RAOF | olli: You should have an email now. | 00:18 |
---|---|---|
RAOF | Hm. How do I go about getting another mir-team PPA? I'm not an administrator on the team, and launchpad doesn't seem to list any administrators either. | 02:30 |
duflu | RAOF: Maybe robert_ancell has access (via https://launchpad.net/~pspmteam/+members#active) | 02:50 |
robert_ancell | RAOF, sure, what name do you want? | 02:50 |
RAOF | robert_ancell: Hows about “system | 02:51 |
RAOF | robert_ancell: Hows about “system-compositor-testing” | 02:51 |
robert_ancell | RAOF, https://launchpad.net/~mir-team/+archive/system-compositor-testing | 02:51 |
RAOF | robert_ancell: Something to do the weeklyish releasesish into. | 02:51 |
RAOF | robert_ancell: Rock on. Thanks. | 02:51 |
RAOF | robert_ancell: Hm. What is ‘lightdm-set-defaults --remove’ supposed to do? | 05:23 |
robert_ancell | RAOF, not sure, I think didrocks did that. I tried it and it seems to set the value to "" which is useless | 05:24 |
RAOF | robert_ancell: Because ‘lightdm-set-defaults --remove --type=unity’ sets ‘type=unity’ if it was not already there, and sets ‘type=’ otherwise. | 05:24 |
robert_ancell | I would change it but I'm not sure if anyone is relying in it in the current form | 05:24 |
RAOF | They can't be, because ‘foo=’ is a syntax error. | 05:25 |
RAOF | If using it breaks your lightdm.conf, then no one can be using it in anger :) | 05:25 |
robert_ancell | Yeah, I figure so | 05:26 |
robert_ancell | 'foo=' is foo="" right? It's syntatically valid, though I can't think off hand of any useful lightdm keys that are useful with empty strings | 05:26 |
RAOF | Hm. At least if you set “type=” then lightdm fails to start. | 05:28 |
robert_ancell | RAOF, yeah, because it looks for the "" type and there is none. It would probably make sense for the daemon to ignore the "" value | 05:32 |
robert_ancell | It's the same behaviour as type=lksdhflkajsf | 05:32 |
RAOF | I guess I could do the "" is treated as default thing. | 05:33 |
RAOF | Oh, also, are you currently in the process of removing the hardware cursor from unity-system-compositor? | 05:33 |
RAOF | Or have any current plans to? | 05:33 |
robert_ancell | RAOF, no, but that sounds like something you'd like to do :) | 05:34 |
RAOF | Yes indeed :) | 05:34 |
duflu | RAOF: What would be a nicer alternative? mir_set_cursor_image() ? | 05:45 |
duflu | Obviously we should use a hardware cursor if we can | 05:45 |
didrocks | RAOF: robert_ancell: it was supposed to remove the default if the corresponding type was set | 05:49 |
robert_ancell | didrocks, right, so it should delete the key right? Not set it to black? | 05:52 |
robert_ancell | blank | 05:52 |
didrocks | robert_ancell: indeed, it was doing that at the time :) | 05:52 |
RAOF | duflu: I could go a mir_set_cursor_image. | 05:53 |
duflu | RAOF: I was thinking about it yesterday but told myself to stop daydreaming about features we don't yet need | 05:54 |
duflu | Now we need it?... :) | 05:54 |
RAOF | Well, "need" is obviously a movable feast :) | 05:54 |
RAOF | I *could* just turn it off and use the X software cursor. | 05:54 |
RAOF | But I would use it right now, if it happened to exist. | 05:55 |
alf_ | RAOF: can't hear half of what you are saying | 06:57 |
RAOF | hikiko: https://code.launchpad.net/~raof/mir/separate-graphics-buffer-and-display is the branch. | 07:14 |
alan_g | hikiko: Is the dead? It's a week since you changed it: https://code.launchpad.net/~hikiko/mir/mir.fix-virt-destructorS/+merge/167705 | 09:28 |
hikiko | oh you are right alan_g | 09:29 |
hikiko | I was looking at the other task | 09:29 |
hikiko | sorry | 09:29 |
hikiko | I will clean it up now | 09:29 |
alan_g | hikiko: That wasn't an order. (I just don't want dead MPs sitting in the list forever.) | 09:31 |
hikiko | no it's a good idea to do it now since I have many questions on the other task | 09:32 |
hikiko | and I emailed the list | 09:32 |
alan_g | I saw the email - but don't know most of the answers | 09:32 |
hikiko | yes, it's not an easy task, so better to finish this branch today :) | 09:33 |
alan_g | (But alf_ does) | 09:33 |
hikiko | thanks for the reminder! | 09:33 |
hikiko | yes, alf_ suggested to start the discussion in the mailing list where anyone can make suggestions | 09:35 |
hikiko | because we discussed some things already in the meeting but still there are some parts that are not very clear | 09:36 |
hikiko | anyway :) I ll fix the destructors first that is quite simpler :) | 09:36 |
