=== sil2100_ is now known as sil2100 [11:56] Saviq: so, turns out this vivid+overlayppa laptop doesn't boot into Unity8 because it was missing EGL_PLATFORM=mir [11:57] which seems to not be usually needed [11:57] so either some hw requires that env var, or there's another miissing pkg [11:57] that makes that env var useless [12:00] faenil, interesting, still a #ubuntu-mir topic, though :) [12:01] * Saviq has no idea what EGL_PLATFORM=mir means :P [12:02] Saviq: pfff, being pushed over here and there :D [12:02] faenil, you're a hot potato, what can I say ;) [12:02] faenil, but the thing says "mir", that must be a hint :D [12:02] Saviq: that I am indeed [12:03] Saviq: I didn't know about ubuntu-mir, sorry [12:03] faenil, nw [12:03] faenil, so you got it working after all? [12:04] Saviq: yes, it was just that damn env var...well now, I have the problem that the dash doesn't show apps with newer frameworks because the system doesn't have new frameworks [12:04] even more fun on the horizon! [12:09] faenil, why wouldn't it have new frameworks? vivid+overlay == rc-proposed? [12:10] Saviq: because they're seeded in the ubuntu-touch seed [12:10] so desktop doesn't get them :( [12:10] which makes me wonder how people are supposed to test unity8 on desktop without them... [12:10] faenil, oh well, looks like it's time for ubuntu-converged... [12:10] seed [12:11] Saviq: it is indeed... [12:12] I'm puzzled :| [12:13] Saviq: do you if unity dash filters apps if the system doesn't have one framework? [12:13] faenil, click scope does [12:13] well, click scope yeah [12:13] ok, so it's that... [12:13] faenil, it sends the available framework to the store when querying [12:13] faenil, but [12:13] I see.. [12:13] makes sense, just breaks my plans :) [12:13] what about the packagekit deb vs. click backend? did you manage to solve that? [12:13] I'll just install the frameworks stuff [12:14] no I'm using click install --user= --allow-unauthenticated [12:14] works ok on xenial [12:14] faenil, right, so why do you care about the store? to download into ~/.cache/* [12:14] ? [12:15] no I don't care about the store, I care that I installed dekko with --force-missing-frameworks [12:15] and now it doesn't show in the apps scope [12:15] faenil, aah, right :) [12:15] :P [12:15] faenil, that I don't know if it does, but anyway, yes, frameworks should be installed [12:16] (although I think Calendar shows on Xenial even if xenial doesn't have 15.04.3) [12:16] so I don't know if it's a bug or a feature :D [12:38] Saviq: my simple qml app doesn't run, says "qmlscene: failed to check version of file 'share/qml/blabla/Main.qml", could not open... [12:38] any idea? [12:38] I just took a working app [12:38] and downgraded the framework [12:38] so that it would show on vivid apps scope [12:39] can that break it? [12:39] (in that way) [12:40] faenil, doubt it, pastebin full output? [12:40] there's also QIBusPlatformInputContext: invalid bus [12:40] Saviq: nothing interesting oherwise, the usual stuff.. [12:41] oh also "QSocketNotified: Can only be used with threads started with QThread" [12:41] notifier* [12:41] UbuntuKeyboardInfo - socket error [12:42] I wonder if that can cause a qml file error though, I don't think so :/ [12:42] faenil, that last one is from unity8.log 'innit? [12:43] yeah [12:43] was having a look there as well [12:44] faenil, did it actually say "share/qml/blabla/Main.qml"? ;) [12:44] I've never seen "failed to check version of file" before... [12:44] well s/blabla/nameoftheproject [12:44] faenil, check if the file it mentions exists? [12:44] it does [12:44] if the path is relative to app's folder, that is [12:44] faenil, and if you run qmlscene yourself? [12:47] Saviq: what did I need aside from desktop file hint? [12:48] faenil, just that [12:48] faenil, assuming you're doing that from the terminal [12:48] faenil, if from VT or something, a bit more [12:48] I don't have terminal, it requires newer framework :P [12:48] so it doesn't show in apps scope [12:49] faenil, well just install it [12:49] why wouldn't you [12:49] I mean newer framework [12:49] yeah, let's