// Builds an LL(1) parse table using the First and Follow sets for the grammar
// returns null if CFG is not LL(1)
public ParseTable CreateLL1ParseTable()
{
ComputeFirstSets();
ComputeFollowSets();
if (Program.debug)
{
Console.WriteLine("=========================================================");
}
ParseTable parseTable = new ParseTable();
// for each nonterminal in the grammar and each production rule for the nonterminal
foreach (Nonterminal var in productions.Keys)
{
foreach (ISymbol[] varProduction in productions[var])
{
if (Program.debug)
{
Console.WriteLine("CFG: LL(1) table gen, var = " + var + " prod = " + SymbolsToString(varProduction));
}
ISet <Terminal> productionFirst = First(varProduction);
// add an entry to the parse table for each terminal in the first set of the production
foreach (Terminal term in productionFirst)
{
if (term != Terminal.EPSILON)
{
if (parseTable.Get(var, term) != null) // LL(1) violation
{
if (Program.debug)
{
Console.WriteLine("CFG: LL(1) violation at (First) [" + var + "," + term + "]");
Console.WriteLine("\tWas: " + SymbolsToString(parseTable.Get(var, term)));
Console.WriteLine("\tNew: " + SymbolsToString(varProduction));
}
return(null);
}
else
{
parseTable.Add(var, term, varProduction);
}
}
}
// if epsilon can be derived from the production
// add an entry to the parse table for each terminal in the follow set of the nonterminal as well
if (productionFirst.Contains(Terminal.EPSILON))
{
foreach (Terminal term in Follow(var))
{
if (parseTable.Get(var, term) != null) // LL(1) violation
{
if (Program.debug)
{
Console.WriteLine("CFG: LL(1) violation at (Follow) [" + var + "," + term + "]");
Console.WriteLine("\tWas: " + SymbolsToString(parseTable.Get(var, term)));
Console.WriteLine("\tNew: " + SymbolsToString(varProduction));
}
return(null);
}
else
{
parseTable.Add(var, term, varProduction);
}
}
}
}
}
if (Program.debug)
{
foreach (Nonterminal var in nonterminals)
{
foreach (Terminal term in terminals)
{
if (parseTable.Get(var, term) != null)
{
Console.WriteLine("CFG: LL(1) Table[" + var + "][" + term + "] = " + SymbolsToString(parseTable.Get(var, term)));
}
}
}
}
return(parseTable);
}
// Builds an LL(1) parse table using the First and Follow sets for the grammar
// returns null if CFG is not LL(1)
public ParseTable CreateLL1ParseTable()
{
ComputeFirstSets();
ComputeFollowSets();
if (Program.debug)
Console.WriteLine("=========================================================");
ParseTable parseTable = new ParseTable();
// for each nonterminal in the grammar and each production rule for the nonterminal
foreach (Nonterminal var in productions.Keys)
{
foreach (ISymbol[] varProduction in productions[var])
{
if (Program.debug)
Console.WriteLine("CFG: LL(1) table gen, var = " + var + " prod = " + SymbolsToString(varProduction));
ISet<Terminal> productionFirst = First(varProduction);
// add an entry to the parse table for each terminal in the first set of the production
foreach (Terminal term in productionFirst)
{
if (term != Terminal.EPSILON)
{
if (parseTable.Get(var, term) != null) // LL(1) violation
{
if (Program.debug)
{
Console.WriteLine("CFG: LL(1) violation at (First) [" + var + "," + term + "]");
Console.WriteLine("\tWas: " + SymbolsToString(parseTable.Get(var, term)));
Console.WriteLine("\tNew: " + SymbolsToString(varProduction));
}
return null;
}
else
{
parseTable.Add(var, term, varProduction);
}
}
}
// if epsilon can be derived from the production
// add an entry to the parse table for each terminal in the follow set of the nonterminal as well
if (productionFirst.Contains(Terminal.EPSILON))
{
foreach (Terminal term in Follow(var))
{
if (parseTable.Get(var, term) != null) // LL(1) violation