alf_ | alan_g: I don't have all the answers either... I do have a lot of suggestions, though ;) | 09:37 |
alan_g | will it help if I read the questions and suggestions and then we jump on hangout? | 09:38 |
hikiko | sure alan_g thanks a lot, I will do the destr in the meantime | 09:39 |
* alan_g first needs to wade through swapper-factory-interfaces and see what gardener is doing. | 09:40 | |
hikiko | hahaha yes and I'll have a quick lunch in a while :) we can do it later | 09:41 |
hikiko | take your time :) | 09:41 |
=== alan_g is now known as alan_g|afk | ||
=== alan_g|afk is now known as alan_g | ||
alan_g | hikiko: I'll be leaving for lunch in 20min - hangout when I get back? Or do it now? | 11:26 |
hikiko | alan_g, enjoy your lunch | 11:34 |
hikiko | and we do it afterwards | 11:34 |
hikiko | :) | 11:34 |
=== alan_g is now known as alan_g|lunch | ||
hikiko | is there any bazaar command that can give you the differences between 2 different branches? (like an extended bzr diff) | 11:55 |
alf_ | hikiko: bzr diff --old/--new | 12:04 |
hikiko | alf_, does this work for too completely different branches as well? eg mir.foo mir.bar ? | 12:13 |
hikiko | two* | 12:14 |
alf_ | hikiko: I haven't tried (but possibly). In the worst case just cd into one of the branches. | 12:14 |
hikiko | hmm it seems that old-new works with one file :s | 12:21 |
hikiko | I'd like to have a diff like what we get when we propose for merge | 12:22 |
hikiko | with the differences between the 2 branches | 12:22 |
hikiko | is this possible? | 12:22 |
alf_ | hikiko: bzr diff --old/--new works fine for me on the whole tree | 12:33 |
hikiko | i did bzr diff --old mir.foo --new mir.bar | 12:35 |
hikiko | and see no differences | 12:35 |
hikiko | :S | 12:35 |
hikiko | lol let me try on trunk just in case that foo and bar have 0 differences :p | 12:35 |
hikiko | cool... it works for the trunk alf_ :p | 12:41 |
hikiko | <--- feels stupid | 12:41 |
hikiko | it was that there were no differences :p | 12:41 |
=== alan_g|lunch is now known as alan_g | ||
hikiko | alan_g, I fixed the destructors | 13:16 |
hikiko | when you aren't busy do you want to do the hangout? | 13:16 |
* alan_g wishes there were not so many MPs to review | 13:17 | |
hikiko | sorry :( do you want me to remove you from the reviewers? | 13:18 |
hikiko | I have mir-team anyway | 13:18 |
hikiko | (I tried to keep it much much simpler this time alan_g very few fixes outside src) | 13:19 |
alan_g | hikiko: this looks like somewhere you could have used "default": | 13:20 |
alan_g | 139- ~Display() = default; | 13:20 |
alan_g | 140+ virtual ~Display() {} | 13:20 |
alan_g | did you try that? | 13:20 |
hikiko | I removed the default from there | 13:20 |
hikiko | because of the nothrow issue | 13:20 |
hikiko | if I add default I'll have to create an empty destructor in every child | 13:21 |
hikiko | like that: virtual ~ChildDisplay() noexcept {} | 13:22 |
hikiko | it's http://stackoverflow.com/questions/11497252/default-destructor-nothrow this problem :/ | 13:23 |
alan_g | Wouldn't the existing code hit the same snag if that | 13:23 |
alan_g | Wouldn't the existing code hit the same snag? | 13:23 |
hikiko | it didn't hit it because the destructors weren't virtual | 13:23 |
hikiko | looser throw specifier for 'virtual Y::~Y()' | 13:24 |
hikiko | error: overriding 'virtual X::~X() noexcept(true)' | 13:24 |
hikiko | that's the error I was getting | 13:24 |
hikiko | I can retry though to get 100% sure | 13:24 |
hikiko | but that was the reason I removed the default | 13:24 |
alan_g | ok - I'm still not entirely sure I understand when the problem exists. | 13:25 |
hikiko | alan_g, see comment 15 in the link above | 13:25 |
hikiko | it's a compiler specific issue | 13:26 |
hikiko | I was using clang when I first changed that branch and I was not getting any errors | 13:27 |
hikiko | then I saw that jenkins fails with that looser throw specifier error | 13:27 |
hikiko | and switched to gcc and got it as well | 13:28 |
hikiko | alan_g, also another question (sorry!) I saw your email, why do you suggest that we have a graphics static library and we load it at runtime? I understand what you suggest but I didn't understand very well which problem it solves :S | 13:42 |
alan_g | s/static/shared/ and it makes more sense. | 13:43 |
alan_g | hangout time? hikiko? alf_ ? | 13:44 |
alf_ | alan_g: sure | 13:44 |
hikiko | sorry shared | 13:45 |
hikiko | alan_g, I am in | 13:45 |
alan_g | https://plus.google.com/hangouts/_/b984b9ad51b94388fdf039db9aa6e0aa01d46d90?hl=en-GB | 13:45 |
