| RAOF | Oh, great. It's a racy test. | 01:18 |
|---|---|---|
| duflu | Umm what is r2962? :) | 01:38 |
| duflu | A bit of a mess it seems | 01:39 |
| duflu | RAOF: ^ that's not the proposal I remember from yesterday | 01:39 |
| RAOF | Not my proposal :) | 01:40 |
| RAOF | Or, rather, mine plus a bunch of other things to make it pass CI, I presume. | 01:40 |
| === marcusto_ is now known as marcustomlinson | ||
| === chihchun_afk is now known as chihchun | ||
| duflu | Oh for the days when the display server was a different process to the shell | 08:36 |
| alan_g | burn the heretic! | 08:49 |
| alan_g | ;) | 08:49 |
| davmor2 | alan_g: ah bless you must be new to the whole heretic thing, we don't burn them now it's ecologically unsound, that's what the dunking stall is for, if they drown they were good, if they float they are heretics and you drive a stake through their heart | 08:53 |
| alan_g | duflu: do you want to leave lp:~alan-griffiths/mir/fix-1496849 pending API rework, or shall we land it as a bugfix? | 09:48 |
| duflu | alan_g: Not worth blocking | 09:48 |
| anpok_ | oh another chapter in the book of module_deleter.. | 10:31 |
| === alan_g is now known as alan_g|lunch | ||
| === chihchun is now known as chihchun_afk | ||
| === alan_g|lunch is now known as alan_g | ||
| === dandrader is now known as dandrader|afl | ||
| === dandrader|afl is now known as dandrader|afk | ||
| Saviq | hey folks, I was able to upgrade qtmir without adding some new needed packages | 15:00 |
| Saviq | so ended up with broken unity8: | 15:00 |
| Saviq | ERROR: /build/mir-MEFZIJ/mir-0.16.0+15.10.20150921.1/src/platform/graphics/platform_probe.cpp(59): Throw in function std::shared_ptr<mir::SharedLibrary> mir::graphics::module_for_device(const std::vector<std::shared_ptr<mir::SharedLibrary> >&, const mir::options::ProgramO | 15:00 |
| Saviq | ption&) | 15:00 |
| Saviq | Dynamic exception type: boost::exception_detail::clone_impl<boost::exception_detail::error_info_injector<std::runtime_error> > | 15:00 |
| Saviq | std::exception::what: Failed to find platform for current system | 15:00 |
| Saviq | there seems to be some dependency missing, as apt was happy with that situation | 15:00 |
| Saviq | dist-upgrade resolved it fortunately | 15:00 |
| Saviq | but still, sounds like there's some link missing | 15:00 |
| Saviq | and will cause boottest failures | 15:07 |
| Saviq | because boottest is stuck on an image pre-mir-0.16 due to a broken NM | 15:08 |
| AlbertA | Saviq: yeah.. it's because of the drivers | 15:54 |
| AlbertA | Saviq: they are pulled by the seeds... | 15:54 |
| AlbertA | Saviq: the issue is different platform need different drivers... and that's not exactly expressible in debian packaging deps | 15:56 |
| === dandrader|afk is now known as dandrader | ||
| alan_g | AlbertA: the assumption that cursors have contiguous buffers is spread over a lot of our code. @https://code.launchpad.net/~alan-griffiths/mir/fix-nested-cursor/+merge/272011/comments/686746 | 16:33 |
| AlbertA | alan_g: It's not the cursor but the underlying buffer stream buffer. in android you may get a stride that is bigger than the width of the image | 16:44 |
| AlbertA | alan_g: but I see it's pre-existing... ummm I guess it may not work correctly in some phones | 16:45 |
| alan_g | AlbertA: I'm willing to log it and fix it. Just think its wider than the code this MP touches. | 16:46 |
| === alan_g is now known as alan_g|EOW | ||
| === dandrader is now known as dandrader|afk | ||
| BlackJohnny | bschaefer, hi | 18:38 |
| BlackJohnny | bschaefer, about the compiling of the cocos2d engine with mir for ubuntu touch | 18:39 |
| BlackJohnny | bschaefer, i get this message "OpenGL version too old ..." and i suspect is an EGL message | 18:41 |
| BlackJohnny | bschaefer, do i need also an EGL build with mir support? | 18:44 |
