=== chihchun is now known as chihchun_afk | ||
=== chihchun_afk is now known as chihchun | ||
=== chihchun is now known as chihchun_afk | ||
=== marcust__ is now known as marcustomlinson | ||
=== 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 | ||
willcooke | kgunn, imported: https://code.launchpad.net/~mlankhorst/xmir/xmir | 09:04 |
---|---|---|
mlankhorst | goodie | 09:04 |
kgunn_ | \o/ | 09:04 |
kgunn_ | @vogons ^ | 09:04 |
kgunn_ | @unity ^ | 09:04 |
kgunn_ | bregma: bschaefer ^ | 09:05 |
mlankhorst | but please use the git repo :P | 09:05 |
willcooke | mlankhorst, looks like it failed actually: http://launchpadlibrarian.net/196684280/mlankhorst-xmir-xmir.log | 09:07 |
mlankhorst | .. :/ | 09:08 |
mlankhorst | bug in bzr import it seems, looks like it can't handle GPG signed releases.. | 09:09 |
mlankhorst | just use raw git :P | 09:13 |
willcooke | kgunn_, can we do that in the short term? | 09:17 |
kgunn_ | willcooke: what's that ? sorry...in case i disconnected | 09:18 |
willcooke | kgunn_, there seems to be an issue with LP pulling in the Git repo. | 09:18 |
kgunn_ | sure | 09:18 |
willcooke | kgunn_, can we work with it in Git for now? | 09:19 |
kgunn_ | we'll just have to get the packages uploaded manually using the trainguards | 09:19 |
willcooke | Or we could some kind of manual push in to LP | 09:19 |
kgunn_ | which is simple | 09:19 |
willcooke | ok! | 09:19 |
willcooke | nice | 09:19 |
kgunn_ | yeah...so mlankhorst if you have a new set of packages that you'd say is ready to be put into a ppa & part of our demo lemme know | 09:20 |
mlankhorst | oke | 09:42 |
mlankhorst | when can I expect mir 0.11? :P | 09:42 |
mlankhorst | or is there a ppa somewhere with the cursor changes? | 09:43 |
kgunn | mlankhorst: so the ppa i shared, will have all those changes | 09:44 |
mlankhorst | oke | 09:44 |
kgunn | mlankhorst: but i think that mirevent2 landed as part of mir0.11 | 09:44 |
kgunn | however | 09:44 |
kgunn | the qtmir utilization of it is in lag | 09:44 |
kgunn | e.g. we'll test and land that next | 09:44 |
mlankhorst | oke I'll build 0.11 myself then.. | 09:44 |
dandrader | kgunn, the fix for the silo 0 https://code.launchpad.net/~alan-griffiths/qtmir/port-to-msh-Shell/+merge/247844/comments/615755 | 09:45 |
kgunn | mlankhorst: oh...actually, there is a ppa for testing that has just mir0.11 | 09:45 |
kgunn | not trunk (i think) | 09:45 |
dandrader | kgunn, gotta apply this patch to this MP http://paste.ubuntu.com/10069953/ | 09:45 |
kgunn | one sec | 09:45 |
kgunn | dandrader: ack | 09:45 |
kgunn | mlankhorst: ah yeah...so this silo/ppa | 09:46 |
kgunn | http://people.canonical.com/~platform/citrain_dashboard/#?distro=ubuntu&q=landing-012 | 09:46 |
kgunn | is the testing silo/ppa for mir0.11 | 09:46 |
mlankhorst | oke | 09:46 |
mlankhorst | and if I use that with unity-system-compositor I get the mouse events? | 09:46 |
kgunn | mlankhorst: mmm....on desktop yeah i thikn you just get it....on andrdoid devices, you have to use the u-s-c that's in the ppa for the demo (vivid silo 0) | 09:47 |
kgunn | anpok: ^ right ? | 09:47 |
kgunn | mlankhorst: difference being that hw cursor is used on desktop, and android doesn't suport hw cursor | 09:48 |
alf_ | mlankhorst: Hi! If I have a patch for the mesa mir egl backend, what's the best way to propose? Diff against ubuntu branch on debian git, debdiff, some other way? | 10:01 |
mlankhorst | alf_: ask raof.. | 10:02 |
alf_ | mlankhorst: OK, forget I mentioned the mesa mir egl detail :) I mean in general, how would do you prefer me to propose a package update for mesa. | 10:05 |