=== alan_g is now known as alan_g|tea | ||
kdub | hello all | 15:08 |
alf_ | kdub: Hi! | 15:11 |
=== mzanetti is now known as mzanetti|food | ||
=== alan_g|tea is now known as alan_g | ||
alan_g | hello one | 15:15 |
greyback | racarr: awake yet? | 15:29 |
=== mzanetti|food is now known as mzanetti | ||
=== mmrazik is now known as mmrazik|afk | ||
tvoss | alf_, ping | 15:32 |
alf_ | tvoss: pong | 15:32 |
tvoss | alf_, can you give me the link for the instructions for installing the mir staging ppa, please? | 15:33 |
alf_ | tvoss: http://unity.ubuntu.com/mir/installing_prebuilt_on_pc.html | 15:34 |
alf_ | tvoss: http://unity.ubuntu.com/mir/installing_prebuilt_on_android.html | 15:35 |
kdub | alf_, are you ok with swapper-factory-interfaces after updating the members/functions to match the classes, and renaming shutdown to force_client_completion? | 15:57 |
alf_ | kdub: alan_g: Is https://code.launchpad.net/~alan-griffiths/mir/fix-1189443/+merge/168639 still relevant if are going to revert to the old way of unblocking the client (for non-swapper-switching cases)? | 16:26 |
alan_g | alf_: no - sorry should have WPI'd it | 16:27 |
alf_ | alan_g: ok | 16:28 |
olli_ | guys... if I want to look at performance on native x vs xmir... what would you advice I use to measure | 16:48 |
olli_ | tvoss, kdub alan_g ^ | 16:49 |
olli_ | RAOF, racarr ^ | 16:49 |
alan_g | olli_: what performance matters to you? | 16:50 |
olli_ | that is a good question | 16:50 |
olli_ | I think atm something like glxgears/fps would keep me happy | 16:51 |
olli_ | but then I know there is more to look at | 16:51 |
olli_ | so you tell me | 16:51 |
kdub | glmark is probably the best choice, works on mir, surfaceflinger, x | 16:51 |
olli_ | I want to get a quick feel for whether xmir has any impact | 16:51 |
kdub | plus alf's an in-team expert on the suite :) | 16:52 |
olli_ | do we have benchmark results for the 3 DS? | 16:52 |
olli_ | yeah, just missed him by 8min | 16:52 |
racarr | Hi all sorry...here now. | 17:01 |
racarr | Hit my head hard pretty hard yesterday and took extra rest | 17:01 |
alan_g | racarr: time for me to leave | 17:01 |
alan_g | Ouch! | 17:01 |
=== olli_ is now known as olli | ||
racarr | alan_g: I'm sure it will stimulate the regrowth of neural connections | 17:02 |
racarr | ...:) | 17:02 |
* alan_g resists temptation | 17:02 | |
=== alan_g is now known as alan_g|life | ||
racarr | See you :) | 17:03 |
racarr | kdub: Yesterday at EOD ricmm was running in to some issues with mir on flipped container android | 17:03 |
racarr | did you hear about these? | 17:03 |
kdub | racarr, yeah | 17:03 |
kdub | well, at least with the issue that we don't use eglGetProcAddress() for eglCreateImageKHR and the like | 17:04 |
racarr | Oh there is a new one | 17:04 |
racarr | /home/phablet/mir/build/src/server/graphics/android/default_framebuffer_factory.cpp(80): Throw in function virtual std::shared_ptr<mir::graphics::android::DisplaySupportProvider> mir::graphics::android::DefaultFramebufferFactory::create_fb_device() const | 17:04 |
racarr | 23:06 <ricmm> Dynamic exception type: boost::exception_detail::clone_impl<boost::exception_detail::error_info_injector<std::runtime_error> > | 17:04 |
racarr | 23:06 <ricmm> std::exception::what: display factory cannot create fb display | 17:04 |
racarr | unless it's solveable by immediate magic I will figure out how to get flipped container on my phone | 17:05 |
ogra_ | wait for the currently running build to finish (max 1h) ... it has a bunch of udev fixed | 17:05 |
ogra_ | *fixes | 17:05 |
racarr | kdub: Is udev involved? I don't actually know where the graphics devices live in android world... | 17:06 |
ogra_ | well, if you use devices on the ubuntu side it surely is involved (since it sets all permissions there) | 17:07 |
kdub | seems a different error | 17:07 |
ogra_ | k | 17:07 |
kdub | the udev structure should be 'normal'... what i'd suspect is the filesystem positioning of the graphics libraries | 17:07 |
racarr | Mm, that was my first guess but I dont know what is supposed to be where | 17:07 |
ogra_ | well. up to today ueventd (androids "udev") and udev were fighting for the device ownership | 17:08 |
ogra_ | android GL libs usually live somewhere in /vendor/lib | 17:08 |
racarr | libgralloc_* or whatever is what we must not be ablet o get here | 17:09 |
ogra_ | (/vendor being a link or bindmount to /system/vendor) | 17:09 |
racarr | I guess first it would be best to check if hw_get_module fails or framebuffer_open | 17:09 |
