[03:28] Good morning [04:55] hi [04:56] Hey hikiko! [04:57] Did you manage to get compiz to use gles? [04:58] hi RAOF ! [04:58] not yet :/ I built unity & nux with gles2 but I can't run them properly [05:01] Huh. It was working perfectly back in 2012. But I forget the details :) [05:01] :D [05:03] fortunately gles2 is a subset of opengl2 so, I can at test it works at least [05:03] Or was that just EGL?.... [05:03] but I'd like to be able to run it [05:04] Oh, yes it was GLES. I remember the VertexBuffer hack and friends [05:05] Sam liked to call it "modern" GL. Even though it's no more modern than full OpenGL, and significantly more limited [05:05] yes that's what I am doing today :) converting fixed function to shaders [05:05] to vertex buffers* [05:05] sorry [05:06] GL_QUADS to triangles... [05:06] :p [05:06] hikiko: Oh I see. We got sloppy somewhere in Compiz and forgot to keep supporting it [05:06] Well, it's much more like GL 3.0 than old GL. [05:06] yes, I wonder why they remove and remove features from opengl though [05:06] You can't use any of the fixed-function support in a 3.0 context. [05:06] Because the hardware that the fixed-function support was providing an interface to no longer exists. [05:07] hikiko: For more limited devices (e.g. mobile) [05:07] duflu: No, actually for more powerful devices :) [05:08] RAOF: "OpenGL for Embedded Systems (OpenGL ES or GLES) is a subset[2] of the OpenGL computer graphics rendering application programming interface (API) for rendering 2D and 3D computer graphics such as those used by video games, typically hardware-accelerated using a graphics processing unit (GPU). It is designed for embedded systems like smartphones, computer tablets, video game consoles and PDAs. " [05:08] maybe some features are difficult to implement for example display lists might need a lot of memory [05:08] duflu: Right, but OpenGL 3.0 *also* loses all the fixed-function stuff, and *that's* for more powerful devices. [05:08] GL 3.0 looks a lot more like GLES 2.0 than GL 2.0 [05:09] true [05:09] hikiko: Yes, phones (particularly old phones) have tiny brains [05:09] Newer ones don't though. They're much more powerful than the desktops of a few years ago [05:10] yes [05:10] hikiko: BTW I think alf wrote VertexBuffer originally [05:19] basically, I won't use it because the shader will only run in GLES2 case (it performs a slower fake clipping in pixel shader because gles doesnt have any way to do clipping in view space) so, I'll use the fixed function everywhere in the plugin and just use a float* varray in gles2 [05:20] I would use the opengl plugin classes if I had to port the whole plugin to GLES2 [05:20] hikiko: GLES does have scissors still. I used it in the GLES port which is why we can still switch workspaces (Ctrl+Alt+left/right) [05:21] scissors is for 2d clipping, I want to fake the clipping of the 3d windows that now is performed in viewing space [05:21] hikiko: Oh yeah. I remember those bugs now (you've been updating recently) [05:21] so what I ve done and works is to calculate in the pixel shader which pixels would correspond to vertices that would be clipped in view space [05:21] and discard them [05:22] the visual effect is the same [05:22] but it's slower [05:22] Cool [05:22] * duflu needs to see the outside world for a little while [05:22] so there's no point to use that trick everywhere only in gles2 devices where there's no alternative :) [05:22] :D [05:22] enjoy your day duflu [05:23] evening * [05:23] hikiko: Thanks. And good luck [05:23] thanks :) [05:24] * RAOF wonders why you need to clip verticies in view space. Can't you just draw them, and if they're outside of view space they'll get discarded? [05:28] RAOF: Because sometimes you want the clipping line to be in the middle of the screen [05:28] Fair chop. [05:29] You are in a QMaze of Qthings, all reimplementing the STL with slightly different names and subtly different behaviour. [06:09] good morning [06:16] good morning desktopers [06:17] re seb128 [06:18] re didrocks [06:23] bonjour mes amis français ! [06:24] bonjour pitti [06:40] hey pitti, wie gehts? [06:40] seb128: viel besser, danke! [06:41] great! [07:51] morning all [07:51] hey willcooke [07:53] hey willcooke [07:53] morning seb128 [08:02] hey hey [08:03] hey Laney [08:03] hey seb128! [08:03] how's it going? [08:04] morning Laney [08:04] Laney, good! you? [08:05] great [08:05] beer festival tomorrow :-) [08:05] hey didrocks! [08:05] what is up homie? [08:06] nothing special ;) yourself? Getting ready for beer festival? [08:06] stretching the liver [08:07] I've got a mini beer festival tomorrow with aquarius \o/ [08:07] in that there is a beer festival on at the local pub, not just going out drinking beer [08:08] http://beerfestival.nottinghamcamra.org/Assets_pdf/BF2015_Beer_Prog_Notes.pdf [08:08] Blue Monkey is good, that's a regular at ours [08:09] ya [08:09] they are near here [08:09] got a pub at canning circus [08:09] oh nice!" [08:10] Laney, there is a Brew Dog pub somewhere around Kings Cross which does food (I'm told) if you fancy it next week? [08:10] pitti: ok, I think I won't go further for now (the real coverage is more like 94% because some code are ran for i386 on another machine but not consolidated): https://jenkins.qa.ubuntu.com/job/udtc-trusty-tests-collect/1950/label=ps-trusty-desktop-amd64-1/artifact/html-coverage/index.html :) [08:10] willcooke: haha [08:10] I'm sure the food is what you're going for ;-) [08:10] time to cut a release! [08:10] natch [08:10] but yeah [08:11] didrocks: wow, that looks awesome! [08:11] I am up for that if true [08:11] didrocks: what's ftp_adapter? [08:11] didrocks, nice! [08:12] pitti: streaming support for python requests module (done by a contributor). I didn't push too much on testing it, but that should be doable (just delayed for now) [08:15] Laney, Ohhhhh... https://www.brewdog.com/bars/uk/two-bit [08:15] popey, ^^^^ [08:16] do they do real food? [08:17] cereal killer is probably nearby... :) [08:17] Laney, it got burned down [08:17] oh, "down" might be an overstatement [08:17] http://www.theguardian.com/uk-news/2015/sep/27/shoreditch-cereal-cafe-targeted-by-anti-gentrification-protesters [08:17] haha [08:17] yeah I saw that [08:19] lol [08:19] https://github.com/auchenberg/volkswagen [08:19] "Volkswagen detects when your tests are being run in a CI server, and makes them pass." [08:19] wouldn't say no to a curry if brewdog doesn't do real food [08:19] complainers can get steak themselves another day :P [08:20] Laney, they say the do a "range" of food. I think a sausage roll and a scotch egg probably don’t count as dinner in your book though. [08:21] liquid dinner? [08:21] There was a decent curry house popey and I went to after msm's leaving do which is walking distance from the hotel. [08:21] seb128, ha, that VW thing is excellent [08:21] :-) [08:21] ahah https://github.com/auchenberg/volkswagen/blob/master/index.js#L18 [08:21] HAHAHAH! [08:22] oh, team dinner is getting booked today. There will be steak [08:22] willcooke, the place at the corner of bluefin is alright, if we don't find anything else [08:22] I like that the guy has tests to check that failing the tests are working :) [08:22] didrocks, hehe [08:23] good ... aehm ... morning?! [08:24] hey larsu -- Maaahlzeit! :-) [08:24] hey larsu [08:24] hihi [08:24] morning larsu [08:25] hi larsu [08:25] too many pink killers yes? [08:25] hey larsu [08:25] willcooke: sausage roll, scotch egg and plenty of liquid bread [08:26] this is fun I should rise late more often :D [08:26] morning pitti didrocks willcooke Laney seb128! [08:26] davmor2, :) [08:26] davmor2, breakfast of champions [08:26] Laney: no pink killer involved [08:26] man my quads feel all tight still [08:26] willcooke: is it the usual "thursday for team dinner"? [08:26] Laney: but maybe (maybe!) the one or other alcoholic beverage [08:26] they were like this yesterday but I assumed it was due to riding after being off the bike for a week [08:26] Laney: wear looser jeans [08:26] but that should be better now? [08:27] didrocks, do we usually have those on thursday? [08:27] * seb128 never noticed [08:27] didrocks, wednesday [08:27] it felt random to me [08:27] we had it on Tuesday before as well [08:27] seb128: I think the vast majority of time (like 80%+ yeah) [08:27] k [08:27] willcooke: noted [08:27] * Laney didn't remember [08:28] well % is falling it seems :p [08:28] you guys, no attention to detail! :p [08:28] quick! Everyone look it up in their expense reports! [08:28] this. is. important. [08:28] http://yoga.about.com/od/anatomicalfocus/tp/Stretch-Your-Quads-With-These-Yoga-Poses.htm [08:36] morning [08:36] hey Trevinho [08:37] :) [08:37] * Trevinho cleanups kernels... sooooo many old packages [08:37] autoremove man [08:38] Laney: I do, but it doesn't always work for everyone [08:38] hey Trevinho [08:38] Trevinho, that's bug #1439769 [08:38] bug 1439769 in update-manager (Ubuntu Vivid) "various linux packages being marked as manually installed, still prevents 'apt-get autoremove' from doing the right thing for kernels" [Critical,Triaged] https://launchpad.net/bugs/1439769 [08:38] slangasek/mvo are on it [08:38] nice [08:40] * Laney nods [08:41] though it seems Steve asked from input from glatzor/aptdaemon maintainers [08:41] but I'm not sure there is anybody active in that