How to define your own polymorfic type in Mercury
Fergus Henderson
fjh at cs.mu.OZ.AU
Sat Apr 10 19:29:36 AEST 1999
On 09-Apr-1999, Henk Vandecasteele <Henk.Vandecasteele at cs.kuleuven.ac.be> wrote:
> Dear mercury-users(@cs.mu.OZ.AU?)
>
> I have the following interface of a mutable object:
>
> :- module mutable.
>
> :- interface.
>
> :- type mutable(T). /* a polymorfic mutable object */
>
> :- pred mutable__init(mutable(T), T).
> :- mode mutable__init(out, in) is det.
> /* create a new mutable object with the initial contents */
>
> :- pred mutable__overwrite(mutable(T), T).
> :- mode mutable__overwrite(in, in) is det.
> /* Overwrite the current contents of the mutable object with
> a new data. When the system backtracks over this operation
> the old data willl be restored */
Those two predicates should be declared `impure'.
:- impure pred mutable__init(mutable(T), T).
:- impure pred mutable__overwrite(mutable(T), T).
If you don't do that, the compiler might (depending on the exact
compilation options) optimize away calls to them.
> :- pred mutable__get(mutable(T), T).
> :- mode mutable__get(in, out) is det.
> /* Get the current contents of the mutable object */
That one should be declared `semipure', for the same reason.
:- semipure pred mutable__get(mutable(T), T).
> I had an implementation which is based on the array-module in
> the library. I do not understand half wat happened in there,
> I suppose this is not the way to do it. I added the code in the end
> of the mail. It worked up to version 0.8, but it fails
> with "rotd-1999-04-04". Of course I could again see what happened
> with array.m (of course I already did) and mimic it again.
> But is there I better documented way? I searched without luck.
In general, the way to declare a type which is actually implemented
in C is to use the `c_pointer' type, e.g.
:- type mutable(T) ---> mutable(c_pointer).
But in this particular case, there is already a library in the
mercury-extras distribution which does exactly the same thing
as your `mutable' module. See extras/references/reference.m.
--
Fergus Henderson <fjh at cs.mu.oz.au> | "I have always known that the pursuit
WWW: <http://www.cs.mu.oz.au/~fjh> | of excellence is a lethal habit"
PGP: finger fjh at 128.250.37.3 | -- the last words of T. S. Garp.
More information about the users
mailing list