import java.lang.reflect.*;
import java.lang.ref.*;
-/** juxtaposition; zero or more adjacent Elements; can specify a rewriting */
+/** <font color=green>juxtaposition; zero or more adjacent Elements; can specify a rewriting</font> */
public abstract class Sequence extends Element implements Iterable<Element> {
+ protected final Element[] elements;
+
+ final HashSet<Sequence> hated = new HashSet<Sequence>();
+
+ final HashSet<Sequence> needs = new HashSet<Sequence>();
+ final HashSet<Sequence> hates = new HashSet<Sequence>();
+
+ final Position firstp;
+
+ public Atom follow = null;
+ public final Topology follow() { return follow; }
+
// Static Constructors //////////////////////////////////////////////////////////////////////////////
+ abstract Sequence _clone();
+ Sequence dup() {
+ Sequence ret = _clone();
+ for(Sequence s : needs) { ret.needs.add(s); }
+ for(Sequence s : hates) { ret.hates.add(s); s.hated.add(ret); }
+ ret.follow = follow;
+ return ret;
+ }
+
/** the empty sequence (matches the empty string) */
public static final Sequence empty = new Sequence.Constant.Empty();
/** after matching the sequence, do not add anything to the output tree */
- public static Sequence drop(Element[] e, HashSet<Sequence> and, HashSet<Sequence> not, boolean lame) { return new Constant.Drop(e, and, not, lame); }
+ public static Sequence drop(Element[] e) { return new Constant.Drop(e); }
/** after matching the sequence, insert a constant into the output tree */
- public static Sequence constant(Element[] e, Object o, HashSet<Sequence> and, HashSet<Sequence> not) { return new Constant(e, o, and, not); }
+ public static Sequence constant(Element[] e, Object o) { return new Constant(e, o); }
/** after matching the sequence, place the result of the <tt>idx</tt>th match in the output tree */
- public static Sequence singleton(Element[] e, int idx, HashSet<Sequence> and, HashSet<Sequence> not) { return new Singleton(e, idx, and, not); }
+ public static Sequence singleton(Element[] e, int idx) { return new Singleton(e, idx); }
+ public static Sequence singleton(Element e) { return singleton(new Element[] { e }, 0); }
/**
* after matching the sequence, create the specified output tree
* @param tag the tag for the output tree
* @param e the elements to match
- * @param drops only elements of <tt>e</tt> whose corresponding <tt>boolean</tt> in <tt>drops</tt> is <i>false</i> will be included in the output tree
+ * @param drops only elements of <tt>e</tt> whose corresponding <tt>boolean</tt> in <tt>drops</tt>
+ * is <i>false</i> will be included in the output tree
**/
- public static Sequence rewritingSequence(Object tag, Element[] e, boolean[] drops, HashSet<Sequence> and, HashSet<Sequence> not) {
- return new RewritingSequence(tag, e, drops, and, not); }
-
- ////////////////////////////////////////////////////////////////////////////////
+ public static Sequence rewritingSequence(Object tag, Element[] e, boolean[] drops) {
+ return new RewritingSequence(tag, e, drops); }
- public Element noFollow = null;
- public final Topology noFollow() { return noFollow==null ? null : noFollow.toAtom(); }
+ public static Sequence regionRewritingSequence(Functor<Input.Region,Object> tagfunctor, Element[] e, boolean[] drops) {
+ return new RegionRewritingSequence(tagfunctor, e, drops); }
- Topology toAtom() {
- if (elements.length!=1) throw new RuntimeException("cannot invoke toAtom() on a Sequence with " + elements.length + " elements: " + this);
- return elements[0].toAtom();
- }
+ ////////////////////////////////////////////////////////////////////////////////
- protected final Element[] elements;
+ public Sequence and(Sequence s) { Sequence ret = dup(); ret.needs.add(s); return ret; }
+ public Sequence not(Sequence s) { Sequence ret = dup(); ret.hates.add(s); s.hated.add(ret); return ret; }
- final HashSet<Sequence> needs;
- final HashSet<Sequence> hates;
- boolean lame = false;
+ public Iterable<Sequence> needs() { return needs; }
+ public Iterable<Sequence> hates() { return hates; }
- final Position firstp;
Position firstp() { return firstp; }
public Iterator<Element> iterator() { return new ArrayIterator<Element>(elements); }
- protected Sequence(Element[] elements, HashSet<Sequence> and, HashSet<Sequence> not) {
- this.needs = and==null ? new HashSet<Sequence>() : and;
- this.hates = not==null ? new HashSet<Sequence>() : not;
+ protected Sequence(Element[] elements) {
this.elements = elements;
this.firstp = new Position(0);
}
// DO NOT MESS WITH THE FOLLOWING LINE!!!
- private Forest.Ref epsilonForm = null;
- private boolean eps = false;
+ private Forest.Many epsilonForm = null;
Forest epsilonForm() {
- if (epsilonForm==null) {
- epsilonForm = new Forest.Ref();
- Forest fo = firstp().rewrite(null);
- epsilonForm.merge(fo);
- }
+ if (epsilonForm!=null) return epsilonForm;
+ epsilonForm = new Forest.Many();
+ epsilonForm.merge(firstp().rewrite(null, false));
return epsilonForm;
}
- protected abstract <T> Forest<T> postReduce(Token.Location loc, Forest<T>[] args);
+ protected abstract <T> Forest<T> postReduce(Input.Region loc, Forest<T>[] args, Position p);
// Position //////////////////////////////////////////////////////////////////////////////
/** the imaginary position before or after an element of a sequence; corresponds to an "LR item" */
- public class Position {
+ public class Position implements IntegerMappable {
+
+ private Forest zero = null;
+ public Forest zero() {
+ if (zero != null) return zero;
+ if (pos > 0) throw new Error();
+ return zero = rewrite(null);
+ }
+
final int pos;
private final Position next;
}
boolean isFirst() { return pos==0; }
- boolean isRightNullable(Walk.Cache cache) {
- if (isLast()) return true;
- if (!element().possiblyEpsilon(cache)) return false;
- return next().isRightNullable(cache);
- }
/** the element immediately after this Position, or null if this is the last Position */
public Element element() { return pos>=elements.length ? null : elements[pos]; }
/** true iff this Position is the last one in the sequence */
public boolean isLast() { return next()==null; }
- // Reduction /////////////////////////////////////////////////////////////////////////////////
+ // Position /////////////////////////////////////////////////////////////////////////////////
- final <T> Forest<T> rewrite(Token.Location loc) {
- if (this==firstp() && eps) return epsilonForm;
- eps = true;
+ final <T> Forest<T> rewrite(Input.Region loc) { return rewrite(loc, true); }
+ private final <T> Forest<T> rewrite(Input.Region loc, boolean epsilonCheck) {
+ if (epsilonCheck && this==firstp()) return epsilonForm();
for(int i=0; i<pos; i++) if (holder[i]==null) throw new Error("realbad " + i);
for(int i=pos; i<elements.length; i++) {
if (holder[i]==null) holder[i] = elements[i].epsilonForm();
if (holder[i]==null) throw new Error("bad " + i);
}
- Forest<T> ret = Sequence.this.postReduce(loc, holder);
- for(int k=0; k<pos; k++) holder[k] = null; // to help GC
- if (this==firstp()) { if (epsilonForm==null) epsilonForm=new Forest.Ref(); epsilonForm.merge(ret); return epsilonForm; }
+ Forest<T> ret = Sequence.this.postReduce(loc, holder, this);
+ //for(int k=0; k<pos; k++) holder[k] = null; // to help GC
return ret;
}
for(Position p = Sequence.this.firstp(); p != null; p = p.next()) {
ret.append(' ');
if (p==this) ret.append(" | ");
- if (p.element()!=null) ret.append(p.element().possiblyEpsilon(null) ? "["+p.element()+"]" : p.element());
+ if (p.element()!=null) ret.append(p.element());
else ret.append(' ');
}
ret.append("}>");
return ret.toString();
}
+ private final int idx = master_position_idx++;
+ public int toInt() { return idx; }
}
-
+ private static int master_position_idx = 0;
// toString //////////////////////////////////////////////////////////////////////////////
sb.append(elements[i].toString());
sb.append(' ');
}
+ if (follow != null) {
+ sb.append("-> ");
+ sb.append(follow);
+ }
return sb;
}
static class Constant extends Sequence {
private final Object result;
- public Constant(Element[] e, Object result, HashSet<Sequence> and, HashSet<Sequence> not) { super(e, and, not); this.result = result; }
- public <T> Forest<T> postReduce(Token.Location loc, Forest<T>[] args) {
- return (Forest<T>)Forest.leaf(loc, result, this);
+ public Constant(Element[] e, Object result) { super(e); this.result = result; }
+ Sequence _clone() { return new Constant(elements, result); }
+ public <T> Forest<T> postReduce(Input.Region loc, Forest<T>[] args, Position p) {
+ return (Forest<T>)Forest.create(loc, result, null, false);
}
static class Drop extends Constant {
- public Drop(Element[] e, HashSet<Sequence> and, HashSet<Sequence> not, boolean lame) {
- super(e, null, and, not);
- this.lame = lame;
- }
+ Sequence _clone() { return new Drop(elements); }
+ public Drop(Element[] e) { super(e, null); }
}
- static class Empty extends Sequence.Constant.Drop { public Empty() { super(new Element[] { }, null, null, false); } }
+ static class Empty extends Sequence.Constant.Drop {
+ Sequence _clone() { return new Empty(); }
+ public Empty() { super(new Element[] { }); } }
}
static class Singleton extends Sequence {
private final int idx;
- public Singleton(Element e, HashSet<Sequence> and, HashSet<Sequence> not) { this(new Element[] { e }, 0, and, not); }
- public Singleton(Element[] e, int idx, HashSet<Sequence> and, HashSet<Sequence> not) { super(e, and, not); this.idx = idx; }
- public <T> Forest<T> postReduce(Token.Location loc, Forest<T>[] args) { return (Forest<T>)Forest.singleton(loc, args[idx], this); }
+ public Singleton(Element e) { this(new Element[] { e }, 0); }
+ public Singleton(Element[] e, int idx) { super(e); this.idx = idx; }
+ public <T> Forest<T> postReduce(Input.Region loc, Forest<T>[] args, Position p) { return args[idx]; }
+ Sequence _clone() { return new Singleton(elements,idx); }
}
- public static class Unwrap extends Sequence {
+ public static Unwrap unwrap(Element[] e, Object tag, boolean[] drops) { return new Unwrap(e, tag, drops); }
+ static class Unwrap extends Sequence {
private boolean[] drops;
- public Unwrap(Element[] e, HashSet<Sequence> and, HashSet<Sequence> not) { super(e, and, not); this.drops = null; }
- public Unwrap(Element[] e, boolean[] drops, HashSet<Sequence> and, HashSet<Sequence> not) { super(e, and, not); this.drops = drops; }
- public <T> Forest<T> postReduce(Token.Location loc, Forest<T>[] args) {
+ private final Object tag;
+ public Unwrap(Element[] e, Object tag) { super(e); this.drops = null; this.tag = tag; }
+ public Unwrap(Element[] e, Object tag, boolean[] drops) { super(e); this.drops = drops; this.tag = tag; }
+ Sequence _clone() { return new Unwrap(elements, drops); }
+ public <T> Forest<T> postReduce(Input.Region loc, Forest<T>[] args, Position p) {
for(int i=0; i<args.length; i++) if (args[i]==null) throw new Error();
- if (drops==null) return Forest.create(loc, null, args, this, true, false);
+ if (drops==null) return Forest.create(loc, (T)tag, args, true);
int count = 0;
for(int i=0; i<drops.length; i++) if (!drops[i]) count++;
Forest<T>[] args2 = new Forest[count];
int j = 0;
for(int i=0; i<args.length; i++) if (!drops[i]) args2[j++] = args[i];
- return Forest.create(loc, null, args2, this, true, false);
+ return Forest.create(loc, (T)tag, args2, true);
+ }
+ }
+
+
+
+ static class RegionRewritingSequence extends RewritingSequence {
+ private Functor<Input.Region, Object> tagf;
+ public RegionRewritingSequence(Functor<Input.Region,Object> tagfunctor, Element[] e, boolean[] drops) {
+ super(null, e, drops);
+ this.tagf = tagfunctor;
+ }
+ public <T> Forest<T> postReduce(Input.Region loc, Forest<T>[] args, Position p) {
+ this.tag = tagf.invoke(loc);
+ Forest<T> ret = super.postReduce(loc, args, p);
+ this.tag = null;
+ return ret;
}
}
static class RewritingSequence extends Sequence {
- /*private*/public final Object tag;
+ /*private*/public /*final*/ Object tag;
private final boolean[] drops;
private int count = 0;
- public RewritingSequence(Object tag, Element[] e, HashSet<Sequence> and, HashSet<Sequence> not) { this(tag, e, null, and, not); }
- public RewritingSequence(Object tag, Element[] e, boolean[] drops, HashSet<Sequence> and, HashSet<Sequence> not) {
- super(e, and, not);
+ Sequence _clone() { return new RewritingSequence(tag, elements, drops); }
