| robert_ancell | Any idea why glGetStringi(GL_EXTENSIONS, i) is returning NULL but glGetString (GL_EXTENSIONS) is not? | 03:09 |
|---|---|---|
| robert_ancell | This is in silo0 and seems to be why XMir is no longer working. | 03:09 |
| robert_ancell | Have we changed GL version at any point? (on Nexus 4) | 03:10 |
| RAOF | robert_ancell: I don't think so? | 03:13 |
| duflu | robert_ancell: glGetStringi is regular OpenGL, not OpenGLES. You will find it does not exist in ES (https://www.khronos.org/opengles/sdk/docs/man/). But Mesa is lazy and mixes both the APIs, so unsupported functions in ES may still be linkable, even if not functional. | 03:25 |
| robert_ancell | Interesting | 03:26 |
| duflu | Is you're using full OpenGL instead of ES then I don't know. But it's meant to not be present in ES | 03:26 |
| duflu | *If | 03:26 |
| robert_ancell | Looks like XMir is using a full OpenGL context | 03:27 |
| robert_ancell | ah, no, that was XWayland. XMir does do both... | 03:28 |
| duflu | robert_ancell: OK, but even on desktop it will only work with very new GPUs (OpenGL 3.0 and later): https://www.khronos.org/opengles/sdk/docs/man3/html/glGetString.xhtml | 03:28 |
| duflu | So probably don't use that | 03:28 |
| duflu | Ideally | 03:28 |
| robert_ancell | duflu, it's being used in libexpoxy if it detects OpenGL 3.0 | 03:29 |
| robert_ancell | I guess this used to work fine, so not sure what changed | 03:29 |
| duflu | Oops, that's ES 3.0 | 03:29 |
| RAOF | When you say “very new GPUs”, what you probably mean is “everything from at least the last 5 years”. (Mesa support is somewhat newer, obviously ☺) | 03:36 |
| duflu | RAOF: Perhaps, but Ubuntu needs to work on a lot more systems than that | 03:43 |
| duflu | Or else we lose a very large chunk of the user base | 03:43 |
| RAOF | I agree that there's no reason to require 3.0 support for this. | 03:44 |
| duflu | Although I would like to build a shell in future, I don't want Unity to have such high requirements that my shell becomes a necessary fallback for lots of people | 03:45 |
| duflu | We already see that in Unity7 where Compiz supports more GPUs than Unity does :/ | 03:46 |
| robert_ancell | gedit on the phone sure looks weird: http://imgur.com/RdxOB85 | 04:15 |
| RAOF | That looks surprisingly reasonable to me? | 04:16 |
| robert_ancell | Just odd to see it there :) | 04:16 |
| RAOF | With the possible exception of using a crazy-arse default theme in the absence of the usual X atoms, I guess. | 04:16 |
| robert_ancell | Anyone remember who you have to ask to get a PPA to build ARM binaries? | 04:19 |
| RAOF | I believe the vanguard on #launchpad can hook you up. | 04:19 |
| robert_ancell | RAOF, on the freenode channel? | 04:20 |
| RAOF | Yah. | 04:20 |
| RAOF | You could probably also prod on Canonical IRC, but I don't think it's necessary. | 04:21 |
| === c74d3 is now known as _8680_ | ||
| === _8680_ is now known as c74d | ||
| duflu | robert_ancell: Yeah if you could ask the GTK guys what we have to set to get the default Ubuntu theme then GTK apps on Mir would look nicer | 04:46 |
| duflu | I suspect the packages are not installed on phone for starters :) | 04:46 |
| duflu | light-themes | 04:46 |
| robert_ancell | duflu, We should be able to hard-code that in to GTK+ mir | 04:46 |
| robert_ancell | not on this phone anyway | 04:47 |
| duflu | OTOH we could just declare we're copying the mistakes of iOS7 design, because the rest of the world did already | 04:47 |
| duflu | (thinner fonts in GTK without the Ubuntu theme) | 04:48 |
| duflu | robert_ancell: Also let me know if you can confirm any of these? https://bugs.launchpad.net/ubuntu/+source/gtk+3.0/+bugs?field.tag=snappyrdp | 05:04 |
| * duflu goes to get sunshine for a bit | 05:04 | |
| robert_ancell | duflu, I'm not working on GTK+ Mir at the moment, you probably want to poke attente | 05:07 |