| BlackJohnny | bschaefer, i also get this GLFW error GLFWError #65542 Happen, EGL: Failed to initialize EGL: EGL is not or could not be initialized | 18:48 |
| bschaefer | BlackJohnny, well thats a good start | 18:49 |
| bschaefer | BlackJohnny, not sure about the OpenGL version to old bit though | 18:50 |
| BlackJohnny | i think egl gets somehow initialized later and not with gles... but lets ignore that | 18:50 |
| bschaefer | BlackJohnny, but everything is compiling? | 18:53 |
| bschaefer | you could try to run a GLFW3 example to make sure GLFW3 is working on mir | 18:53 |
| BlackJohnny | yes | 18:53 |
| BlackJohnny | did not compile samples, but I will do that thank you | 18:53 |
| BlackJohnny | bschaefer, -- NOTE: Examples and tests require OpenGL | 18:58 |
| BlackJohnny | bschaefer, not possible to link against GLES? | 18:58 |
| BlackJohnny | bschaefer, actually i have compiled them and got the libGL not found msg | 18:58 |
| BlackJohnny | bschaefer, compiled and executed | 18:59 |
| * bschaefer thought it ran through egl | 19:00 | |
| BlackJohnny | bschaefer, for the record, i do not use the latest code because I had to revert to the version of mir that is installed with ota 7 (0.12?) | 19:00 |
| bschaefer | maybe it just uses EGL to get a gl context | 19:00 |
| BlackJohnny | ota6 | 19:00 |
| bschaefer | BlackJohnny, yeah it looks like it using EGL to get a GL context | 19:01 |
| bschaefer | soo opengl can be used :( | 19:01 |
| bschaefer | for those examples, maybe we can disable opengl | 19:01 |
| bschaefer | in glfw when compiling | 19:01 |
| bschaefer | GLFW_USE_EGL=ON | 19:02 |
| bschaefer | BlackJohnny, but that should be turned on when you do -DGLFW_USE_MIR=ON | 19:02 |
| * bschaefer wonders how he got the examples working on mir now looking at the cmake | 19:02 | |
| BlackJohnny | bschaefer, it is, i have checked | 19:02 |
| bschaefer | since EGL=ON means no tests | 19:02 |
| bschaefer | or examples | 19:02 |
| bschaefer | actually | 19:03 |
| bschaefer | if (${GLFW_CLIENT_LIBRARY} STREQUAL "opengl") | 19:03 |
| bschaefer | err | 19:03 |
| BlackJohnny | i have changed that to gles2 | 19:03 |
| bschaefer | o i see | 19:03 |
| bschaefer | BlackJohnny, yeah then the examples wont work hmm | 19:03 |
| BlackJohnny | bschaefer, at runtime you mean because they get compiled | 19:04 |
| bschaefer | BlackJohnny, this is a compile time check | 19:04 |
| BlackJohnny | bschaefer, hmm i understand why u say that | 19:04 |
| bschaefer | you can only have OPENGL OR glesv2 OR glesv1 | 19:04 |
| BlackJohnny | maybe those binaries ar from a previous build :S | 19:04 |
| BlackJohnny | sorry | 19:04 |
| bschaefer | BlackJohnny, check: src/glfw_config.h:80:/* #undef _GLFW_USE_GLESV2 */ | 19:05 |
| bschaefer | build/src/glfw_config.h | 19:05 |
| bschaefer | to see what was selected | 19:05 |
| * bschaefer does not have glesv2 selected | 19:05 | |
| BlackJohnny | bschaefer, you are right | 19:05 |
| BlackJohnny | bschaefer, with glesv2 the samples are not built | 19:05 |
| bschaefer | right, which is expected since the examples use OpenGL | 19:06 |
| bschaefer | BlackJohnny, but you had that already right? | 19:06 |
| BlackJohnny | bschaefer, i guess so ... with ldd i see the deps on mir | 19:06 |
| bschaefer | i think you need | 19:06 |
| bschaefer | -DGLFW_CLIENT_LIBRARY=glesv2 | 19:06 |
| * bschaefer tries | 19:07 | |
| BlackJohnny | bschaefer, i have changed this line: set(GLFW_CLIENT_LIBRARY "glesv2" CACHE STRING | 19:07 |
| BlackJohnny | bschaefer, is the define a better way? | 19:07 |
| bschaefer | cmake .. -DCMAKE_INSTALL_PREFIX=<PREFIX> -DGLFW_USE_MIR=ON -DGLFW_CLIENT_LIBRARY=glesv2 | 19:07 |
| bschaefer | BlackJohnny, now the issue, i've glesv2 built... but how do i test it hmm | 19:08 |
| BlackJohnny | bschaefer, i've missed this error msg "libEGL warning: DRI2: xcb_connect failed" | 19:11 |
