Closed Bug 163538 Opened 23 years ago Closed 23 years ago

Mozilla sometimes hangs after using Java applet for hours

Categories

(Core Graveyard :: Java: OJI, defect)

Product:
Core Graveyard
Component:
Java: OJI
Platform:
PowerPC
macOS
Type:
defect
Priority:
Not set
Severity:
critical

Tracking

(Not tracked)

Status:
RESOLVED INVALID

People

(Reporter: josh, Assigned: beard)

References

(
URL
)

Details

(Keywords: hang)

Attachments

(1 file)

Output from sampler
23.87 KB, text/plain
Details
User-Agent: Mozilla/5.0 (Macintosh; U; PPC Mac OS X; en-US; rv:1.1b) Gecko/20020722 Build Identifier: Mozilla/5.0 (Macintosh; U; PPC Mac OS X; en-US; rv:1.1b) Gecko/20020722 Eventually, the spinner (Mac equivalent of the hour glass comes on and won't go away. I hit force quit and restart. It seems to be a timing issue as it happens sooner and more often with other apps running. Reproducible: Always Steps to Reproduce: 1.go to http://www.freecell.com/f/c/javagame.html, login 2.play freecell for a few 100 games until it happens 3. Actual Results: Just hangs, other applications unaffected. Expected Results: Detected that the application can no longer function and die gracefully or find some way of preventing the problem to start with.
Probably memory related, which means Mozilla or Java leaks. My bet is on Java, since this also happens with IE when playing on games.yahoo.com. How much RAM do you have?
In response to the request, "how much memory do you have?", here is the output from the top of top. I have 512M of Ram and as far as I can tell I almost never get to were I run out of memory. Processes: 63 total, 3 running, 60 sleeping... 151 threads 21:06:41 Load Avg: 2.17, 2.04, 1.89 CPU usage: 79.4% user, 20.6% sys, 0.0% idle SharedLibs: num = 124, resident = 17.6M code, 1.33M data, 4.72M LinkEdit MemRegions: num = 4493, resident = 173M + 3.37M private, 44.1M shared PhysMem: 45.9M wired, 170M active, 197M inactive, 413M used, 99.3M free VM: 1.23G + 48.5M 36828(0) pageins, 28817(0) pageouts
Reassigning to OJI.
Assignee: idk → joe.chou
Severity: minor → critical
Component: Java-Implemented Plugins → OJI
Keywords: hang
QA Contact: avm → pmac
Josh, can you attach some Sampler data taken from Mozilla during the hang?
Summary: Sooner or later, sometimes after many hours, mozilla hangs and I must force quit → Mozilla sometimes hangs after using Java applet for hours
Greg, I can sit and play freecell for hours until it hangs, but how would I capture the data. This is my first bug report for Mozilla, so I am unfamiliar with any debugging tools which may be embedded. Is there a log someplace? Just tell me how to get the Sampler data.
Re-assign to Babak.
Assignee: joe.chou → babak.mahbod
Josh, Sampler is an application included with Apple's OS X Developer tools.
I know now that I have sampler. Thank you. I am still unfamiliar with it so I'll need a little be more help getting info to you. I started up Mozilla using Sampler. Running this game with Sampler running is enough to cause Mozilla to hang. Rather, than send the output I got, I'm going to ask you to give me slightly more explicit instructions. When or how should I start sampler? What settings should I use and what output do you want?
Run Mozilla and play the game until Mozilla hangs. Then attach Sampler to it and record a few seconds of data.
I did as suggested and generated the output using the generate report option from Sampler. For some reason I can't seem to get the output pasted in here. Let me know what else I need to do, or if I missed something obvious about attaching a file. The top part: Report #0 - Sampling Mozilla (process 17557) Stacks at 2002-09-09 00:09:51 -0400 Samples (displayed/total): 3051/3051 Call graph: 113 Thread_3e1b 113 main 113 main1(int, char **, nsISupports *) 113 nsAppShellService::Run(void) 113 nsAppShell::Run(void) 113 nsMacMessagePump::DoMessagePump(void) 113 nsMacMessagePump::GetEvent(EventRecord &) 113 WaitNextEvent 113 WNEInternal 113 GetNextEventMatchingMask 113 GetOrPeekEvent 113 CallEventDispatchHook 113 ToolboxEventDispatcher 113 SendEventToEventTarget 113 SendEventToEventTargetInternal 113 DispatchEventToHandlers 113 ToolboxEventDispatcherHandler 113 HandleMouseEvent 113 HandleMouseEventForWindow 113 SendEventToEventTarget 113 SendEventToEventTargetInternal 113 DispatchEventToHandlers 113 0x4092a70 113 Java_com_apple_mrj_internal_jdirect_LibraryMacOSX_1bundle_findSymbol 113 JVM_ClassDepth 113 JVM_MonitorNotifyAll and the bottom part: 113 Thread_7d03 113 _pthread_body 113 pthread_exit 113 _pthread_become_available 