Open
Bug 584162
Opened 14 years ago
Updated 2 years ago
Running SSL WebSocket test crashes Firefox on debug Windows only
Categories
(Core :: General, defect)
Tracking
()
NEW
People
(Reporter: jgriffin, Unassigned)
References
(Blocks 1 open bug)
Details
When you apply the patch https://bug573803.bugzilla.mozilla.org/attachment.cgi?id=462476 from bug 573803, which enables SSL WebSocket tests in mochitest, and run the test content/base/test/test_websocket_ssl_hello.html, Firefox will assert and terminate, but only if you're running a debug Windows build with --enable-libxul (the same config used by tinderbox). This problem isn't seen on Mac or Linux, nor on Windows opt builds, nor on Windows debug builds without --enable-libxul. The stack trace of the assert is: KernelBase.dll!76fd22a1() [Frames below may be incorrect and/or missing, no symbols loaded for KernelBase.dll] nspr4.dll!PR_Assert(const char * s=0x681f1c8c, const char * file=0x681f1cb0, int ln=322) Line 563 C nspr4.dll!_PR_MD_PR_POLL(PRPollDesc * pds=0x09bb7980, int npds=2, unsigned int timeout=65535000) Line 322 + 0x1c bytes C nspr4.dll!PR_Poll(PRPollDesc * pds=0x09bb7980, int npds=2, unsigned int timeout=65535000) Line 173 + 0x11 bytes C xul.dll!nsSocketTransportService::Poll(int wait=1, unsigned int * interval=0x0a01fc88) Line 357 + 0x12 bytes C++ xul.dll!nsSocketTransportService::DoPollIteration(int wait=1) Line 668 + 0x10 bytes C++ xul.dll!nsSocketTransportService::OnProcessNextEvent(nsIThreadInternal * thread=0x09c41058, int mayWait=1, unsigned int depth=1) Line 547 + 0xd bytes C++ xul.dll!nsThread::ProcessNextEvent(int mayWait=1, int * result=0x0a01fd0c) Line 519 C++ xul.dll!NS_ProcessNextEvent_P(nsIThread * thread=0x09c41058, int mayWait=1) Line 250 + 0x16 bytes C++ xul.dll!nsSocketTransportService::Run() Line 589 + 0xb bytes C++ xul.dll!nsThread::ProcessNextEvent(int mayWait=1, int * result=0x0a01fdac) Line 547 + 0x19 bytes C++ xul.dll!NS_ProcessNextEvent_P(nsIThread * thread=0x09c41058, int mayWait=1) Line 250 + 0x16 bytes C++ xul.dll!nsThread::ThreadFunc(void * arg=0x09c41058) Line 263 + 0xb bytes C++ nspr4.dll!_PR_NativeRunThread(void * arg=0x00a46c28) Line 426 + 0xf bytes C nspr4.dll!pr_root(void * arg=0x00a46c28) Line 122 + 0xf bytes C msvcr80d.dll!_callthreadstartex() Line 348 + 0xf bytes C msvcr80d.dll!_threadstartex(void * ptd=0x09bc8c38) Line 331 C kernel32.dll!76af3677() ntdll.dll!77cf9d42() ntdll.dll!77cf9d15() The mozconfig I used to reproduce this is: ac_add_options --enable-application=browser ac_add_options --enable-jemalloc ac_add_options --disable-optimize ac_add_options --enable-debug ac_add_options --enable-libxul ac_add_options --enable-trace-malloc ac_add_options --enable-tests # For NSS symbols export MOZ_DEBUG_SYMBOLS=1 # Needed to enable breakpad in application.ini export MOZILLA_OFFICIAL=1 The assert seems to happen as soon as the WebSocket server returns HTTP 200 to the original HTTP CONNECT request: CONNECT example.com:443 HTTP/1.1 Host: example.com:443 HTTP/1.1 200 Connected Connection: keep-alive The problem is not 100% reproducible on my machine, it seems to happen around 50% of the time.
Updated•2 years ago
|
Severity: normal → S3
You need to log in
before you can comment on or make changes to this bug.
Description
•