Date: Mon, 2 Feb 1998 22:55:39 -0700 From: Nate Williams <nate@mt.sri.com> To: Ian Grigg <iang@systemics.com> Cc: freebsd-java@FreeBSD.ORG Subject: Re: jdk115 JNI requires ELF, gcc makes a.out Message-ID: <199802030555.WAA00791@mt.sri.com> In-Reply-To: <34D6A37E.29B3C1AE@systemics.com> References: <199802021801.MAA06681@damon.com> <34D6A37E.29B3C1AE@systemics.com>
next in thread | previous in thread | raw e-mail | index | archive | help
> will soon rival the JDK :-). I'm now seeing useful debugging, and I get
> this:
>
> cannot stat
> "/usr/local/apps/java/java/bin/../lib/i386/green_threads//home/iang/src/helplets/current/native/build/x86-freebsd/libBlowfish.so"
> : No such file or directory
> (/home/iang/src/helplets/current/native/build/x86-freebsd/libBlowfish.so)
> java.lang.UnsatisfiedLinkError:
> /home/iang/src/helplets/current/native/build/x86-freebsd/libBlowfish.so
> ...
Hmm, I have an idea what might be causing this, but my *very simple*
test code appears to work fine loading libraries:
public class Lib {
public static void main(String args[]) {
new Lib();
}
static {
try {
System.load("/usr/lib/libz.so.2.0");
System.out.println("Library loaded...");
} catch (Exception e) {
System.err.println("Error:" + e);
}
}
}
> System.load("/home/iang/src/helplets/current/native/build/x86-freebsd/libBlowfish.so");
>
> (sorry about the long names...).
No problem.
> It is insisting on turning the absolute pathname into one relative to
> /usr/local/apps/java/java/bin/../lib/i386/green_threads/
I have an idea what might be causing that, but I wouldn't think it would
be doing that in the Java code.
> Now, I can fix this by setting LD_LIBRARYPATH to the current dir and
> using a relative path:
>
> $ LD_LIBRARY_PATH=$PWD java Blowfish.BlowfishNative
> build/x86-freebsd/libBlowfish.so
>
> BlowNat: load: build/x86-freebsd/libBlowfish.so
> BlowNat: lib loaded successfully.
> BlowNat: Loaded.
> BlowNat: FAILED: java.lang.UnsatisfiedLinkError: getLibMajorVersion
>
> Which gets us nicely up to (my) next bug.
I'm assuming the UnsatisfiedLinkErro is yours and not ours, right?
> However, that is the relative
> case, not the absolute case, the absolute still fails:
As above, right.
> Perhaps I'd better throw some questions in:
>
> * is "complete" synonymous with absolute?
Totally qualified, and not relative. (Absolute works too.)
> * is this a bug, or a misinterpretation?
Possibly the former.
> * should I be using System.loadLibrary?
No.
> * can I use absolute, relative, or what?
Absolute (according to the docs.)
> Note that at this stage I can survive, as a relative solution is fine.
> But I'd like to know what the end verdict is, for the doco.
However, I *believe* that it will pull things out of the LD_LIBRARY_PATH
if you don't specify the entire path. So, using my above if I set the
LD_LIBRARY_PATH to /usr/lib, I can do:
System.load("libz.so.2.0");
And it will work. (If I don't set the LD_LIBRARY_PATH, by default it is
set to /usr/local/jdk1.1.5/lib/i386/green_threads).
But, if you give it the absolute path, it *should* work.
Any chance of having you simplifying your test program into less than
100 lines? *grin*
Nate
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199802030555.WAA00791>