| RAOF | duflu: Do you have a working mako handy? | 02:47 |
|---|---|---|
| RAOF | duflu: Mine is in flat-battery purgatory for the moment, and I was wondering if you could try running https://code.launchpad.net/~raof/mir/fix-deadlock-in-glib-alarm/+merge/266682 on it locally to see if you can reproduce the crash in CI? | 02:48 |
| duflu | RAOF: Yes and no. My mako is heavily hacked and I'm not wanting to change anything until I finish my performance work on it. Although I could run local builds... | 03:00 |
| RAOF | duflu: That's fine. I'll try building in a vivid chroot and deploying... | 03:01 |
| shengchieh | hi, guys. I met a mir exception error(0.12.1) when bootup on a arm device. so i try to rebuild mir with ./cross-sompile-chroot.sh but the built library can not be lunched functional on arm target. Does any one has the same experience and know how to solve it? | 03:26 |
| duflu | shengchieh: Got a log of what the exception says? | 03:30 |
| duflu | shengchieh: Also, feel free to log bugs and any issues you have can be discussed and resolved offline: https://bugs.launchpad.net/mir/+filebug | 03:31 |
| shengchieh | the log is : $/usr/share/ubuntu-touch-session/usc-wrapper --file'/run/mir_socket' --from-dm-fd 9 --to-dm-fd 13 --vt 1 | 03:37 |
| shengchieh | Warning:ignoring unrecognised arguments: --vt 1 | 03:37 |
| shengchieh | [527.561310]Server: Starting | 03:37 |
| shengchieh | [527.562040]Loader: Loading modules from: /usr/lib/arm-linux-gnueabihf/mir/server-platform | 03:37 |
| shengchieh | [527.562228]Loader: Loading module:/usr/lib/arm-linux-gnueabihf/mir/server-platform/graphics-mesa.so.1 | 03:37 |
| shengchieh | [527.567219]Loader: Loading module:/usr/lib/arm-linux-gnueabihf/mir/server-platform/graphics-dummy.so | 03:37 |
| shengchieh | [527.576294]Loader: Loading module:/usr/lib/arm-linux-gnueabihf/mir/server-platform/graphics-android.so.1 | 03:37 |
| shengchieh | [527.578435]Platform Loader: Selected driver: android (version 0.11.0) | 03:37 |
| shengchieh | Killed | 03:37 |
| duflu | shengchieh: It sounds like you have built Mir 0.12 but have 0.11 installed on the system. That's OK, but if you're running a binary from 0.12 then it needs to find the drivers for 0.12 and not the system ones. Try setting environment MIR_SERVER_PLATFORM_PATH=(path to server-modules/) and MIR_CLIENT_PLATFORM_PATH=(path to client-modules/) | 03:40 |
| duflu | Also, Mir 0.14 is released. We recommend using that. | 03:43 |
| shengchieh | Hi duflu, thanks for your suggestion, but we overwirte the previous .so files and use the same folder. | 03:44 |
| duflu | shengchieh: Then you also need to overwrite /usr/lib/arm-linux-gnueabihf/mir/... with your newly built drivers | 03:44 |
| duflu | Although other parts of the phone will then fail (broken ABIs) | 03:45 |
| shengchieh | We can not use Ubuntu's mir, because we have an integration issue with our devices.... Our EGL query an information which mir does not support | 03:47 |
| shengchieh | That's what we need to modify somethin in mir to fit our egl | 03:47 |
| shengchieh | Here comes original error log: [157678.315339] <ERROR> mircommon: Caught exception at Mir/EGL driver boundary (in query): /build/buildd/mir-0.12.1+15.04.20150324/src/platforms/android/server/server_render_window.cpp(90): Throw in function virtual int mir::graphics::android::ServerRenderWindow::driver_requests_info(int) const | 03:47 |
| shengchieh | Dynamic exception type: N5boost16exception_detail10clone_implINS0_19error_info_injectorISt13runtime_errorEEEE | 03:48 |
| shengchieh | std::exception::what: driver requests info we dont provide. key: 10 | 03:48 |
| shengchieh | So, we need to build a modified graphics-android.so.1 for our devices. | 03:48 |
| duflu | shengchieh: Good news: grapphics-android.so is not device specific. It's just Mir-version-specific | 03:48 |
| duflu | shengchieh: It appears your problem is that mir-team broke the ABI of graphics-android.so.1 between releases 0.11 and 0.12 but failed to bump the ABI number. So that's a problem. Although overwriting the binary using your 0.12 build is probably the only option | 03:51 |
| shengchieh | So, should I use mir 0.11 and have a build to test it again? I want to make sure the so files can be executed successfuly first. | 03:52 |
| shengchieh | But I am confused about my mir version, becase based on the log, it looks like 0.12 | 03:53 |
| shengchieh | /build/buildd/mir-0.12.1+15.04.20150324/src/platforms/android/server/server_render_window.cpp | 03:53 |
| duflu | shengchieh: Yes, a recent Ubuntu Touch installation should have no Mir version older than 0.12 on it. The 0.11 issue sounds like it's from a stale package. Try uninstalling the driver package that is versioned 0.11* and replace it with the 0.12 driver package | 03:55 |
| duflu | It should be called mir-platform-graphics-android1 | 03:57 |
| shengchieh | Sorry, I am newbie for ubuntu touch, can I knnow how to uninstall the driver? thanks. | 03:59 |
