[05:07] Good morning [05:22] when I am trying to build nautilus from source: I got this: gcc: error: @UNITY_CFLAGS@: No such file or directory [05:22] guys, any suggestion? [05:56] morning === adam3 is now known as Steve_Jobs [06:50] * popey sees https://sites.google.com/a/android.com/tools/recent/androidstudio0810released and looks at didrocks [06:50] Mirv, hey, totally off-topic, I saw you were replying to a thread about tearing in xbmc/sdl in 14.04, that should be fixed in 14.10 already shouldn't it? do you know what I'd need to set up to get vblank on intel? [06:50] popey: canary channel man! only for my laptop (already updated :p) ;) [06:50] ☻ [07:09] Saviq: it's fixed in both 14.04 via SRU and 14.10 [07:09] Saviq: the original problem at least [07:10] Saviq: and I can confirm it on my 14.04 xbmc + haswell setup [07:10] Mirv, 14.10, haswell + xbmc here, tearing :/ [07:10] * Mirv haz tear free, 24p output [07:10] Mirv, but it might be some configuration issue [07:10] Saviq: :( might be config sure, too [07:11] of course, the problem field of tear free video is sufficiently complex to have numerous problem points [07:11] it's been borken quite a few times [07:11] hmm glxgears is vblanked at least [07:11] * Saviq drops .xbmc [07:11] Mirv, thanks [07:13] Saviq: so you go middle of nowhere but get your xbmc with you? ;) [07:13] Mirv, I've to do *something* there don't I ;D [07:14] hehe [07:16] morning folks [07:17] Saviq: my possibly relevant settings (settings level expert) system | vertical blank sync: always enabled, video | sync playback to display: off (used to be on, was useful to set off in 14.04 to get 24p output), render method: advanced shaders (glsl), decoding method: hw - enabled vaapi + mpeg-4 vaapi [07:18] Mirv, thanks a bunch [07:21] what hardware Mirv,saviq? Or specially, what video hardware [07:22] willcooke: both have intel haswell. I've 14.04 and all great, Saviq has tearing on 14.10 [07:22] willcooke, haswell, hd4000 or so [07:22] cool [07:22] I was playing with vlc and VA api the other day [07:22] worked nicely, once I'd jumped through some hoops [07:23] and to that end, I'm planning a session at our sprint to discuss getting HW accel. enabled by default for GStreamer etc [07:23] my biggest problem on 14.04 was that screen refresh rate didn't change to match the video, but turning that sync playback to display off helped (and I tested that lips seem to still move naturally compared to the speech heard) [07:23] Saviq: Anything that's fullscreen in Ubuntu should get "unredirected" (the same thing that we call "bypass"). In Xorg this also means the app has to do wait-for-vsync. In Mir, we have no such archaic requirement :) [07:23] willcooke: cool! it's about time to get that enabled everywhere by default. [07:24] * willcooke invites Mirv to the session ;) [07:24] hey willcooke [07:24] morning didrocks [07:24] didrocks, good ideas coming in over G+ last night [07:24] duflu: I think compiz is probably ok, the regressions happen on x.org server+drivers, libraries when things change. like that SDL saviq originally asked me about. [07:24] willcooke: ok :) [07:25] Mirv: Well it /should/ tear in fullscreen by design :) [07:25] Unless you disable unredirect in CCSM [07:25] willcooke: seeing that! loving it. I want that we discuss that more in detail with stuart live, I guess some parts should be on the base platform and not in udtc itself [07:25] didrocks, excelelnt [07:25] ... or add a blacklist entry. [07:25] * duflu now remembers what he did just before leaving the Compiz project [07:25] willcooke: oh, btw, not sure if you noted, but we start having volonteers and as well a lot of suggestions on the bug tracker: https://github.com/didrocks/ubuntu-developer-tools-center/issues [07:26] volounteers* [07:26] didrocks, awesomesauce :) [07:27] duflu: right, but then the libraries like SDL come into play and handle the syncing [07:29] Mirv: Yeah I call it "archaic" but sometimes useful if used right. I saw a video recently where John Carmack mentioned doing the same thing on the Samsung Oculus project. [07:30] I've been thinking about giving Mir the same archaic "option" [07:35] Although Carmack's hack may only be equivalent to Mir's bypass feature. I don't know Android enough to be sure. [07:45] Mirv, ha, seems to be something in my session, all's nice and smooth in a guest session [07:45] Mirv, I can see the UI being quite slow here, not so in the guest session [07:45] * Saviq needs a new $HOME [07:46] hah, so unity being slow maybe is the same problem [07:51] Saviq: Sanity check: Is the dash still transparent/blurred when slow? [07:52] duflu, it is, but late [07:52] actually no, not any more, seems fine [07:52] Saviq: I think it would be opaque still if in software mode. But software mode would also explain tearing [07:53] Aha! Software mode [07:53] You have no GPU support [07:53] duflu, hmm... glxinfo says haswell :/ [07:53] Saviq: Yes, but they're different processes. Compiz/Unity7 can choose to use software rendering while clients still probe the DRI implementation [07:54] And in fact clients can still be accelerated while Compiz/Unity is software rendering === Sweetsha1k is now known as Sweetshark [07:54] Just not accelerated in composition [07:54] duflu, right, any way to check why would unity do software? and how to check? [07:54] Saviq: ~/.xsession-errors [07:55] Various users have reported spurious fallbacks to software mode. Not all explained [07:55] duflu, hmm I might've just fixed it with dconf reset [07:55] no [07:57] duflu, http://pastebin.ubuntu.com/8325328/ [07:57] nothing interesting in .xsession-errors [07:58] and it seems to have been faster before I launched xbmc [07:58] like now I'm framedropping when switching workspaces [07:58] Saviq: compiz (core) - Info: Unity is fully supported by your hardware. [07:58] mind you, I'm on 1920x1080+1920x1200 [07:59] Saviq: Yes there are a few long outstanding performance bugs with Unity7 on multimonitor, and generally with the unity-shell plugin [07:59] Just make Unity8+Mir work :)( [08:01] duflu, indeed [08:01] duflu, but even if I disable the external monitor, things don't get better :/ [08:02] Saviq: I used to have to spend many hours every day to keep up with Compiz/Unity bugs. Now I'm a few years out of date, I know nothing [08:03] duflu, actually now it did help, might be I need to restart unity to be able to get it in a working state on a single screen [08:03] no tearing [08:03] thanks guys [08:05] gooooooooood morning [08:07] Hey Laney [08:11] yo darkxst [08:11] wie gehts? [08:12] I'm good [08:12] hey Laney [08:12] welcome back Laney! [08:13] hey darkxst [08:13] hey willcooke et didrocks [08:13] hey didrocks [08:13] * didrocks likes the "et" :-) [08:13] I trust the world broke without my input [08:13] Laney: how were your holidays? :) [08:13] Laney: sure, did you plan to take the week-end off? :p [08:13] Laney, it was a close call with the GStreamer stuff :) [08:14] oh right, the gstreamer one… [08:14] the one that was something else? [08:14] holidays were nice! [08:15] Saviq: Oh, also Xorg is incapable (by design?) of syncing to vblank on multiple monitors... [08:15] good sun, good beer, food, seaside [08:15] Not sure if that was ever fixed or fixable [08:15] duflu, yeah, that I knew, I just wasn't able to get no tearing across a single screen [08:16] Saviq: See? Mir is so much better already [08:17] Laney: enjoyed Portland? [08:17] duflu, it is! [08:18] so yeah, seems unity just can't deal with multimonitor on this GPU, either screen is fine, but both leads to degrading performance [08:20] didrocks: it's a fun place [08:20] they have some big cool forested parks [08:20] oh nice, I didn't stay long enough to see those parks [08:20] Saviq: note that I have only single screen attached to that haswell mini-itx pc [08:21] makes things slightly less complicated [08:21] Mirv, yeah indeed [08:21] Mirv, managed to get no tearing after all, thanks [08:22] Saviq: \o/ [08:22] oh, I see [08:22] restart unity after disconnecting external monitor [08:48] Trevinho, after poking my users a little bit, feedback has been very positive, only one has said that they have seen the problem again after installing the packages, but I suspect they hadn't restarted the session. [08:49] Trevinho, unfortunately, some of the users report that compiz crashes for them after unlocking. I have a bunch of crash files that I'll look into later today, but could it be that this change makes it more likely that compiz will crash? [08:52] moin [09:03] hey Sweetshark [09:07] https://bugs.freedesktop.org/show_bug.cgi?id=81861 *sigh* already collecting some fun to look into for Ubuntu 15.04 ... [09:07] Freedesktop bug 81861 in Libreoffice "The "File -> Open" dialog box is broken in LO 4.4 Ubuntu's Unity" [Major,New] [09:07] willcooke: heya [09:09] (OTOH a really good sign that we get bug reports for something to be released in January 2015 in July 2014 -- really shows that we have decent manual test coverage now). [09:09] +1 [09:09] :) [09:10] complain early, complain often ;) === MacSlow is now known as MacSlow|lunch === dbarth is now known as dbarth-afk === MacSlow|lunch is now known as MacSlow === mvo__ is now known as mvo [14:57] chrisccoulson, lulz - Chuckle Brothers :) [14:58] hah :) [14:59] You know how Apple have these ridiculously photogenic people in their screen shots of the Contacts in an iPhone. Well I move that Barry & Paul Chuckle should feature on the Ubuntu ones [15:16] I think we should have the desktop team doing their best smouldering [15:21] Laney, do you have merge rights to the right places? ;) [15:21] My fingers have been known to slip from time to time [15:24] * willcooke . . o o O O ( Evil thoughts ) [15:24] I've sometimes thought about slipping an easter egg in :-) [15:24] there are people who read the upload diffs though, spoilsports [15:26] I remember when the desktop wallpaper got changed [15:26] that was good lulz [15:50] marga: there shouldn't be many changes causing crashes, but we might have lose something (or I might, when cherry picking the changes from trunk, as most of them are already in utopic, and we didn't experience many troubles there). Anyway if you've crash files I can look at, it would be nice (although probably they're with no debug symbols). [15:56] Trevinho, some are very big crash files. [15:56] I'll look into them and tell you if I find anything interesting [15:57] marga: I seel, let me know === dbarth-afk is now known as dbarth [16:21] Oh seb128, where art thou? [16:33] Sweetshark, seb's on holiday, back on Monday [16:40] willcooke_: np, just wondered on how to do the next libreoffice bump. it will work out ;) [17:02] * Laney EOWs with an upload of the utopic wallpaper contest winners [17:10] Laney, you slipped in the extra one? [17:11] * willcooke_ -> EOW. toodles === oCrazyLem is now known as CrazyLemon === yp is now known as ypwong === zoktar_ is now known as zoktar === bdrung_ is now known as bdrung === Laney is now known as Guest14360 === zoktar is now known as 77CAAFYCY === dobey_ is now known as dobey === jpds is now known as 6JTAAFS52 === hggdh is now known as 6JTAAFWBI