[m-dev.] shims

[Mark Brown]

On Mon, Sep 15, 2014 at 3:20 PM, Julien Fischer <jfischer at opturion.com> wrote:
>
>
> On Fri, 12 Sep 2014, Peter Wang wrote:
>
>> On Thu, 11 Sep 2014 17:26:04 +1000, Peter Wang <novalazy at gmail.com> wrote:
>>>
>>>
>>> Ideally you could update code immediately but maintain a fallback for
>>> older libraries.  I don't think there is a way to do it currently, so
>>> perhaps it is time for another pragma, e.g.
>>>
>>>     :- pragma shim(p/1, other.p/1).
>>>
>>>     p(X) :- ...
>>
>>
>> This could not cope with predicates which just have their arguments
>> reordered (hopefully a rare occurrence).

If you require that other.p/1 not only exists, but has the same
signature as p/1, then I think it will cope so long as the arguments
don't have the same type.

(If the signatures don't change, my proposal wouldn't cope either.)

>
>
> Over the past few years it's been a very common occurrence in Mercury's
> standard library.  (I'll grant that those reorderings were primarily
> legacy issues and that APIs built against a current version of Mercury
> are unlikely to have the same sort of problems.)

In most of these cases the arguments have different types so the
signature has changed, right?

Cheers,
Mark.