mlankhorst | feel free to upload and I'll merge the changes back to the packaging | 10:06 |
alf_ | mlankhorst: upload where? | 10:07 |
mlankhorst | archive? | 10:07 |
mlankhorst | or do you want a review :P | 10:08 |
alf_ | mlankhorst: I don't have ubuntu archive access, if that's what you mean. I could push to a PPA though. | 10:08 |
mlankhorst | oke then just send me a diff .. | 10:08 |
alf_ | mlankhorst: ok, thanks. I will do so after some wider testing (probably will push to PPA first anyway). | 10:09 |
alf_ | mlankhorst: (after 0.11 is released) | 10:09 |
kgunn | alan_g: hey mornin' would you mind updating your qtmir branch based on dandrader's feedback ? | 10:25 |
kgunn | https://code.launchpad.net/~alan-griffiths/qtmir/port-to-msh-Shell/+merge/247844/comments/615755 | 10:25 |
kgunn | ....we _think_ it's the last thing to change in order to test/rel mir0.11 | 10:25 |
alan_g | kgunn: just compiling it before pushing | 10:25 |
kgunn | thank you sir | 10:26 |
* alan_g wishes there were automated acceptance tests that pick this up. | 10:27 | |
alan_g | kgunn: dandrader - done | 10:29 |
kgunn | cool | 10:29 |
mlankhorst | alf_: hm do I need a even newer version of mir to kill of the thread mir creates? | 10:29 |
kgunn | thank you | 10:29 |
dandrader | alan_g, thanks | 10:29 |
mlankhorst | I want to kill the thread proxy in Xmir | 10:29 |
mlankhorst | makes input handling slightly easier for me.. | 10:30 |
alf_ | mlankhorst: It's now my turn to tell you 'ask RAOF' :) But assuming what you want is an event fd, that didn't make it into 0.11 AFAICT. | 10:32 |
mlankhorst | aw.. | 10:32 |
mlankhorst | yeah probably not | 10:32 |
alf_ | mlankhorst: https://code.launchpad.net/~raof/mir/provide-event-fd/+merge/247794 , in case you want to check if this is good enough for your use cases | 10:33 |
mlankhorst | it probably is | 10:35 |
mlankhorst | the main thing I worry about is if my hack that copies the full event will hold until mir advanced enough for that one.. | 10:36 |
mlankhorst | it's a lot easier to rework events after that provide-event-fd code is merged | 10:36 |
=== alan_g is now known as alan_g|reboot | ||
Saviq | hey guys, does that crash ring a bell for you bug #1418456? mardy can't get unity8 to start on an Atom netbook | 10:55 |
ubot5 | bug 1418456 in unity8 (Ubuntu) "unity8 assert failure: *** Error in `unity8': double free or corruption (fasttop): 0xb20055a0 ***" [Medium,New] https://launchpad.net/bugs/1418456 | 10:55 |
ogra_ | its not only free sowftware but even double free software :P | 10:56 |
alan_g | alf_: you were looking at a heap corruption recently ^ | 10:57 |
alf_ | alan_g: doesn't look familiar, it seems that an exception(?) happens and we are unwinding the stack when we get the memory error | 11:10 |
alan_g | alf_: ack | 11:11 |
mardy | alan_g, alf_: I've that machine running, if you want feel free to ask me more questions for debugging, I can give you the results right away | 12:01 |
alf_ | mardy: could you run unity8 under gdb? | 12:04 |
mardy | alf_: sure; any specific parameters or env variables to set? | 12:04 |
alf_ | mardy: no params, but type 'catch throw' in gdb before running | 12:05 |
alf_ | mardy: (so it will stop when an exception is thrown) | 12:05 |
mlankhorst | so why is my touchpad in unity-session-compositor seen as a touch event? :s | 12:07 |
mardy | alf_: but I should export QT_QPA_PLATFORM=ubuntumirclient, right? Otherwise it seems it's trying to talk to X | 12:07 |
mlankhorst | 12:08 | |
mardy | alf_: or maybe export QT_QPA_PLATFORM=mirserver? | 12:09 |
