WebKit Bugzilla
New
Browse
Log In
×
Sign in with GitHub
or
Remember my login
Create Account
·
Forgot Password
Forgotten password account recovery
RESOLVED WORKSFORME
Bug 16807
gmail crashes webkit browsers
https://bugs.webkit.org/show_bug.cgi?id=16807
Summary
gmail crashes webkit browsers
Tom Wright
Reported
2008-01-09 14:49:03 PST
gmail will crash any webkit browser i have tried it with (epiphany-webkit, midori, the gtk test browser) webkit is definetly at the latest version (just checkted out of SVN and compiled) i am running ubuntu 7.10
Attachments
crash dump from XP gmail crash
(43.81 KB, text/plain)
2008-01-09 19:20 PST
,
Cameo Wood
no flags
Details
stack-trace webkit r30868, OpenSolaris, GtkLauncher & Epiphany
(15.23 KB, text/plain)
2008-03-07 00:38 PST
,
kappasolaris
no flags
Details
View All
Add attachment
proposed patch, testcase, etc.
Pierre-Luc Beaudoin
Comment 1
2008-01-09 15:38:04 PST
On
r29341
, I don't crash but I get 'problem with the SSL CA cert (path? access rights?)'. What is the SVN version you used? Did you try it with the GtkLauncher built from your SVN checkout? Can you rebuild it with debug and see if you hit an assertion?
Cameo Wood
Comment 2
2008-01-09 19:05:18 PST
I've tested the lastest nightly (WebKit-
r29336
) on XP, and this is confirmed. I'm attaching a crash dump from xp.
Cameo Wood
Comment 3
2008-01-09 19:20:04 PST
Created
attachment 18358
[details]
crash dump from XP gmail crash
Mark Rowe (bdash)
Comment 4
2008-01-09 20:46:08 PST
I'm not sure that the initial problem reported on Linux is the same as what Cameo mentioned seeing on Windows. It's possible they're unrelated.
Mark Rowe (bdash)
Comment 5
2008-01-09 20:47:01 PST
Tom, are you able to provide a stack trace for the crash you're seeing on Linux? That would allow us to confirm whether the Windows issue Cameo mentioned is in fact the same crash.
Tom Wright
Comment 6
2008-01-10 08:55:48 PST
same problem with gtklauncher i am building with debug now it outputs this error message to the command line on crash: tom@tom-laptop:~/Desktop/webkit/WebKit/WebKitTools/Scripts$ ./run-launcher --gtk Starting webkit launcher with LD_LIBRARY_PATH set to point to built WebKit in /home/tom/Desktop/webkit/WebKit/WebKitBuild/Release/lib. console message:
https://ssl.google-analytics.com/urchin_beta.js
@398: ReferenceError: Can't find variable: ActiveXObject Segmentation fault (core dumped)
Tom Wright
Comment 7
2008-01-10 09:38:41 PST
sorry, i can't figure out how generate a stack trace (i don't have much programming/debugging experience)
Adam Roben (:aroben)
Comment 8
2008-01-10 13:40:07 PST
(In reply to
comment #3
)
> Created an attachment (id=18358) [edit] > crash dump from XP gmail crash
Which nightly were you using when you got this crash?
Cameo Wood
Comment 9
2008-01-10 13:41:46 PST
(In reply to
comment #8
)
> (In reply to
comment #3
) > > Created an attachment (id=18358) [edit] > > crash dump from XP gmail crash > > Which nightly were you using when you got this crash? >
WebKit-
r29336
Adam Roben (:aroben)
Comment 10
2008-01-10 13:48:34 PST
(In reply to
comment #4
)
> I'm not sure that the initial problem reported on Linux is the same as what > Cameo mentioned seeing on Windows. It's possible they're unrelated.
Looking at the backtrace, I think Mark is right that the crashes are unrelated:
> kernel32.dll!_c_PmapEntries_mpr() - 0x3032f bytes
WebKit.dll!__delayLoadHelper2(const ImgDelayDescr * pidd=0xc06d007e, int (void)* * ppfnIATEntry=0x00000000) Line 331 C++ WebKit.dll!__delayLoadHelper2(const ImgDelayDescr * pidd=0x00000028, int (void)* * ppfnIATEntry=0x0012e768) Line 331 C++ WebKit.dll!__tailMerge_QTMovieWin_dll() + 0xd bytes Unknown It looks like the nightly needs to either have <video> support disabled or include QtMoveWin.dll
Adam Roben (:aroben)
Comment 11
2008-01-10 13:50:48 PST
The Windows issue is now covered by
bug 16825
.
