Bug 186279

Summary: metacity crashes on enabling transparency
Product: [Fedora] Fedora Reporter: Andrea Santilli <sacntct>
Component: metacityAssignee: Søren Sandmann Pedersen <sandmann>
Status: CLOSED WONTFIX QA Contact:
Severity: medium Docs Contact:
Priority: medium    
Version: 5CC: felipe_alfaro, filip.tsachev, kem, marius.andreiana, ralph+rh-bugzilla
Target Milestone: ---   
Target Release: ---   
Hardware: i386   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2006-09-08 02:48:06 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
metacity 2.14.0-1 backtrace none

Description Andrea Santilli 2006-03-22 17:52:47 UTC 
Description of problem:
I tested both aiglx (a bit slow) and exa compositions with kwin and everything
is ok.
Then i wanted to test the new effects with metacity but as soon as I do:
gconftool-2 -s '/apps/metacity/general/compositing_manager' --type bool true
metacity crashes.
Starting metacity from xterm, when i enable the composition and the wm crashes
there can be read:

composite all adds window (this line is repeated a lot of times)
....
composite all ends
Avviso del window manager: Log level 6: Unexpected X error:
GLXUnsupportedPrivateRequest serial 1612 error_code 162 request_code 143
minor_code 16)
Please notice that this happens on both aiglx and exa.


Version-Release number of selected component (if applicable): 2.14.0-1


How reproducible:
Always


Steps to Reproduce:
1. Enable exa or aiglx and restart X
2. login in gnome and enable metacity's compositor by gconf2-editor
3. metacity crashes (you can also see metacity restarting and re-crashing for a
while)
  
Actual results:
metacity crashes on composition

Expected results:
to not crash

Additional info:
I don't know if this can help you. My videocard is a
Sapphire Radeon 9000 Atlantis [RV250]
Comment 1 Andrea Santilli 2006-03-23 11:26:41 UTC 
i cant do the backtrace because gdb locks the screen until i kill it
Comment 2 Andrea Santilli 2006-03-23 12:30:25 UTC 
Created attachment 126539 [details]
metacity 2.14.0-1 backtrace 

I've managed to get the backtrace from metacity.
Comment 3 Marius Andreiana 2006-03-26 14:35:39 UTC 
I confirm the same behaviour (FC5, same metacity version). 

I've added this to xorg.conf:
Section "Extensions"
        Option "Composite"      "Enable"
EndSection
Comment 4 Andrea Santilli 2006-03-27 19:56:47 UTC 
This bug seems to be metacity-only related: as i said before kwin works
perfectly; moreover i've rebuilt the xfwm rpm to include composite support and
it works good too.
Comment 5 Andrea Santilli 2006-03-30 12:45:38 UTC 
I've managed to make Metacity work (still buggy) following the instructions
described at http://fedoraproject.org/wiki/RenderingProject/aiglx
I've installed the Xair server, changed the /usr/bin/X link from Xorg to Xair,
added the lines:
Option      "RenderAccel" "true"
Option	    "AccelMethod" "aiglx"
Option	    "XAANoOffscreenPixmaps"
in the Device section and enabled the Composite, RENDER and DAMAGE extensions.

After doing all this, metacity kept on crashing in runlevel 5 but worked when i
launched startx from command line (with wonderful but ugly and unconfigurable
effects)

So ive understood that the xserver included in FC5 isnt supposed to make aiglx
work but metacity crashes because there have been included some experimental
effect. Is it possible have plain metacity compositing effects? I mean such as
the ones included in kwin that just do shadows and transparent windows instead
of those ugly floating windows
Comment 6 Andrea Santilli 2006-04-09 16:10:02 UTC 
Metacity crashes even on a GeForce FX5200. So it isn't a hardware related problem.
Comment 7 Filip Tsachev 2006-04-10 11:51:43 UTC 
I can confirm with GB Radeon 9200, does it need "radeon" or "vesa" driver? I
think I tried both
Comment 8 Søren Sandmann Pedersen 2006-09-08 02:48:06 UTC 
Closing compositor related bugs