User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.1) Gecko/20021003 Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.1) Gecko/20021003 This is related to bugs 102474, 85181, 115523, and 173179. This is a case of at best lazy, at worst unprofessional, design. Whenever the OJI plugin is unhappy, it calls exit. Let's define our terms. There are three parts to Applet execution in Mozilla: - The Mozilla plugin interface code, in the Mozilla tree, part of the Mozilla process - The JVM interface code, in the JVM tree, loaded into the Mozilla process in libjavaplugin_oji.so - The JVM code, in the JVM tree, a separate process launched by the previous component. Now, in the second component, you have code like this: ...after having forked... if (tracing) { execlp(fullprogname, progname, "-t", 0); } else { execlp(fullprogname, progname, "", 0); } plugin_error("Exec of \"%s\" failed: %d\n<", progname, errno); exit(6); Where progname is java_vm. So if the program is not found, we exit...Mozilla. Now that is indefensible! I have access to the SUN code, version 1.4.x . I will be redesigning the thing so that it is more gracious, for example, displaying the ALT message. Any better idea for the error handling? The idea is, a Java plugin that fails to initialize is as good as none, and otherwise, in any other case, absolutely do not exit the browser. --- There are tons of others cases. I'll be fixing them all (hopefully). See the previous bug numbers, which only list some. plugin_error actually calls exit. What we want is really a warning. The extra exit call above possibly demonstrates some confusion about the whole thing. When is the OJI redesign announced on the dead OJI mailing list due? Can someone give me the right ocntact at SUN? What is the level of interest in OJI? Reproducible: Always Steps to Reproduce: 1.do away with the java_vm program in the jre tree associated with libjavaplugin_oji.so 2. 3. Actual Results: System Error: : No such file or directory INTERNAL ERROR on Browser End: Could not start JavaVM! Expected Results: not exited This is a code redesign issue

Olivier Cahagne (wolruf@free.fr) ask me to be on the cc

related: bug 156493.

This issue has been filed in Sun's bug database. Number: 4782716

added SUN contact to cc: list

Jay Yan asked to be kept posted

->joshua.xia@sun.com

I am sorry to report that my company is no longer interested in this project.

both JPI/OJI redesign

I first started experiencing exits after trying to use the plugins from my /home/sbrown/.mozilla/plugins directory. I copied everything from /usr/local/mozilla/plugins because I didn't want to have to reinstall the plugins every time I installed the latest build. After getting kicked out of Mozilla the first time, I deleted the .mozilla/plugins directory I created. I still get kicked out, though. If I install 1.3b, it works OK -- until I install another 'latest build'. On my machine at home, which I did not try to use plugins from my home directory, everything works OK.

->kyle

Hi i have been having the same problem for the past few days in my attempt to geting java aplets work with Mozilla 1.2.1 using j2re-1_3_1_10... Here are the steps i have tryed... Create a link in /usr/bin/mozilla/plugins, /usr/bin/mozilla1.2.1/plugins from /usr/java/jre1.3.1_10/plugin/i386/ns600/libjavaplugin_oji.so using the "ln /usr/java/jre1.3.1_10/plugin/i386/ns600/libjavaplugin_oji.so" command with no success. The problem was i was trying to create a hard link but when i created a symbolic link the problem was solved. Here is a list of directories and the command i have used to create the symbolic link Directories... /usr/bin/mozilla/plugins /usr/bin/mozilla1.2.1/plugins /~.mozilla/plugins/ Command.. ln -s /usr/java/jre1.3.1_10/plugin/i386/ns600/libjavaplugin_oji.so I can now run java applets without a problem! If anyone would like to contact me for help or information my email address is mozun@hotpop.com or miracleman_me@hotmail.com

mass reassign to Alfred