Merge lp://staging/~thumper/unity/regional-damage-sru into lp://staging/unity/5.0

Proposed by Tim Penhey
Status: Merged
Approved by: Łukasz Zemczak
Approved revision: no longer in the source branch.
Merged at revision: 2394
Proposed branch: lp://staging/~thumper/unity/regional-damage-sru
Merge into: lp://staging/unity/5.0
Diff against target: 784 lines (+353/-117)
12 files modified
manual-tests/Dash.txt (+15/-0)
manual-tests/Launcher.txt (+15/-0)
manual-tests/Panel.txt (+17/-0)
manual-tests/WindowManagement.txt (+20/-0)
plugins/unityshell/src/AbstractLauncherIcon.h (+1/-0)
plugins/unityshell/src/Launcher.cpp (+11/-0)
plugins/unityshell/src/Launcher.h (+4/-0)
plugins/unityshell/src/LauncherIcon.cpp (+3/-0)
plugins/unityshell/src/PanelController.cpp (+15/-0)
plugins/unityshell/src/PanelController.h (+1/-0)
plugins/unityshell/src/unityshell.cpp (+239/-110)
plugins/unityshell/src/unityshell.h (+12/-7)
To merge this branch: bzr merge lp://staging/~thumper/unity/regional-damage-sru
Reviewer Review Type Date Requested Status
Łukasz Zemczak Approve
Marco Trevisan (Treviño) Approve
Daniel van Vugt Approve
Review via email: mp+116796@code.staging.launchpad.net

Commit message

Backport the fixes for regional damage, and the regressions that were introduced by it.
Fixes lots of bugs.

Description of the change

This branch backports the fixes for regional damage, and the regression that were introduced by it.

Fixes lots of bugs.

To post a comment you must log in.
Revision history for this message
Daniel van Vugt (vanvugt) wrote :

Please note: This proposal has two external prerequisite fixes:
  bug 1012956 (already fixed in lp:compiz-core/0.9.7)
  bug 1014610 (already fixed in lp:nux/2.0)

It appears neither of these fixes have hit ubuntu yet.

Revision history for this message
Daniel van Vugt (vanvugt) wrote :

Looks OK but keeps crashing when I run it on nvidia-295.49. The crash does not happen with lp:unity/5.0 !?

(gdb) bt
#0 0x00007fb402b65e1e in ?? () from /usr/lib/nvidia-current-updates/libnvidia-glcore.so.295.49
#1 0x00007fb402c09da1 in ?? () from /usr/lib/nvidia-current-updates/libnvidia-glcore.so.295.49
#2 0x00007fb402e0ada1 in ?? () from /usr/lib/nvidia-current-updates/libnvidia-glcore.so.295.49
#3 0x00007fb402e081f3 in ?? () from /usr/lib/nvidia-current-updates/libnvidia-glcore.so.295.49
#4 0x00007fb3ff4df22a in nux::GraphicsEngine::QRP_GLSL_1Tex(int, int, int, int, nux::ObjectPtr<nux::IOpenGLBaseTexture>, nux::TexCoordXForm&, nux::color::Color const&) ()
   from /usr/lib/libnux-graphics-2.0.so.0
#5 0x00007fb3ff4c98e9 in nux::GraphicsEngine::QRP_1Tex(int, int, int, int, nux::ObjectPtr<nux::IOpenGLBaseTexture>, nux::TexCoordXForm&, nux::color::Color const&) ()
   from /usr/lib/libnux-graphics-2.0.so.0
#6 0x00007fb3ff822891 in nux::WindowCompositor::PresentBufferToScreen(nux::ObjectPtr<nux::IOpenGLBaseTexture>, int, int, bool, bool, float, bool) () from /usr/lib/libnux-2.0.so.0
#7 0x00007fb3ff825990 in nux::WindowCompositor::RenderTopViews(bool, std::list<nux::ObjectWeakPtr<nux::BaseWindow>, std::allocator<nux::ObjectWeakPtr<nux::BaseWindow> > >&, bool) ()
   from /usr/lib/libnux-2.0.so.0
#8 0x00007fb3ff825cca in nux::WindowCompositor::Draw(bool, bool) () from /usr/lib/libnux-2.0.so.0
#9 0x00007fb3ff82ccb1 in nux::WindowThread::RenderInterfaceFromForeignCmd(nux::Rect*) () from /usr/lib/libnux-2.0.so.0
#10 0x00007fb4003e8cc4 in unity::UnityScreen::paintDisplay (this=0x1f07220, region=..., transform=..., mask=65536) at /home/dan/bzr/unity/tim/plugins/unityshell/src/unityshell.cpp:829

review: Needs Fixing
Revision history for this message
Daniel van Vugt (vanvugt) wrote :

And yes I did try compiling the latest lp:nux/2.0 as well. But when I use it, I get NO readable stacktrace of the crash.

