This is the mail archive of the gcc-bugs@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Bug libgcj/7938] java-rmi-problem


PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=7938



------- Additional Comments From norbertf at gmx dot net  2003-08-04 02:16 -------
gcj-rmi/jdk-rmi interoperability:

i did play with gcj-rmi/jdk-rmi interoperability a bit longer (i am using the
gcj mainline branch, which i have downloaded a couple of days ago).

this time i started testsrv (gcj) on port 8000 and started the client with jdk
1.4.0_01 several times. the strange thing: the rmi-call worked in about one out
of 10 cases. see the output a) and b)!

this random behavior lets me think, that there must be a timing or threading
problem - or some problem in the basic network classes. but it also means, that
gcj-rmi and jdk-rmi are not far from playing together (the lookup of the object
in the registry obviously worked).

the other way round (gcj calling jdk) the behavior was more deterministic.
"status" and "com3" always worked.

a) -----------------------------
$ java RMIClient //192.168.2.9:8000/rmitestsrv com3
Exception in thread "main" java.rmi.UnmarshalException: Error unmarshaling
return header; nested exception is:
        java.io.EOFException
        at sun.rmi.transport.StreamRemoteCall.executeCall(StreamRemoteCall.java:203)
        at sun.rmi.server.UnicastRef.invoke(UnicastRef.java:133)
        at RMITestServer_Stub.com3(RMITestServer_Stub.java:57)
        at RMIClient.main(RMIClient.java:46)
Caused by: java.io.EOFException
        at java.io.DataInputStream.readByte(DataInputStream.java:278)
        at sun.rmi.transport.StreamRemoteCall.executeCall(StreamRemoteCall.java:189)
        ... 3 more

b) ----------------------------
$ java RMIClient //192.168.2.9:8000/rmitestsrv com3
method com3
writeValue() valueClass=boolean
method com3
true
 
---------------------------

other issues:

1) the RMI-url //:port/xxxx is currently not accepted by gcj-rmi. I think it
should treated as //0.0.0.0:port/xxxx.

2) rmic -v1.1 or -vcompat does not work - it never returns.

 
norbert


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]