=== tmpRAOF is now known as RAOF [09:16] 5.4.1 migrated \o/ [09:40] Saviq: :) [09:40] would be happier if QT CI wasn't blocked my mac tests not passing though :D [09:42] Saviq: how's our landing going? [09:42] Cimi: what happened with the zooming stuff then? [09:42] tsdgeos, I gave up [09:43] tsdgeos, it worked fine on the phone, but the test fails [09:43] Cimi: so the regression is still there [09:43] tsdgeos, in the test [09:43] Cimi: well since you gave up [09:43] yes [09:43] the regression is still there and the test passes, no? [09:44] tsdgeos, with the changes you did for 5.4, the test pass but the panning the image doesn't work anymore [09:44] tsdgeos, there is no test for the panning [09:44] so we have a regression [09:44] no? [09:45] yes [09:45] did you report it? [09:45] if I revert your change, zoom and pan seems to work fine [09:45] on the phone [09:45] but the test fails [09:45] I didn't [09:45] so can you create a bug so we don't forget to fix it? [09:45] let me report [09:46] cool :) [09:48] tsdgeos, reconfiguring now [09:48] https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1433506 [09:48] Ubuntu bug 1433506 in unity8 (Ubuntu) "Pan doesn't work anymore after 5.4 transition" [Undecided,New] [09:48] tsdgeos, need to rebuild and retest after the Qt landing === vrruiz_ is now known as rvr [09:55] mzanetti, put dpr in silo: http://people.canonical.com/~platform/citrain_dashboard/#?distro=ubuntu&q=landing-015 [09:55] Saviq, ok, thanks [09:56] so will look at that soon, couldn't get to it yesterday [10:13] ok, let's see how the mother of all silos copes... [10:19] whoa, 24 MPs and no conflicts... we're getting good at this :) [10:33] Cimi: i can't get pinch and pan to work on the phone with the old code [10:33] Cimi: would you mind sharing which code you were using on the phone? [10:34] Cimi: wait my mistake [11:54] /food [12:01] @unity: here we go: http://notyetthere.org/ === paulliu1 is now known as paulliu [12:01] it's getting serious now [12:02] :) [12:03] mzanetti, top! :) [12:04] mzanetti, but you took him the pleasure of unboxing! :) [12:04] yeah :D [12:04] couldn't resist [12:04] will wrap it again [12:05] mzanetti, the folded info-card is so nice! [12:06] yeah. They've written "Ubuntu Edition" everywhere [12:06] really cool [12:06] * MacSlow is a bit envious :) === Malsasa_ is now known as Malsasa [12:43] /reallyfood === alan_g is now known as alan_g|lunch === MacSlow is now known as MacSlow|lunch [13:26] mzanetti, you got a reviewer for lp:~mzanetti/unity8/pinlock-scaling ? === alan_g|lunch is now known as alan_g [13:44] Saviq, MacSlow|lunch said he'll look at it === MacSlow|lunch is now known as MacSlow [13:44] mzanetti, kk [13:44] Saviq, looking into it atm === dandrader is now known as dandrader|afk [14:10] hi, what is the replacement for putting scripts in /etc/X11/Xsession.d/ in unity 8? === dandrader|afk is now known as dandrader === boiko_ is now known as boiko [14:28] seb128, is there a bug reported about apps not running under systemd in unity8 session? [14:29] Saviq, is that on an upgraded system? [14:29] seb128, I believe so, I had to switch to upstart to have them work [14:30] will try and reboot to sd again [14:30] Saviq, systemctl enable cgmanager [14:31] * Saviq reboots [14:31] Saviq, no need [14:31] you can just start it manually [14:31] seb128, need, on upstart now [14:31] k [14:41] seb128, indeed, wasn't enabled, do we have an idea (or a bug) tracking why? [14:42] Saviq, it was not enabled by default earlier in the cycle and the enable/disable don't change on upgrade [14:42] so if you had this version you need to manually fix it [14:42] Cimi: can you do https://code.launchpad.net/~aacid/unity8/pan_regression/+merge/253375 ? [14:42] meh, k [14:42] known issue with systemd packaging in debian/ubuntu [14:42] didrocks is looking into fixing it [14:43] it's not going to happen for new installs or people upgrading now [14:43] since cgmanager is enabled by default now [14:43] yeah understood [14:50] mzanetti, can I reproduce the pixel ratio issue on desktop somehow? [14:51] Saviq, only when using ubuntumirclient I'd say [14:51] Saviq: lp:~saviq/unity8/wrap-qmltestrunner: would delete 7.85+14.10.20140428.2-0ubuntu1 [14:51] so yes, in the desktop-session-mir it should be the same [14:51] lp:~alecu/unity8/payments-button: is the evil [14:51] has zillions of tags [14:52] alecu: what's the deal with that branch ↑↑ is it still meaningful? [14:53] tsdgeos, yeah, that branch needs updating first of all [14:55] tsdgeos, very good [14:58] tsdgeos, it'd add some wait after each commit, so we see a bit of animation too [14:59] tsdgeos, and apparently is more reliable daniel said [14:59] Cimi: i totally see the animation here [14:59] tsdgeos, cool [14:59] tsdgeos: I think it was merged into this other branch: https://code.launchpad.net/~unity-team/unity8/payments-button [14:59] tsdgeos, I meant on the test [14:59] tsdgeos: so, I'm setting it to abandoned [15:00] tsdgeos, do you see in the test? [15:00] Cimi: yes i see it pinch, pan and unpinch [15:00] alecu: oki :) [15:00] Cimi: you don't? [15:00] tsdgeos, ok, not on my machine, weird [15:00] tsdgeos, nevermind [15:01] tsdgeos: actually, I was able to set the status of the MP, but I can't change the status of the branch from "Development". Can you do it? [15:01] let me try [15:02] alecu: nope [15:02] alecu: it's weird you cna't edit your own branch [15:02] I can [15:02] for whatever reason [15:02] alecu, tsdgeos, done [15:02] :) [15:03] tsdgeos: doh, I was not logged into launchpad in that browser window :P [15:03] he he [15:03] tsdgeos: now, what about it having "zillions of tags"? what are those, and why are they bad? [15:03] mzanetti, so most of it happens in the QPA, not the SDK? how does that affect our apps running on X11? [15:04] alecu: bzr tags, they came from misterious places, and they replicate a lot, they are wrong because they point to broken revisions of the code [15:04] Saviq, well, it happens in both, the QPA and the SDK [15:04] alecu: if you have any local checkout either remove it and checkout again or run http://people.canonical.com/~msawicz/unity8/strip-u8-tags.py over it [15:04] Saviq, yes, X11 is unchanged atm [15:05] Saviq, actually, X11 already supports QT_DEVICE_PIXEL_RATIO, while Mir doesn't [15:05] tsdgeos: ack, thanks. [15:06] tsdgeos: "they came from misterious places, and they replicate a lot" <- they sound like Gremlins [15:06] he he :D [15:11] mzanetti, well, right, but are you not doing something special in mir QPA to "inflect" DEVICE_PIXEL_RATIO from GRID_UNIT_PX? [15:12] Saviq, I do that in the qtubuntu qpa, yes, not in the X11 one [15:13] mzanetti, but if I set DEVICE_PIXEL_RATIO in X11, everything will Just Work™ anyway, right? [15:13] yes, mostly.. [15:14] Saviq, this is how I start qtcreator nowadays: QT_DEVICE_PIXEL_RATIO=2 GRID_UNIT_PX=16 qtcreator [15:15] Saviq, worth noting that before some recent upgrade I had to use "QT_DEVICE_PIXEL_RATIO=2 GRID_UNIT_PX=8 qtcreator" for the same result [15:15] mzanetti, the "new" way seems correct [15:15] or does it.. [15:16] * Saviq no gets why we'd not get the loop under X11 then [15:16] I still haven't made up my mind about that tbh... given that on Mir we won't have QT_DEVICEL_RATIO in the end [15:16] QT_DEVICEL_RATIO :D [15:16] mzanetti, Hi! [15:16] mzanetti, when does this land https://code.launchpad.net/~mzanetti/unity8/fix-laggy-launcher-dnd/+merge/251567 ? [15:16] om26er, my last state is "when QA manages to test the silo" [15:17] mzanetti, well, we kind of will, only it will by default be based off of GRID_UNIT_PX, right? [15:17] mzanetti, otherwise QWidget apps would be bad [15:17] Saviq, exactly [15:17] we should allow for overriding btw [15:17] like if you do have DEVICE_PIXEL_RATIO in env, that should take precedence [15:17] Saviq, yeah, in the end both mechanisms override the "virtual qreal QPlatformWindow::devicePixelRatio() const" [15:18] Saviq, it just depends what qgetenv we do in there [15:18] mzanetti, I mean that DPR should take precedence over GUX [15:18] Saviq, I don't think we want to regard the DPR env at all [15:19] dunno, maybe [15:19] on one hand that means that someone exports DPR and expects it to work and it doesn't [15:19] Saviq, note that the env var is something specific to the X11 QPA, Windows, Mac etc doesn't use that either [15:20] oh [15:20] not sure about the wayland QPA [15:20] thought that was Qt [15:20] aaanyway, trying the desktop session to see if I can get it to die [15:20] yeah well, the X11 qpa is from upstream Qt, so it kind a is [15:20] yeah you know what I mean [15:20] mzanetti, btw, on flo + DPR, fonts huge, that expected? [15:21] really... I did test it on flo, was working fine [15:21] mzanetti, check out silo 15 [15:21] ack [15:21] have to do a bit for the alt+tab stuff atm, but I'll come back to this at latest tomorrow [15:21] Saviq: hi, do you know where i can put a startup script for a unity 8 desktop session? [15:22] attente_, we use upstart for the user session http://upstart.ubuntu.com/cookbook/#session-job [15:22] attente_, what sort o script? [15:23] Saviq, seb128: /etc/X11/Xsession.d script [15:23] what is the script doing? [15:23] we should probably avoid adding X11 scripts [15:23] seb128: it's the im-config script that launches the user's im framework [15:24] don't we already have that? or what do you want to change? [15:24] seb128: we have it, but it doesn't run under u8 [15:24] oh, right, no X11 there [15:25] you don't want to add a script to Xsession.d [15:25] you want to replace one by an upstart job? [15:25] seb128: sure. does it make sense to do a systemd one instead though? [15:25] shouldn't that be done for non unity8 sessions as well and in im-config itself? [15:25] we don't use