JDK-4802619 : plugin exits mozilla on slightest error
  • Type: Bug
  • Component: deploy
  • Sub-Component: plugin
  • Affected Version: solaris_11,1.4.1_01
  • Priority: P4
  • Status: Closed
  • Resolution: Cannot Reproduce
  • OS: solaris_nevada,solaris_10
  • CPU: generic
  • Submitted: 2003-01-14
  • Updated: 2010-10-14
  • Resolved: 2010-10-14
The Version table provides details related to the release that this issue/RFE will be addressed.

Unresolved : Release in which this issue/RFE will be addressed.
Resolved: Release in which this issue/RFE has been resolved.
Fixed : Release in which this issue/RFE has been fixed. The release containing this fix may be available for download as an Early Access Release or a General Availability Release.

To download the current JDK release, click here.
JDK 8
8Resolved
Related Reports
Relates :  
JDK-4800337 - java 1.4.1's java_vm dumps core under mozilla 1.2 and libumem
Relates :  
JDK-4706369 - Revisit JPI code to improve exception handling
Relates :  
JDK-4746614 - java_vm dumps core inside mozilla1.1
Description
[dep, 13Jan2003]

  The browser side of the java plugin will call plugin_error, which in
  turn calls exit(2), for all manner of problems.  While sometimes it
  is in response to situations which could only be caused by a sick
  browser or plugin (in which case we should really be dumping core),
  and some reflect a legitimate fatal error, many seem to be just a
  half-hearted response to something which should cause just the plugin
  to fail but not the entire browser.
  
  For example, many people have been encountering bugs recently which
  involve the jvm process crashing.  Unfortunately, the plugin's
  reaction apparently is to call plugin_error, resulting in this
  message:
  
  > INTERNAL ERROR on Browser End: Pipe closed during read? State may be corrupt
  > System error?:: Resource temporarily unavailable

  and a dead mozilla, as well.
  
  Curiously, in several places we attempt to clean up after a call to
  plugin_error, despite its fatal nature.
Comments
EVALUATION Not reproducible if plugin2 is used (available since 6u10).
14-10-2010
EVALUATION We may have to fix these issues as part of refactoring the code in tiger ###@###.### 2003-01-13 Part of the Plugin code is loaded into browser address space. The plugin browser side code uses a pipe to talk to java vm running in the different process. If for some reason, java_vm dies, this will also cause another end of the pipe dead which in turn cause browser exit. I am not sure we can recover from this kind of failure if the pipe is broken. ###@###.### 2003-01-17 [dep, 21Jan2003] I'm not saying the java plugin should recover. I'm saying it should fail gracefully. Unless the relationship between the jvm and browser is a lot more incestuous than it appears, I can't think of any good reason why the jvm dying would mean that the browser also has to exit. This bug is a dup of 4706369. ###@###.### 2003-06-18
18-06-2003