group :-/ [08:42] * Laney thinks mvo is as close to one as anyone else :) [08:51] seb128, Laney: yeah, sorry, I pushed the patch but I'm not yet super happy with the api, at the same time I think its good enough and we should just push it out [08:52] mvo, +1, the current situation is a bit ridiculous, filing your disk with old kernels [08:52] so a fix even not perfect is already a good improvement [08:52] mvo, thanks for working on that! === alan_g is now known as alan_g|lunch [11:58] hi desktop [11:58] hey desrt [11:59] * desrt reads the daily intel report, doesn't like what she sees === alan_g|lunch is now known as alan_g [12:02] I was looking at the mouse-not resizing on HiDPI... I've found a solution, but it means to basically reload (recreate) all the cursors when default size changes... Now, I'm wondering whether there's a better way of doing that, since this would probably need some compiz ABI changes otherwise (to do it with less hacks) [12:03] morning desrt [12:03] hello seb128, willcooke [12:03] desrt, what is in there? [12:03] seb128: my area is a bit more green than i would like it to be this morning :) [12:04] lol [12:04] you are addict ;-) [12:04] seb128: since your eyes are already here (well, actually I know you've multiple pairs of eyes :)) do you know anything about that ^? [12:08] Trevinho, not really sorry [12:08] no worries [12:11] Trevinho, https://bugzilla.gnome.org/show_bug.cgi?id=306795 [12:11] Gnome bug 306795 in Backend: X11 "Make XCursor themes instantly applicable" [Normal,Resolved: notgnome] [12:11] can be useful [12:12] seb128: awesome, thanks a lot [12:13] Trevinho, yw [12:16] mh, theme seems to be unrelated to size, actually.. .But still we've the similar issue. As theme changes in gtk, but not in compiz (and root window) [12:22] seb128: oh, who's that guy who sent the patch 10yrs ago? https://bugzilla.gnome.org/show_bug.cgi?id=308104 [12:22] Gnome bug 308104 in [obsolete] settings-daemon "expose cursor theme settings as xsettings" [Normal,Resolved: fixed] [12:22] err, committed the patch [12:22] hehe :-) [12:45] so, yeah there's no real other way than what I found out by myself (set new the default size/theme and reload every cursor)... [12:46] seb128: so... to fix this I need to add the xcursor dependency on compiz (and maybe some ABI changes too), i guess this is for X at this point, right? [12:47] Trevinho, right [12:54] seb128, Laney, have you guys tried cross-compiling gtk+3.0 for armhf? [12:54] nope [12:57] mh, maybe compiz ABI changes aren't necessary as per your fix, but XCursor new dependency yes... [12:57] It's likely that the build-deps will need working on [12:58] ok [12:59] G'morning. [12:59] i'm just wondering how there are armhf builds on launchpad without running into this error [13:00] they are native builds [13:01] hey qengho [13:02] attente: Looks like libicu-dev and gobject-introspection need attention, at least [13:02] for their g++ and build-essential dependencies respectively [13:02] they break cross-build-dep installation === alan_g is now known as alan_g|lunch [13:25] mh, maybe compiz ABI changes aren't necessary as per your fix, but XCursor new dependency yes... [13:35] * Trevinho leaves for an icecream [13:40] morning [13:44] hey andyrock [13:44] attente, no; sorry, I never really used crossbuilding [13:44] xnox and Laney did look at cross building installability in the past and know about it [13:44] but I see that Laney already replied === alan_g|lunch is now known as alan_g [14:02] * Laney meows [14:08] * seb128 pets Laney [14:11] willcooke, good pick for the team dinner ;-) [14:12] name seems familiar... [14:12] seb128, we have to chose from a limited menu because there are so many of us, just waiting for details of that menu to be confirmed. It will have steak on it. [14:12] is that the place where the wine + coke happened? [14:12] I think so [14:12] oh, could be [14:12] wasn't really IIRC [14:12] good* [14:12] the map location matches [14:13] didrocks, what wasnt good, the wine or the cocaine ? [14:13] :P [14:13] I don't remember what I had there [14:13] ogra_: well, some people have horrible ideas about putting coke in the wine [14:13] that was bad… [14:13] eeek [14:13] but the steak and such wasn't really good as well IIRC [14:14] also, it's where attente tried blue cheese for the first time [14:14] I remember to find it expensive for what it was [14:14] but it's London... [14:14] seb128: yep… "fancy place in London" [14:14] its cheaper than pretty much everywhere else I could find within a reasonable distance [14:14] I guess there is a reason for that [14:14] :) [14:14] what Laney said about not the last place because he wanted to try