/// <summary>
/// Initializes new instance of Parser class.
/// </summary>
/// <param name="textReader">TextReader instance to read data from.</param>
/// <param name="grammar">Grammar with parsing tables to parser input stream.</param>
public Parser(TextReader textReader, Grammar grammar)
{
if (textReader == null)
{
throw new ArgumentNullException("textReader");
}
if (grammar == null)
{
throw new ArgumentNullException("grammar");
}
m_TextReader = textReader;
m_BufferSize = MinimumBufferSize;
m_Buffer = new Char[m_BufferSize + 1];
m_LineLength = Undefined;
ReadBuffer();
m_InputTokens = new Token[MinimumInputTokenCount];
m_LRStack = new LRStackItem[MinimumLRStackSize];
m_Grammar = grammar;
// Put grammar start symbol into LR parsing stack.
m_LRState = m_Grammar.InitialLRState;
var start = new LRStackItem { Token = { Symbol = m_Grammar.StartSymbol }, State = m_LRState };
m_LRStack[m_LRStackIndex] = start;
m_CommentSymbols = new List<Regex>();
m_ReadOnlyCommentSymbols = new ReadOnlyCollection<Regex>(m_CommentSymbols);
m_CommentEndSymbols = new Stack<Regex>();
m_ReductionCount = Undefined; // there are no reductions yet.
}
private const int Undefined = -1; // Used for undefined int values.
#endregion
#region Constructors
/// <summary>
/// Initializes new instance of Parser class.
/// </summary>
/// <param name="textReader">TextReader instance to read data from.</param>
/// <param name="grammar">Grammar with parsing tables to parser input stream.</param>
public Parser(TextReader textReader, Grammar grammar)
{
if (textReader == null)
{
throw new ArgumentNullException("textReader");
}
if (grammar == null)
{
throw new ArgumentNullException("grammar");
}
m_textReader = textReader;
m_bufferSize = MinimumBufferSize;
m_buffer = new char[m_bufferSize + 1];
m_lineLength = Undefined;
ReadBuffer();
m_inputTokens = new Token[MinimumInputTokenCount];
m_lrStack = new LRStackItem[MinimumLRStackSize];
m_grammar = grammar;
// Put grammar start symbol into LR parsing stack.
m_lrState = m_grammar.InitialLRState;
LRStackItem start = new LRStackItem();
start.m_token.m_symbol = m_grammar.StartSymbol;
start.m_state = m_lrState;
m_lrStack[m_lrStackIndex] = start;
m_reductionCount = Undefined; // there are no reductions yet.
}
public void SetSourceCode(string sourceCode)
{
m_buffer = sourceCode;
m_charIndex = 0;
m_lineStart = 0;
m_lineNumber = 1;
m_lineLength = Undefined;
// Put grammar start symbol into LR parsing stack.
m_lrState = m_grammar.InitialLRState;
m_lrStack[m_lrStackIndex] = new LRStackItem {
m_token = { m_symbol = m_grammar.StartSymbol }
};
m_reductionCount = Undefined; // there are no reductions yet.
}
private TokenParseResult ParseToken()
{
LRStateAction stateAction = m_lrState.m_transitionVector[m_token.m_symbol.m_index];
if (stateAction != null)
{
//Work - shift or reduce
if (m_reductionCount > 0)
{
int newIndex = m_lrStackIndex - m_reductionCount;
m_lrStack[newIndex] = m_lrStack[m_lrStackIndex];
m_lrStackIndex = newIndex;
}
m_reductionCount = Undefined;
switch (stateAction.Action)
{
case LRAction.Accept:
m_reductionCount = 0;
return(TokenParseResult.Accept);
case LRAction.Shift:
m_lrState = m_grammar.m_lrStateTable[stateAction.m_value];
LRStackItem nextToken = new LRStackItem();
nextToken.m_token = m_token;
nextToken.m_state = m_lrState;
if (m_lrStack.Length == ++m_lrStackIndex)
{
LRStackItem[] larger_m_lrStack = new LRStackItem[m_lrStack.Length + MinimumLRStackSize];
Array.Copy(m_lrStack, larger_m_lrStack, m_lrStack.Length);
m_lrStack = larger_m_lrStack;
}
m_lrStack[m_lrStackIndex] = nextToken;
return(TokenParseResult.Shift);
case LRAction.Reduce:
//Produce a reduction - remove as many tokens as members in the rule & push a nonterminal token
int ruleIndex = stateAction.m_value;
Rule currentRule = m_grammar.m_ruleTable[ruleIndex];
//======== Create Reduction
LRStackItem head;
TokenParseResult parseResult;
LRState nextState;
if (m_trimReductions && currentRule.m_hasOneNonTerminal)
{
//The current rule only consists of a single nonterminal and can be trimmed from the
//parse tree. Usually we create a new Reduction, assign it to the Data property
//of Head and push it on the stack. However, in this case, the Data property of the
//Head will be assigned the Data property of the reduced token (i.e. the only one
//on the stack).
//In this case, to save code, the value popped of the stack is changed into the head.
head = m_lrStack[m_lrStackIndex];
head.m_token.m_symbol = currentRule.m_nonTerminal;
head.m_token.m_text = null;
parseResult = TokenParseResult.ReduceEliminated;
//========== Goto
nextState = m_lrStack[m_lrStackIndex - 1].m_state;
}
else
{
//Build a Reduction
head = new LRStackItem();
head.m_rule = currentRule;
head.m_token.m_symbol = currentRule.m_nonTerminal;
head.m_token.m_text = null;
m_reductionCount = currentRule.m_symbols.Length;
parseResult = TokenParseResult.ReduceNormal;
//========== Goto
nextState = m_lrStack[m_lrStackIndex - m_reductionCount].m_state;
}
//========= If nextAction is null here, then we have an Internal Table Error!!!!
LRStateAction nextAction = nextState.m_transitionVector[currentRule.m_nonTerminal.m_index];
if (nextAction != null)
{
m_lrState = m_grammar.m_lrStateTable[nextAction.m_value];
head.m_state = m_lrState;
if (parseResult == TokenParseResult.ReduceNormal)
{
if (m_lrStack.Length == ++m_lrStackIndex)
{
LRStackItem[] larger_m_lrStack = new LRStackItem[m_lrStack.Length
+ MinimumLRStackSize];
Array.Copy(m_lrStack, larger_m_lrStack, m_lrStack.Length);
m_lrStack = larger_m_lrStack;
}
m_lrStack[m_lrStackIndex] = head;
}
else
{
m_lrStack[m_lrStackIndex] = head;
}
return(parseResult);
}
else
{
return(TokenParseResult.InternalError);
}
}
}
//=== Syntax Error! Fill Expected Tokens
m_expectedTokens = new Symbol[m_lrState.ActionCount];
int length = 0;
for (int i = 0; i < m_lrState.ActionCount; i++)
{
switch (m_lrState.GetAction(i).Symbol.SymbolType)
{
case SymbolType.Terminal:
case SymbolType.End:
m_expectedTokens[length++] = m_lrState.GetAction(i).Symbol;
break;
}
}
if (length < m_expectedTokens.Length)
{
Symbol[] newArray = new Symbol[length];
Array.Copy(m_expectedTokens, newArray, length);
m_expectedTokens = newArray;
}
return(TokenParseResult.SyntaxError);
}
private TokenParseResult ParseToken()
{
LRStateAction stateAction = m_lrState.m_transitionVector[m_token.m_symbol.m_index];
if (stateAction != null)
{
//Work - shift or reduce
if (m_reductionCount > 0)
{
int newIndex = m_lrStackIndex - m_reductionCount;
m_lrStack[newIndex] = m_lrStack[m_lrStackIndex];
m_lrStackIndex = newIndex;
}
m_reductionCount = Undefined;
switch (stateAction.Action)
{
case LRAction.Accept:
m_reductionCount = 0;
return TokenParseResult.Accept;
case LRAction.Shift:
m_lrState = m_grammar.m_lrStateTable[stateAction.m_value];
LRStackItem nextToken = new LRStackItem();
nextToken.m_token = m_token;
nextToken.m_state = m_lrState;
if (m_lrStack.Length == ++m_lrStackIndex)
{
LRStackItem[] larger_m_lrStack = new LRStackItem[m_lrStack.Length + MinimumLRStackSize];
Array.Copy(m_lrStack, larger_m_lrStack, m_lrStack.Length);
m_lrStack = larger_m_lrStack;
}
m_lrStack[m_lrStackIndex] = nextToken;
return TokenParseResult.Shift;
case LRAction.Reduce:
//Produce a reduction - remove as many tokens as members in the rule & push a nonterminal token
int ruleIndex = stateAction.m_value;
Rule currentRule = m_grammar.m_ruleTable[ruleIndex];
//======== Create Reduction
LRStackItem head;
TokenParseResult parseResult;
LRState nextState;
if (m_trimReductions && currentRule.m_hasOneNonTerminal)
{
//The current rule only consists of a single nonterminal and can be trimmed from the
//parse tree. Usually we create a new Reduction, assign it to the Data property
//of Head and push it on the stack. However, in this case, the Data property of the
//Head will be assigned the Data property of the reduced token (i.e. the only one
//on the stack).
