14:09] input filtering a Mir concept, I've just wrapped it for use in QML [14:09] Yar, just hoping it was easy [14:09] Still, feels weird to add this complexity to USC just for the wizard [14:10] * mterry wishes wizard didn't need OSK [14:10] think mir::shell::surface has a installInputArea method, which takes a rectangle. [14:10] greyback, how ugly do you feel it would be to make wizard use unity-mir? [14:10] inside that rectangle, the surface gets the events. Else it doesn't. [14:11] mterry: then you're making the wizard a mir server. Which would work, yeah [14:11] greyback, we'd have to name the wizard surface like the shell does. But it sounds like besides that, it'd work? [14:12] mterry: think so, yes. Maliit needs to connect to the wizard's mir socket, Then all should be good [14:12] Cimi, ^ what do you think? [14:13] * Cimi reads [14:15] mterry, I like your simpler solution :D [14:15] Cimi, "simpler" :) [14:19] Cimi, so to do that... unity8 has src/main.cpp which dlopens unity-mir [14:19] mterry, already saw it [14:19] Cimi, we probably don't need all the complexity it has (in particular, we don't need to worry about non-mir instances) [14:20] Cimi, and check unity-mir source. Somewhere in there, it checks a surface name that we'll want to match to unity8 [14:21] * Cimi looks [14:28] kgunn: greyback: https://code.launchpad.net/~afrantzis/mir/non-blocking-swap-buffers-spike , it's a spike not ready for an MP just yet, but it should work. [14:28] have a look at lp:~gerboland/+junk/qml-demo-shell/ for a simple exmaple [14:29] alf__: wow nice! Will test straight away [15:30] jenkins traffic jam === chihchun is now known as chihchun_afk [16:55] anyone know what's going on with autolanding? [16:56] kdub: yeah, there's problems with the phones [16:57] alan_g, thanks [16:58] I asked on #ubuntu-ci-eng and found that it had been announced on a mailing list I didn't know about === alan_g is now known as alan_g|EOD [18:19] * mterry is seeing odd input bug with mir/devel + unity8... [18:19] I'm testing further, but if anyone has played with full stack lately, let me know [19:23] mterry: got a description ? [19:23] racarr: anpok ^ [19:23] kgunn, I was seeing the greeter session come up but wasn't able to interact with it [19:23] kgunn, still not sure which component's fault it is yet. [19:24] mterry: so like osk wouldn't invoke ? as well, you couldn't swipe? [19:25] kgunn, right, couldn't swipe. didn't get as far as OSK [19:25] kgunn, nm, can't reproduce now... :-/ [19:25] kgunn, I have some updates for silo 002 if you have time [19:26] sure [19:27] hey...crap, meant to ask you mterry about [19:27] https://code.launchpad.net/~mterry/gsettings-ubuntu-touch-schemas/volume/+merge/209158 [19:27] i have it in the landing sheet...but, i can't recall...did you want it landed in isolation or not ? [19:27] wasnt that one of the issues that crept in and got fixed with the surface restructuing? the thing that racarr looked at? [19:27] +r [19:28] hm most current greeter stuff is now in a different silo? [19:28] kgunn, it landed with indicator-sound in trusty already. We can drop it [19:29] kgunn, for silo 002... add ~mterry/platform-api/mir-changes & ~alan-griffiths/unity-mir/compatibility-with-mir-changes & rebuild platform-api, unity-mir, and ubuntu-touch-session [19:29] kgunn, we can remove any indicator-sound or schema stuff in silo 002 now too, if any are lfet [19:30] anpok, we moved to silo 002 yeah [19:30] (from 004) [19:30] k [19:31] anpok, silo doesn't work right now due to mir churn, but it's about to because kgunn is a silo wizard [19:49] mterry: building now [19:51] gotta reboot.. [19:52] racarr, do you know about any changes that would affect raising/focusing sessions? [19:52] Or anybody really [19:54] mterry, yeah, SurfaceRanker is gone [19:54] the raise() method moved to SurfaceConfigurator [19:56] * kdub always thought "surface rancor" when reading that one [19:57] mm [19:57] My USC changes to raise the spinner session don't seem to be working anymore [19:57] someone must have landed unity-mir/usc updates or it just wouldnt build right [19:57] hmm [19:58] racarr, my copy of mir/devel must be between re-organizations then [20:00] racarr, I don't see raise() in SurfaceConfigurator in mir/devel [20:00] maybe [20:01] I think kdub is right with investigate SurfaceRAnker...lots landing so might be easier to just try and see what is going on with GDB rather than guess at what change could have caused it [20:01] surface wanker removal has not landed yet [20:01] I can help tomorrow but right now am reloading chromium in to my head and doing updates etc [20:01] racarr, no worries [20:02] https://code.launchpad.net/~alan-griffiths/mir/make-use-of-scene-Surface/+merge/213894 [20:02] anpok: Top 10 typos of 2014 [20:02] Mm [20:02] :) [20:02] maybe something doesnt work [20:02] in the intermediate state [20:02] though [20:03] now i see it .. oh dear.. [20:04] Hahahahaha [20:04] mterry: but it could also be related to other changes.. [20:04] well, landing soon then :) [20:52] bregma: so...i'm sure i did somethin' wrong...but, i added your unity8 desktop ppa, but i only installed indicator-session....as i wanted to test the usc thingy... [20:52] and i am working off the assumption/knowledge that qtubuntu got update [20:52] d [20:52] I believe it did, let me check on my most updated machine.... [20:52] and that lightdm is fixed for cursor [20:53] bregma: yeah...i just didn't have an "option" on the greeter ? [20:53] ( i assume its not guest...but that it would say "unity8 desktop" or some such) [20:54] is unity8-desktop-session-mir installed? The one in universe is current [20:56] I have the qtubuntu-desktop installed from universe (0.54+14.04.20140402-0ubuntu1) and it's working for me [20:57] lightdm is from main 1.9.14-0ubuntu1 [20:58] unity8-desktop-session-mir is from universe 1.0.10+14.04.20140402-0ubuntu1 [21:02] bregma: well i feel stupid.... [21:03] just installed [22:00] bregma: still no luck...here's my usc log [22:00] https://pastebin.canonical.com/107762/ [22:00] i gotta run... [22:00] but you might want to give usc from the silo ppa a shot [22:00] https://launchpad.net/~ci-train-ppa-service/+archive/landing-011 [22:01] if it works for you...then its me [22:01] if it doesn't then something else went wonky with usc [22:12] good old "Failed to set DRM crtc" [22:25] unrelated :) but, its a real benefit for our interfaces that we have to look at things from the two platform's perspectives === chihchun_afk is now known as chihchun