Open GoogleCodeExporter opened 9 years ago
PS! We have the same problem with our custom plugin
Original comment by janek.pr...@gmail.com
on 11 Feb 2010 at 7:09
Another crash report. Looks different.
Original comment by janek.pr...@gmail.com
on 11 Feb 2010 at 7:17
Attachments:
I believe this to be related to issue 24; could you try editing
src/NpapiPlugin/NPObjectAPI.cpp and adding a second Retain in the constructor?
Don't
release twice, just retain twice. See if that makes the issue go away. (yes,
I know
this is a hack, but let's see if we can confirm what the problem is)
Original comment by taxilian
on 11 Feb 2010 at 8:02
Tested. Still crashes. The crash log is different every time. For example:
Exception Type: EXC_BAD_ACCESS (SIGBUS)
Exception Codes: KERN_PROTECTION_FAILURE at 0x0000000000000005
Crashed Thread: 0 Dispatch queue: com.apple.main-thread
Thread 0 Crashed: Dispatch queue: com.apple.main-thread
0 com.FBDevTeam.FBTestPlugin 0x00616ed7 NP_Shutdown + 25
(np_macmain.cpp:69)
1 com.apple.WebKit.PluginHost 0x00005b77 0x1000 + 19319
2 com.apple.WebKit.PluginHost 0x00003fd3 0x1000 + 12243
3 com.apple.WebKit.PluginHost 0x0000d6bf 0x1000 + 50879
4 com.apple.WebKit.PluginHost 0x000071e5 0x1000 + 25061
5 com.apple.CoreFoundation 0x93515e31 __CFRunLoopDoBlocks + 353
6 com.apple.CoreFoundation 0x934c6de3 __CFRunLoopRun + 3715
7 com.apple.CoreFoundation 0x934c5864 CFRunLoopRunSpecific + 452
8 com.apple.CoreFoundation 0x934c5691 CFRunLoopRunInMode + 97
9 com.apple.HIToolbox 0x97a66f0c RunCurrentEventLoopInMode + 392
10 com.apple.HIToolbox 0x97a66cc3 ReceiveNextEventCommon + 354
11 com.apple.HIToolbox 0x97a66b48
BlockUntilNextEventMatchingListInMode + 81
12 com.apple.AppKit 0x957b1ac5 _DPSNextEvent + 847
13 com.apple.AppKit 0x957b1306 -[NSApplication
nextEventMatchingMask:untilDate:inMode:dequeue:] + 156
14 com.apple.AppKit 0x9577349f -[NSApplication run] + 821
15 com.apple.WebKit.PluginHost 0x0000625f 0x1000 + 21087
16 com.apple.WebKit.PluginHost 0x00002035 0x1000 + 4149
***
I also managed to crashed the plugin on initialization!
Exception Type: EXC_BAD_ACCESS (SIGBUS)
Exception Codes: KERN_PROTECTION_FAILURE at 0x0000000000000028
Crashed Thread: 0 Dispatch queue: com.apple.main-thread
Thread 0 Crashed: Dispatch queue: com.apple.main-thread
0 com.apple.HIToolbox 0x97a534a0 IsHandlerOverridden + 67
1 com.apple.HIToolbox 0x97a53415
HIView::IsOurHandlerOverridden(unsigned long, unsigned long,
unsigned long, bool) + 125
2 com.apple.HIToolbox 0x97a5412b
HIView::SendOwningWindowChanged(unsigned long,
OpaqueWindowPtr*, OpaqueWindowPtr*) + 51
3 com.apple.HIToolbox 0x97a540be
HIView::NotifySubtreeWindowChanged(OpaqueWindowPtr*,
OpaqueWindowPtr*, unsigned char) + 90
4 com.apple.HIToolbox 0x97a53ec1
HIView::SetWindowRef(OpaqueWindowPtr*) + 107
5 com.apple.HIToolbox 0x97a51664
WindowData::FinishConstruction(unsigned long, unsigned long
long, unsigned long, Rect const*, unsigned char*, unsigned int,
OpaqueWindowPtr*, OpaqueControlRef*) +
1642
6 com.apple.HIToolbox 0x97a50fb5
WindowData::Initialize(OpaqueEventRef*) + 555
7 com.apple.HIToolbox 0x97a50d49
AppleWindowDef::Initialize(OpaqueEventRef*) + 33
