]> I am thinking of a system where I would be dynamically loading files, repeatedly in a long-lived OCaml process. But I am a bit uncertain how things are handled. Let's just assume Dynlink.loadfile_private for now - since loadfile seems to open up a can of worms I'd rather not deal with, including segmentation faults if interfaces change. What is actually changed when I do Dynlink.loadfile_private? What happens when I do Dynlink.loadfile_private again on the same file? Assuming also that the file could have changed in the meanwhile. Atleast something is left behind right now. If I do: ,---- | Dynlink.init ();; | Gc.print_stat stdout;; | for i = 0 to 50000 do | Dynlink.loadfile_private "test.cmo"; | done;; | Gc.compact ();; | Gc.print_stat stdout;; `---- I find that the process eats 20 megs of memory after loading the file 50000 times. And if I change that to be 100000 times, I get a fatal exception of Stack_overflow. This is a bit worrying, since it might have to scale up to 100000 different files some day as well. So what gives? Or should I just read the source more? -- Naked ------------------- 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