[m-dev.] for review: move tabling builtins into new module

[Fergus Henderson]

On 20-Sep-2000, Simon Taylor <stayl at cs.mu.OZ.AU> wrote:
>  
> > Index: compiler/dead_proc_elim.m
> > ===================================================================
> > RCS file: /home/mercury1/repository/mercury/compiler/dead_proc_elim.m,v
> > retrieving revision 1.59
> > diff -u -d -r1.59 dead_proc_elim.m
> > --- compiler/dead_proc_elim.m	2000/08/09 07:46:22	1.59
> > +++ compiler/dead_proc_elim.m	2000/09/19 05:37:45
> > @@ -790,6 +790,9 @@
> >  			% Don't eliminate preds from builtin.m or
> >  			% private_builtin.m, since polymorphism.m
> >  			% needs unify/2 and friends.
> > +			% (Should the same apply for other builtin
> > +			% modules, such as table_builtin? 
> > +			% I don't think so...)
> >  			mercury_public_builtin_module(PredModule)
> >  		;
> >  			mercury_private_builtin_module(PredModule)
> 
> This does need to check for predicates from table_builtin.m.
> Any predicates for which the compiler may introduce calls
> must be checked for here. This code is only run
> with --intermodule-optimization.

OK, fixed (diff below).

I also forgot to include my change to intermod.m in the diff
that I posted.  I've enclosed that below.

I've committed this whole change now.

----------

compiler/intermod.m:
        Add the builtin module dependencies.
	This is needed in case a .opt file contains a tabling pragma.

Index: intermod.m
===================================================================
RCS file: /home/mercury1/repository/mercury/compiler/intermod.m,v
retrieving revision 1.83
retrieving revision 1.84
diff -u -d -u -r1.83 -r1.84
--- intermod.m	2000/09/08 12:20:34	1.83
+++ intermod.m	2000/09/21 00:23:49	1.84
@@ -2014,7 +2014,11 @@
 		% Figure out which .int files are needed by the .opt files
 		%
 	{ get_dependencies(OptItems, NewImportDeps0, NewUseDeps0) },
-	{ list__append(NewImportDeps0, NewUseDeps0, NewDeps0) },
+	globals__io_get_globals(Globals),
+	{ get_implicit_dependencies(OptItems, Globals,
+		NewImplicitImportDeps0, NewImplicitUseDeps0) },
+	{ NewDeps0 = list__condense([NewImportDeps0, NewUseDeps0,
+		NewImplicitImportDeps0, NewImplicitUseDeps0]) },
 	{ set__list_to_set(NewDeps0, NewDepsSet0) },
 	{ set__delete_list(NewDepsSet0, [ModuleName | OptFiles], NewDepsSet) },
 	{ set__to_sorted_list(NewDepsSet, NewDeps) },
 
----------

Estimated hours taken: 0.25

Fix an issue with my change to move the tabling builtins to a new module
that was identified in Simon Taylor's review comments.

compiler/dead_proc_elim.m:
	Use any_mercury_builtin_module rather than just
	testing for mercury_public_builtin_module and
	mercury_private_builtin_module.

Workspace: /home/pgrad/fjh/ws/hg
Index: compiler/dead_proc_elim.m
===================================================================
RCS file: /home/mercury1/repository/mercury/compiler/dead_proc_elim.m,v
retrieving revision 1.60
diff -u -d -r1.60 dead_proc_elim.m
--- compiler/dead_proc_elim.m	2000/09/20 12:11:53	1.60
+++ compiler/dead_proc_elim.m	2000/09/20 12:14:26
@@ -787,15 +787,11 @@
 			% polymorphism.
 			code_util__compiler_generated(PredInfo)
 		;
-			% Don't eliminate preds from builtin.m or
-			% private_builtin.m, since polymorphism.m
-			% needs unify/2 and friends.
-			% (Should the same apply for other builtin
-			% modules, such as table_builtin? 
-			% I don't think so...)
-			mercury_public_builtin_module(PredModule)
-		;
-			mercury_private_builtin_module(PredModule)
+			% Don't eliminate preds from builtin modules,
+			% since later passes of the compiler may introduce
+			% calls to them (e.g. polymorphism.m needs unify/2
+			% and friends).
+			any_mercury_builtin_module(PredModule)
 		;
 			% Don't attempt to eliminate local preds here, since we
 			% want to do semantic checking on those even if they 
-- 
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.
--------------------------------------------------------------------------
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
--------------------------------------------------------------------------