language-Modula2-0.1: examples/Modula-2_Libraries/C.-Lins_Modula-2_Software_Component_Library/Vol1/SETS/SETSBMI.DEF
(*
12 The Bounded Set
The first of three implementations for the set abstraction described in the previous chapter
is presented here. This particular form has the properties: Sequential, Bounded, Managed,
and Iterator, describing specific aspects of the implementation as follows:
Sequential: Can only be used in a non-tasking environment, or by only one task.
Bounded: The maximum size of a set is given when the set is created.
Managed: Memory space for items and objects is returned to the system when no
longer needed.
Iterator: Routines for looping over each of a set's items are provided.
As has been the convention for all other abstractions, the interface to the module defining
set operations and exceptions is included here as well in Section 12.1. The bounded set
interface follows in Section 12.2 while the actual implementation appears in Section
12.3.
12.2 SetSBMI Interface
*)
DEFINITION MODULE SetSBMI;
(*==========================================================
Version : 1.00 30 Apr 1989 C. Lins
Compiler : TopSpeed Modula-2 Compiler
Component: Monolithic Structures - Set
Sequential Bounded Managed Iterator
INTRODUCTION
This module provides the interface to the bounded
implementation Set abstraction for generic Items.
REVISION HISTORY
v1.00 30 Apr 1989 C. Lins
Initial implementation for TopSpeed Modula-2.
(C) Copyright 1989 Charles A. Lins
==========================================================*)
FROM Items IMPORT
(*--Type*) Item, AccessProc, LoopAccessProc;
FROM ErrorHandling IMPORT
(*--Proc*) HandlerProc;
FROM SetEnum IMPORT
(*--Type*) Exceptions;
FROM TypeManager IMPORT
(*--Type*) TypeID;
(*-----------------------*)
TYPE Set;
TYPE SizeRange = [ 1 .. 8000 ];
CONST NullSet = Set(NIL);
(*
12.2.1 Exceptions
The ModuleID is used by the exception handling mechanism to distinguish this module
from other modules.
SetError returns the exception code from the most recent set operation. A result of noerr
indicates successful completion of the operation.
GetHandler returns the exception handler routine associated with the given exception.
Though the routine is a function procedure returning a procedure as its result, the
HandlerProc may not be called from within the GetHandler call itself. The procedure
result must be first assigned to a procedure variable before invocation. Exception handlers
are given an initial value of ExitOnError except for the handler for noerr which is
initialized to the null exception handler.
SetHandler associates an exception handler routine with the given exception and is the
inverse of GetHandler. This routine may be used to override the default settings for the
exception handlers.
*)
CONST ModuleID = 3;
PROCEDURE SetError () : Exceptions (*-- out *);
PROCEDURE GetHandler ( theError : Exceptions (*-- in *))
: HandlerProc (*-- out *);
PROCEDURE SetHandler ( theError : Exceptions (*-- in *);
theHandler : HandlerProc (*-- in *));
(*
12.2.2 Constructors
The Create operation attempts to generate a new, empty set of the given maximum size
(theSize) and Item operations associated with the given data type identifier (theType).
For the bounded form of set, it is necessary to define the maximum number of items
desired. This is done with theSize parameter.
The TypeID is the method by which Items of any data type are supported. Using the
TypeID, the component can assign one item to another and release any dynamically
allocated resources associated with an Item without knowledge of the Item's internal
composition. In this way, one may create a set whose items consist of other
(dynamically allocated) structures, as well as sets consisting of the basic data types.
Create will return the new set upon successful completion of the routine. If it not
possible for the set to be created, the constant NullSet will be returned instead.
Destroy takes the given set, clears it of any items, and then destroys the set itself.
Where Create makes a defined set, Destroy is its inverse, making the set undefined.
Clear takes the given set and removes all of its items. theType attribute of the set
(assigned when the set was created) is used to retrieve the item deallocation routine for the
items of the set. Clearing the set returns it to the empty state.
The Assign operation attempts to generate a duplicate of the source set (theSet) in the
target set (toSet). The target set is automatically created using the size and data type
attributes of the source set, if necessary. If this step is unnecessary, (the target set has
already been previously created), the target is cleared of its present contents, its data type
is set to that of the source set but the size is left unchanged.
There is no guarantee that the client module would desire the target set to be defined with
the same size as the source. The minimum requirement for the target set size is that it
must be capable of storing all of the items present in the source set. It may be desirable
that the target set size be greater than the source set size. Such a situation could occur
during error recovery of a bounded set overflow caused by the set's cardinality
encountering the set size, and the client module may effectively attempt to increase the
the set size using the assignment mechanism.
