On Fri, May 16, 2014 at 2:49 AM, Thomas Voß <1319...@bugs.launchpad.net> wrote:
> Trying to clarify on your latest comment: With app you are referring to
> an actual application running in the emulator? If so, I would propose to
> start digging into qtubuntu and see if we fail to call
> eglDestroyS
** Changed in: mir
Importance: Undecided => High
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1319582
Title:
'Failed to start RenderThread' after opening/closing applications
To manage notifica
Obviously you shouldn't use SIGKILL unless you've concluded an app is
broken and not responding. The polite thing to do before that is SIGTERM
(which is at least catchable).
I agree Mir needs some kind of "quit" event. We need something to
communicate from the shell to the app/surface that the use
Trying to clarify on your latest comment: With app you are referring to
an actual application running in the emulator? If so, I would propose to
start digging into qtubuntu and see if we fail to call
eglDestroySurface. Mid-term, I would think that Mir/UnityMir should make
sure that all resources ar
After discussing this bug with ricmm, we came down into two possible solutions:
1 - Changing Mir/Unity8 and app lifecycle to gracefully terminate the app
instead of sending sigkill;
2 - Changing Mir to make sure that every surface used is properly released;
So, opening a bugtask for mir as well.
Recording irc convo...
rsalveti, the buffer that comes out of the OMX layer that should
contain the video metadata is corrupt
rsalveti, and so when the encoder tries to detect what type of format
is embedded in that metadata, it segfaults
rsalveti, but it's deeper than that...when I try and f
The bug happens because eglDestroySurface never gets called when the app
is closed. That function is the one responsible to tell the emulator
stack that the surface should be destroyed.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
h
After adding a bit of more debug entries, it's clear that the window
surface gets allocated when the app is instantiated, but it's never
released.
When opening a new app:
Framebuffer::createWindowSurface
Framebuffer::createColorBuffer
When destrying the app:
Framebuffer::closeColorBuffer
In this
The error actually happens when RendorControl tries to start a new thread
(host), so it can handle the new GLES surface:
shared/OpenglOsUtils/osThreadUnix.cpp
bool
Thread::start()
{
pthread_mutex_lock(&m_lock);
m_isRunning = true;
int ret = pthread_create(&m_thread, NULL, Thread::threa