Brian Rogoff <bpr@best.com> writes:

> > > <hint for next ocaml ;)> > 
> >   It would be very nice to be able to rely on Marshal as safely as on
[...]
> > </hint for next ocaml ;)>
> 
> A more type safe marshaling framework is way up there on my list of
> desired enhancements. I think this will be part of the extensional
> polymorphism enhancements that you can see now in G'Caml. 

In my opinion, they are two different kinds of marshaling that one could
expect: the safe one and the correct one.

Supose that I have :

 type t = A | B of int | C of float

To be safe is just to ensure that if I marshal a B(3), then if I receive
a B(2.6) it will raise an exception. In other words, the marshaled data
structure should contain enough information to ensure that once
unmarshaled, it won't trigger a segfault. That's all. Basically, it
means to encode in some way the type constructor of sent data structure
(in my example, something like (type t: first: nothing; second: int;
third: float)), and to check those constructors at unmarshaling time.

The correct encoding is of course to ensure that the A of my program is
the same A of the .mli you are using to send to me a data structure. It
is of course much more difficult.

I have probably missed something on the difficulty of marshaling. I
tried to look at the mailing list archive, but did not find out the
relevant messages. Any pointer?

Best regards,
d.
-- 
 David.Mentre@inria.fr
 Opinions expressed here are only mine.
-------------------
Bug reports: http://caml.inria.fr/bin/caml-bugs  FAQ: http://caml.inria.fr/FAQ/
To unsubscribe, mail caml-list-request@inria.fr  Archives: http://caml.inria.fr