CNI changes (Was: Binary Compatibility)

Tom Tromey tromey@redhat.com
Mon Aug 4 17:19:00 GMT 2003


>>>>> "Jeff" == Jeff Sturm <jsturm@one-point.com> writes:

Jeff> If _Jv_LookupJNIMethod has no important side effects here, my
Jeff> hunch is that the PR should be closed.

It doesn't have any side effects we need to be concerned with.

(It might allocate memory, and it acquires a lock, both of which are
side effects.  And in theory there is a race where another thread
could load another .so that implements a different form of the JNI
method -- but this is pathological and probably incorrect on the
user's part anyway.)

Tom



More information about the Java mailing list