alf_ | mardy: I haven't tried in a while but unity8 env. variables I used for manual invocation are: MIR_SERVER_NAME=session-0 MIR_SOCKET=/run/mir_socket QT_QPA_PLATFORM=mirserver | 12:10 |
mardy | alf_: anyway, looks like it's quitting immediately, no crashes (but I see a "Gtk-WARNING **: cannot open display:" message (Gtk??) | 12:11 |
mardy | alf_: ok, let me try with those | 12:11 |
alf_ | mardy: perhaps MIR_SOCKET is different now | 12:11 |
mardy | alf_: even with those variables, it exits immediately, with that Gtk warning | 12:13 |
alf_ | mardy: oh, are you running USC? | 12:13 |
mardy | alf_: so, I've been running this commands from a text terminal, after doing a "sudo service lightdm stop" | 12:14 |
alf_ | mardy: so USC has stopped too | 12:14 |
alf_ | mardy: let's go back a step and try with the demo servers only | 12:15 |
alf_ | mardy: Install the mir-demos package (if you don't have it already) | 12:15 |
mardy | alf_: done | 12:17 |
mardy | alf_: I even tried the mir_demo_server before, but it took control of the terminal and I didn't know how to quit or switch to another terminal | 12:18 |
alan_g | mardy: the mir servers will exit on Ctrl+Alt+BkSp | 12:23 |
alan_g | Or SIGTERM etc | 12:24 |
alan_g | It is easiest if you can get a terminal session from another device - e.g. ssh | 12:28 |
greyback_ | mardy: this is my command to start unity8 on desktop: http://pastebin.ubuntu.com/10072333/ | 12:32 |
alf_ | mardy: be careful to run *sudo* mir_demo_server | 12:33 |
willcooke | mlankhorst, so still issues with the pointer in xmir? | 12:33 |
mardy | greyback_: it still quits immediately, with that Gtk warning | 12:33 |
mardy | alf_: ah, let me try | 12:33 |
mlankhorst | seems to right now, but I don't know if a fix has made it in to qtmir yet.. | 12:33 |
alf_ | mardy: do you have ssh access to the machine? | 12:33 |
mardy | alf_: yep | 12:34 |
greyback_ | mardy: that command depends on USC running | 12:34 |
mardy | alf_: so, with the demo server I get a black screen and a working mouse pointer | 12:34 |
mardy | greyback_: ok, I'll try that later as soon as we confirm with alf_ that the mir things are working | 12:34 |
alf_ | mardy: ok keep that running and ssh to the machine | 12:35 |
mardy | alf_: yep, done | 12:36 |
alf_ | mardy: as first test run in ssh "sudo mir_demo_client_egltriangle" | 12:36 |
alf_ | mardy: you should see a rotating triangle on screen | 12:36 |
mardy | alf_: yes, it works | 12:37 |
alf_ | mardy: ok, stop it (ctrl-c), and run 'sudo mir_demo_server --host /tmp/mir_socket -f /tmp/mir_nested' | 12:38 |
mardy | alf_: ok, no changes on screen, but the thing is running | 12:38 |
alf_ | mardy: ok, new ssh and run 'sudo mir_demo_client_egltriangle -m /tmp/mir_nested' | 12:39 |
mardy | alf_: works fine, too | 12:39 |
alf_ | mardy: ok, good, so base mir is working | 12:40 |
alf_ | mardy: now kill every mir related process | 12:41 |
alf_ | mardy: and 'sudo start lightdm' | 12:41 |
alf_ | mardy: that should start usc, and try to run unity8 and fail | 12:41 |
alf_ | mardy: but usc should stay running | 12:42 |
willcooke | anpok, were you working on the input stuff in qtmir for pointer events etc? | 12:42 |
alf_ | mardy: then ssh and use e.g. greyback's invocation to start unity | 12:42 |
mardy | alf_: so, I should login using the unity8 session, right? | 12:42 |
alf_ | mardy: actually before you log in do a 'ps -Af | grep unity', does it show anything? | 12:44 |
mardy | alf_: ops, too late, I already logged in. It shows USC, unity8-dash and a couple of processes related to unity-scopes | 12:45 |
