[Caml-list] on the design of Dynlink

[Dave Mason <dmason@sarg.ryerson.ca>]

I am curious about a design feature of Dynlink.

Assume module C uses module B and module B uses module A.

If A loads B and C, then it seems that C can see all of A and
everything it depends on.  This seems like a very unfortunate outcome,
because it means (for example) that C can now use Dynlink itself to
load things, and there is no way for A or B to provide an interface to
hide things from the modules that (transitively) use them.

Now, I understand how to work around this (provide a Registry module
that A calls to register the functions that it wants to provide to B
and C and then they call the Registry to get those values).  But that
is not a very intuitive style, and can make it difficult to extend B
in a completely general way unless, for each module B, I provide a
B'Registry and always load the pair.

Have I misunderstood something, or if I have it right, is there any
hope for this to work differently in future?

Thanks  ../Dave
-------------------
To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr
Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners