Closed Bug 1124744 Opened 9 years ago Closed 9 years ago
crash in mozilla::net::Http
Channel Parent Listener::On Start Request(ns IRequest*, ns ISupports*)
This bug was filed from the Socorro interface and is report bp-5fded522-462b-442e-9b6e-a13a92150122. ============================================================= This is completely reproducible for me when I have NoScript 126.96.36.199 enabled in Nightly.
This is the problem of the NoScript addon. It is not calling redirect properly.
I don't have NoScript installed and yet this signature crash happened to me on visiting a link from this mozillazine post: http://forums.mozillazine.org/viewtopic.php?p=13989263#p13989263
There is 2- 3 separate problems. Two of them were causing crash Bug 1106396: 99% of the time it was noscript (which I am not interesting in fixing) and 1% of the time another error that I was trying fix, but I do not have a reproducible case. I checked in a patch that was trying to make this more visible but it made nightly not usable for users using noscript. The patch will be backed out. And the patch found another error in our code :) This is probably your error Zlip792. can you send me crash code to check which error it is :)
Bug 1106396 was backed out in https://hg.mozilla.org/integration/mozilla-inbound/rev/60c7f9da3e16 but we'll need to get inbound merged to m-c before this fixes things for Nightly users.
(In reply to Zlip792 from comment #2) > I don't have NoScript installed and yet this signature crash happened to me > on visiting a link from this mozillazine post: > > http://forums.mozillazine.org/viewtopic.php?p=13989263#p13989263 Can you provide a crash report?
Definitely Yes. https://crash-stats.mozilla.com/report/index/8fc49028-e069-4d89-a7f4-a56352150122 If you guys want any test for me to run or try test builds, I am good to go. Want Profile or Remote access to Firefox. Do let me know. :-)
Here is a crash report using a new blank profile, with e10s on and with only the development build of HTTPS Everywhere installed. https://crash-stats.mozilla.com/report/index/4c959875-6c34-44d7-9065-247802150122 https://www.eff.org/files/https-everywhere-devel.xpi The crash is 100% reproducible for me by going to www.amazon.com.
I can consistently reproduce this assertion failure on OS X with the Shumway 0.9.3632 add-on enabled (with e10s either on or off): Assertion failure: mRedirectChannelId == 0 (The redirect must be finished before OnStartRequest is called!), at /builds/slave/m-cen-osx64-ntly-0000000000000/build/src/netwerk/protocol/http/HttpChannelParentListener.cpp:51 #01: _ZN7mozilla3net13nsHttpChannel33ContinueAsyncRedirectChannelToURIE12tag_nsresult[/Applications/Nightly.app/Contents/MacOS/XUL +0x263200] #02: _ZN7mozilla3net13nsHttpChannel24OnRedirectVerifyCallbackE12tag_nsresult[/Applications/Nightly.app/Contents/MacOS/XUL +0x26eab3] #03: _ZN34nsAsyncVerifyRedirectCallbackEvent3RunEv[/Applications/Nightly.app/Contents/MacOS/XUL +0x1195f0] #04: _ZN8nsThread16ProcessNextEventEbPb[/Applications/Nightly.app/Contents/MacOS/XUL +0xab66d] #05: _Z23NS_ProcessPendingEventsP9nsIThreadj[/Applications/Nightly.app/Contents/MacOS/XUL +0xcbc21] #06: _ZN14nsBaseAppShell19NativeEventCallbackEv[/Applications/Nightly.app/Contents/MacOS/XUL +0x1b13527] #07: _ZN10nsAppShell18ProcessGeckoEventsEPv[/Applications/Nightly.app/Contents/MacOS/XUL +0x1b59342] #08: __CFRUNLOOP_IS_CALLING_OUT_TO_A_SOURCE0_PERFORM_FUNCTION__[/System/Library/Frameworks/CoreFoundation.framework/Versions/A/CoreFoundation +0x80661] #09: __CFRunLoopDoSources0[/System/Library/Frameworks/CoreFoundation.framework/Versions/A/CoreFoundation +0x727ed] #10: __CFRunLoopRun[/System/Library/Frameworks/CoreFoundation.framework/Versions/A/CoreFoundation +0x71e1f] #11: CFRunLoopRunSpecific[/System/Library/Frameworks/CoreFoundation.framework/Versions/A/CoreFoundation +0x71838] #12: RunCurrentEventLoopInMode[/System/Library/Frameworks/Carbon.framework/Versions/A/Frameworks/HIToolbox.framework/Versions/A/HIToolbox +0x2e43f] #13: ReceiveNextEventCommon[/System/Library/Frameworks/Carbon.framework/Versions/A/Frameworks/HIToolbox.framework/Versions/A/HIToolbox +0x2e1ba] #14: _BlockUntilNextEventMatchingListInModeWithFilter[/System/Library/Frameworks/Carbon.framework/Versions/A/Frameworks/HIToolbox.framework/Versions/A/HIToolbox +0x2dffb] #15: _DPSNextEvent[/System/Library/Frameworks/AppKit.framework/Versions/C/AppKit +0x246d1] #16: -[NSApplication nextEventMatchingMask:untilDate:inMode:dequeue:][/System/Library/Frameworks/AppKit.framework/Versions/C/AppKit +0x23e80] #17: -[GeckoNSApplication nextEventMatchingMask:untilDate:inMode:dequeue:][/Applications/Nightly.app/Contents/MacOS/XUL +0x1b58946] #18: -[NSApplication run][/System/Library/Frameworks/AppKit.framework/Versions/C/AppKit +0x17e23] #19: _ZN10nsAppShell3RunEv[/Applications/Nightly.app/Contents/MacOS/XUL +0x1b59b1d] #20: _ZN12nsAppStartup3RunEv[/Applications/Nightly.app/Contents/MacOS/XUL +0x235a849] #21: _ZN7XREMain11XRE_mainRunEv[/Applications/Nightly.app/Contents/MacOS/XUL +0x23a4f69] #22: _ZN7XREMain8XRE_mainEiPPcPK12nsXREAppData[/Applications/Nightly.app/Contents/MacOS/XUL +0x23a521f] #23: XRE_main[/Applications/Nightly.app/Contents/MacOS/XUL +0x23a5584] #24: main[/Applications/Nightly.app/Contents/MacOS/firefox +0x1c7d]
the crash depends on e10s being enabled and Noscript installed. I disabled e10s, installed Noscript, and everything was loading fine. I enabled e10s, triggered a restart, and firefox crashed when loading a website. I disabled e10s, triggered a restart, firefox crashed and started using a different named profile. I'm using multiple versions of firefox each started via the -P option with its own named profile.
Pushlog: https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=48409acb6e98&tochange=d11387c52d67 Bug 1106396 seems to trigger the crash.
I got this crash 3 or so times rapidly, without any extensions installed, but haven't had any further trouble.
(In reply to Alice0775 White from comment #11) > Pushlog: > https://hg.mozilla.org/integration/mozilla-inbound/ > pushloghtml?fromchange=48409acb6e98&tochange=d11387c52d67 > > Bug 1106396 seems to trigger the crash. And it was backed out. Should be fixed in the next Nightly.
WFM with todays build on WinXP, E10S enabled, Noscript 188.8.131.52 enabled BuildID 20150123030203 https://hg.mozilla.org/mozilla-central/rev/494632b9afed
Fixed? Should we keep it for tracking any new assertions we might want to add?
I will marked it as fix. You have already open another bug to investigate redirect in e10s.
Status: NEW → RESOLVED
Closed: 9 years ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.