k-billy
10-01-2003, 11:17 AM
Hi,
I have some new information regarding a previous thread that discusses Stack Overflows when compiling Laszlo files.
We got a much more detailed stack trace on a Windows XP machine running the 1.4.1_02 JVM (see attached file for the stack trace).
The problem appears to be occuring in the com.thaiopensource.relaxng.impl package, which is in the jing.jar file in lps webapp.
On the Windows and Mac OSX (using the 1.4.1 JVM) the only way we can workaround this problem is to pass the JVM the -Xint option, forcing it to run in interpreted mode. The problem with this is that the -Xint option slows down the JVM considerably and result in very long compile times (3-4 minutes or more). This problem is not happening on the Linux machine we have running.
Regards.
I have some new information regarding a previous thread that discusses Stack Overflows when compiling Laszlo files.
We got a much more detailed stack trace on a Windows XP machine running the 1.4.1_02 JVM (see attached file for the stack trace).
The problem appears to be occuring in the com.thaiopensource.relaxng.impl package, which is in the jing.jar file in lps webapp.
On the Windows and Mac OSX (using the 1.4.1 JVM) the only way we can workaround this problem is to pass the JVM the -Xint option, forcing it to run in interpreted mode. The problem with this is that the -Xint option slows down the JVM considerably and result in very long compile times (3-4 minutes or more). This problem is not happening on the Linux machine we have running.
Regards.