+ public RewritingSequence(Object tag, Element[] e) { this(tag, e, null); }
+ public RewritingSequence(Object tag, Element[] e, boolean[] drops) {
+ super(e);
this.tag = tag;
this.drops = drops == null ? new boolean[e.length] : drops;
for(int i=0; i<this.drops.length; i++) if (!this.drops[i]) count++;
}
- public <T> Forest<T> postReduce(Token.Location loc, Forest<T>[] args) {
+ public <T> Forest<T> postReduce(Input.Region loc, Forest<T>[] args, Position p) {
Forest<T>[] args2 = new Forest[count];
int j = 0;
for(int i=0; i<args.length; i++) if (!drops[i]) args2[j++] = args[i];
//System.out.println("reduce \""+tag+"\"");
- return Forest.create(loc, (T)tag, args2, this, false, false);
+ return Forest.create(loc, (T)tag, args2, false);
}
public StringBuffer toString(StringBuffer sb, boolean spacing) {
int len = sb.length();
+ if (tag != null)
+ sb.append("\""+StringUtil.escapify(tag.toString(),"\"\r\n")+"\":: ");
super.toString(sb, spacing);
len = sb.length()-len;
if (spacing) for(int i=0; i<50-len; i++) sb.append(' ');
- sb.append(" => ");
- sb.append(tag);
return sb;
}
}
+
+ // Repeat //////////////////////////////////////////////////////////////////////////////
+
+ /** repeat zero or one times */
+ public static Element maybe(Element e) { return new Repeat(e, true, false, null, null); }
+ public static Element maybe(Element e, Object tag) { return new Repeat(e, true, false, null, tag); }
+ /** repeat zero or more times */
+ public static Element many0(Element e) { return new Repeat(e, true, true, null, null); }
+ public static Element many0(Element e, Object tag) { return new Repeat(e, true, true, null, tag); }
+ /** repeat zero or more times, separated by <tt>sep</tt> */
+ public static Element many0(Element e, Element sep) { return new Repeat(e, true, true, sep, null); }
+ public static Element many0(Element e, Element sep, Object tag) { return new Repeat(e, true, true, sep, tag); }
+ /** repeat one or more times */
+ public static Element many1(Element e) { return new Repeat(e, false, true, null, null); }
+ public static Element many1(Element e, Object tag) { return new Repeat(e, false, true, null, tag); }
+ /** repeat one or more times, separated by <tt>sep</tt> */
+ public static Element many1(Element e, Element sep) { return new Repeat(e, false, true, sep, null); }
+ public static Element many1(Element e, Element sep, Object tag) { return new Repeat(e, false, true, sep, tag); }
+
+ /** repeat zero or more times, matching a maximal sequence of atoms */
+ public static Element maximal0(Atom e) { return new Repeat.Maximal(e, true, true, null); }
+ public static Element maximal0(Atom e, Object tag) { return new Repeat.Maximal(e, true, true, tag); }
+ /** repeat one or more times, matching a maximal sequence of atoms */
+ public static Element maximal1(Atom e) { return new Repeat.Maximal(e, false, true, null); }
+ public static Element maximal1(Atom e, Object tag) { return new Repeat.Maximal(e, false, true, tag); }
+ /** repeat one or more times, separated by an atom <tt>sep</tt>, matching a maximal sequence */
+ public static Element maximal1(Element e, Atom sep) { return new Repeat.Maximal(e, false, true, sep, null); }
+ public static Element maximal1(Element e, Atom sep, Object tag) { return new Repeat.Maximal(e, false, true, sep, tag); }
+
+ public static Element repeatMaximal(Atom e, boolean zero, boolean many, Object tag) {
+ return new Repeat.Maximal(e, zero, many, tag); }
+ public static Element repeatMaximal(Element e, boolean zero, boolean many, Atom sep, Object tag) {
+ return new Repeat.Maximal(e, zero, many, sep, tag); }
+ public static Element repeat(Element e, boolean zero, boolean many, Object tag) {
+ return new Repeat(e, zero, many, tag); }
+ public static Element repeat(Element e, boolean zero, boolean many, Element sep, Object tag) {
+ return new Repeat(e, zero, many, sep, tag); }
}
projects / sbp.git / blobdiff