113 mach_msg 113 mach_msg_overwrite_trap 113 mach_msg_overwrite_trap [STACK TOP] Total number in stack (recursive counted multiple, when >=5): 32 __builtin_new 26 0x3e10068c 26 _pthread_body 23 mach_msg 23 mach_msg_overwrite_trap 22 JNI_CreateJavaVM_Impl 19 __floatdidf 17 JVM_GetInheritedAccessControlContext 16 JVM_GetCPMethodClassNameUTF 16 JVM_Read 16 __ashrdi3 16 bad_cast virtual table 15 JVM_InternString 15 JVM_IsNaN 12 JVM_MonitorNotify 12 JVM_MonitorWait 8 0x3e1005cc 7 0x3e26b17c 6 0x3e102668 6 0x3e26e9c0 5 0x3e10086c Sort by top of stack, same collapsed (when >= 5): mach_msg_overwrite_trap [STACK TOP] 2485 semaphore_wait_signal_trap [STACK TOP] 339 semaphore_timedwait_signal_trap [STACK TOP] 113 syscall [STACK TOP] 113
Josh, click "Create a New Attachment" to attach the file.
Attached file Output from sampler —
Output from sampler. I started playing freecell by signing in, http://www.frecell.com/f/c/javagame.html, played until the game hung, attached sampler, and sampled very briefly. Hopefully this will help.
I just upgraded my iMac to 10.2 and mozilla on the same iMac to build 2002091014. The problem not only persists, but seems to no longer require the machine to be remotely busy. With Mozilla as the only app running the bug manifests itself instantaneously; I don't need to sit and play for a while for it to crash Mozilla. It seems to happen, in most cases before I can even start a single game. I have no idea if this has to do with Mozilla or if Apple has a new or different Java that Mozilla is using. Should I run sampler again.
Josh, please also retest with a recent nightly build. Thanks.
Greg I am tring to do what you suggest but I'm not quite sure which build will match as there seem to be a whole slew of what look to me to be different OS X builds. I'm trying a CVS checkout, with configure and a make but I'm missing a lot of inof on what to build and how. I could follow notes for Fizzila with Carbon but even that leaves two choices. How do I build the OS X app nightly build? I'm at a loss. I've got a CVS checkout as of 20021015 at about 8PM. configure and make produce errors. What next?
Josh, you can just download a nightly build from: ftp://ftp.mozilla.org/pub/mozilla/nightly/latest/
Hmmm, I wonder how come I couldn't find the url when I was looking? OK, I downloaded the ...trunk.smi.bin -- there are stukk about three and it is not clear which one is the standard OSX build, but I think I got the right one.. I was at work and downloaded it to my laptop. It behaved the same as before as far as playing Freecell and crashing the app.
Chris Petersen is a new QA contact for oji component. His email is: petersen@netscape.com
Assignee: babak.mahbod → petersen
Status: UNCONFIRMED → NEW
fixing small error for pmac@netscape.com (filter with : SPAMMAILSUCKS)
Assignee: petersen → beard
QA Contact: pmac → petersen
I've tried some newer versions of Mozilla and just for laughs tried MSIE. Both fail in exactly the same way; for instance if Mozilla fails just after signing in and before the buttons for cards get lit up so will MSIE. This is a evidence that Apple's implementation of Java is hanging the machine. So the solution is either to get Apple to fix their Java plug in or find an alternative plug in to use. I've no idea how to contact Apple and tell them their plug-in is broken, and I've no idea if I have an option for an alternative plug-in that is less buggy than apples.. I put on the Java Console, but never saw any messages except when things were working normally. I also tested a Solaris Machine which can play the game without a problem for what that is worth.
Confirming this on latest trunk (2002-10-23-05) OS X branch under 10.2.1. Also , happens with IE 5.2.
I tried to test if the bug still exists with Apple's JDK 1.4.1 but it doesn't support being a plug-in. (According to Apple's notes there is no plug-in for MSIE, I don't know a work around, they suggested that jave used by plug-ins could be tested from the command line, but I didn't really follow it or figure out what command they were suggesting I could use.)
We can't fix this.
Status: NEW → RESOLVED
Closed: 23 years ago
Resolution: --- → INVALID
Apple has a new beta version of JVM 1.4.1 which is spotted and used by their Safari browser automatically when the DP9 version of their Java is downloaded and installed from Apples http://connect.apple.com, that seems to work where the old Java fails, but I haven't figured out how to get Mozilla to use it yet.
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: