Closed Bug 217296 Opened 21 years ago Closed 20 years ago

Crash when set display=none to div with Flash inside [@ nsXPCWrappedJSClass::CallMethod]

Categories

(Core Graveyard :: Plug-ins, defect)

x86
All
defect
Not set
critical

Tracking

(Not tracked)

RESOLVED DUPLICATE of bug 246892

People

(Reporter: mgalli, Assigned: peterlubczynski-bugs)

References

()

Details

(Keywords: crash, testcase, topcrash)

Crash Data

Attachments

(6 files)

Crashes Netscape 7.1. There is a flash animation inside a DIV. This DIV is block by its own nature. At the end of the flash animation, the flash animation sends an event to the JS, calling a function which works fine. This function is trying to: Disable the DIV via setting its style.display CSS property: document.getElementById("mask").style.display="none"; PPS: Mask is the DIV with the flash. I am filing against trunk however I just confirmed with N7.1. So need help to check with Mozilla.
wfm 2003081904 on Win2k + flash 7.0.249. anyone who crashes: please submit data from Netscape Feedback Agent then mention the generated Talkback ID here "mozilla/bin/components/talkback.exe". If you crash, mention Flash version and whether or not you have a flashplayer.xpt in your components/ directory (which enables Flash-to-JavaScript communication, apparently used by this URL).
Severity: normal → critical
Keywords: crash, stackwanted
wfm Mozilla/5.0 (Windows; U; Win98; en-US; rv:1.5b) Gecko/20030824 Shockwave Flash 6.0 r79
Additional information about the crash: When using the demo, please click the image. It will trigger the animation. - Scriptability is working in this case fine, is the JS CSS trying to display=none that crashes. - Talback ID is: 23063634. - Version is: Shockwave Flash 6.0 r79 - Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.4) Gecko/20030622 Netscape/7.1 (ax)
Version: Trunk → 1.4 Branch
I've reproduced the crash here, I've clicked to start the animation and after the end of the animation the browser crashed. Mozilla 1.4 Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.4) Gecko/20030624 Shockwave Flash 7.0 d249 my Talkback ID is: TB23066086H
crashes on a celeron 333, Win98, Flash 6.0r79 with Netscape 7.1 TB23065221W Mozilla 1.4 TB23065289K Mozilla 1.3.1 TB 23065336H no crash on trunk BuildID 2003060824, see my comment #3 Crashes on a Athlon XP1600, Win98SE, Flash 6.0r79 with Mozilla trunk, BuildID 2003082204, TB230658524 no crash with BuildID 2003082505 All tests done by clicking on that animation. The keyhole opens to a various degrees, then DocWatson comes up, tells about a crash in XPC3250.DLL
Didn´t change my working installation of BuildID 2003082505 installed from sea.exe Unzipped zip-Build BuildID 2003082204 to retest, as my sea.exe 2003082204 had crashed on this Win98SE. Started Mozilla from where it was unzipped, didn´t make it to default browser, didn´t crash. Maybe that is because it was using the plugins from the 203082505 default installation, but that had been installed on top of 2003082204, without reinstallation of plugins. installed and tested last three working Firebird nightlies, no crash.
Marcio, can you produce a reduced testcase that show this issue (without the whole mozilla.org page) ?
Whiteboard: TB230658524
Version: 1.4 Branch → Trunk
There are stack traces (TB23065998, TB23065852) referring to this bug: nsXPCWrappedJSClass::CallMethod [c:/builds/seamonkey/mozilla/js/src/xpconnect/src/xpcwrappedjsclass.cpp line 1412] nsXPCWrappedJS::CallMethod [c:/builds/seamonkey/mozilla/js/src/xpconnect/src/xpcwrappedjs.cpp line 429] PrepareAndDispatch [c:/builds/seamonkey/mozilla/xpcom/reflect/xptcall/src/md/win32/xptcstubs.cpp line 119] SharedStub [c:/builds/seamonkey/mozilla/xpcom/reflect/xptcall/src/md/win32/xptcstubs.cpp line 147] NPSWF32.DLL + 0x4ed2a (0x1004ed2a) NPSWF32.DLL + 0x4bb65 (0x1004bb65) NPSWF32.DLL + 0x41bd7 (0x10041bd7) NPSWF32.DLL + 0x3c7ff (0x1003c7ff) NPSWF32.DLL + 0x3ce53 (0x1003ce53) NPSWF32.DLL + 0x4d364 (0x1004d364) NPSWF32.DLL + 0x4b22f (0x1004b22f) KERNEL32.DLL + 0x363b (0xbff7363b) KERNEL32.DLL + 0x24407 (0xbff94407) 0x00658b66 Source File : c:/builds/seamonkey/mozilla/js/src/xpconnect/src/xpcwrappedjsclass.cpp line no: 1412
Keywords: stackwanted
Summary: Crash when set display=none to div with Flash inside → Crash when set display=none to div with Flash inside [@ nsXPCWrappedJSClass::CallMethod]
Whiteboard: TB230658524
sounds like this isn't crashing on the trunk, right? Can someone help narrow down what checkin fixed this on the trunk so we can migrate it to the stable 1.4 branch?
Crashes noted in comment #9 are from build 2003082204/W98 http://ftp.mozilla.org/pub/data/crash-data/detailed-crash-analysis-all.html Peter: there was your bug 169951 between 1.4a and 1.4b with same signature, but it's long time.