{
if (Program.debug)
{
Console.WriteLine("CFG: LL(1) violation at (Follow) [" + var + "," + term + "]");
Console.WriteLine("\tWas: " + SymbolsToString(parseTable.Get(var, term)));
Console.WriteLine("\tNew: " + SymbolsToString(varProduction));
}
return null;
}
else
{
parseTable.Add(var, term, varProduction);
}
}
}
}
}
if (Program.debug)
foreach (Nonterminal var in nonterminals)
foreach (Terminal term in terminals)
if (parseTable.Get(var, term) != null)
Console.WriteLine("CFG: LL(1) Table[" + var + "][" + term + "] = " + SymbolsToString(parseTable.Get(var, term)));
return parseTable;
}
// Parse the given stream of tokens
public ParseTree Parse(IEnumerable <Token> tokenSource)
{
isValidParseTree = true;
errors = new List <Error>();
Stack <ISymbol> symbolStack = new Stack <ISymbol>();
symbolStack.Push(Terminal.EOF);
symbolStack.Push(start);
ParseTree parseTree = new ParseTree(start);
Stack <IParseNode> treeStack = new Stack <IParseNode>();
treeStack.Push(new ParseLeaf(Terminal.EOF));
treeStack.Push(parseTree);
IEnumerator <Token> tokenStream = tokenSource.GetEnumerator();
tokenStream.MoveNext();
while (symbolStack.Count > 0)
{
if (Program.debug)
{
Console.WriteLine("=========================================================");
Console.WriteLine(" PARSE: Stack " + SymbolsToString(symbolStack));
Console.WriteLine(" PARSE: expecting " + symbolStack.Peek());
Console.WriteLine(" PARSE: token " + tokenStream.Current);
}
// ignore error tokens
if (tokenStream.Current.Type == TokenType.ERROR)
{
if (Program.debug)
{
Console.WriteLine(" PARSE: skipping error token");
}
errors.Add(new LexicalError(tokenStream.Current));
tokenStream.MoveNext();
continue;
}
if (symbolStack.Peek() is Terminal)
{
Terminal term = symbolStack.Peek() as Terminal;
ParseLeaf leaf = treeStack.Peek() as ParseLeaf;
if (term == Terminal.EPSILON)
{
// epsilon production was used, exclude from parse tree
if (Program.debug)
{
Console.WriteLine(" PARSE: ignore epsilon");
}
symbolStack.Pop();
treeStack.Pop();
}
else if (term.Matches(tokenStream.Current))
{
// current token matches the top of the parse stack, add it to parse tree
if (Program.debug)
{
Console.WriteLine(" PARSE: Terminal match");
}
leaf.Token = tokenStream.Current;
tokenStream.MoveNext();
symbolStack.Pop();
treeStack.Pop();
}
else
{
// current token does no match, recover from error
if (Program.debug)
{
Console.WriteLine(" PARSE: Error, Terminal mismatch");
}
errors.Add(new SyntaxError(tokenStream.Current));
Synchronize(symbolStack, treeStack, tokenStream);
}
}
else // top of stack is a nonterminal
{
Nonterminal var = symbolStack.Pop() as Nonterminal;
IParseNode popped = treeStack.Pop();
ParseTree subtree = popped as ParseTree;
ISymbol[] production = table.Get(var, tokenStream.Current);
if (production == null)
{
// cannot derive the current token from the nonterminal at the top of the stack
if (Program.debug)
{
Console.WriteLine(" PARSE: Error, No such production");
}
symbolStack.Push(var);
treeStack.Push(popped);
errors.Add(new SyntaxError(tokenStream.Current));
Synchronize(symbolStack, treeStack, tokenStream);
}
else
{
// use the production specified by the parse table, add node to parse tree
if (Program.debug)
{
Console.WriteLine(" PARSE: Using production " + SymbolsToString(production));
}
for (int i = production.Length - 1; i >= 0; i--)
{
IParseNode treeChild;
if (production[i] is Terminal)
{
treeChild = new ParseLeaf(production[i] as Terminal);
}
else
{
treeChild = new ParseTree(production[i] as Nonterminal);
}
subtree.Children.Insert(0, treeChild);
treeStack.Push(treeChild);
symbolStack.Push(production[i]);
}
}
}
}
if (Program.debug)
{
Console.WriteLine(parseTree);
}
return(parseTree);
}