| bschaefer | o... hmm try | 19:11 |
| bschaefer | EGL_PLATFORM=mir | 19:11 |
| BlackJohnny | building glfw? | 19:11 |
| bschaefer | BlackJohnny, when you run it | 19:12 |
| bschaefer | or just export EGL_PLATFORM=mir (env var) | 19:12 |
| === dandrader|afk is now known as dandrader | ||
| BlackJohnny | is it ok to modify the EXEC line in the desktop file with this? Exec=export EGL_PLATFORM=mir && bin/hexpuzzle | 19:13 |
| bschaefer | BlackJohnny, you can make the exec in the desktop file point to an exe wrapper | 19:14 |
| bschaefer | ie. hexpuzzle.sh | 19:14 |
| bschaefer | which will run the export + bin/hexpuzzle | 19:14 |
| BlackJohnny | ok | 19:14 |
| BlackJohnny | bschaefer, Segmentation fault (core dumped) :) | 19:18 |
| bschaefer | BlackJohnny, nice, soo hmm now why | 19:19 |
| BlackJohnny | bschaefer, i will check wihtout the export to check if that is the real reason | 19:19 |
| bschaefer | you can also just try | 19:20 |
| bschaefer | EGL_PLATFORM=mir bin/hexpuzzle | 19:20 |
| bschaefer | but hmm | 19:20 |
| BlackJohnny | bschaefer, without that is gives the same errors as before | 19:20 |
| bschaefer | well we will need to batch gdb | 19:20 |
| BlackJohnny | bschaefer, but it might be that it works but then i implemented wrongly the mir connection or something | 19:20 |
| BlackJohnny | bschaefer, same error as before = OpenGL related | 19:21 |
| bschaefer | BlackJohnny, http://paste.ubuntu.com/12558073/ | 19:21 |
| bschaefer | BlackJohnny, well the mir connection is glfw3 doing | 19:21 |
| BlackJohnny | bschaefer, i also do that with cocos2d to get the display parameters | 19:22 |
| bschaefer | --args /usr/games/hexpuzzle $ARGS | 19:22 |
| bschaefer | BlackJohnny, o open the mir sever? | 19:22 |
| bschaefer | try to remove all that extra stuff for now (or just dont do anything) | 19:23 |
| bschaefer | lets just try to get glfw main loop spinning | 19:23 |
| bschaefer | even if it doesnt display anything | 19:23 |
| BlackJohnny | bschaefer, /opt/click.ubuntu.com/hexpuzzle/0.1/bin/hexpuzzle: error while loading shared libraries: libglfw.so.3: cannot open shared object file: No such file or directory | 19:27 |
| bschaefer | hmmm | 19:27 |
| BlackJohnny | bschaefer, probably it does not check the lib path? | 19:27 |
| bschaefer | BlackJohnny, you can force the path with export LD_LIBRARY_PATH in the *.sh | 19:28 |
| bschaefer | but im not super good with clicks :) | 19:28 |
| BlackJohnny | bschaefer, ok | 19:28 |
| BlackJohnny | sec | 19:28 |
| BlackJohnny | bschaefer, my .sh, is it correct?http://paste.ubuntu.com/12558221/ | 19:31 |
| bschaefer | BlackJohnny as long as thats where the libglfw lib lives :) | 19:32 |
| bschaefer | BlackJohnny, if it still fails, we'll have to use strace to check *where* its looking for the lib | 19:33 |
| BlackJohnny | bschaefer, it is ok, I will continue myself because u did more than enough | 19:33 |
| BlackJohnny | bschaefer, i will ask you when I get really stucked | 19:33 |
| bschaefer | BlackJohnny, good luck! | 19:34 |
| BlackJohnny | bschaefer, i suspect a missing dep | 19:34 |
| BlackJohnny | will check | 19:34 |
| BlackJohnny | bschaefer, nope ldd says everything is ok | 19:35 |
| bschaefer | :( | 19:35 |
| bschaefer | BlackJohnny, you can try ldd when you run it :) | 19:35 |
| bschaefer | otherwise check strace when you run it | 19:35 |
| BlackJohnny | bschaefer, hmm and it is not linked to libGL directly | 19:35 |
| bschaefer | strace <exe> | 19:35 |
| bschaefer | it shouldnt use libGL anymore (should be using gles2 | 19:36 |
| bschaefer | ) | 19:36 |
| BlackJohnny | bschaefer, that is ok but then the libGL is ref by EGL | 19:36 |
| BlackJohnny | bschaefer, thank you and have a good day | 19:37 |
| bschaefer | BlackJohnny, np, and you as well | 19:38 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!