| duflu | shengchieh: Or download the latest 0.12 packages for armhf here: https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-013/+build/7105384 | 03:59 |
| duflu | shengchieh: Make sure you have a working internet connection on the device, then: sudo mount -o remount,rw / ; sudo apt-get update ; sudo apt-get dist-upgrade | 04:00 |
| duflu | Should fix it completely actually | 04:00 |
| === chihchun_afk is now known as chihchun | ||
| === chihchun is now known as chihchun_afk | ||
| === guest42345 is now known as o | ||
| === o is now known as Guest57566 | ||
| === chihchun_afk is now known as chihchun | ||
| === chihchun is now known as chihchun_afk | ||
| === chihchun_afk is now known as chihchun | ||
| === chihchun is now known as chihchun_afk | ||
| alan_g | greyback: just a FYI - I've pushed my WIP on window management to lp. There's a bunch of prerequisites already queued but you've seen most of them. https://code.launchpad.net/~alan-griffiths/qtmir/window-management/+merge/267181 | 12:05 |
| greyback | alan_g: many thanks, shall have a loo | 12:06 |
| greyback | k | 12:06 |
| === alan_g is now known as alan_g|lunch | ||
| === chihchun_afk is now known as chihchun | ||
| dandrader | anpok_, ping | 12:42 |
| === alan_g|lunch is now known as alan_g | ||
| kenvandine | greyback, http://paste.ubuntu.com/12013841/ | 14:04 |
| greyback | kenvandine: why is mouse-primary-button-enum != primary-button-enum ? | 14:05 |
| kenvandine | whoops | 14:06 |
| kenvandine | i think that was a mistake | 14:06 |
| kenvandine | greyback, http://bazaar.launchpad.net/~ken-vandine/unity8/mouse_touchpad_schema/view/head:/data/com.canonical.Unity8.gschema.xml | 14:09 |
| greyback | kenvandine: ok, looks reasonable, now need to see what mir supports! | 14:19 |
| kenvandine | greyback, great, thanks | 14:20 |
| greyback | anpok_: hey, you've knowledge of the input stack, can you have a look at Ken's xml file above and see if Mir would have any built-in support for those settings? | 14:27 |
| alan_g | camako: can you confirm whether this is OK with your Mir-on-X work? https://code.launchpad.net/~andreas-pokorny/mir/disable-android-input-stack-when-platform-available/+merge/267022 | 14:33 |
| camako | alan_g, will do | 14:33 |
| === dandrader is now known as dandrader|afk | ||
| === chihchun is now known as chihchun_afk | ||
| === chihchun_afk is now known as chihchun | ||
| === chihchun is now known as chihchun_afk | ||
| === dandrader_ is now known as dandrader | ||
| AlbertA | cool a short example exhibits the same GCC5 armhf bug... | 16:50 |
| AlbertA | http://pastebin.ubuntu.com/12014642/ | 16:50 |
| AlbertA | now to report the issue... | 16:50 |
| alan_g | AlbertA: it /could/ be simpler. Vis: "int main()" | 16:53 |
| alan_g | thanks for tracking it down | 16:53 |
| AlbertA | alan_g: heh true...force of habit | 16:54 |
| greyback | AlbertA: hey, could I get you to rebase lp:~albaguirre/qtubuntu/use-mir-surface-apis on top of lp:~dandrader/qtubuntu/busySwap-lp1473720. I want to land both, but they conflict | 16:56 |
| AlbertA | greyback: np | 16:56 |
| greyback | ta | 16:56 |
| bschaefer | AlbertA, nice find! | 17:07 |
| AlbertA | greyback: ok I merged daniel's branch into https://code.launchpad.net/~albaguirre/qtubuntu/use-mir-surface-apis/+merge/262257 | 17:14 |
| greyback | AlbertA: could you please re-propose the MR, with daniel's branch as a prereq. Just to keep reviews clear | 17:15 |
| AlbertA | greyback: done | 17:18 |
| greyback | thank you | 17:18 |
| anpok_ | greyback: will take a look | 17:43 |
| greyback | anpok_: thanks. I know libinput probably has support for more bits, but we're stuck with android input on phone, yeah? | 17:44 |
| anpok_ | nope.. | 17:50 |
| anpok_ | we can settle with one | 17:50 |
| anpok_ | interesting part is.. we might have to apply that per session.. | 17:50 |
| greyback | eh? use-case? | 17:51 |
| greyback | we don't want clients of unity8 to have their own input config | 17:52 |
| anpok_ | and thus either expose the settings to usc, and let usc apply it... or apply those settings within nested mir... | 17:52 |
| greyback | I vote for USC | 17:52 |
| anpok_ | just a matter of where to put the code.. and when to switch settings | 17:52 |
| greyback | or make a special mir_client interface for neste servers, which adds bits like this | 17:52 |
| anpok_ | exposing the setting to usc.. might mean that there is client api for config.. yes.. | 17:53 |
| anpok_ | greyback: how are those speed settings defined? | 17:59 |
| anpok_ | i mean... are there more detailed descriptions available or is this just a list of required settings and we shall figure out which ranges are reasonable? | 18:00 |
| greyback | anpok_: it's just a list of required settings really | 18:10 |
| === dandrader is now known as dandrader|afk | ||
| === dandrader|afk is now known as dandrader | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!