JDK-4419275 : Hotspot croaks from within JBuilder 4
  • Type: Bug
  • Component: core-svc
  • Sub-Component: debugger
  • Affected Version: 1.4.0
  • Priority: P3
  • Status: Closed
  • Resolution: Cannot Reproduce
  • OS: solaris_8
  • CPU: x86
  • Submitted: 2001-02-27
  • Updated: 2001-05-10
  • Resolved: 2001-05-10
Related Reports
Relates :  
Description
I get this message when trying to use the debugger from withing JBuilder 4 on Merlin b53.

C:\Users\bcbeck\swing-merlin\build\win32\bin\javaw -classpath "C:\Users\bcbeck\swing-merlin\ColorChooser;C:\Users\bcbeck\swing-merlin\build\win32\classes;C:\Users\bcbeck\swing-merlin\build\win32\lib\i18n.jar;C:\Users\bcbeck\swing-merlin\build\win32\lib\ext\jce.jar;C:\Users\bcbeck\swing-merlin\build\win32\lib\ext\sunjce_provider.jar;C:\Users\bcbeck\swing-merlin\build\win32\lib\ext\jsse.jar;C:\Users\bcbeck\swing-merlin\build\win32\lib\ext\jcert.jar;C:\Users\bcbeck\swing-merlin\build\win32\lib\ext\jnet.jar;C:\Users\bcbeck\swing-merlin\build\win32\lib\ext\sunrmijsse.jar;C:\Users\bcbeck\swing-merlin\build\win32\lib\ext\dnsns.jar;C:\Users\bcbeck\swing-merlin\build\win32\lib\ext\jgss_kerberos.jar;C:\Users\bcbeck\swing-merlin\build\win32\lib\ext\ldapsec.jar;C:\Users\bcbeck\swing-merlin\build\win32\lib\sunrsasign.jar;C:\Users\bcbeck\swing-merlin\build\win32\demo\jfc\Java2D\Java2Demo.jar"  -Xdebug -Xnoagent -Djava.compiler=NONE -Xrunjdwp:transport=dt_shmem,address=javadebug,suspend=y ColorChooser 
#
# HotSpot Virtual Machine Error, Internal Error
# Please report this error at
# http://java.sun.com/cgi-bin/bugreport.cgi
#
# Error happened during: full generation collection
#
# Error ID: 53414645504F494E540E4350500124
#
# Problematic Thread: prio=5 tid=0x008498A8 nid=0x2b4 runnable 
#


----------------------

The problem continues.  After spending some time trying the JBuilder debugger in various ways, I've seen at least three behaviors.  The first is that messages like the one above get thrown.  The second is that the whole JBuilder tool hangs and needs to be killed.  The third is that everything works OK.  For a given test program and build version, the behavior seems constant.  Changing the test program or making changes in the build being debugged can cause the behavior to switch to one of the others.   There is no descernable pattern at this point
brian.beck@Eng 2001-03-26


daniel.daugherty@Eng 2001-03-27

The above error ID translates to:

src/share/vm/runtime/safepoint.cpp, 292 which is in
SafepointSynchronize::block():

          fatal1("Deadlock in safepoint code. stopped at 0x%p", stop_pc)

------------------------

One thing I should have pointed out in the original description.  I am running JBuilder (which is a Java program) on the VM that JBuilder comes with (1.3.0-C).  The program I am trying to debug is running in a 1.4 VM.  While it may be interesting to try this out when JBuilder itself is running on 1.4, I think that is a less likely scenario in the field.  People don't care what version of Java JBuilder is run on as long as it works.  They do care about the version of Java they are developing against.

One more thing, all my testing has been on Windows 2K.  I can't get JBuilder to run at all on Solaris due to performance issues.

brian.beck@Eng 2001-04-26

Comments
EVALUATION IDEs need to work for beta janet.koenig@Eng 2001-03-21
21-03-2001