Version 1.0 Alpha3

Collecting Trace Information

By including trace information in your bug report, you greatly increase our chances of fixing the bug. HotJava and other Java programs can produce three kinds of trace information:

Getting Diagnostic Information

HotJava prints its diagnostic information to a file named ~username/.hotjava/weblog. When HotJava starts up, it renames the existing weblog file (if any) to weblog.bak. If either weblog or weblog.bak has any information that seems like it might be relevant, include it in your bug report.

Getting a Java Stack Dump

If HotJava (or any other Java program) is hung, you can get its Java stack dump. You can usually do this by pressing Control-\ in the shell that you invoked hotjava or java from. The stack dump information will appear in the shell window.

If Control-\ doesn't work, use the ps and kill commands instead. The Java stack dump will most likely be displayed in the console window. 

    %ps -ax | grep hotjava
      1920 pts/2    S  0:00 grep hotjava
      1902 ?        S  0:37 /usr/local/java/bin/../bin/sun4/java hotjava.we
    %kill -QUIT 1902

Getting a C Stack Dump

If the program has dumped core (exited, leaving behind a file named core), you can get its C stack dump if you have the dbx debugger installed.

First, you must find out where the java executable file for your architecture is. This file is one directory below the java script file. You can get it by inserting the directory for your machine's architecture (such as '/sun4') after 'bin' in the string that 'which java' returns. For example, if 'which java' returns '/usr/local/java/bin/java', then the java executable file for the sun4 architecture is /usr/local/java/bin/sun4/java.

Now you get the C stack dump by running dbx, specifying the java executable file and the core file, as shown below. 

    %dbx /usr/local/java/bin/sun4/java core
    (dbx) where

Back to the bug report form

Send your comments or questions to java@java.sun.com