Re: [3.3] Patch ping

[Ranjit Mathew <rmathew at hotmail dot com>]

Ranjit Mathew wrote:
> > [ ] The patch either fixes an existing testcase, or adds a new testcase.
>
> Got me there...the "libjava/testsuite/libjava.jni" testcases should have
> JNIEXPORT and JNICALL in the function definitions as per the JNI ABI.
I have posted a patch for the "libjava.jni" testcases to fix the above
issue:

    http://gcc.gnu.org/ml/java-patches/2003-q1/msg00165.html

Since I do a crossed-native build of GCC for MinGW from Linux, I run
the testcases manually on Windows. The following is the result with
the "libjava.jni/invoke" testcase which tests the basic ability of
Java methods to call C functions using JNI - this is against the
current GCC 3.3 on a Win98 box running MinGW/MSYS (target "mingw32").

First with GCJ (compiler and interpreter) without my patches:
------------------------------------- 8< -------------------------------------
/c/temp/libjava.jni > /j/wingcc/bin/gcj -fjni --main=invoke invoke.java
/c/temp/libjava.jni > a
trying...
Exception in thread "main" java.lang.ExceptionInInitializerError
   <<No stacktrace available>>
Caused by: java.lang.ArrayIndexOutOfBoundsException
   <<No stacktrace available>>
/c/temp/libjava.jni >
/c/temp/libjava.jni > /j/wingcc/bin/gij invoke
trying...

abnormal program termination
/c/temp/libjava.jni >
------------------------------------- 8< -------------------------------------

Now with GCJ (compiler and interpreter) with my JNI patches applied:
------------------------------------- 8< -------------------------------------
/c/temp/libjava.jni > gcj -fjni --main=invoke invoke.java
/c/temp/libjava.jni > a
trying...
loaded
23
/c/temp/libjava.jni > gij invoke
trying...
loaded
23
/c/temp/libjava.jni >
------------------------------------- 8< -------------------------------------

Sincerely Yours,
Ranjit.

--
Ranjit Mathew        Email: rmathew AT hotmail DOT com
Bangalore,
INDIA.               Web: http://ranjitmathew.tripod.com/