Revision history for this message
Daniel van Vugt (vanvugt) wrote :

Uninstalled nvidia-295.49 and tested with nouveau. Now it works well.

Can anyone else confirm the nvidia crash?

review: Needs Information
Revision history for this message
Daniel van Vugt (vanvugt) wrote :

Verified; works on intel too. Why is only nvidia broken?!

Revision history for this message
Brenton Hall (brent-hall-27) wrote :

Compiled it. Running fine on ATI, using the open drivers.

Revision history for this message
Brenton Hall (brent-hall-27) wrote :

Been through the new manual tests for this branch. They all passed for ATI.

Revision history for this message
Daniel van Vugt (vanvugt) wrote :

Tested Unit 6.0 and Nux 3.0 again. They now crash in the same way on nvidia. So it's nothing to do with this branch.

I also upgraded nvidia 295.40 to 295.49 but no difference (other than the stack traces are now completely unreadable).

I suspect the nvidia problem could be related to bug 982485 or bug 980298. When run under valgrind, I get a rubbish stack, but it happens immediately after bug 988880, which might be relevant.

Anyway, this branch seems good but I can't test it on nvidia-current/nvidia-current-updates.

review: Approve
Revision history for this message
Daniel van Vugt (vanvugt) wrote :

Actually, bug 982626 seems to describe the exact crash I'm getting now. Let's make that a separate discussion to this.

Revision history for this message
Marco Trevisan (Treviño) (3v1n0) wrote :

I can confirm the crash on a Nvidia GeForce GT 650M, the first time I launched this unity branch it worked and it also seemed to fix the bug #915265, however it hanged and after reloading compiz (one or two times), I can't run it anymore.

Gdb complains issues related to the nvidia bloab, so... It's not helpful at all.

The same happens with trunk with this change enabled.

Revision history for this message
Łukasz Zemczak (sil2100) wrote :

I would probably like to have this fixed before releasing this as an SRU. I know we have the same bug in trunk and quantal, but Precise is a LTS, so we probably shouldn't release anything that has any known ill side-effects. So fixing bug #982626 becomes rather a priority here.

Revision history for this message
Marco Trevisan (Treviño) (3v1n0) wrote :

FYI this is the stacktrace of the crash that I'm getting http://paste.ubuntu.com/1121725/

Revision history for this message
Marco Trevisan (Treviño) (3v1n0) wrote :

A weird thing about this crash is that if I replace an unity sesssion, from unity it will always crash.
If I replace it after a tty switch, then there's no crash...

Revision history for this message
Marco Trevisan (Treviño) (3v1n0) wrote :

Using a bad workaround like this, seems to bypass the crash... http://paste.ubuntu.com/1122495/
Checking something better.

Revision history for this message
Marco Trevisan (Treviño) (3v1n0) wrote :

Ok, thanks to the Sam's tip, I got the proper fix for the crash that was affecting the nvidia cards (using the nvidia driver).

I've made a branch that fixes this, and that uses this one as prerequisite: lp:~3v1n0/unity/fix-nvidia-glDrawArrays-crash-5.0/+merge/117553

So, at this point I can approve this branch too.

review: Approve
Revision history for this message
Łukasz Zemczak (sil2100) wrote :

Since I already tested the branch, along with Alan and the Nice guys, I think it's safe to apply as the nvidia bug is getting handled as well.

review: Approve
Revision history for this message
Unity Merger (unity-merger) wrote :

No commit message specified.

Preview Diff

[H/L] Next/Prev Comment, [J/K] Next/Prev File, [N/P] Next/Prev Hunk
The diff is not available at this time. You can reload the page or download it.

Subscribers

People subscribed via source and target branches

to all changes: