java/decl.c patch for clear_binding_level

Zack Weinberg zackw@stanford.edu
Sat Mar 17 14:00:00 GMT 2001


Just a note - with this patch series, the libjava test suite is now a
lot happier.  I'm at

                === libjava Summary ===

# of expected passes            1348
# of unexpected failures        96
# of unexpected successes       12
# of expected failures          264

Two days ago that was more like 200 unexpected and 500 expected
failures (it uses XFAIL for execution tests when the program failed to
compile - shouldn't that be UNTESTED?)  Note this is with Alexandre's
patch to work around libtool thinking gcj doesn't support -c with -o.

I am still seeing this sort of failure:

Executing on host:
/home/zack/src/b/gcc_vanilla/i686-pc-linux-gnu/libjava/testsuite/../libtool --silent --tag=GCJ --mode=link /home/zack/src/b/gcc_vanilla/gcc/gcj -B/home/zack/src/b/gcc_vanilla/gcc/ -B/home/zack/src/b/gcc_vanilla/i686-pc-linux-gnu/libjava/testsuite/../  /home/zack/src/b/gcc_vanilla/i686-pc-linux-gnu/libjava/testsuite/A.class /home/zack/src/b/gcc_vanilla/i686-pc-linux-gnu/libjava/testsuite/PR260.class /home/zack/src/b/gcc_vanilla/i686-pc-linux-gnu//libjava/libgcj.la /home/zack/src/b/gcc_vanilla/i686-pc-linux-gnu//boehm-gc/libgcjgc.la /home/zack/src/b/gcc_vanilla/i686-pc-linux-gnu//zlib/libzgcj.la   --main=PR260 -g -L/home/zack/src/b/gcc_vanilla/i686-pc-linux-gnu//libjava/.libs -L/home/zack/src/b/gcc_vanilla/i686-pc-linux-gnu//boehm-gc/.libs -L/home/zack/src/b/gcc_vanilla/i686-pc-linux-gnu//zlib/.libs  -lm   -o /home/zack/src/b/gcc_vanilla/i686-pc-linux-gnu/libjava/testsuite/PR260    (timeout = 300)
jc1: No such file or directory: can't open ^?ELF^A^A^A
jc1: No such file or directory: can't open ^?ELF^A^A^A
jc1: No such file or directory: can't open ^?ELF^A^A^A

caused by libgcj.so, libgcjgc.so, and libzgcj.so being fed to jc1
instead of the linker.  This appears to happen every time gcj is asked
to link multiple .class files together.  There are usually bytecode
verifier errors associated.

There are also a number of failures like so:

PASS: Thread_Join -O compilation from source
libgcj failure: only thread already running
FAIL: Thread_Join -O execution from source compiled test
XFAIL: Thread_Join -O execution from source compiled test
PASS: Thread_Join byte compilation

(why does that get both a FAIL and an XFAIL?); and a very few
'compilation from bytecode' failures with no error message attached,
e.g. 

Executing on host: /home/zack/src/b/gcc_vanilla/i686-pc-linux-gnu/libjava/testsuite/../libtool --silent --tag=GCJ --mode=link /home/zack/src/b/gcc_vanilla/gcc/gcj -B/home/zack/src/b/gcc_vanilla/gcc/ -B/home/zack/src/b/gcc_vanilla/i686-pc-linux-gnu/libjava/testsuite/../ /home/zack/src/gcc_vanilla/libjava/testsuite/libjava.compile/assignment_2.java /home/zack/src/b/gcc_vanilla/i686-pc-linux-gnu//libjava/libgcj.la /home/zack/src/b/gcc_vanilla/i686-pc-linux-gnu//boehm-gc/libgcjgc.la /home/zack/src/b/gcc_vanilla/i686-pc-linux-gnu//zlib/libzgcj.la   --main=assignment_2 -g -L/home/zack/src/b/gcc_vanilla/i686-pc-linux-gnu//libjava/.libs -L/home/zack/src/b/gcc_vanilla/i686-pc-linux-gnu//boehm-gc/.libs -L/home/zack/src/b/gcc_vanilla/i686-pc-linux-gnu//zlib/.libs  -lm   -o /home/zack/src/b/gcc_vanilla/i686-pc-linux-gnu/libjava/testsuite/assignment_2    (timeout = 300)
PASS: assignment_2 compilation from source
FAIL: assignment_2 byte compilation
XFAIL: assignment_2 compilation from bytecode

zw



More information about the Java-patches mailing list