[mercury-users] Maths and contexts.

[Mark Brown]

On 21-Apr-2012, Julien Fischer <juliensf at csse.unimelb.edu.au> wrote:
>
> On Sat, 21 Apr 2012, Mark Brown wrote:
>
>> On 20-Apr-2012, Michael Richter <ttmrichter at gmail.com> wrote:
>>> On 20 April 2012 23:33, Julien Fischer <juliensf at csse.unimelb.edu.au> 
>>> wrote:
>>>
>>>>   :- module test.
>>>>   :- interface.
>>>>   :- import_module io.
>>>>   :- pred main(io::di, io::uo) is det.
>>>>   :- implementation.
>>>>   :- import_module decimal2.
>>>>   :- import_module bool, integer.
>>>>
>>>>   main(!IO) :-
>>>>      Context1 = make_context(integer(30), round_floor, signal(no, no)),
>>>>      Context2 = make_context(integer(40), round_up, signal(no, no)),
>>>>      X = to_number(Context1, "1000000"),
>>>>      Y = to_number(Context2, "2900000"),
>>>>      Z = X `add` Y,
>>>>      io.write_string(to_scientific_**string(Z), !IO),
>>>>      io.nl(!IO).
>>>>
>>>> Attempting to compile the above will result in the following compilation
>>>> error:
>>>>
>>>>   Making Mercury/cs/test.c
>>>>   test.m:013: In clause for predicate `main'/2:
>>>>   test.m:013:   in unification of variable `Z'
>>>>   test.m:013:   and term `add(X, Y)':
>>>>   test.m:013:   type error in argument(s) of functor `add/2'.
>>>>   test.m:013:   Argument 1 (X) has type `(some [C] 
>>>> decimal2.decimal(C))',
>>>>   test.m:013:   expected type was `decimal2.decimal(C)';
>>>>   test.m:013:   argument 2 (Y) has type `(some [C] 
>>>> decimal2.decimal(C))',
>>>>   test.m:013:   expected type was `decimal2.decimal(C)'.
>>>>   ** Error making `Mercury/cs/test.c'.
>>>>
>>>
>>> OK, thanks, Julien.  That's looking like a very promising line of 
>>> inquiry.
>>>  Adding the context to the creation of a number was the line I was going 
>>> to
>>> originally investigate but I was concerned at the need for run-time
>>> checking.  Your solution eliminates that problem for me.
>>
>> Julien said it gives you compile-time checking, not that it eliminates
>> run-time checking.  You'll still need run-time checks if two contexts
>> don't originate from the same call to make_context.
>
> What runtime checks are you performing there?  That the two contexts
> are in fact semantically identical?

That would be sufficient to cover all the cases that Paul's version can
check at compile-time.

Cheers,
Mark.

--------------------------------------------------------------------------
mercury-users mailing list
Post messages to:       mercury-users at csse.unimelb.edu.au
Administrative Queries: owner-mercury-users at csse.unimelb.edu.au
Subscriptions:          mercury-users-request at csse.unimelb.edu.au
--------------------------------------------------------------------------