for review: use '=>' for existential type class constraints

[Fergus Henderson]

DJ, could you please review this one?

--------------------

Estimated hours taken: 2

Change the syntax for existential type class constraints to use
`=>' instead of `&'. 

We eventually decided that `=>' was better, even though semantically
speaking it's more like conjunction than implication, because of the
symmetry with `<=' for universal type class constraints, and because
the precedence of `=>' matches what we want. 
[However, the associativity might not.]

compiler/prog_io.m:
	Update the code that parses existential type class constraints.

compiler/mercury_to_mercury.m:
	Update the code that prints existential type class constraints.

tests/hard_coded/typeclasses/existential_type_classes.m:
	Update the test case that tests existential type class constraints.
	Also add a test of passing typeclass_infos to C code.

Index: compiler/mercury_to_mercury.m
===================================================================
RCS file: /home/mercury1/repository/mercury/compiler/mercury_to_mercury.m,v
retrieving revision 1.147
diff -u -r1.147 mercury_to_mercury.m
--- mercury_to_mercury.m	1998/09/22 16:13:32	1.147
+++ mercury_to_mercury.m	1998/10/28 06:59:23
@@ -1186,14 +1186,14 @@
 		io__write_string("(")
 	),
 
-	% we need to quote ';'/2, '{}'/2, '&'/2, and 'some'/2
+	% we need to quote ';'/2, '{}'/2, '=>'/2, and 'some'/2
 	{ list__length(Args, Arity) },
 	(
 		{ Arity = 2 },
 		{ Name = unqualified(";")
 		; Name = unqualified("{}")
 		; Name = unqualified("some")
-		; Name = unqualified("&")
+		; Name = unqualified("=>")
 		}
 	->
 		io__write_string("{ ")
@@ -1216,7 +1216,7 @@
 		{ Name = unqualified(";")
 		; Name = unqualified("{}")
 		; Name = unqualified("some")
-		; Name = unqualified("&")
+		; Name = unqualified("=>")
 		}
 	->
 		io__write_string(" }")
@@ -1224,7 +1224,7 @@
 		[]
 	),
 
-	mercury_output_class_constraint_list(Constraints, VarSet, "&"),
+	mercury_output_class_constraint_list(Constraints, VarSet, "=>"),
 	(
 		{ ExistQVars = [] }
 	->
@@ -1439,7 +1439,7 @@
 
 mercury_output_class_context(ClassContext, ExistQVars, VarSet) -->
 	{ ClassContext = constraints(UnivCs, ExistCs) },
-	mercury_output_class_constraint_list(ExistCs, VarSet, "&"),
+	mercury_output_class_constraint_list(ExistCs, VarSet, "=>"),
 	( { ExistQVars = [], ExistCs = [] } -> 
 		[] 
 	; 
Index: compiler/prog_io.m
===================================================================
RCS file: /home/mercury1/repository/mercury/compiler/prog_io.m,v
retrieving revision 1.175
diff -u -r1.175 prog_io.m
--- prog_io.m	1998/07/25 13:36:23	1.175
+++ prog_io.m	1998/10/28 06:53:27
@@ -1035,7 +1035,7 @@
 parse_decl_attribute("semipure", [Decl], purity(semipure), Decl).
 parse_decl_attribute("<=", [Decl, Constraints],
 		constraints(univ, Constraints), Decl).
-parse_decl_attribute("&", [Decl, Constraints],
+parse_decl_attribute("=>", [Decl, Constraints],
 		constraints(exist, Constraints), Decl).
 parse_decl_attribute("some", [TVars, Decl],
 		quantifier(exist, TVarsList), Decl) :-
@@ -1075,7 +1075,7 @@
 attribute_description(quantifier(exist, _), "existential quantifier (`some')").
 attribute_description(constraints(univ, _), "type class constraint (`<=')").
 attribute_description(constraints(exist, _),
-	"existentially quantified type class constraint (`&')").
+	"existentially quantified type class constraint (`=>')").
 
 %-----------------------------------------------------------------------------%
 
@@ -1432,7 +1432,7 @@
 			term__contains_var_list(Args, Var),
 			\+ list__member(Var, ExistQVars)
 		->
-			Result = error("type variables in class constraints introduced with `&' must be explicitly existentially quantified using `some'",
+			Result = error("type variables in class constraints introduced with `=>' must be explicitly existentially quantified using `some'",
 					Body)
 		;
 			(
@@ -1658,15 +1658,17 @@
 	%	1. universal quantifiers	all		950
 	%	2. existential quantifiers	some		950
 	%	3. universal constraints	<=		920
-	%	4. existential constraints	&		1020	[*]
+	%	4. existential constraints	=>		920	[*]
 	%	5. the decl itself 		pred or func	800
 	%
 	% When we reach here, Attributes0 contains declaration attributes
 	% in the opposite order -- innermost to outermost -- so we reverse
 	% them before we start.
 	%
-	% [*] Note that the precedence of `&' is not quite what we want for
-	% this purpose -- the user will have to put in explicit parentheses.
+	% [*] Note that the semantic meaning of `=>' is not quite
+	%     the same as implication; logically speaking it's more
+	%     like conjunction.  Oh well, at least it has the right
+	%     precedence.
 	%
 	% In theory it could make sense to allow the order of 2 & 3 to be
 	% swapped, or (in the case of multiple constraints & multiple
@@ -1763,7 +1765,7 @@
 get_existential_constraints_from_term(ModuleName, PredType0, PredType,
 		MaybeExistentialConstraints) :-
 	(	
-		PredType0 = term__functor(term__atom("&"), 
+		PredType0 = term__functor(term__atom("=>"), 
 			[PredType1, ExistentialConstraints], _)
 	->
 		PredType = PredType1,
Index: tests/hard_coded/typeclasses/existential_type_classes.m
===================================================================
RCS file: /home/mercury1/repository/tests/hard_coded/typeclasses/existential_type_classes.m,v
retrieving revision 1.2
diff -u -r1.2 existential_type_classes.m
--- existential_type_classes.m	1998/07/08 20:59:12	1.2
+++ existential_type_classes.m	1998/10/29 06:57:56
@@ -27,11 +27,11 @@
 
 	% my_univ_value(Univ):
 	%	returns the value of the object stored in Univ.
-:- some [T] (func my_univ_value(univ) = T & fooable(T)).
+:- some [T] func my_univ_value(my_univ) = T => fooable(T).
 
-:- some [T] (func call_my_univ_value(univ) = T & fooable(T)).
+:- some [T] func call_my_univ_value(my_univ) = T => fooable(T).
 
-:- some [T] (func my_exist_t = T & fooable(T)).
+:- some [T] func my_exist_t = T => fooable(T).
 
 :- pred int_foo(int::in, int::out) is det.
 int_foo(X, 2*X).

-- 
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.