Closed Bug 102474 Opened 23 years ago Closed 11 years ago

Killing the Java plugin subprocess kills the Mozilla process.

Categories

(Core Graveyard :: Java: OJI, defect)

x86
Linux
defect
Not set
normal

Tracking

(Not tracked)

RESOLVED INCOMPLETE

People

(Reporter: hniksic, Assigned: alfred.peng)

References

Details

(Keywords: dataloss, Whiteboard: redesign, error-handling)

Due to the unstable nature of the Java plugins under Linux, as well as some
problems related to the use of external resources (such as audio devices), it is
sometimes necessary to kill the `java_vm' process spawned by the Java plugin. 
Note that simply closing the Mozilla window with the plugin doesn't help because
resources can remain occupied.

However, doing a `pkill java_vm' has the effect of killing the whole Mozilla
process, which is decidedly not what I wanted.  The browser windows should
remain running, only the Java VM should die.

It would be really good if Mozilla could handle the death of the java_vm
subprocess gracefully.  In general, and it is likely outside the scope of this
bug, it might be good to have a way to forcibly unload or disable a plugin in
the running browser.
This is using Mozilla 0.9.4, sorry for omitting it before.
confirming
Status: UNCONFIRMED → NEW
Ever confirmed: true
-->OJI
Assignee: av → edburns
Component: Plug-ins → OJI
QA Contact: shrir → pmac
*** Bug 102748 has been marked as a duplicate of this bug. ***
*** Bug 109140 has been marked as a duplicate of this bug. ***
Blocks: 167474
Chris Petersen is a new QA contact for oji component. His email is:
petersen@netscape.com
Assignee: edburns → petersen
fixing small error for pmac@netscape.com (filter with : SPAMMAILSUCKS)
Assignee: petersen → joe.chou
QA Contact: pmac → petersen
I just killed a rogue java_vm process and it didn't take down the brower.  I'm
using j2sdk-1.4.0_02-fcs and Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.1)
Gecko/20020827.

Of course, the rogue process is a pain in the first place (buring cycles when no
currently loaded page has applets) but I don't see what Moz can do about it.

Anyway, consider this a weak WORKSFORME.
reassign to me
Assignee: joe.chou → joshua.xia
*** Bug 185803 has been marked as a duplicate of this bug. ***
Confirm on Linux(RH8.0) mozilla1.2 JRE1.4.1.
This bug depends on oji/jpi redesign
Add "redesign" on Whiteboard
Whiteboard: redesign
Whiteboard: redesign → redesign, error-handling
I've got this too. It's a real pain, since sometimes the java_vm can use 100 %
CPU, and even if the mozilla tab is closed on the website that invoked (and
infinite-looped) java, it still doesn't exit.
*** Bug 210232 has been marked as a duplicate of this bug. ***
Status: NEW → ASSIGNED
This seems like a specific case of bug 156493, in which a buggy/unstable plugin
can bring down the entire Mozilla process.  There is a great demand for running
plugins as separate processes.  If this was done, Mozilla could be tolerant of
plugin malfunctions, and continue to run.  If bug 156493 is fixed, then this bug
would also be fixed.

I very much hope that the Mozilla developers can find time to work on bug 156493 :)
->kyle
Assignee: joshua.xia → kyle.yuan
Status: ASSIGNED → NEW
I saw this issue on Firefox today, and lost an email written in a web form on
Webmail because of it. Adding "dataloss" keyword.
Keywords: dataloss
Depends on: 156493
Assignee: yuanyi21 → pete.zha
mass reassign to Alfred
Assignee: zhayupeng → alfred.peng
Product: Core → Core Graveyard
Mass-closing bugs in the "OJI" component: OJI plugin integration was replaced with npruntime long ago, and these bugs appear to be irrelevant now. If there is in fact a real bug that remains, please file it new in the "Core" product, component "Plug-ins".
Status: NEW → RESOLVED
Closed: 11 years ago
Resolution: --- → INCOMPLETE
You need to log in before you can comment on or make changes to this bug.