Created attachment 521108 [details]
I have this hang caught in GDB right now if anyone wants
debug data. Please let me know what data you want from it.
I'll terminate the session in an hour.
It never returns from frame #25 js::Interpret
which repeatedly calls js::MonitorLoopEdge, the return value
from that function is js::MONITOR_NOT_RECORDING every time
it seems. The file name in frame 26 is
It's a nightly Linux x86-64 build, I have downloaded symbols for it.
Mozilla/5.0 (X11; Linux x86_64; rv:2.0b13pre) Gecko/20110316 Firefox/4.0b13pre
I think the JS Engine should be the same as in 4.0 final.
Created attachment 521109 [details]
Mozilla/5.0 (X11; Linux i686; rv:10.0) Gecko/20100101 Firefox/10.0
STR (slow PC is required)
2. Click "LAUNCH DEMO" red button
3. Click "single_light" (at 2nd row/5th column)
4. Click "Ray Trace!" blue button
5. Close this tab just before slow script dialog pops up (15-20second)
6. Wait until the tab is completely closed
7. Slow script dialog will pops up, click Stop
8. Repeat step3 to step7 if browser does not crash.
Browser crashes w/ crash report
Looks like some sort of web-worker-related bustage.
(In reply to David Mandelin from comment #3)
> Looks like some sort of web-worker-related bustage.
Hm, I don't see any worker stuff in the stacks... Looks like all main-thread-only stuff.
I think Alice is saying it still repros now. I see WorkerRun (I'm assuming that's what starts up a worker) in both of her stacks in comment 2.
comment 2 links to a mozilla-beta rev which still contains jstracer.cpp
The str comment#2 cannot be reprodiced anymore on
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:12.0a1) Gecko/20120126 Firefox/12.0a1 ID:20120126031113
And comment#2 was fixed by Bug 675078.
However,I cannot confirm whether original comment#0 was fixed.
Wow, unexpected bug fix #2 for bug 675078.
I haven't seen this hang in a very long time so I think you can close this bug now.