Comment author: administrator

Dear Markus,

using the latest CVS-release, the following program behaves differently
when interpreted with "ocaml" or when compiled (byte- or native code):

---

exception E of char
let _ = raise (E 'a')

The interpreter prints:

Exception: E 'a'.

Running the compiled program yields:

Fatal error: exception Foo.E(97)

This behavior has been there for quite a while. The toplevel has
access to typing information, and therefore can print the exception
value correctly. The "Fatal error" message of the run-time system
has access only to the machine representations of the arguments to the
exceptions, and thus cannot print them accurately, e.g. it can't
distinguish an integer from a character. I know it's a bit
surprising, but sometimes this behavior of the run-time system is more
helpful than not printing exception arguments at all.

Best wishes,

• Xavier Leroy

Comment author: administrator

not a bug DD 2004-04-29