alf_ | mardy: no problem, just wanted to make sure USC wasn't already running | 12:45 |
mardy | alf_, greyback_: always the same: http://pastebin.ubuntu.com/10072557/ | 12:46 |
alf_ | mardy: hmm, let's see where the mir_socket is created at, try 'find /var /tmp /run -name mir_socket' | 12:48 |
greyback_ | mardy: could you put 'import QtQuick 2.2; Rectangle{ color: "blue" }' into a test.qml file, and do the above command, but replacing /usr/bin/unity8 with "qmlscene test.qml" | 12:49 |
=== alan_g is now known as alan_g|lunch | ||
mardy | alf_: not found | 12:49 |
mardy | greyback_: same output as above; so I guess this is due to the missing mir socket | 12:51 |
greyback_ | mardy: yeah, tho am surprised it's not printing an error | 12:51 |
alf_ | mardy: try 'cat /proc/$(pidof unity-system-compositor)/cmdline' | 12:52 |
alf_ | mardy: and 'cat /proc/$(pidof unity-system-compositor)/environ' | 12:52 |
alf_ | mardy: perhaps they will give us a clue | 12:52 |
alf_ | greyback_: The invocation you pasted is for a nested unity8, right? | 12:53 |
greyback_ | alf_: correct | 12:53 |
mardy | alf_: the first says /usr/sbin/unity-system-compositor--disable-inactivity-policy=true--on-fatal-error-abort--file/run/lightdm-mir-0--from-dm-fd12--to-dm-fd21--vt8--enable-hardware-cursor=true | 12:53 |
mardy | alf_: the second is XDG_SEAT=seat0XDG_VTNR=8PWD=/ | 12:54 |
greyback_ | alf_: to test qtmir is working, stop unity-system-compositor, run "DESKTOP_SESSION=unity8-mir QT_QPA_PLATFORM=mirserver qmlscene test.qml" | 12:54 |
alf_ | mardy: ok, so the file you are looking for is | 12:54 |
alf_ | mardy: '/run/lightdm-mir-0' | 12:55 |
mardy | alf_: yes, it's there, it's a socked owned by root with 777 permissions | 12:56 |
alf_ | mardy: hmm, but that was already correct in greyback's invocation... | 12:56 |
mardy | greyback_: should I run your command as root? | 12:56 |
greyback_ | mardy: yes | 12:57 |
mardy | ahh | 12:57 |
greyback_ | sorry, I always forget that | 12:57 |
greyback_ | and yeah, you need ssh access to stop it | 12:57 |
mardy | greyback_: nope, nothing changed, same output | 12:58 |
greyback_ | mardy: ok that's not expected | 12:58 |
greyback_ | mardy: it crashes right? | 12:59 |
greyback_ | could you grab a backtrace? | 12:59 |
mlankhorst | I'm not a mir-dev.. but shouldn't input_event_type_to_string handle type_pointer ? | 12:59 |
mardy | greyback_: no, it looks like a clean exit "[Inferior 1 (process 12590) exited with code 01]" | 12:59 |
mardy | greyback_: and it's even claiming that your process is inferior ;-) | 13:00 |
greyback_ | :) | 13:00 |
mardy | greyback_: is that Gtk warning expected? | 13:00 |
greyback_ | mardy: no | 13:01 |
alf_ | greyback_: can we (try) to connect one of the mir demos to usc, to check if at least that part is OK? | 13:01 |
greyback_ | alf_: go for it | 13:01 |
mardy | greyback_, alf_: do you think that a strace can help? | 13:01 |
greyback_ | mardy: I'm extremely confused, so yeah, give me anything you think of | 13:01 |
alf_ | greyback_: Well, it was mostly a question of how to make USC accept it? What environment should we set? MIR_SERVER_NAME=session-0 DESKTOP_SESSION=unity8-mir ? | 13:02 |
greyback_ | alf_: MIR_SOCKET=$XDG_RUNTIME_DIR/mir_socket \ | 13:03 |
greyback_ | MIR_SERVER_PROMPT_FILE=1 \ | 13:03 |
greyback_ | MIR_SERVER_HOST_SOCKET=/run/lightdm-mir-0 \ | 13:03 |
greyback_ | MIR_SERVER_FILE=$XDG_RUNTIME_DIR/mir_socket \ | 13:03 |
greyback_ | MIR_SERVER_NAME=session-0 \ | 13:03 |
greyback_ | those should do the trick | 13:04 |
