corection - i was setting the LD_LIBRARY_PATH to /usr/local/java/jre/lib/i386/ not
/usr/local/java/jre/lib/i386/classic
Aaron Stromas wrote:
> Hello again,
>
> I managed to build my library using the the GNU linker
> (ld -G -f /usr/lin/libdb.so -o libkeys.so) but another problem surfaced - I now get
>java.lang.UnsatisfiedLinkError: /home/ams/work/keys/libkeys.so: undefined
> symbol: GetStringUTFChars
>
> I pass -Djava.library.path=/home/ams/work/keys which according to everything i read
>should be sufficient. Besides, setting
> LD_LIBRARY_PATH=/usr/local/java/jre/lib/i386/classic doesn't help either.
>
> Any idea what is going on? TIA,
>
> -a
>
> Joi Ellis wrote:
>
> > "Aaron M. Stromas" wrote:
> > >
> > > greetings,
> > >
> > > I need a little assistance with building a shared library for JNI under
> > > Linux 2.2.
> > > This is how I went about creating the library:
> > >
> > > compiling: g++ -I /usr/local/java/include
> > > -I/usr/local/java/include/linux -shared -fPIC -o keys.o keys.c
> > >
> > > linking: gcc -shared -Wl,-soname,libkeys.so -o libkeys.so.0 keys.o -lc
> > >
> > > $ java -Djava.library.path=/home/ams/work/keys T
> > > Exception in thread "main" java.lang.UnsatisfiedLinkError:
>/home/ams/work/keys/libkeys.so: keys.o: cannot open shared object file: No such file
>or directory
> > > at java.lang.ClassLoader$NativeLibrary.load(Native Method)
> > > at java.lang.ClassLoader.loadLibrary0(ClassLoader.java:1357)
> > > at java.lang.ClassLoader.loadLibrary(ClassLoader.java:1281)
> > > at java.lang.Runtime.loadLibrary0(Runtime.java:469)
> > > at java.lang.System.loadLibrary(System.java:774)
> > > at T.<clinit>(T.java:5)
> > >
> > > I have also tried another way of building the library (ld -G -f
> > > /lib/libuuid.so.1 keys.o -o libkeys.so) but got the same error. I have
> > > to mention that the native code accesses the libdb.so library which i
> > > copied to /home/ams/work/keys. Why doesn't java find my library? TIA,
> >
> > I found that JNI code can compile and link without error and still be
> > incorrect, and the only error you do get is the one above. I diagnose
> > this with 'ldd -r /path/to/library/file' and see if there are any
> > unresolved references that don't point to an existing library.
> >
> > Your library is looking for a routine "keys" which does not exist in
> > that library, or any of the other .o files you sent through the linker.
> >
> > Here's bits of a Makefile with the gcc/library commands I use:
> >
> > JAVA_HOME=/usr/local/java/jdk1.2.2
> > JAVA_INC= -I$(JAVA_HOME)/include -I$(JAVA_HOME)/include/linux
> #
> > # gcc configurable compile options of interest
> > # -H list include header names
> > # -g debugging code enabled
> > GCCOPT= -g
> > LDOPT=-Xlinker -t -Xlinker -Map -Xlinker /tmp/linkmap
> > GCCDEF=-DIP_SUPPORT -DUNIX
> > LIBOUT=$(HOME)/lib
> > SRC=$(PROJ)/src
> >
> > $(LIBOUT)/libjsystem.so: \
> > $(SRC)/JNsApiSubsystemInterfaces.h \
> > $(SRC)/JNsApiSubsystemInterfaces.c \
> > $(CLASSES)/com/aravox/nsapi/subsystem/JNsApiSubsystemInterfaces.class
> > - rm -f $@
> > @echo Creating $@...
> > gcc $(GCCOPT) $(GCCDEF) $(JAVA_INC) -I$(SRC) \
> > $(LDOPT) \
> > -shared -o $@ $(SRC)/JNsApiSubsystemInterfaces.c \
> > $(LIBOUT)/lib*.so
> > /home/joi/bin/anyUndefined $(LIBOUT)/lib*.so
> >
> >
> > anyUndefined is a bash script which inspects the resulting .so for
> > unresolved references and exits with an error code make will detect.
> >
> > #!/bin/bash
> > tmp=/tmp/anyUndefined.$$
> > ldd -r $1 >$tmp 2>&1
> > echo "----------"
> > cat $tmp
> > echo "----------"
> > if
> > grep undefined $tmp;
> > then
> > exit 1
> > else
> > exit 0
> > fi
> >
> > --
> > Joi Ellis Software Engineer
> > Aravox Technologies [EMAIL PROTECTED], [EMAIL PROTECTED]
> >
> > No matter what we think of Linux versus FreeBSD, etc., the one thing I
> > really like about Linux is that it has Microsoft worried. Anything
> > that kicks a monopoly in the pants has got to be good for something.
> > - Chris Johnson
>
> --
> Aaron Stromas | "Tick-tick-tick!!!... ja, Pantani is weg..."
> Oracle Corp | BRTN commentator
> +1 703.708.68.21 | L'Alpe d'Huez
> 1995 Tour de France
--
Aaron Stromas | "Tick-tick-tick!!!... ja, Pantani is weg..."
Oracle Corp | BRTN commentator
+1 703.708.68.21 | L'Alpe d'Huez
1995 Tour de France
----------------------------------------------------------------------
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