somewhere else? :) [14:17] willcooke, the refinery is better imho [14:17] and not more expensive [14:17] but that's just me [14:17] (from the team dinner we had that the steaks were good iirc) [14:17] that's the one at the corner of the office building [14:22] seb128: btw, I can typical heisenbug for the keybinding issue, I didn't get it once since I enabled upstart bustle [14:22] seb128: I think it might slow down startup, and so, hide the race… [14:22] could be :-/ [14:23] * didrocks removes and upgrades, can't keep on old install forever… [14:24] let's see if the fix works at least ;-) [14:24] yeah ;) [14:25] and better to upgrade today than tomorrow before leaving for the sprint [14:43] * Sweet5hark mumbles ETOOMANYCALLS ... [14:45] Laney, good job fixing that typeahead nautilus bug! do you think that's something worth a trusty SRU? [15:23] larsu, any luck with the gtk/notify-osd/greeter issue? [15:36] seb128: could do, might want to wait a few days to see it doesn't break anything [15:36] k, seems good here from testing [15:36] does somebody fancy looking at a python exception easy to trigger in software-properties? ;-) [15:37] can handle this, might be a nice distraction [15:37] didrocks, bug #1372086 [15:37] bug 1372086 in software-properties (Ubuntu) "add-apt-repository hits softwareproperties.ppa.PPAException when trying to add a ppa which doesn't exist" [High,Confirmed] https://launchpad.net/bugs/1372086 [15:37] basically [15:37] sudo add-apt-repository ppa:user/name [15:38] yep [15:38] thanks [15:38] I'll look first at why it regressed [15:38] and what was the message that we used to display [15:38] I think it's either http://bazaar.launchpad.net/~ubuntu-core-dev/software-properties/main/revision/904 [15:38] or http://bazaar.launchpad.net/~ubuntu-core-dev/software-properties/main/revision/905 [15:38] "make _get_suggested_ppa_message work without launchpadlib (as its not available for py3)" [15:39] ahah, yeah, probably [15:39] will have a look during the sprint at worse, should be easy to fix [15:39] thanks [15:39] yw [16:16] Laney, I would appreciate if you could review some of my unity-settings-daemon 'backport upstream commit' fixes, if you are too busy let me know and I can try to email robert_ancell about it [16:16] (setting up a silo to test/prepare thing, I would like to land that tomorrow ideally) [16:22] * didrocks waves good evening and good night [16:26] seb128: will try to look shortly, not here tomorrow [16:26] oh, right [16:27] Laney, if you can't don't worry, maybe robert_ancell do some or I can try to catch larsu tomorrow [16:27] oh, robert_ancell is off according to calendar [16:33] I broke upstart, might be a bit /o\ [16:34] urg [16:34] what did you do? [16:35] tried to write some code :P [16:36] lol [16:38] was attempting to fix the alt-f2 exec thing [16:39] oh, good one [16:40] xnox did it already [16:40] was just adding some exceptions [16:40] so maybe it isn't my branch that broke it ... [16:40] xnox!!! [16:40] Laney: i don't believe my branch was ever fully working ;-) [16:41] lolz [16:41] Laney: it was more of work in progress stab as to what should happen. [16:41] I knew I shouldn't have tried it on my real system!!!! [16:41] Laney: what? did anyone blidly uploaded it? [16:41] HAHA no [16:41] ok, good [16:41] cause like, my code is not that good by default. [16:42] I would try to get whoever we have that passes for an upstart maintainer to review [16:42] but yeah, doesn't work [16:42] I had "Handling dbus event" spammed many times [16:43] maybe that is normal... [16:43] laney@raleigh> grep -c "Handling dbus event" startups ~ [16:43] 14282 [16:55] Laney: about your blacklisted_environment_variables... does upstart provides a way to put these variables in a place so that they're not hardcoded? [16:56] what kind of place? [16:56] (not atm) [16:56] Laney: I imagine a configuration file, or I don't know.. [16:56] ok [16:56] you'll note that I didn't propose it [16:57] Laney: yeah, sure... I just saw it and I was curious about the new changes :) === alan_g is now known as alan_g|EOD [17:08] Clearly not fixing that this week [17:08] seb128: I got to some of them, a couple are left for the next lucky person [17:08] Laney, thanks! [17:08] see you all Sunday/Monday! [17:08] cya Laney [17:08] have a fun w.e [17:08] see you on sunday [17:08] safe travels [17:27] hmmmm, https://bugzilla.mozilla.org/show_bug.cgi?id=1209932 [17:27] Mozilla bug 1209932 in General "Disable all 32-bit linux testing" [Normal,New] [17:27] That wouldn't be good for us [17:45] * Trevinho has to make pizza... 🍕 [17:46] chrisccoulson, can we discuss next week? Do we have time? === \b is now known as benonsoftware