This is the mail archive of the java@gcc.gnu.org mailing list for the Java 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]

Exception handling unclear


Exception handling is discussed in VMSpecs, but there is no discussion
what happens when error/exception of exception/error happens, when circular
exceptions happen. I guess the VM must detect this. I've tried this with Sun's VM
and it clearly does not handle even the simplest of such complications, basically
unpredictable behavior. This are fatal errors for sure, same with OutOfMemory
on a non-GC collected memory. (Clearly cannot create instance when there is no
memory). So there is a set of errors that are fatal. There is info what should happen
then. I think there should be a whole chapter by itself on this, what they are and
what to do when they happen. An Example would be to try to load a class, which is
not there - > ClassNotFound, but ClassNotFound has ClassFormatError, and ClassFormatError
is also not found ->ClassNotFound, etc...

Also, could a user program return with a fatal error to a native method ? When the
user wants to see if ExceptionOccured, then there is no reall instance of that exception/error.
And it will return no exception/error.

I guess implications then are that the thread must perform proper cleanup (ex:release monitors).

Anyway, its interesting to see what you guys think about this issues.

Martin.



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