The new feature of the rmid tool in J2SE 5.0 to support launching from
inetd/xinetd leads to the following fatal behavior:
If rmid is started via rcmd from another computer it behaves as if it was
started by inetd. It writes the error message "rmid: rmid was launched
from inetd with an invalid status (must be wait)" followed by the usage
message into a file in the tmpdir and terminates. After looking at the source
code we think that this seems to happen because rmid checks whether file
descriptor 0 (stdin) is a socket and if yes decides that it must have been
started by inetd. When rmid has been started via rcmd this assumption is
obviously wrong and leads to the abnormal termination.
###@###.### 2005-04-26 13:52:33 GMT
inetd/xinetd leads to the following fatal behavior:
If rmid is started via rcmd from another computer it behaves as if it was
started by inetd. It writes the error message "rmid: rmid was launched
from inetd with an invalid status (must be wait)" followed by the usage
message into a file in the tmpdir and terminates. After looking at the source
code we think that this seems to happen because rmid checks whether file
descriptor 0 (stdin) is a socket and if yes decides that it must have been
started by inetd. When rmid has been started via rcmd this assumption is
obviously wrong and leads to the abnormal termination.
###@###.### 2005-04-26 13:52:33 GMT