| * robert_ancell EOD | 05:07 | |
| === 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 | ||
| duflu | I have a hunch that some simple sleep calls might make the phone smoother | 07:59 |
| duflu | Depending on how long Unity8 holds buffers for compositing | 07:59 |
| === mr2515_ is now known as sturmflut-work | ||
| duflu | greyback__: It occurs to me Mir should have a report or something for how long compositor buffers are held. Do you have any idea for Unity8's performance there? | 09:41 |
| duflu | (should have a report but doesn't) | 09:41 |
| greyback__ | duflu: such a report would be of use for me (and other mir-based compositors) | 09:42 |
| duflu | greyback__: Yeah in the past I've just used temporary logging to get an idea | 09:42 |
| * duflu just discovered today that bypass/overlays actually hold each buffer for almost 32ms (which is surprisingly correct) | 09:42 | |
| greyback__ | I don't think unity8's code manages buffers in the most optimal way currently. I've prototyped better solutions, but numbers always help | 09:43 |
| greyback__ | numbers which I can compare with Mir would be ideal | 09:43 |
| === greyback__ is now known as greyback | ||
| === dandrader is now known as dandrader|afk | ||
| === dandrader|afk is now known as dandrader | ||
| === chihchun is now known as chihchun_afk | ||
| === dandrader is now known as dandrader|afk | ||
| === dandrader|afk is now known as dandrader | ||
| racarr | Morning | 14:49 |
| camako | o/ | 14:50 |
| seb128 | hey racarr | 14:51 |
| seb128 | racarr, what's the status of being able to change the physical keyboard layout under Mir/unity8 (setxkbmap style)? In Brussels you said it should be easy and you could do it in the next weeks | 14:53 |
| racarr | seb128: There is server side API in Mir now so | 14:54 |
| racarr | unity8 could implement per surface keymap changes | 14:54 |
| seb128 | oh, nice | 14:54 |
| seb128 | or as a first step per session one | 14:55 |
| seb128 | so it's one for the unity8 team? | 14:55 |
| alan_g | greyback: I'm trying to build qtmir on a tablet and getting some undefined references I don't see when cross compiling. Like "/home/phablet/spike-using-WindowManager/src/modules/Unity/Application/mirbuffersgtexture.cpp:56: undefined reference to `glDeleteTextures'. Is that something you recognise? | 14:55 |
| racarr | seb128: afaik | 14:55 |
| racarr | alan_g: QTMIR_USE_GLES or | 14:55 |
| racarr | alan_g: just a sec | 14:55 |
| seb128 | mzanetti, greyback, ^ do you know? | 14:56 |
| racarr | alan_g: cmake -DUSE_OPENGLES=1 | 14:56 |
| greyback | alan_g: yeah, see the switch used in the debian/rules file | 14:56 |
| === sturmflut-work is now known as sturmflut2 | ||
| racarr | alf_: alan_g: https://code.launchpad.net/~mir-team/mir/pluck-low-hanging-event-fruit/+merge/258213 maybe you guys can update review comments? thinking I could start landing the pipe... | 15:42 |
| racarr | not sure if we want to get the client ABI branch up first... | 15:42 |
| alf_ | racarr: sure | 15:43 |
| * alan_g thinks RAOF and duflu will have opinions about breaking ABI | 15:43 | |
| racarr | thanks :) | 15:47 |
| racarr | I really like listening to the UOS summits in the background while I code... | 15:47 |
| racarr | the smooth sounds of willcooke listing default applications | 15:53 |
| === dandrader is now known as dandrader|lunch | ||
| === alan_g is now known as alan_g|EOD | ||
| === dandrader|lunch is now known as dandrader | ||
| racarr | RAOF: http://bbcamerica.tumblr.com/post/118281271151/announcing-jonathan-strange-mr-norrell I still havent gotten to reading it yet :( | 18:11 |
| racarr | stuck in this weird book 2666 atm lol | 18:11 |
| === ljp is now known as lpotter | ||
| racarr | poor jenkins seems pretty busy... | 18:17 |
| RAOF | racarr: Shall I bring my copy to Dallas for you? :) | 23:25 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!