racarr | kdub: It is framebuffer_open not hw_get_module | 17:22 |
racarr | and we are pretty sure surface flinger is dead :( | 17:22 |
ogra_ | how did you kill it ... there is a certain process on the flipped images to change android service startup | 17:23 |
kdub | android graphics libraries live in /system/lib/hw, /system/lib/egl, /system/lib | 17:23 |
kdub | or /vendor/lib/hw /vendor/lib /vendor/lib/egl | 17:23 |
kdub | depending on device | 17:23 |
racarr | ogra_: Moved the binary and killed it | 17:23 |
kdub | racarr, nexus 4? | 17:24 |
racarr | kdub: ricmm is testing on galaxy actually, I am setting it up on my 4 now. | 17:24 |
racarr | flipped container jubilee | 17:24 |
racarr | ...sorry lingering effects of the head bang. | 17:24 |
ogra_ | racarr, cp /var/lib/lxc/android/rootfs/init.rc /var/lib/lxc/android/overrides/ | 17:25 |
ogra_ | then edit init.rc to not contain any surfaceflinger blocks | 17:25 |
ogra_ | (in the overrides dir) | 17:25 |
racarr | ogra_: Ok. We ust have a new theory though and maybe udev is it XD | 17:25 |
ogra_ | that way it will not be started on boot | 17:25 |
racarr | on raring, where its working I have /dev/graphics/fb* | 17:26 |
racarr | flipped container has broken symlinks there atm? | 17:26 |
racarr | I dunno | 17:26 |
racarr | ricardo is testing something it seems | 17:26 |
ogra_ | it shouldnt if your container started properly | 17:26 |
ogra_ | see the lxc-android-config upstart job ... it creates the links | 17:27 |
racarr | XD. ok apparently the symlinks are fine | 17:27 |
racarr | HMMMMMMMMMMM | 17:28 |
ogra_ | ah, good | 17:28 |
racarr | ok guys, well we had a good run | 17:28 |
racarr | but i guess it's time to give up now | 17:28 |
racarr | pack up head home | 17:28 |
ogra_ | but there are other files in /dev you might need | 17:28 |
racarr | :p | 17:28 |
ogra_ | and not all of them have udev rules yet | 17:28 |
racarr | the failing code path is pretty small, hw_get_module works (i.e. it loads gralloc), framebuffer_open fails | 17:29 |
ogra_ | http://cdimage.ubuntu.com/ubuntu-touch/daily-preinstalled/current/ has the most recent flipped image btw ... with a lot less races | 17:29 |
kdub | racarr, fire up the integration tests on the device | 17:30 |
kdub | they will drive some tests that will test gralloc-only operations | 17:30 |
kdub | with some gralloc+openGL tests too | 17:30 |
racarr | Ohh good idea | 17:30 |
racarr | ogra_: Flashing the latest ones now :) | 17:31 |
ogra_ | good, tell me how it goes ... they should boot you into the shell and apps should run properly now | 17:31 |
ogra_ | if i know they are usable i will announce them to the ubuntu-phone list as experimental developer images :) | 17:32 |
* ogra_ is still downloading here ... slow DSL | 17:32 | |
racarr | :) | 17:35 |
racarr | flasssshin | 17:35 |
* kdub is curious too, will give it a try | 17:39 | |
racarr | ogra_: Boot to black screen :( | 17:42 |
ogra_ | what device ? | 17:42 |
racarr | nexus 4 | 17:42 |
racarr | adb shell works, but still drops me in android and there is ubuntu_chroot | 17:42 |
racarr | which I wasn't expecting | 17:43 |
racarr | is that right? | 17:43 |
ogra_ | maguro is the only one thats actually known to work | 17:43 |
ogra_ | no, that isnt right | 17:43 |
racarr | :) That didn't sound very right | 17:43 |
ogra_ | how did you flash it ? | 17:43 |
racarr | http://cdimage.ubuntu.com/ubuntu-touch/daily-preinstalled/current/ | 17:43 |
ogra_ | right, and then ? | 17:43 |
racarr | downloaded armel+mako.zip | 17:43 |
racarr | and touch-armhf.zip | 17:44 |
racarr | puhed to sdcard | 17:44 |
ogra_ | ok | 17:44 |
racarr | intalled from recovery | 17:44 |
racarr | do I need the bootimg file? | 17:44 |
ogra_ | no | 17:44 |
ogra_ | but the order is essential | 17:44 |
ogra_ | armel+$subarch needs to go first | 17:44 |
racarr | ah I did the opposite and I was doing that I was like | 17:44 |
racarr | ...hmm I wonder | 17:44 |
racarr | ...ok | 17:44 |
ogra_ | as soon as i'm done testing here i'll mail out instructions | 17:45 |
racarr | flashing...2.0! | 17:46 |
ogra_ | (if that rsync ever finishes that is) | 17:46 |
ricmm | kdub: E/IMGSRV ( 7010): :0: PVRSRVCreateDCSwapChain: Error - 10 returned | 17:50 |
ricmm | E/IMGSRV ( 7010): :0: framebuffer_device_open: Failed to create flip chain; retrying | 17:50 |
ogra_ | did it boot to the shell before you killed SF ? | 17:50 |
racarr | ogra_: Ok I sitll get boot to black but adb-shell | 17:50 |
ricmm | yes | 17:50 |