systemd for user session yet [15:25] so no [15:25] ok [15:26] seb128: it's already there, but i'm wondering how to get it working under u8 [15:26] so i guess adding the upstart job is what we should be doing [15:26] yes [15:26] replace the Xsession script by an upstart job [15:26] ok, thanks [15:27] mzanetti, yeah, and can't repro anything wrong under desktop session [15:29] Saviq, you commented that you were having problems with the cancel-pam-harder branch (https://code.launchpad.net/~mterry/unity8/cancel-pam-harder/+merge/251174/comments/628754) [15:29] Saviq, were you using the u8 session? or just doing ./run.sh [15:29] mterry, real session [15:30] mterry, and I couldn't get past the lockscreen [15:30] Saviq, ok === boiko_ is now known as boiko === boiko_ is now known as boiko === dandrader is now known as dandrader|lunch [16:31] tsdgeos, ok. tested your branch now. looks ok. anything special I should test? [16:32] mzanetti: nothing special comes to mind [16:32] ok. approving then [16:32] the code is centered around updates of the model [16:32] you can try to pull a few times in a few scopes [16:32] see nothing breaks [16:32] tsdgeos, yeah, I tried switching departments and refreshing like mad for a few minutes now [16:32] pull => refresh [16:32] besides that nothing else comes to mind [16:32] ok then. works for me [16:33] let's hope it fixes it [16:33] that's the nice thing about errors.ubuntu.com [16:33] you can see if the crash continues happening or not [16:34] besides images can anyone think of something we download from unity8-dash itself? [16:34] because everything else i can think of comes from backends [16:35] tsdgeos, for music there's a play feature for audio... not sure how much that qualifies for "downloading" though [16:35] mzanetti: does it come from http and uses QNAM? [16:35] and is done in the unity8-dash process? [16:36] tsdgeos, hmm... I think we get a url from the backend, not sure if that's local or not... but it's been a while since I did that.. [16:37] mzanetti: ok no worries [16:43] mzanetti: my branch has "less" failures because it compares to previous branch [16:43] which was one from dednick that needs a change in some other component [16:43] and thus was failing to compile some tests [16:43] so it seems mine "fixes" stuff [16:43] when really doesn't [16:44] tsdgeos, huh... it says +24 failures here [16:44] relaly? [16:44] ok that's weird [16:44] mzanetti: https://jenkins.qa.ubuntu.com/job/unity-phablet-qmluitests-vivid/608/ ? [16:44] -7 failures [16:44] tsdgeos, yeah, AP [16:44] mzanetti: ah === dandrader|lunch is now known as dandrader === alan_g is now known as alan_g|EOD === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader === dpm_ is now known as dpm-afk [21:21] got a server that a customer is running unity with remote desktop on [21:21] (using like a terminal server setup, multiple sessions independently) [21:22] would I need logs of all the sessions to see about improving performance? [21:23] right now load is around 3.0 (which is totally fine), mostly caused by compiz processes... [21:23] just looking for next steps if performance get's bad with compiz/unity [21:25] (already have it set to LOW graphics mode) [21:26] gQuigs, unity relies heavily on GPU acceleration, which without a GPU happens on the CPU, hopefully via the best thing out ther - LLVM / Gallium, but that's still heavy on the CPU because it's not geared well to be doing that [21:27] gQuigs, you could try disabling some plugins in ccsm, but not much else can be done [21:28] Saviq: what's used for performance tuning? [21:30] gQuigs, there's basically no "performance tuning" that I know of you can do on Unity/Compiz, they assume OpenGL is available and that's basically it, all the things are (hopefully) done in the best performing way [21:32] Saviq: thanks for insight.. hopefully there won't be any issues... thanks again [21:32] gQuigs, it's just that it's not geared to be run remotely [21:32] Saviq: yea, I know.. customer was nervous about upgrading from 12.04-> 14.04 because of it (us dropping 2d unity) [21:33] and 16.04 will basically have 2d unity back.. so 14.04 puts us in a weird place [21:33] so far, so good though [21:33] gQuigs, well, I wouldn't say it will have 2d unity back ;) [21:33] just the fact that it's built on QML doesn't make it "unity 2d" [21:34] unity8, which I assume you're referring to, relies on GPU/OpenGL just as heavily as unity7 [21:34] there's no CPU renderer for QML in upstream Qt now [21:34] so you can't run anything modern QML without OpenGL [21:35] Saviq: well from my testing on an unaccelerated VM, unity8 feels quick, and unity7 doesn't.. but I head you... [21:35] gQuigs, sure, unity8 is just leaner still, and we'll try and keep that as we build more desktop functionality in it [21:35] :) [21:36] but yeah, the primary target being mobile, we have to keep it quick