mlankhorst | willcooke: https://bugs.launchpad.net/qtmir/+bug/1417650 seems this bug is related? | 13:04 |
ubot5 | Launchpad bug 1417650 in QtMir "Pointer motion and crossing events are missing" [Undecided,In progress] | 13:04 |
alf_ | mardy: when you are done strace-ing, you could try 'mir_demo_client_egltriangle' with the environment greyback pasted above to see if at least the demo client can connect to usc | 13:04 |
mardy | greyback_, alf_: https://launchpadlibrarian.net/196699566/unity8-strace.log | 13:05 |
willcooke | mlankhorst, aha! So it's being worked on already, right? | 13:05 |
willcooke | I will speak to kgunn and the guys here shortly.... | 13:05 |
=== chihchun_afk is now known as chihchun | ||
mlankhorst | I'm not sure if it's worked on specifically.. | 13:05 |
mlankhorst | but lets hope so | 13:05 |
mardy | alf_: nope: "Can't get connection" | 13:05 |
alf_ | mardy: sudo? | 13:06 |
alf_ | mardy: I noticed in strace that you are running 32-bits (i386), perhaps that could be significant (or not) | 13:07 |
mardy | alf_: yes, I started it from within a "sudo -i" environment | 13:08 |
greyback_ | mardy: could you run the qmlscene line in gdb, break on gdk_init and backtrace it please | 13:09 |
mardy | alf_: if I add the "-m /run/lightdm-mir-0" option to the egltriangle test, then it connects and apparently runs, though nothing is visible on screen | 13:09 |
alf_ | mardy: ah, right, this is a client -m is the correct switch | 13:10 |
willcooke | mlankhorst, if we need it looking at I can ask | 13:10 |
alf_ | mardy: ok, that's expected, usc doesn't bring it to the foreground | 13:11 |
alf_ | mardy: so the socket works fine | 13:11 |
mardy | greyback_: so, it doesn't hit gdk_init | 13:11 |
greyback_ | mardy: how about gdk_display_open | 13:12 |
mardy | greyback_: gtk_init() is hit, instead | 13:12 |
greyback_ | ah | 13:12 |
alf_ | mardy: I leave you in greyback_'s capable hands for now (he is more familiar with the qt/unity8 side of things anyway), got to get some lunch, be back in a bit | 13:12 |
mardy | alf_: ok, thanks a lot | 13:13 |
mardy | greyback_: http://paste.ubuntu.com/10072964/ | 13:13 |
mlankhorst | willcooke: well it looks from the diff like it's useful.. | 13:13 |
mardy | greyback_: I hope there is not a switch based on the CPU architecture in there... | 13:13 |
greyback_ | mardy: there isn't in my code anyway | 13:14 |
greyback_ | mardy: could you muck with /usr/lib/i386-linux-gnu/qt5/plugins/platformthemes/libappmenu-qt5.so to prevent it being loaded? | 13:14 |
greyback_ | tho oddly it is loaded here and unity8 comes up just fine | 13:16 |
mardy | greyback_: ok, now I get the "double free or corruption" | 13:18 |
mardy | greyback_: wait, let me try with qmlscene first | 13:18 |
mardy | greyback_: yep, I get the memory corruption with qmlscene as well | 13:19 |
greyback_ | mardy: backtrace again please | 13:19 |
mardy | greyback_: http://paste.ubuntu.com/10073018/ | 13:20 |
mardy | greyback_: as you can see I typed "catch throw", but it seems like there was no exception | 13:20 |
mlankhorst | willcooke: hm it's probably necessary but not sufficient | 13:21 |
mlankhorst | afaict | 13:21 |
greyback_ | mardy: could you add the missing symbols? | 13:21 |
greyback_ | that code is more mir's | 13:22 |
willcooke | mlankhorst, just spoke to kgunn and greyback_, someone will get on to this asap. greyback_ will probably work on it | 13:24 |
mlankhorst | oke | 13:26 |
mlankhorst | just tested and nested mir works as intended | 13:26 |
mlankhorst | just unity8 is borked | 13:27 |
