Petter Reinholdtsen <[EMAIL PROTECTED]> writes:
> I ran into similar problems. I'm using JDK 1.2, and had to increase
> the max heap value to javac to get it to compile. I then used javah
> to generate the header files, and it seems like the headers are
> generated based on the JDK classes, not Classpath.
>
I gave JDK 1.3 beta for Linux a try tonight and ran into a few
problems trying to compile things. I'll take a look again later this
week when I get a chance to find and fix these problems if it is
possible.
> The return value is different. The same thing happends with
> FileOutputStream. As far as I can see from the documentation, this
> function is not a public function in JDK. Is this difference
> intentional?
>
> While compiling, I rad across this error:
>
> Index: java_io_ObjectInputStream.c
> ===================================================================
> RCS file: /cvs/classpath/native/java.io/java_io_ObjectInputStream.c,v
> retrieving revision 1.7
> diff -u -w -r1.7 java_io_ObjectInputStream.c
> --- java_io_ObjectInputStream.c 2000/03/17 20:59:11 1.7
> +++ java_io_ObjectInputStream.c 2000/06/20 03:40:18
> @@ -43,9 +43,10 @@
> "()Ljava/lang/ClassLoader;" );
>
> if( id == NULL )
> - return;
> + return 0;
>
> (*env)->CallObjectMethod( env, loader, id );
> + return loader; /* XXX Not quite sure what should be returned here. */
> }
>
>
> The function have return type jobject, but did not return anything.
>
> BTW: The classpath/javax/ subdirectory is missing / empty. Is this
> normal?
Yes, sort of. I sent email to cvs-hackers tonight asking how to get
the swing project files moved back to this directory location from the
project directory since we don't have access to the CVS repository in
an administrative type way.
Brian
--
Brian Jones <[EMAIL PROTECTED]>
