[m-rev.] for review: fix typeclass bug

[Fergus Henderson]

On 12-Mar-2003, David Overton <dmo at cs.mu.OZ.AU> wrote:
> 
> compiler/typecheck.m:
> 	In `eliminate_constraint_by_class_rules', if `type_unify_list'
> 	fails for a subclass, remove that subclass from the list of
> 	constraints being examined rather than aborting.
> 	`type_unify_list' may fail if one of the arguments to the
> 	superclass constraint is not a type variable and does not unify
> 	with the corresponding argument of the subclass.
> 	In such cases, this particular superclass-subclass relationship
> 	is not relevant to the constraint we are trying to prove so it is safe
> 	to remove it.

Could you explain in more detail?  Perhaps with an example?
I'm not yet completely convinced that removing such constraints is safe.

> tests/valid/Mmakefile:
> tests/valid/typeclass_constraint_nonvar_bug.m:
> 	Test case.

The test case does not seem to be a proper regression test, since this
test passes with the existing compiler (I tested with rotd-2003-03-08).

> Index: compiler/typecheck.m
> ===================================================================
> RCS file: /home/mercury1/repository/mercury/compiler/typecheck.m,v
> retrieving revision 1.331
> diff -u -r1.331 typecheck.m
> --- compiler/typecheck.m	28 Feb 2003 06:40:43 -0000	1.331
> +++ compiler/typecheck.m	12 Mar 2003 03:50:54 -0000
> @@ -4629,7 +4629,8 @@
>  		% Convert all the subclass_details into class_constraints by
>  		% doing the appropriate variable renaming and applying the
>  		% type variable bindings.
> -	list__map(subclass_details_to_constraint(TVarSet, SuperClassTypes),
> +	list__filter_map(subclass_details_to_constraint(TVarSet,
> +			SuperClassTypes),

Please update the comment here.  This is now doing more than just
converting things, it is also removing some contraints.
(E.g. with text similar to that in the log message.)

>  		SubClasses, SubClassConstraints),
>  
>  	(
> @@ -4720,7 +4721,7 @@
>  	).
>  
>  :- pred subclass_details_to_constraint(tvarset::in, list(type)::in,
> -		subclass_details::in, class_constraint::out) is det.
> +		subclass_details::in, class_constraint::out) is semidet.

Please add some comments here indicating when the procedure will succeed
and when it will fail.

-- 
Fergus Henderson <fjh at cs.mu.oz.au>  |  "I have always known that the pursuit
The University of Melbourne         |  of excellence is a lethal habit"
WWW: <http://www.cs.mu.oz.au/~fjh>  |     -- the last words of T. S. Garp.
--------------------------------------------------------------------------
mercury-reviews mailing list
post:  mercury-reviews at cs.mu.oz.au
administrative address: owner-mercury-reviews at cs.mu.oz.au
unsubscribe: Address: mercury-reviews-request at cs.mu.oz.au Message: unsubscribe
subscribe:   Address: mercury-reviews-request at cs.mu.oz.au Message: subscribe
--------------------------------------------------------------------------