Re: irritability of structure name conflicts

From: John Skaller (skaller@maxtal.com.au)
Date: Wed Jul 14 1999 - 05:04:36 MET DST


Message-Id: <3.0.6.32.19990714130436.00987970@mail.triode.net.au>
Date: Wed, 14 Jul 1999 13:04:36 +1000
To: Dave Mason <dmason@sarg.Ryerson.CA>, caml-list@inria.fr
From: John Skaller <skaller@maxtal.com.au>
Subject: Re: irritability of structure name conflicts
In-Reply-To: <199907131759.NAA00710@sarg.Ryerson.CA>

At 13:59 13/07/99 -0400, Dave Mason wrote:
>Is there some way I can force it to interpret the type constructor as
>a particular type so that it chooses the labels from the correct set?

This problem pervades ocaml. Another case of it is variants:

        type A = X | Y;
        type B = X | Z;

which all compiles fine, but then X is a constructor for
the B variant.

C also has the same problem with enumerations:

        enum X {a,b,c}; /* a,b,c are globals */

The problem can be solved, clumbsily, by wrapping inside a module.
(In C++, inside a class or namespace)

While I'm also expressing irritation, when I compile
a file (as opposed to using the interactive system);
rebinding of symbols shouldn't be allowed:

        let a = 1;;
        ...
        let a = 1.0;;
        ...
        print_int a;;

        (* what does that strange type error mean!
              a is an integer! It is right there at the top of the file!
        *)

Another similar case is

        open List
        open Map
        iter ...

Woops! I meant List.iter, not Map.iter :-)

There are several solutions, but the one that could
be implemented now with least impact would be to
issue warnings whenever a symbol is rebound
in the same scope during batch compilation.

(No warning in the interactive system; rebinding
to fix mistakes is common)

-------------------------------------------------------
John Skaller email: skaller@maxtal.com.au
                http://www.maxtal.com.au/~skaller
                phone: 61-2-96600850
                snail: 10/1 Toxteth Rd, Glebe NSW 2037, Australia



This archive was generated by hypermail 2b29 : Sun Jan 02 2000 - 11:58:23 MET