racarr | is container flipped now :) | 17:50 |
racarr | oh | 17:50 |
racarr | hey heres the shell | 17:51 |
racarr | nice | 17:51 |
ogra_ | yay | 17:51 |
racarr | fuck only 69 minutes | 17:51 |
ogra_ | thanks | 17:51 |
racarr | of talk time left | 17:51 |
ricmm | ogra_: yea, it booted to shell before I killed SF | 17:51 |
racarr | nice folding launcher :) | 17:51 |
racarr | unortunately at first glance the network indicator doesn't seem to be working | 17:52 |
ogra_ | hmm, it should | 17:52 |
ogra_ | its the same as on unflipped saucy now ... | 17:52 |
ogra_ | try phablet-network-setup if you cant get it to work | 17:52 |
racarr | ogra_: Worked on reboot | 17:54 |
ogra_ | ah, good | 17:55 |
ricmm | ogra_: any idea on my weird GL error? | 17:55 |
ogra_ | ricmm, not really, no | 17:55 |
racarr | its before GL | 17:55 |
ogra_ | looks like a permission thing though | 17:55 |
ogra_ | but i'm only wildly guessing | 17:55 |
racarr | a permission that root wouldnt have? | 17:56 |
kdub | it could be any misuse or mis-setup of the driver | 17:56 |
ogra_ | yeah | 17:56 |
kdub | it still is the android kernel, right? o.O | 17:57 |
ogra_ | yes | 17:57 |
ogra_ | same kernerl as in raring | 17:57 |
racarr | ok installing all sorts of boost headers and such on my phone, back in 10 | 17:58 |
racarr | fascinating, someone else has produced | 18:18 |
racarr | the same error http://forum.imgtec.com/discussion/2611/pvrsrvcreatedcswapchain-error-on-android | 18:18 |
racarr | (im still installing build dependencies at 41.9 kbs) | 18:18 |
kdub | an my n4, i see a firmware problem | 18:19 |
gotwig | hey there | 18:19 |
racarr | kdub: Firmware problem? | 18:20 |
racarr | Hello :) | 18:20 |
gotwig | Is Mir going to be hold as an independend software project, independend from Unity/ubuntu? | 18:20 |
gotwig | so other software projects could reuse it | 18:20 |
kdub | gotwig, our primary target is to support unity, but if other projects take an interest i don't think anyone would object :) | 18:21 |
tvoss | gotwig, rephrasing it a bit: mir does not assume Unity to be there, but it caters towards its goals | 18:22 |
kdub | racarr, i see that the driver is complaining about not finding the gpu firmware | 18:24 |
racarr | kdub: Oh interesting...where should it bet? | 18:24 |
tvoss | gotwig, does that help? | 18:24 |
racarr | system/vendor/firmware? | 18:24 |
racarr | ogra_: till trying to understand this flipped container... | 18:25 |
racarr | is there a notion of chrooting in to android now? | 18:25 |
gotwig | thanks | 18:26 |
racarr | Apparently rebooting the device after killing sf fixed it on galaxy nexus :) | 18:26 |
gotwig | right now I only see baby basic stuff in Mir | 18:26 |
gotwig | is this normal :D? | 18:26 |
kdub | no one believes me that surfaceflinger is evil | 18:26 |
gotwig | nothing usable :/ | 18:26 |
ogra_ | racarr, i ahip a script for that, yeah ... | 18:26 |
ogra_ | racarr, android-chroot | 18:26 |
gotwig | what can you do with mir right now | 18:27 |
gotwig | I just tested the client demos | 18:27 |
kdub | gotwig, there are qt/sdl branches and stuff floating around | 18:28 |
kdub | but you have to be motivated to track them down if you're interested :) | 18:28 |
racarr | Mm. | 18:28 |
gotwig | how can I get this? https://www.youtube.com/watch?v=E9AzRxsnfTE | 18:29 |
racarr | I mean you can run xmir, qt, sdl, etc. | 18:29 |
gotwig | so, qt is the future for ubuntu? no GTK at all? | 18:29 |
gotwig | because I am/was a GTK developer (haha -..-) | 18:29 |
tvoss | gotwig, gtk will still be around, but unity8 is qt/qml and so is the ubuntu sdk | 18:30 |
racarr | Well, we are focused on Qt (in particular QML yeah) | 18:30 |
gotwig | tvoss, only? | 18:30 |
racarr | but no one is planning to rewrite libreoffice in QML either, so there will be some GTK around for a while ;) | 18:30 |
tvoss | gotwig, only as in ubuntu sdk or unity8? | 18:30 |
gotwig | this new Mir thing | 18:31 |
gotwig | Unity Next (?) | 18:31 |
gotwig | can you tell me how to get a setup like in the video I posted plase | 18:31 |
gotwig | *please | 18:31 |
gotwig | he says he is using mir, but I dont believe it | 18:31 |
tvoss | gotwig, well, I'm pretty sure that it's unity8 on mir :) | 18:32 |
tvoss | gotwig, you should talk to greyback_ if he is around | 18:32 |
racarr | gotwig: Sorry XD, it's not an easy question to answer because a lot is currently in flux for those | 18:32 |
racarr | gotwig: If you were really intent on it, you could build mir, then build ~robertcarr/platform-api/mir2 with | 18:33 |
