I'm running into an assert firing at startup of running `./mach awsy-test --quick` > xul.dll!mozilla::net::nsHttpConnectionMgr::nsHalfOpenSocket::OnOutputStreamReady(nsIAsyncOutputStream * out) Line 4463 C++ > xul.dll!mozilla::net::nsSocketOutputStream::OnSocketReady(nsresult condition) Line 554 C++ > xul.dll!mozilla::net::nsSocketTransport::OnSocketDetached(PRFileDesc * fd) Line 2373 C++ > xul.dll!mozilla::net::nsSocketTransportService::DetachSocket(mozilla::net::nsSocketTransportService::SocketContext * listHead, mozilla::net::nsSocketTransportService::SocketContext * sock) Line 308 C++ > xul.dll!mozilla::net::nsSocketTransportService::DoPollIteration(mozilla::BaseTimeDuration<mozilla::TimeDurationValueCalculator> * pollDuration) Line 1223 C++ > xul.dll!mozilla::net::nsSocketTransportService::Run() Line 969 C++ > xul.dll!nsThread::ProcessNextEvent(bool aMayWait, bool * aResult) Line 1093 C++ > xul.dll!NS_ProcessNextEvent(nsIThread * aThread, bool aMayWait) Line 519 C++ > xul.dll!mozilla::ipc::MessagePumpForNonMainThreads::Run(base::MessagePump::Delegate * aDelegate) Line 364 C++ > xul.dll!MessageLoop::RunHandler() Line 320 C++ > xul.dll!MessageLoop::Run() Line 300 C++ > xul.dll!nsThread::ThreadFunc(void * aArg) Line 427 C++ > nss3.dll!_PR_NativeRunThread(void * arg) Line 406 C > nss3.dll!pr_root(void * arg) Line 138 C > [External Code] The line is question is [1]. [1] https://searchfox.org/mozilla-central/rev/5ff2d7683078c96e4b11b8a13674daded935aa44/netwerk/protocol/http/nsHttpConnectionMgr.cpp#4463 This is on a clean recent mozilla-central build (2d95d70298e2). I have a minidump I can share if desired, and can reproduce it pretty easily. > ac_add_options --enable-application=browser > ac_add_options --target=x86_64-pc-mingw32 > ac_add_options --host=x86_64-pc-mingw32 > ac_add_options --disable-webrtc > mk_add_options MOZ_OBJDIR=@TOPSRCDIR@/obj/opt64 > ac_add_options --enable-optimize --enable-debug --enable-profiling --enable-crashreporter --enable-dmd Note, I've been able to get around this by commenting out the assert for now.

Dragana, this callback seems to be coming from mBackupStreamOut, not mStreamOut.

Can you make a http log: https://developer.mozilla.org/en-US/docs/Mozilla/Debugging/HTTP_logging thanks!

This is the only log that had any contents. > export MOZ_LOG=timestamp,rotate:200,nsHttp:5,cache:2:5,nsSocketTransport:5,nsHostResolver:5,sync

(In reply to Ryan Hunt [:rhunt] from comment #3) > Created attachment 8974497 [details] > moz.log.0 > > This is the only log that had any contents. > > > export MOZ_LOG=timestamp,rotate:200,nsHttp:5,cache:2:5,nsSocketTransport:5,nsHostResolver:5,sync Have you maybe misspelled nsSocketTransport, because I do not see the log... I also do not see nsHostResolver either...

Ah, the problem was 'cache:2:5' should have been 'cache2:5' and it ignored everything after that. Here's the updated logs.

Can you reproduce the problem with this patch. The patch just adds a one line log to print the error code.

With that patch applied.

(In reply to Ryan Hunt [:rhunt] from comment #8) > Created attachment 8976973 [details] > crash-log.zip > > With that patch applied. Dragana, does this help?

For some reason PR_Connect does not call nspr connect. I do not know why. I will need to reproduce it myself and look into when PR_Connect call gets back. I usually do not develop on windows and I have not had time to set up a vm and configure everything to make a build and debug it on window. Is this making a lot of trouble? I can disable TFO, we plan to disable it anyway shortly.

this is fix with disabling tfo for proxies and also tfo is disabled as well