kappasolaris
Comment 12
2008-03-07 00:33:03 PST
I'm also having issues with gmail & latest webkit (latest I tried this on is
r30868
): partial trace: core 'core' of 36: ./GtkLauncher fe7e6e57 bool WebCore::JSDOMWindowBase::allowsAccessFromPrivate(const KJS::ExecState*,WebCore::SecurityOrigin::Reason&)const (f99f1fa0, 80459d0, 804579c, fe7e239c, f8705368, fea99f66) + 67 fe7e23b5 bool WebCore::JSDOMWindowBase::allowsAccessFromNoErrorMessage(KJS::ExecState*)const (f99f1fa0, 80459d0, 80457e8, fe7e774c) + 25 fe7e782a bool WebCore::JSDOMWindow::customGetOwnPropertySlot(KJS::ExecState*,const KJS::Identifier&,KJS::PropertySlot&) (f99f1fa0, 80459d0, f8703ff8, 8045820) + ea fe6e8fe4 bool WebCore::JSDOMWindow::getOwnPropertySlot(KJS::ExecState*,const KJS::Identifier&,KJS::PropertySlot&) (f99f1fa0, 80459d0, f8703ff8, 8045820) + 24 fe697c69 KJS::JSValue*KJS::ResolveNode::evaluate(KJS::ExecState*) (f8703ff0, 80459d0, 8045878, fe6dc1de) + 49 fe6dc226 KJS::JSValue*KJS::BracketAccessorNode::inlineEvaluate(KJS::ExecState*) (f8703fe0, 80459d0, 80458c8, fe698adc, f9e42600) + 56 fe698af1 KJS::JSValue*KJS::BracketAccessorNode::evaluate(KJS::ExecState*) (f8703fe0, 80459d0, 8045908, fe6a200e) + 21 fe6a204e bool KJS::LessNode::evaluateToBoolean(KJS::ExecState*) (f8703fd0, 80459d0, f8703c9c, fe6a90fe) + 4e fe6a9124 KJS::JSValue*KJS::WhileNode::execute(KJS::ExecState*) (f87058e8, 80459d0, 8045998, fe6a8c1e) + 34 fe6a8c55 KJS::JSValue*KJS::BlockNode::execute(KJS::ExecState*) (f87046c0, 80459d0, 8045a38, fe6abdec, 80459d0) + 45 fe6abe0f KJS::JSValue*KJS::FunctionBodyNode::execute(KJS::ExecState*) (f87046c0, 80459d0, 8045ac0, fe65d1ae) + 2f fe65d1f4 KJS::JSValue*KJS::FunctionImp::callAsFunction(KJS::ExecState*,KJS::JSObject*,const KJS::List&) (f84f4780, 8045c30, f805ac20, 8045ac0) + 54 fe6b32b1 KJS::JSValue*KJS::JSObject::call(KJS::ExecState*,KJS::JSObject*,const KJS::List&) (f84f4780, 8045c30, f805ac20, 8045ac0) + 41 fe6db156 KJS::JSValue*KJS::FunctionCallDotNode::inlineEvaluate(KJS::ExecState*) (f8705558, 8045c30, 8045b68, fe69a7cc, f99f1fa0) + 1b6 fe69a7e1 KJS::JSValue*KJS::FunctionCallDotNode::evaluate(KJS::ExecState*) (f8705558, 8045c30, f99f1fa0, fe6a8cda) + 21 fe6a8cf5 KJS::JSValue*KJS::ExprStatementNode::execute(KJS::ExecState*) (f8705540, 8045c30, f8703b9c, fe6a8c1e) + 25 fe6a8c55 KJS::JSValue*KJS::BlockNode::execute(KJS::ExecState*) (f9e42680, 8045c30, f86fe708, fe6a8eca) + 45 fe6a8f04 KJS::JSValue*KJS::IfElseNode::execute(KJS::ExecState*) (f9e42640, 8045c30, 8045bf8, fe6a8c1e) + 44 fe6a8c55 KJS::JSValue*KJS::BlockNode::execute(KJS::ExecState*) (f87047e0, 8045c30, 8045c98, fe6abdec, 8045c30) + 45 fe6abe0f KJS::JSValue*KJS::FunctionBodyNode::execute(KJS::ExecState*) (f87047e0, 8045c30, 8045d20, fe65d1ae) + 2f fe65d1f4 KJS::JSValue*KJS::FunctionImp::callAsFunction(KJS::ExecState*,KJS::JSObject*,const KJS::List&) (f84f4720, 8045e30, f805ac20, 8045d20) + 54 fe6b32b1 KJS::JSValue*KJS::JSObject::call(KJS::ExecState*,KJS::JSObject*,const KJS::List&) (f84f4720, 8045e30, f805ac20, 8045d20) + 41 fe6db156 KJS::JSValue*KJS::FunctionCallDotNode::inlineEvaluate(KJS::ExecState*) (f8701e10, 8045e30, 8045de8, fe69a7cc, f8701e28) + 1b6 fe69a7e1 KJS::JSValue*KJS::FunctionCallDotNode::evaluate(KJS::ExecState*) (f8701e10, 8045e30, f76d69a0, fe6a8cda) + 21 fe6a8cf5 KJS::JSValue*KJS::ExprStatementNode::execute(KJS::ExecState*) (f8701df8, 8045e30, 8045df8, fe6a8c1e) + 25 fe6a8c55 KJS::JSValue*KJS::BlockNode::execute(KJS::ExecState*) (f8704c60, 8045e30, 8045e98, fe6abdec, 8045e30) + 45 fe6abe0f KJS::JSValue*KJS::FunctionBodyNode::execute(KJS::ExecState*) (f8704c60, 8045e30, 8045f20, fe65d1ae) + 2f
kappasolaris
Comment 13
2008-03-07 00:38:57 PST
Created
attachment 19581
[details]
stack-trace webkit
r30868
, OpenSolaris, GtkLauncher & Epiphany This happens when you are Signing in to gmail, you see "Loading..." to the left followed at 90% of progress with a crash.
Luka Napotnik
Comment 14
2009-07-31 14:10:44 PDT
I can't reproduce this using the latest WebKit.
Jan Alonzo
Comment 15
2009-07-31 16:20:00 PDT
I can't reproduce this or it's already been fixed in recent WebKitGtk releases. Closing. If this still happens, feel free to reopen and we also appreciate if you can attach a crash dump. Thanks!
Note
You need to
log in
before you can comment on or make changes to this bug.
Top of Page
Format For Printing
XML
Clone This Bug