In order to allow Items of any data type, a method must be provided to assign the contents
of one item to another item. This is accomplished, as above for Create, through the
TypeID of the source set.
The Include operation adds items to the given set. If the cardinality of the set is already
at its maximum size the overflow exception will be raised and the set remains unchanged.
The Exclude operation removes the specified item from the given set. If the given set is
empty on entry to Exclude, or the given item is not a member of the set the notinset
exception may be raised and the set remains unchanged.
Union, Intersection, Difference, and SymDifference (symmetric difference) operations all
implement the standard set operations of the same name.
*)
PROCEDURE Create ( theSize : SizeRange (*-- in *);
theType : TypeID (*-- in *))
: Set (*-- out *);
PROCEDURE Destroy (VAR theSet : Set (*-- inout *));
PROCEDURE Clear (VAR theSet : Set (*-- inout *));
PROCEDURE Assign ( theSet : Set (*-- in *);
VAR toSet : Set (*-- inout *));
PROCEDURE Include ( theItem : Item (*-- in *);
VAR inSet : Set (*-- inout *));
PROCEDURE Exclude ( theItem : Item (*-- in *);
VAR fromSet : Set (*-- inout *));
PROCEDURE Union ( left : Set (*-- in *);
right : Set (*-- in *);
VAR toSet : Set (*-- inout *));
PROCEDURE Intersection ( left : Set (*-- in *);
right : Set (*-- in *);
VAR toSet : Set (*-- inout *));
PROCEDURE Difference ( left : Set (*-- in *);
right : Set (*-- in *);
VAR toSet : Set (*-- inout *));
PROCEDURE SymDifference ( left : Set (*-- in *);
right : Set (*-- in *);
VAR toSet : Set (*-- inout *));
(*
12.2.3 Selectors
IsDefined attempts to determine whether the given set is valid, e.g., has been created and
not yet destroyed. How this is accomplished may be as simple or complicated as the
implementer desires and the requirements of the application.
IsEmpty returns true if the given set contains no items; in other words, its cardinality is
zero. Undefined sets are always considered to be empty.
IsEqual returns true if the left and right sets contain the same items. Obviously, both
must also have the same data type and have been created. An undefined set is not equal to
any other set, including itself.
The routines SizeOf and TypeOf return the values given the set when it was created, and
both are provided so the user of the module need not maintain separate variables recording
this information.
NumMembers returns the number of items present on the given set. Undefined sets are
considered to have a cardinality of zero.
IsAMember returns true if the given item is present in the given set, and false otherwise.
IsSubset and IsProperSubset both implement standard logical set operations as defined in
the Set Abstraction (chapter 11).
*)
PROCEDURE IsDefined ( theSet : Set (*-- in *))
: BOOLEAN (*-- out *);
PROCEDURE IsEmpty ( theSet : Set (*-- in *))
: BOOLEAN (*-- out *);
PROCEDURE IsEqual ( left : Set (*-- in *);
right : Set (*-- in *))
: BOOLEAN (*-- out *);
PROCEDURE SizeOf ( theSet : Set (*-- in *))
: CARDINAL (*-- out *);
PROCEDURE TypeOf ( theSet : Set (*-- in *))
: TypeID (*-- out *);
PROCEDURE NumMembers ( theSet : Set (*-- in *))
: CARDINAL (*-- out *);
PROCEDURE IsAMember ( theItem : Item (*-- in *);
theSet : Set (*-- in *))
: BOOLEAN (*-- out *);
PROCEDURE IsSubset ( left : Set (*-- in *);
right : Set (*-- in *))
: BOOLEAN (*-- out *);
PROCEDURE IsProperSubset( left : Set (*-- in *);
right : Set (*-- in *))
: BOOLEAN (*-- out *);
(*
12.2.4 Iterators
The iterator routine LoopOver provides the facility for looping over some or all items of
a set, with read-only access to each item. theProcess procedure parameter to this routine
returns a BOOLEAN function result where TRUE allows the iteration to proceed to the
next item and FALSE causes the iteration to be terminated.
The Traverse iterator provides the facility for looping over all items of a set, with
read-only access to each item.
Both iterators traverse the given set from the first item towards the last item in ascending
order. Obviously, if given an empty set the processing procedure will not be invoked.
Iterators allowing changes to an item's value are not provided as this would violate our
abstraction (and increase the complexity of the implementation).
*)
PROCEDURE LoopOver ( theSet : Set (*-- in *);
process : LoopAccessProc (*-- in *));
PROCEDURE Traverse ( theSet : Set (*-- in *);
process : AccessProc (*-- in *));
END SetSBMI.