it crashes on trunk, I was wrong in comment 2, I hadn't click on the Flash app. It doesn't seem to crash on Linux (2003082005 + Flash 6.0r79) but this could be because of lack of Flash scriptability (bug 212798). Similar stack in bug 169951.
ok, just looked at the testcase and I see what's going on: it's nasty. Basically, in an FSCommand, there is a script that is causing the destruction of frames who have the plugin as a child. When you do a style.display='none' on the plugin that is executing the FSComamnd, it casues the plugin to be unloaded. Also see: bug 90268, bug 136927, bug 158670, and bug 195116. The author of the page can probably get the desired effect without the crash by trying to toggle CSS visibility and/or width+height. A simplified testcase would also help here.
Attached file testcase
simplified testcase (without Mozilla.org's table layout)
I was crashing with nightbuild from 24th, but after deleting Mozilla binaries and refresh from zipfile, both testcases are WFM (tryed with nightbuilds from 24th and 26th). Could anyone retry with fresh installation of any Mozilla?
Adam, do you have flashplayer.xpt in your components/ directory ? This is required for this testcase to crash afaik.
Olivier: Oops, I lost it when deleting. If I choose "NN4.78, Opera 7" while installing Flash 7beta, flashplayer.xpt isn't present and Mozilla is not crashing (and demo looks fine). If I choose Mozilla plugin directory, flashplayer.xpt is presented and Mozilla starts crashing. (discover I America just now? =))
TB23098851G Mozilla/5.0 (Windows; U; Win98; en-US; rv:1.5b) Gecko/20030826 I didn´t crash, when flashplayer.xpt wasn´t installed. At first install of Flash, I got a menu with Netscape4/Mozilla1.4a(Firebird)/Mozilla1.5b(Trunk) and I didn´t select anything, or maybe all, and flash was working in all browsers, but flashplayer.xpt wasn´t installed. After reading comment #16 I reinstalled (without deinstalling) Flash, this time only Mozilla 1.5b selected. Result: flashplayer.xpt installed, and producing more talkbacks :-(
Keywords: topcrash
Bug is reproducable on Win 2K/Netscape 7.1 and Mozilla 1.6 with Player 7.
Confirmed crash launching testcase. talkbackid: TB9477G Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.7b) Gecko/20040331 Microsoft Windows 2000 Professional 5.00.2195 SP4 Shockwave Flash 7.0 r19
confirming crash FF 20040401 Linux + Flash 6.0r81 (scriptable): TB9969X.
OS: Windows 2000 → All
Whiteboard: TB9477G TB9969X
Keywords: testcase
Whiteboard: TB9477G TB9969X
Attached file TB9477 stacktrace
Confirmed crash on: Win2K, Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.6) Gecko/20040113, Shockwave Flash 7.0 r19 AND Linux Redhat 9, Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.6) Gecko/20040113, Shockwave Flash 7.0 r25
run html packaged in zip file to reproduce the bug. what this flash movie does is send a variable to javascript w/ fscommand and then a javascript function hides the div -- causing the crash. I have included the flash fla (src) and a swf file published to Flash 7, use mozilla with flash 7 plugin for this to work.
yeah, this is becaue "plugins are tied to frames" (search for an old bug on this) and when you set display: none, that destroys the frame, and plugin, right away. Since your going through an FSCommand and your in the code of the plugin, destroying it at that time is bad. One thing you can try as a workaround is putting a setTimeout() someplace in the JS/FSCommand sequence to allow the browser to pop it's stack to service the message loop.
*** Bug 243357 has been marked as a duplicate of this bug. ***
Ignore the duped bug above, that was my mistake. Any updates on this crash? There aren't as many incidents in Talkback data as before, but still enough to take a closer look if we can. If anyone is able to reproduce this with a recent release/nightly, it would also be nice to get a fresh Talkback ID (since old incidents are no longer available after the db upgrade).
I can confirm this crash on Windows 2000 SP4, Mozilla v1.7, Flash 7.0 r19 User Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.7) Gecko/20040616 MultiZilla/1.6.4.0b Also it crashes on WinXP (version of Mozilla and Flash player are same). I have also prepared simplified testcase for this bug.
Attached file Simple Flash movie
Flash movie for testcase
Attached file Testcase page
Testcase page
Attached file Source .fla file
Source .fla file for Flash movie, used in testcase, just in a case anybody is interested in it
*** This bug has been marked as a duplicate of 246892 ***
Status: NEW → RESOLVED
Closed: 20 years ago
Resolution: --- → DUPLICATE
Crash Signature: [@ nsXPCWrappedJSClass::CallMethod]
Product: Core → Core Graveyard
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: