On Sun, 23 Aug 1998, Herouth Maoz wrote:
> At 14:46 +0300 on 21/8/98, Bartlomiej Papierski wrote:
>
>
> > I'm trying to compile JDBC interface for PostgreSQL 6.3 using JDK 1.1.6
> > under Linux 2.0.34.
> > I took JDBC sources from PostgreSQL 6.3 distribution.
> > I also tried with earlier versions of JDK. The problem is that jar
> > produces 'segmentation fault'.
> >
> > SIGSEGV 11* segmentation violation
> > stackbase=0xbffff360, stackpointer=0xbffff264
> >
> > Full thread dump:
> > "Finalizer thread" (TID:0x4049c208, sys_thread_t:0x413a0f04, state:R)
> ...
>
> Very familiar error... I use Java-in-cgi, and when we stop and start the
> CGI server, the same error appears in the Java programs when they run...
> One of the reasons we're moving on to servlets...
Under normal use, the java interpretter does try to connect to the X
server when it starts, even if your application doesn't use any GUI.
With the linux JDK, try setting NS_JAVA to any non-zero length value.
> We didn't narrow the problem down to absolute certainty, but apparently,
> because all the Java tools (java, jre, jar) are, in fact, korn shell (ksh)
> scripts - they are affected oddly by things in your environment.
>
> My suggestion to you is this - start a new shell before you compile, and
> make sure you have only the bare minimum things in your environment. The
> CLASSPATH (minimal), the PATH (also minimal, to find the java stuff and the
> make), the barebones stuff (USER, PWD). Now that you have a clean shell, do
> the make, and may the force be with you...
I also found this was true in the past with some versions of CVSUP. If
there was too much in the environment, it would fail horribly.
--
Peter T Mount peter@retep.org.uk
Main Homepage: http://www.retep.org.uk
PostgreSQL JDBC Faq: http://www.retep.org.uk/postgres
Java PDF Generator: http://www.retep.org.uk/pdf