27064 - BSD: Mozila crashes on startup

Status: VERIFIED INVALID

(Core :: XUL, defect, P3)

Description

[Pascal Hofstee]

On my FreeBSD-4.0-CURRENT with a recent CVS-Mozilla i get the following crash on
startup:

#0  0x29dfb7c8 in nsGCCache::GetGCFromDW (this=0x80e50e0, window=0x82f4380,
gcv=0xbfbfe8a0, flags=32777, clipRegion=0x0)
    at ../../../../gfx/src/gtk/nsGCCache.cpp:122
#1  0x29e02631 in nsGCCache::GetClipGC (this=0x80e50e0, window=0x82f4380,
gcv=0xbfbfe8a0, flags=32777, clipRegion=0x0)
    at ../../../../gfx/src/gtk/nsGCCache.h:51
#2  0x29dfca67 in nsRenderingContextGTK::UpdateGC (this=0x8373f00) at
../../../../gfx/src/gtk/nsRenderingContextGTK.cpp:503
#3  0x29dfd32f in nsRenderingContextGTK::CreateDrawingSurface (this=0x8373f00,
aBounds=0xbfbfe9a8, aSurfFlags=0, 
    aSurface=@0x28b761a4) at
../../../../gfx/src/gtk/nsRenderingContextGTK.cpp:733
#4  0x28b65b70 in nsViewManager::GetDrawingSurface (this=0x82c5e00,
aContext=@0x8373f00, aBounds=@0xbfbfea98)
    at ../../../view/src/nsViewManager.cpp:2188
#5  0x28b60178 in nsViewManager::Refresh (this=0x82c5e00, aView=0x82e3f80,
aContext=0x8373f00, rect=0xbfbfeb68, aUpdateFlags=1)
    at ../../../view/src/nsViewManager.cpp:585
#6  0x28b63f6e in nsViewManager::DispatchEvent (this=0x82c5e00,
aEvent=0xbfbfed44, aStatus=0xbfbfebd4)
    at ../../../view/src/nsViewManager.cpp:1615
#7  0x28b52de6 in HandleEvent (aEvent=0xbfbfed44) at
../../../view/src/nsView.cpp:68
#8  0x29da27b2 in nsWidget::DispatchEvent (this=0x82c4600, aEvent=0xbfbfed44,
aStatus=@0xbfbfecd4)
    at ../../../../widget/src/gtk/nsWidget.cpp:1370
#9  0x29da2359 in nsWidget::DispatchWindowEvent (this=0x82c4600,
event=0xbfbfed44) at ../../../../widget/src/gtk/nsWidget.cpp:1261
#10 0x29da830b in nsWindow::DoPaint (this=0x82c4600, aX=0, aY=0, aWidth=2,
aHeight=1, aClipRegion=0x82eb1e0)
    at ../../../../widget/src/gtk/nsWindow.cpp:577
#11 0x29da84de in nsWindow::Update (this=0x82c4600) at
../../../../widget/src/gtk/nsWindow.cpp:613
#12 0x28b63876 in nsViewManager::Composite (this=0x82c5e00) at
../../../view/src/nsViewManager.cpp:1428
#13 0x28b5e79c in vm_timer_callback (aTimer=0x8206f40, aClosure=0x82c5e00) at
../../../view/src/nsViewManager.cpp:88
#14 0x288bfe85 in nsTimerGtk::FireTimeout (this=0x8206f40) at
../../../../../../widget/timer/src/unix/gtk/nsTimerGtk.cpp:48
#15 0x288c04e1 in nsTimerExpired (aCallData=0x8206f40) at
../../../../../../widget/timer/src/unix/gtk/nsTimerGtk.cpp:165
#16 0x28a2c757 in g_timeout_dispatch () from /usr/local/lib/libglib12.so.3
#17 0x28a2b8df in g_main_dispatch () from /usr/local/lib/libglib12.so.3
#18 0x28a2bef8 in g_main_iterate () from /usr/local/lib/libglib12.so.3
#19 0x28a2c090 in g_main_run () from /usr/local/lib/libglib12.so.3
#20 0x2894faf7 in gtk_main () from /usr/X11R6/lib/libgtk12.so.2
#21 0x29d85c92 in nsAppShell::Run (this=0x81d0290) at
../../../../widget/src/gtk/nsAppShell.cpp:304
#22 0x296625a7 in nsAppShellService::Run (this=0x81a0000) at
../../../../xpfe/appshell/src/nsAppShellService.cpp:403
#23 0x2960b170 in nsProfile::LoadDefaultProfileDir (this=0x820b0f0,
profileURLStr=@0xbfbff238)
    at ../../../profile/src/nsProfile.cpp:314
#24 0x2960ab98 in nsProfile::StartupWithArgs (this=0x820b0f0,
cmdLineArgs=0x819ffc0) at ../../../profile/src/nsProfile.cpp:233
#25 0x804d8e8 in main1 (argc=1, argv=0xbfbff45c, splashScreen=0x0) at
../../../xpfe/bootstrap/nsAppRunner.cpp:520
#26 0x804e3e6 in main (argc=1, argv=0xbfbff45c) at
../../../xpfe/bootstrap/nsAppRunner.cpp:677

Comment 1

[cbegle]

assigning to pav since from 
http://bonsai.mozilla.org/cvsblame.cgi?file=mozilla/gfx/src/gtk/nsGCCache.cpp
it looks like it's his.

Comment 2

[Stuart Parmenter]

change component, give real qa contact.  someone else was seeing this with gcc
2.95.2 (??) i think.

Comment 3

[cls]

Yeah, that was me.  For giggles, I dropped down to egcs 1.1.2 yesterday and that
"fixed" the problem.  I don't know what gcc bug we exposed but I used to have a
working gcc 2.95.2 build before I got wrapped up in moving (about 3-4 weeks
ago).

Comment 4

[Stuart Parmenter]

So the difference I see is that I have binutils 2.9.5.0.16 and rginda has
2.9.1.0.32.

Comment 5

[Stuart Parmenter]

So it looks like if you are using gcc 2.95.2 and an old version of binutils like
2.9.1, you will have problems.  You need to upgrade binutils.

Comment 6

[leger]

Putting crash in the keyword field.

Comment 7

[Pascal Hofstee]

The bug here is triggered by me on a FreeBSD-4.0-CURRENT system ... which means
if my GCC (2.95.2 btw) is updated so will (should) be my binutils.

Comment 8

[Stuart Parmenter]

could you please verify what version of binutils you have?   run 'as --version'

Comment 9

[Pascal Hofstee]

ok... as --version results in:

GNU assembler 2.9.1

Looks like i am in trouble then huh ?

Comment 10

[Stuart Parmenter]

yup.  you want to upgrade.

Comment 11

[Stuart Parmenter]

you can't use new gcc and old binutils.  won't fix.

Comment 12

[Claudius Gayle]

after careful consideration, marking this bug VERIFIED

Comment 13

[Stuart Parmenter]

*** Bug 27556 has been marked as a duplicate of this bug. ***

Comment 14

[Stuart Parmenter]

*** Bug 27556 has been marked as a duplicate of this bug. ***

Comment 15

[Pascal Hofstee]

Ok ... I know it's probably a gcc-2.95.2/gas-2.9.1 issue but still i have one
more question ....

How come i CAN build and successfully run the M13-code with this same
compiler/assembler-combo .... I assume at least Something changed in the
nsGCCache-code then ?

Not asking to re-think this bugs' status ... just curious