greyback_ | mlankhorst: borked how? | 13:28 |
willcooke | mlankhorst, side question: how's your xmir ppa looking atm? Good to try on our devices here? | 13:29 |
mlankhorst | ought to? | 13:30 |
mlankhorst | greyback_: I don't get pointer events.. | 13:30 |
mlankhorst | my pointer events get converted to touch events | 13:30 |
greyback_ | mlankhorst: that issue, ok, yeah I'm on it | 13:30 |
willcooke | thanks guys | 13:32 |
mlankhorst | willcooke: ppa works fine afaik, I was working on moving to new touch events but it can wait | 13:32 |
mardy | greyback_, alf_: http://paste.ubuntu.com/10073198/ | 13:32 |
willcooke | mlankhorst, ack | 13:33 |
greyback_ | mardy: frame 11 could be very useful, libmirserver28-dbgsym is missing? | 13:34 |
=== alan_g|lunch is now known as alan_g | ||
mardy | greyback_, alf_: http://paste.ubuntu.com/10073557/ | 13:53 |
anpok | willcooke: input stuff in unity-system-compositor and mir | 13:54 |
willcooke | anpok, no worries, I think greyback_ is looking at the issue | 13:55 |
greyback_ | mardy: thanks for the paste, I need alf_ to help you with that | 14:05 |
alf_ | greyback_: mardy: back, looking | 14:06 |
alf_ | mardy: could you run qmlscene with valgrind please? | 14:07 |
mardy | alf_: ah! I anticipated you :-) I just did it, let me collect the logs :-) | 14:08 |
=== marcusto_ is now known as marcustomlinson | ||
mardy | alf_: http://paste.ubuntu.com/10073830/ | 14:09 |
mardy | alf_: see the error at the end of the logs: "std::exception::what: Failed to create shared EGL context" | 14:11 |
alf_ | mardy: ah, this is much more helpful, thanks! | 14:17 |
alf_ | mardy: could you please attach this to the bug so it doesn't get lost? | 14:17 |
mardy | alf_: ok | 14:17 |
kgunn | racarr: i owe you a beer....disabling n seems to be working | 14:18 |
kgunn | like life chaning | 14:18 |
kgunn | changing event | 14:18 |
mardy | alf_: done | 14:19 |
alf_ | mardy: thanks | 14:20 |
alf_ | mardy: greyback_: So, the root cause of the bug is the exception "std::exception::what: Failed to create shared EGL context". This causes the stack to unwind and releases objects in a way that causes memory errors, but that's secondary, only a side effect. I will work on fixing the effect, probably the problem is in our Mesa code. | 14:22 |
alf_ | mardy: still we need to find out why the context creation fails... | 14:23 |
alf_ | mardy: could you run qmlscene under gdb and additional environment variables MESA_DEBUG=1 and EGL_LOG_LEVEL=debug | 14:27 |
mardy | alf_: sure | 14:27 |
alf_ | mardy: then in gdb break on _mesa_error() | 14:27 |
alf_ | mardy: before running | 14:27 |
racarr | alan_g: Hey...testing release and running in to some issues I think may originate with the SurfaceConfigurator changes... | 14:28 |
alan_g | racarr: ok? | 14:29 |
racarr | aha sorry got distracted mid sentence loud room | 14:29 |
racarr | alan_g: OSK hiding got broken somewhere between | 14:29 |
racarr | the OSK and the QML the minimized state isn't | 14:29 |
racarr | successfully applying or being pushed through or some such | 14:29 |
racarr | just wondering if you have any ideas :) | 14:30 |
duflu_ | alf_, mardy: Sounds like bug 1408910 :) | 14:32 |
ubot5 | bug 1408910 in unity8 (Ubuntu) "Ubuntu Desktop Next fails to start; just black screen and mouse pointer (unity8.log: std::exception::what: Failed to create shared EGL context)" [Critical,Confirmed] https://launchpad.net/bugs/1408910 | 14:32 |
* alan_g goes to look at the Q_EMIT surfaceAttributeChanged changes... | 14:32 | |
racarr | alan_g: I haven't verified it entirely because Im still getting debug symbols...but we added some print in the QML side on the input method element and | 14:32 |
racarr | the first time it does show that state changed to showed...and | 14:33 |
racarr | set_surface_attribute continues being called but | 14:33 |
racarr | state never changes again...so it may be uh | 14:33 |
racarr | the value of "result" is incorrect or being mangled or some such | 14:33 |
dandrader | anpok, the env var is GRID_UNIT_PX | 14:34 |
alf_ | duflu_: mardy: ah, indeed | 14:34 |
mardy | alf_: this is what I get, the _mesa_error() was not hit: http://paste.ubuntu.com/10074123/ | 14:39 |
alf_ | mardy: ok, please try setting the breakpoint after you have run the program once in gdb and then re-run, sometimes BPs don't work correctly if the symbols is not loaded | 14:43 |
alan_g | racarr: nothing leaps out at me. Is there an image I can try? | 14:46 |
alf_ | duflu_: bug 1418486, I will re-add mir, because although it's strictly not code in the Mir codebase, it's still the "Mir project" and I want it to show in the bug list for Mir. I will remove it when done. | 14:46 |
ubot5 | bug 1418486 in mesa (Ubuntu) "Mir Mesa EGL platform leaks memory with every frame" [High,In progress] https://launchpad.net/bugs/1418486 | 14:46 |
duflu_ | alf_: Yep, OK, standard behaviour | 14:47 |
alf_ | duflu_: I wonder if there is a better way to express the relationship of the two projects in LP... Perhaps have a "Mir project" meta-project? Anyway... | 14:48 |
racarr | alan_g: Yeah its showing up in the release PPA http://people.canonical.com/~platform/citrain_dashboard/#?distro=ubuntu&q=landing-012 | 14:48 |
racarr | alan_g: The whole issue I guess is...the way the osk hides is it does this animation | 14:49 |
duflu_ | alf_: Yeah, not without LP enhancements, which take decades :) Just using "Mir" is OK | 14:49 |
racarr | and then it sets the minimize state | 14:49 |
racarr | and qtmir hides it. | 14:49 |
racarr | so there is an existing bug where sometimes the animation does not finish | 14:49 |
racarr | (it mostly seems to show up on n7...) | 14:49 |
racarr | and hten you will see | 14:49 |
racarr | a little bit of the OSK left over at the bottom | 14:49 |
racarr | because the | 14:50 |
racarr | well, the hiding doesnt work | 14:50 |
=== marcusto_ is now known as marcustomlinson_ | ||
=== marcustomlinson_ is now known as marcustomlinson | ||
racarr | alan_g: We are seeing it may actually just be the wrong version | 14:55 |
racarr | of the package in the silo....so | 14:55 |
racarr | maybe don't spend any time on it now lol...sorry | 14:55 |
alan_g | racarr: OK. Keep me informed. | 14:56 |
racarr | alan_g: Thanks :) | 14:57 |
racarr | alan_g: Ah its still broken even with the correct packages :( (featuring the metatype registration) | 15:54 |
alan_g | racarr: :( | 15:56 |
=== marcusto_ is now known as marcustomlinson_ | ||
greyback_ | mlankhorst: lp:~mir-team/qtmir/port-to-event-2.0/ should enable mouse hover - it needs mir 0.11 though | 16:40 |
alan_g | racarr: I'm still not sure what your saying is going wrong. Is the server is getting the state change notification at set_surface_attribute() but not publishing the surfaceAttributeChanged event? Or is the there something not happening client side? | 16:53 |
mlankhorst | greyback_: oke, last I tried it was borked, is it fixed now? | 17:27 |
mlankhorst | greyback_: I was getting touch events rather than pointer events even with that branch :/ | 17:29 |
=== chihchun is now known as chihchun_afk |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!