tvoss | gotwig, or mzanetti | 18:33 |
racarr | ENABLE_MIRSERVER_IMPLEMENTATION=true ENABLE_HYBRIS_IMPLEMENTATION=false | 18:33 |
* gotwig feels ashamed for being a Ubuntu Member and stopping Ubuntu software development | 18:33 | |
tvoss | gotwig, why stopping ubuntu software development? | 18:33 |
gotwig | tvoss, well... I found elementary | 18:33 |
gotwig | maybe Unity Next brings me back | 18:33 |
gotwig | Once I did some Unity Scopes/Lenses, it was fun... | 18:34 |
greyback_ | gotwig: I hope within a day or two to have a PPA with packages ready so you can easily run unity+mir on your machine. | 18:34 |
racarr | then build ~robertcarr/qtubuntu/mir2 from src/platforms with | 18:34 |
gotwig | greyback_, :3 | 18:34 |
racarr | qmake "CONFIG += mirserver" | 18:34 |
gotwig | racarr, so I have to recompile mir, to get unity 8 running? | 18:34 |
racarr | Then! you can build ~robertcarr/unity/phablet-wednesday | 18:34 |
racarr | with ./build -s | 18:34 |
racarr | and ./run -i -m | 18:35 |
racarr | and it should work XD | 18:35 |
racarr | or you can wait on the PPA | 18:35 |
racarr | gotwig: No Mir from the ppa should work | 18:35 |
gotwig | racarr, can I do this all from TTY or a X session | 18:35 |
racarr | you can do it all from an X session | 18:35 |
gotwig | racarr, oh, these steps are for unity 8 ? | 18:35 |
racarr | Yes for something like the video you linked to | 18:35 |
gotwig | racarr, thank you very much | 18:35 |
racarr | :) | 18:35 |
gotwig | fast and clean support :O | 18:36 |
racarr | gotwig: No worries :) | 18:36 |
racarr | you should probably actually | 18:36 |
greyback_ | lp:~robertcarr/unity/phablet-tuesday | 18:36 |
racarr | when building platform, api and qtubuntu | 18:36 |
gotwig | greyback_, ? | 18:36 |
racarr | gotwig: Err, greyback is correct on the branch | 18:36 |
greyback_ | phablet-wednesday doesn't exist (yet) | 18:36 |
racarr | haha | 18:36 |
gotwig | why do you create branches for every day lol | 18:36 |
racarr | gotwig: Also probably or platform api also do ENABLE_MIRCLIENT_IMPLEMENTATION=true and | 18:36 |
gotwig | or do you do commits every week for every day xD | 18:36 |
greyback_ | gotwig: to match his underwear :D | 18:36 |
racarr | CONFIG += mirserver mirclient for | 18:37 |
racarr | qtubuntu | 18:37 |
racarr | and you can test qt client apps too | 18:37 |
racarr | gotwig: Just because like I said things are ina lot of flux this week haha | 18:37 |
gotwig | When I want to develop something for unity next | 18:37 |
gotwig | do I have to mess with QT? | 18:37 |
gotwig | I liked all this abstraction stuff from Unity, with the dbus interface... | 18:37 |
gotwig | do I need to add the PPA's for build dependencies? | 18:38 |
tvoss | gotwig, the primary languages are qml/qt/c++, for scopes and friends, things are a little different | 18:38 |
tvoss | gotwig, most likely python support and such | 18:39 |
gotwig | what I hate about ubuntu stuff is this extensive use of python -..- its so slow | 18:39 |
tvoss | gotwig, that's going away mostly | 18:40 |
* greyback_ on his bike home | 18:40 | |
tvoss | gotwig, we are actively working towards getting python out of the picture for the ubuntu touch | 18:40 |
tvoss | greyback_, don't bike and irc :) | 18:40 |
gotwig | damn that is sexy https://www.youtube.com/watch?v=oaT5JsZQQyU | 18:40 |
gotwig | so guys... | 18:43 |
gotwig | I got the branch, what now? | 18:43 |
kdub | racarr, my container flip is stuck in a reboot-loop, but does yours have /system/etc/firmware/a300_pm4.fw ? | 18:43 |
gotwig | (I already did run ./build -s | 18:43 |
racarr | kdub: Yeah | 18:44 |
racarr | gotwig: Err, did you already do the first | 18:44 |
racarr | two bits i.e. build and install platform API and build and install | 18:44 |
racarr | qtubuntu? | 18:44 |
gotwig | racarr, I dont know :O I just did build -s | 18:45 |
gotwig | racarr, for the mir definition stuff, variables are enough? | 18:45 |
gotwig | qmake: could not find a Qt installation of '' | 18:47 |
gotwig | please help, this is so awesome xD | 18:48 |
racarr | gotwig: I just tried writing up everything | 18:49 |
racarr | here :) | 18:49 |
racarr | http://studio.sketchpad.cc/gmY0M6iqeh? | 18:49 |
gotwig | do I need all this? | 18:50 |
gotwig | Platform-api? | 18:50 |
racarr | Yep | 18:50 |
racarr | platform-api is the new low level library that sits on top of mir and other stuff, that toolkits, or applications which do not want to use a toolkit | 18:51 |
racarr | should use | 18:51 |