do it [12:49] I think I just forgot in the chaos [12:49] do you know the pkgname already? [12:50] faenil, https://launchpad.net/~vthompson/+archive/ubuntu/ppa [12:50] core_t: ? [12:50] you can install the terminal from there [12:50] core_t: the frameworks pkg [12:51] oh [12:51] core_t: terminal requires 15.04.3 fw [12:51] it won't show if I don't install new fw === alan_g is now known as alan_g|lunch === alan_g|lunch is now known as alan_g [15:07] mterry: see my comment on your MR [15:07] in gerrit [15:07] your code may need adapting [15:08] tsdgeos, yeah I'm noticing [15:08] but i think the comment from Tor is mostly bogus [15:10] tsdgeos, yeah that code snippet you posted seems like a more conservative version of my patch [15:10] tsdgeos, I had forgotten about keyboard layouts that use Alt [15:12] also Alt != AltGr [15:12] mterry, tsdgeos ↑ [15:12] yes [15:12] you won't get Alt at all on the event IIUC [15:12] but still if they are arguing for "logic is wrong" [15:12] i think using the same code that has been deployed for a long time makes snese [15:12] less battles to fight [15:13] ;) [15:20] tsdgeos, naw now they just want to change that old code [15:20] no words [15:21] tsdgeos, Saviq: but... I can change my patch in the silo to be the more conservative version. At least that way we get our fix in a way that doesn't break any other layouts. And we can see what Qt wants to do in the general case? [15:22] mterry, let's fix proper, use shortcuts for shortcuts [15:22] that's bullshit btw [15:22] if they change that code in qwdiget [15:22] 95.67% of the people is going to get their code broken [15:23] meh, ours won't :P [15:23] because we'll be using shortcuts [15:23] tsdgeos, we may argue on the upcoming change for that, then [15:25] Saviq, ok... so drop the qtdeclarative stuff from the silo then. I can look at what it would take to switch webbrowser-app, but that's not likely to be an ota9.5 thing [15:25] mterry, I'm punting to oSoMoN as we speak ;) [15:25] Saviq, great [15:26] oSoMoN, https://bugs.launchpad.net/ubuntu/+source/qtdeclarative-opensource-src/+bug/1537782/comments/10 [15:26] Launchpad bug 1537782 in Canonical System Image "Modifier ignored when pressing a key if TextInput has active focus" [High,In progress] [15:58] Saviq, seen your comment, thanks. I’ll look into this new Shortcut thing, but not an OTA9.5 thing anyway [15:59] oSoMoN, you'll need to go QShortcut for vivid, though [15:59] oSoMoN, yeah, sry about that, we only just got feedback today [15:59] no worries === JanC_ is now known as JanC [16:39] mterry: So I have a full libertine setup now, but still don't have apps launching. [16:39] mterry: The only error I'm seeing right now is an incorrect icon. [16:39] mterry: Will qtmir/unity reject an app for a bad icon? [16:39] tedg, no [16:39] tedg, same error with busted appId? [16:41] mterry, no, it looks pretty clean there https://www.irccloud.com/pastebin/CtsHvTEc/ [16:41] tedg, qtmir.applications: Application::Application - appId= "test-xenial_sakura" [16:42] tedg, that's the line that yesterday had only sakura [16:42] So that's better... [16:42] tedg, do clicks and legacy apps work? [16:42] mterry, yeah, still getting xmir rejected though https://www.irccloud.com/pastebin/wYiYpnxI/ [16:43] Legacy apps do, I don't have any clicks installed. [16:43] tedg, no clicks?! camera-app? [16:43] No, this is on my laptop [16:43] tedg, ah [16:43] * tedg isn't *that* into selfies ;-) [16:43] tedg, no one believes you :) [16:44] tedg, well... I would expect to see a REJECTED message in unity8.log if we were doing the normal rejection of a surface request [16:45] tedg, although, I don't see ApplicationManager::authorizeSession in that lo [16:45] *log [16:46] Hmm, I might know what it is. I think we might be not readding the 0.0 in the case of short ids going to long ids on libertine apps. [16:46] tedg, you might need the whole USC / phone infrastructure for this to work as expected [16:46] ok [16:46] Only checking click for getting version numbers. === pat_ is now known as Guest30437 === alan_g is now known as alan_g|EOD