8 com.apple.HIToolbox 0x97a3a515
HIObject::HandleClassHIObjectEvent(OpaqueEventHandlerCallRef*, OpaqueEventRef*,
void*) + 229
9 com.apple.HIToolbox 0x97a3a3ad
HIObject::EventHook(OpaqueEventHandlerCallRef*,
OpaqueEventRef*, void*) + 55
10 com.apple.HIToolbox 0x97a39e29
DispatchEventToHandlers(EventTargetRec*,
OpaqueEventRef*, HandlerCallRec*) + 1567
11 com.apple.HIToolbox 0x97a390f0
SendEventToEventTargetInternal(OpaqueEventRef*,
OpaqueEventTargetRef*, HandlerCallRec*) + 411
12 com.apple.HIToolbox 0x97a38f4f SendEventToEventTargetWithOptions
+ 58
13 com.apple.HIToolbox 0x97a349dd HIObject::Create(__CFString
const*, OpaqueEventRef*,
HIObject**) + 311
14 com.apple.HIToolbox 0x97a34883 HIObjectCreate + 76
15 com.apple.HIToolbox 0x97a4bf01 NewWindowCommon(WindowData**,
unsigned long,
unsigned long long, WindowDefSpec const*, Rect const*, unsigned char const*,
unsigned char,
OpaqueWindowPtr*, long, unsigned int, unsigned short*, bool) + 1481
16 com.apple.HIToolbox 0x97a96af6
CreateNewWindowInternal(WindowDefSpec const*, unsigned
long, unsigned long long, Rect const*, OpaqueWindowPtr**) + 328
17 com.apple.WebKit.PluginHost 0x0000e0e9 0x1000 + 53481
18 com.apple.WebKit.PluginHost 0x00004145 0x1000 + 12613
19 com.apple.WebKit.PluginHost 0x000070e8 0x1000 + 24808
Original comment by janek.pr...@gmail.com
on 12 Feb 2010 at 5:29
If safari is running in 64 bit, and the plugin is being built only in 32 bit,
it'll likely not
work right. right-click on safari icon in applications, a checkbox off in
"open in 32-bit
mode" tells you if safari is in 64 or 32 bit. if it's in 64 bit, the plugins
will need to built
for, or include 64 bit code. flipping safari to open in 32 bit will probably
tell you if
that's the issue.
j.h.
Original comment by sleeples...@gmail.com
on 17 Feb 2010 at 6:57
The custom plugin is now compiled as:
32 - i386, ppc and SDK 10.4
64 - x86_64 and SDK 10.5
I don't see crashes in 32 bit mode anymore, but see in 64 bit mode. One tester
is also confirming it (claims
10.6.2 and 32b).
Initially I used SDK 10.6
Original comment by janek.pr...@gmail.com
on 17 Feb 2010 at 10:13
Re: claims 10.6.2 and 32b
Was incorrect I think. Crash report contains:
Code Type: X86-64 (Native)
Parent Process: WebKitPluginAgent [2612]
Date/Time: 2010-02-17 12:06:11.587 +0200
OS Version: Mac OS X 10.6.2 (10C540)
Could the problem be base SDK related?
Original comment by janek.pr...@gmail.com
on 17 Feb 2010 at 10:18
Original comment by georg.fritzsche
on 24 Mar 2010 at 1:07
Original comment by taxilian
on 9 Sep 2010 at 2:00
I can't reproduce this anymore; I think it may have been fixed in 1.2 by the
shared_ptr changed, as this could be related to some weird AutoPtr edge cases.
Can someone confirm whether or not this is gone?
Original comment by taxilian
on 16 Sep 2010 at 5:29
Unfortunately the bug still exists after the shared_ptr change. Same process of
steps to reproduce, though it doesn't seem to be consistent.
Original comment by amack...@gmail.com
on 20 Sep 2010 at 3:40
Original comment by taxilian
on 20 Sep 2010 at 3:49
Original issue reported on code.google.com by
janek.pr...@gmail.com
on 11 Feb 2010 at 7:07Attachments: