[m-dev.] for review: improvements for record syntax

[Simon Taylor]

Fergus wrote:
> However, note that `:=' is not referentially transparent either.
> For example, `Y = X^f1, R = (Y^f2 := Z)' is not
> the same as `R = ((X^f1)^f2 := Z)'!

Yes, that's a bit icky, but probably unavoidable.

It is possible to do referentially transparent field
update of nested fields using update functions which apply
a mapping function to the fields, as shown below.

Unfortunately, there is no nice way to determine
which modes of the `map_field' functions should
be generated. The modes to generate would depend
on the mapping functions that the user wanted 
to apply to sub-fields of the field to be updated.

Currently we avoid that problem by converting the
compiler-generated `field/1' and `field :=/2' functions
into unifications before mode-checking, but that won't work
for the `map_field' functions.
	
The restrictions on taking the address of functions with
multiple modes would also cause problems.

Simon.


:- module field_update.

:- interface.

:- import_module io.

:- pred main(io__state::di, io__state::uo) is det.

:- implementation.

:- type has_f
        --->    has_f(f :: has_g).

:- type has_g
        --->    has_g(g :: int).

:- func map_f(func(has_g) = has_g, has_f) = has_f.

map_f(Update, has_f(F)) = has_f(Update(F)).

:- func map_g(func(int) = int, has_g) = has_g.

map_g(Update, has_g(G)) = has_g(Update(G)).

:- func set(T, T) = T.

set(T, _) = T.

main -->
        { X = has_f(has_g(1)) },
        { Z = 2 },
        io__write(map_f(map_g(set(Z)), X)),
        io__nl.

--------------------------------------------------------------------------
mercury-developers mailing list
Post messages to:       mercury-developers at cs.mu.oz.au
Administrative Queries: owner-mercury-developers at cs.mu.oz.au
Subscriptions:          mercury-developers-request at cs.mu.oz.au
--------------------------------------------------------------------------