Re: Questions about __cxa_throw and type_info

[Tom Tromey <tromey at redhat dot com>]

>>>>> "Brian" == Brian Ellis <nullroute@yahoo.com> writes:

Brian> I already have a proof of concept that is able to forge a
Brian> virtual table and instance an object. The virtual methods that
Brian> are overridden by the new forged virtual table enter a thunk
Brian> which invokes the interpreter passing the this pointer and
Brian> virtual function id as parameters.

We do a lot of this same stuff in libgcj -- we make new classes at
runtime; make calls between interpreted and compiled code; can have
interpreted methods in a new class override methods in its superclass;
etc.

Unfortunately for you our approach is much simpler than handling all
of C++ -- we only handle Java, which has single inheritance (it is
more complicated really, but our C++ code doesn't know that), and we
only handle the subset of exceptions that Java can use.  Also we rely
on gcj to emit a lot of metadata describing class and vtable layout.

Doing this for C++ would be a nice library project IMO, say as an
adjunct to libffi.  One nice thing about having a standard C++ ABI is
having higher-level tools to manipulate it.. :-)

Brian> The only major concept that I do not yet have a proof-positive
Brian> for is exception throwing and catching. I'd like for the
Brian> interpreter to be able to catch arbitrary objects thrown to it
Brian> from any c++ function that it has invoked as well as throw
Brian> objects that it has created.

In libgcj we only catch the subset of exceptions that are "java-like".
I looked at this problem a little bit years and years ago but we
punted on it ... we didn't need the full generality and it looked
super hard.

Tom