racarr | qtubuntu has the qtbackends you need for unity (in the same process as mir) and running apps (out of process) | 18:51 |
gotwig | racarr, do you work for canonical :X? | 18:51 |
racarr | gotwig: They keep paying me so I guess. | 18:51 |
gotwig | racarr, cool, do you live in the UK? | 18:51 |
racarr | Nope. San Francisco CA | 18:52 |
gotwig | Greetings from Germany ;D | 18:52 |
racarr | *waves* | 18:52 |
racarr | Also. yay finaly finished fetching build dependencies on nexus aha | 18:52 |
gotwig | -- package 'mircommon' not found | 18:52 |
gotwig | I know, I sound stupid for a dev xD but I never did qt stuff | 18:53 |
racarr | no worries | 18:53 |
racarr | I guess if you installed mir from ppa you need the mir dev packages | 18:53 |
racarr | gotwig: Which are called libmirserver-dev and libmirclient-dev :) | 18:53 |
gotwig | looking forward to packages ;D | 18:55 |
gotwig | ubuntu devs can do that very good :X | 18:55 |
gotwig | CMake Error: your CXX compiler: "CMAKE_CXX_COMPILER-NOTFOUND" was not found. Please set CMAKE_CXX_COMPILER to a valid compiler path or name. | 18:56 |
racarr | Do you have g++ ? | 18:57 |
gotwig | tvoss, I found a video about you on youtube ;D | 18:57 |
tvoss | gotwig, about me? | 18:57 |
tvoss | gotwig, or posted _by_ me? | 18:58 |
gotwig | tvoss, yeah, dat Ubuntu OnAir ;D | 18:58 |
racarr | tvoss: Yeah, "A secret look at the private life of Ubuntu Touch's Architect" | 18:58 |
tvoss | gotwig, yup | 18:58 |
racarr | I watch it every day | 18:58 |
tvoss | what? | 18:58 |
racarr | :p | 18:58 |
gotwig | racarr, I am retarded >_> thx | 18:58 |
kdub | haha | 18:58 |
racarr | im just silly | 18:58 |
gotwig | One Night in Ubuntistan ;D | 18:58 |
gotwig | /usr/include/mirserver/mir/options/program_option.h:25:51: fatal error: boost/program_options/variables_map.hpp: not found :X | 19:01 |
gotwig | I installed libboost-dev | 19:02 |
racarr | oh, thi was a recent fix that probably didnt land in the PPA yet | 19:02 |
gotwig | worx | 19:02 |
racarr | ah | 19:02 |
racarr | yeah you need libboost-programoptions-dev in particular | 19:02 |
racarr | program-options? | 19:02 |
racarr | something | 19:02 |
gotwig | racarr, only libboost-dev | 19:02 |
racarr | probably pulls in everything | 19:03 |
gotwig | racarr, qmake: could not find a Qt installation of '' | 19:03 |
gotwig | 19:03 | |
gotwig | racarr, do I have to download the second branch into the first one? | 19:03 |
racarr | no, it can go anywhere | 19:04 |
gotwig | getting qmake dependencies :X | 19:05 |
gotwig | I need the package for qplatformintegration.h | 19:11 |
racarr | gotwig: Let me try and remember um | 19:11 |
racarr | gotwig: qtbase5-private-dev | 19:11 |
racarr | :) | 19:11 |
gotwig | >_> | 19:12 |
gotwig | thx | 19:12 |
gotwig | did not fix it :O | 19:13 |
racarr | maybe you only had qt4 before and it was trying to build against | 19:14 |
racarr | qt4 | 19:14 |
gotwig | :( | 19:14 |
racarr | maybe recreate the makefile (qmake command again) | 19:14 |
gotwig | I already tried some things | 19:14 |
racarr | there should be something in the output between that and build that makes it clear if it is doing | 19:14 |
racarr | qt5 or qt4 | 19:14 |
gotwig | qmake runs fine | 19:14 |
racarr | do you have | 19:14 |
racarr | qt5-qmake | 19:15 |
racarr | ? | 19:15 |
racarr | If not, I think you want qt5-qmake and (perhaps?) qt5-default | 19:15 |
gotwig | Project MESSAGE: Warning: unknown QT: core-private | 19:15 |
gotwig | Project MESSAGE: Warning: unknown QT: gui-private | 19:15 |
gotwig | Project MESSAGE: Warning: unknown QT: platformsupport-private | 19:15 |
gotwig | oups, 3 lines ;D | 19:15 |
gotwig | do I have to build with qt5? | 19:16 |
racarr | Hmm. | 19:16 |
racarr | Yeah pretty much | 19:16 |
racarr | I'm not ure that qt4 wouldn't work if you changed a little API | 19:16 |
racarr | but qt5 is the target | 19:17 |
gotwig | I guess I miss a dependency | 19:17 |
racarr | gotwig: So ait, you have qt5-qmake, and qt5-default and qtbase5-private-dev | 19:19 |
racarr | and ran the qmake | 19:19 |
racarr | and...what happens? | 19:19 |
gotwig | *magic pills* | 19:19 |
gotwig | nope :X | 19:20 |
gotwig | I wonder why there is no qt4-qtconfig for qt5 | 19:22 |
racarr | I have had weird things with qmake in the past with qt4/qt5 | 19:23 |
racarr | could try making sure qmake -v | 19:23 |
gotwig | oh | 19:23 |
racarr | points at qt5 | 19:23 |
racarr | and deleting the makefiles | 19:23 |
racarr | I really don't understand qmake, it drives me insane :) | 19:23 |
gotwig | whats wrong with make | 19:24 |
