Check libjava completes now, but leaves spurious processes

[Gordon Sadler]

On Sat, Apr 14, 2001 at 12:53:22PM -0600, Tom Tromey wrote:
> >>>>> "Gordon" == Gordon Sadler <gbsadler1@lcisp.com> writes:
> 
> Gordon> WARNING: program timed out.
> Gordon> FAIL: klass execution from source compiled test
> Gordon> WARNING: program timed out.
> Gordon> FAIL: klass execution from bytecode->native test
> Gordon> WARNING: program timed out.
> Gordon> FAIL: klass -O execution from source compiled test
> Gordon> WARNING: program timed out.
> Gordon> FAIL: klass -O execution from bytecode->native test
> 
> Gordon> 		# of unexpected failures        4
> 
> I don't have any good way to debug this, since I can't reproduce this
> failure.
> 
> Does libjava.log say anything interesting?
> What if you run with RUNTESTFLAGS='--verbose'?
> 
> I'm glad you are running the tests.  Unfortunately you're somewhat on
> your own for diagnosing the problem :-(
> 
My libjava.log is showing this for klass.java:

Executing on host: /home/gbsadler/gcc-3.0/i686-pc-linux-gnu/libjava/testsuite/../libtool --silent --tag=GCJ --mode=link /home/gbsadler/gcc-3.0/gcc/gcj -B/home/gbsadler/gcc-3.0/gcc/ -B/home/gbsadler/gcc-3.0/i686-pc-linux-gnu/libjava/testsuite/../ /home/gbsadler/cvs/gcc-3.0/libjava/testsuite/libjava.lang/klass.java /home/gbsadler/gcc-3.0/i686-pc-linux-gnu//libjava/libgcj.la /home/gbsadler/gcc-3.0/i686-pc-linux-gnu//boehm-gc/libgcjgc.la /home/gbsadler/gcc-3.0/i686-pc-linux-gnu//zlib/libzgcj.la   --main=klass -g -L/home/gbsadler/gcc-3.0/i686-pc-linux-gnu//libjava/.libs -L/home/gbsadler/gcc-3.0/i686-pc-linux-gnu//boehm-gc/.libs -L/home/gbsadler/gcc-3.0/i686-pc-linux-gnu//zlib/.libs  -lm   -o /home/gbsadler/gcc-3.0/i686-pc-linux-gnu/libjava/testsuite/klass    (timeout = 300)
PASS: klass compilation from source
Exception in thread "main" WARNING: program timed out.
FAIL: klass execution from source compiled test
UNTESTED: klass output from source compiled test
PASS: klass byte compilation
Executing on host: /home/gbsadler/gcc-3.0/gcc/jv-scan /home/gbsadler/cvs/gcc-3.0/libjava/testsuite/libjava.lang/klass.java  --print-main     (timeout = 300)
klassoutput is:
klass
Executing on host: /home/gbsadler/gcc-3.0/gcc/jv-scan /home/gbsadler/cvs/gcc-3.0/libjava/testsuite/libjava.lang/klass.java  --list-class     (timeout = 300)
klass output is:
klass

Debian unstable.
In order to attempt.. to debug this, I installed this build in a local
directory. Sun JDK1.3 is installed as well as jikes 1.13. I added
gcj-2.95 to the mix as well. All 'java' cmds are Sun JDK VM.

I copied klass.java to my $HOME. Then tried:
javac klass.java (SUN JDK)
java klass
true

jikes klass.java
java klass
true

gcj-2.95 -C klass.java
java klass
true

gcj-2.95 klass.java --main=klass
./a.out
Aborted

gcj-3.0 -C klass.java
java klass
true

gcj-3.0 klass.java --main=klass
./a.out
Exception in thread "main" java.lang.NullPointerException
   at 0x401664ec: _Jv_ThrowSignal (/home/gbsadler/gcc-2001-04-14/lib/libgcj.so.2)
   at 0x4016652d: _Jv_ThrowSignal (/home/gbsadler/gcc-2001-04-14/lib/libgcj.so.2)
   at 0x401840a7: java.lang.Object.clone() (/home/gbsadler/gcc-2001-04-14/lib/libgcj.so.2)
   at 0x08048cd4: klass::main(JArray<java::lang::String*>*) (/home/gbsadler/klass.java:8)
   at 0x4017d40b: gnu.gcj.runtime.FirstThread.run() (/home/gbsadler/gcc-2001-04-14/lib/libgcj.so.2)
   at 0x4018868b: java.lang.Thread.run_(java.lang.Object) (/home/gbsadler/gcc-2001-04-14/lib/libgcj.so.2)
   at 0x4029d625: _Jv_ThreadSetPriority(_Jv_Thread_t, int) (/home/gbsadler/gcc-2001-04-14/lib/libgcj.so.2)
   at 0x40440414: GC_start_routine (/home/gbsadler/gcc-2001-04-14/lib/libgcjgc.so.1)
   at 0x4045c065: pthread_detach (/lib/libpthread.so.0)
   at 0x4054ba4a: __clone (/lib/libc.so.6)

Thought I'd be 'smart' try gdb 5.0 on a.out:
(gdb) r
Starting program: /home/gbsadler/./a.out
[New Thread 1024 (LWP 22236)]
[New Thread 2049 (LWP 22237)]
[New Thread 1026 (LWP 22238)]

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 1026 (LWP 22238)]
0x401840a5 in _ZN4java4lang6Object5cloneEv () from /home/gbsadler/gcc-2001-04-14/lib/libgcj.so.2
(gdb) bt
#0  0x401840a5 in _ZN4java4lang6Object5cloneEv () from /home/gbsadler/gcc-2001-04-14/lib/libgcj.so.2
#1  0x8048cd4 in _ZN5klass4mainEP6JArrayIPN4java4lang6StringEE (args=0x40338200) at klass.java:8
#2  0x4017d40b in _ZN3gnu3gcj7runtime11FirstThread3runEv () from /home/gbsadler/gcc-2001-04-14/lib/libgcj.so.2
#3  0x4018868b in _ZN4java4lang6Thread4run_EPNS0_6ObjectE () from /home/gbsadler/gcc-2001-04-14/lib/libgcj.so.2
#4  0x4029d625 in _Z12really_startPv () from /home/gbsadler/gcc-2001-04-14/lib/libgcj.so.2
#5  0x40440414 in GC_start_routine () from /home/gbsadler/gcc-2001-04-14/lib/libgcjgc.so.1
#6  0x4045c065 in pthread_start_thread () from /lib/libpthread.so.0
#7  0x4045c0ad in pthread_start_thread_event () from /lib/libpthread.so.0
							  
Still mangled here... not sure if gdb 5.0 is designed/able to handle this?

Hmm, at least gcj-2.95 shows same results, not a regression...

Gordon Sadler

PS
 Doing RUNTESTFLAGS='--verbose' next.