gotwig | lol | 19:24 |
gotwig | and cmake | 19:24 |
gotwig | Project ERROR: Unknown module(s) in QT: sensors | 19:24 |
racarr | gotwig: Are you building from | 19:27 |
racarr | src/platforms? | 19:27 |
gotwig | YES | 19:27 |
racarr | gotwig: maybe it just means you need qtsensors5-dev? | 19:28 |
gotwig | yeah D | 19:28 |
gotwig | do you maybe got a dependency list for me lol | 19:31 |
racarr | I guess debian/control in the qtubuntu dir | 19:32 |
racarr | should have them | 19:32 |
gotwig | right | 19:38 |
racarr | kdub: Everything is working fine for me nexus 4 flipped | 19:50 |
kdub | yay | 19:50 |
gotwig | racarr, :O | 19:55 |
bregma | hmm, running xmir on the desktop I get both an X cursor and the big pink dogprick mir cursor following it around (only the mir cursor is in the wrong place) | 19:55 |
bregma | any way to staunch that thing? | 19:55 |
gotwig | UnityCore/Lens.h is missing | 19:57 |
gotwig | where can I get that | 19:57 |
racarr | gotwig: It should come from ./build -s successfully completing | 19:57 |
gotwig | I use 13.10 | 19:57 |
gotwig | is this supported? | 19:57 |
racarr | gotwig: Could try libunity-core-6.0-dev | 19:58 |
racarr | bregma: Uh...I guess --enable-input=false to mir | 19:58 |
racarr | ould do it | 19:58 |
racarr | since currently xmir doesnt use mir input | 19:58 |
gotwig | racarr, no chance | 19:59 |
bregma | racarr, at least that's a clue, thanks | 19:59 |
racarr | gotwig: I dunno. I don't know much about the unity phablet build system | 20:00 |
racarr | only thing I can suggest, is the build scripts are a little finnicky so clean the tree and build -s again (I think it drops some stuff int he directory one above too that you will want to clean) | 20:00 |
racarr | bregma: MIR_SERVER_ENABLE_INPUT=false in environment will also work... | 20:00 |
racarr | We can add a cursor option too | 20:01 |
racarr | you don't like the big pink orange arrow? ;( | 20:01 |
racarr | we put our best visual designers on the job for over 15 seconds. | 20:01 |
gotwig | racarr, build -s as root? | 20:02 |
racarr | I dont think so the scripts normally ask for sudo when they need them to install dependencies etc | 20:03 |
gotwig | the problem was with the update | 20:05 |
gotwig | build -s did not work for me | 20:05 |
gotwig | because some update sources did not work | 20:06 |
gotwig | I replaced apt-get update with echo "blub" and it seems to work :X | 20:06 |
gotwig | (not because of you guys, but because the bumblebee PPA does not work with 13.10, and I did not knew that :/) | 20:06 |
racarr | Ah, great | 20:07 |
gotwig | Unity is so awesome | 20:10 |
gotwig | I really have to say it | 20:10 |
gotwig | smart scopes are the perfect addition | 20:10 |
racarr | :) | 20:12 |
gotwig | too bad the music is stopped when I leave the dash :/ | 20:14 |
gotwig | now I should do some learning for my test.. see you later guys. And thank you very much | 20:15 |
gotwig | I think you convinced me to come back to unity development :X (third party app stuff, huh) | 20:16 |
racarr | gotwig: Good luck! nice meeting you | 20:18 |
racarr | Playing with the component showcase on nexus 4! | 20:19 |
gotwig | something that I did notice | 20:19 |
gotwig | build.sh wants to install libboost-regex-dev | 20:19 |
gotwig | it shouldnt do that | 20:19 |
gotwig | when It does that, it installs version 1.49 I guess, but it needs the newest version 1.53-dev I guess | 20:20 |
gotwig | first it installs one version, than removes packages to get the other, and its a mess | 20:20 |
racarr | oh! thanks, will check | 20:20 |
gotwig | I am on 13.10 | 20:20 |
racarr | ...have you guys ever typed | 20:32 |
racarr | "adb hell" | 20:33 |
racarr | you get ADB but the text background is red | 20:33 |
racarr | and the text is yellow | 20:33 |
racarr | im not loosing it right? | 20:33 |
kdub | racarr, haha! i see it too | 21:35 |
kdub | easter eggs | 21:35 |
racarr | XD yeah | 21:35 |
racarr | I only use adb hell now | 21:35 |
kdub | i guess if we have a request for swapinterval 0 in our client api | 21:36 |
kdub | we need both a sync and non-blocking function call | 21:36 |
kdub | we should 'stop the bleeding' and inject pure virtual interfaces around ms::Surface and msh::Surface | 22:38 |
racarr | kdub: ? | 23:44 |
racarr | kdub: p.s. Happy to report unity8 running well on nexus 4 :) | 23:44 |
kdub | oh really! | 23:49 |
kdub | yay | 23:49 |
kdub | racarr with like... the ability to launch applications? | 23:59 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!