//In this case, to save code, the value popped of the stack is changed into the head.
head = m_lrStack[m_lrStackIndex];
head.m_token.m_symbol = currentRule.m_nonTerminal;
head.m_token.m_text = null;
parseResult = TokenParseResult.ReduceEliminated;
//========== Goto
nextState = m_lrStack[m_lrStackIndex - 1].m_state;
}
else
{
//Build a Reduction
head = new LRStackItem();
head.m_rule = currentRule;
head.m_token.m_symbol = currentRule.m_nonTerminal;
head.m_token.m_text = null;
m_reductionCount = currentRule.m_symbols.Length;
parseResult = TokenParseResult.ReduceNormal;
//========== Goto
nextState = m_lrStack[m_lrStackIndex - m_reductionCount].m_state;
}
//========= If nextAction is null here, then we have an Internal Table Error!!!!
LRStateAction nextAction = nextState.m_transitionVector[currentRule.m_nonTerminal.m_index];
if (nextAction != null)
{
m_lrState = m_grammar.m_lrStateTable[nextAction.m_value];
head.m_state = m_lrState;
if (parseResult == TokenParseResult.ReduceNormal)
{
if (m_lrStack.Length == ++m_lrStackIndex)
{
LRStackItem[] larger_m_lrStack = new LRStackItem[m_lrStack.Length + MinimumLRStackSize];
Array.Copy(m_lrStack, larger_m_lrStack, m_lrStack.Length);
m_lrStack = larger_m_lrStack;
}
m_lrStack[m_lrStackIndex] = head;
}
else
{
m_lrStack[m_lrStackIndex] = head;
}
return parseResult;
}
else
{
return TokenParseResult.InternalError;
}
}
}
//=== Syntax Error! Fill Expected Tokens
m_expectedTokens = new Symbol[m_lrState.ActionCount];
int length = 0;
for (int i = 0; i < m_lrState.ActionCount; i++)
{
switch (m_lrState.GetAction(i).Symbol.SymbolType)
{
case SymbolType.Terminal:
case SymbolType.End:
m_expectedTokens[length++] = m_lrState.GetAction(i).Symbol;
break;
}
}
if (length < m_expectedTokens.Length)
{
Symbol[] newArray = new Symbol[length];
Array.Copy(m_expectedTokens, newArray, length);
m_expectedTokens = newArray;
}
return TokenParseResult.SyntaxError;
}
private bool m_trimReductions = true; // Allowes to minimize reduction tree.
#endregion Fields
#region Constructors
/// <summary>
/// Initializes new instance of Parser class.
/// </summary>
/// <param name="textReader">TextReader instance to read data from.</param>
/// <param name="grammar">Grammar with parsing tables to parser input stream.</param>
public Parser(TextReader textReader, Grammar grammar)
{
if (textReader == null)
{
throw new ArgumentNullException("textReader");
}
if (grammar == null)
{
throw new ArgumentNullException("grammar");
}
m_textReader = textReader;
m_bufferSize = MinimumBufferSize;
m_buffer = new char[m_bufferSize + 1];
m_lineLength = Undefined;
ReadBuffer();
m_inputTokens = new Token[MinimumInputTokenCount];
m_lrStack = new LRStackItem[MinimumLRStackSize];
m_grammar = grammar;
// Put grammar start symbol into LR parsing stack.
m_lrState = m_grammar.InitialLRState;
LRStackItem start = new LRStackItem();
start.m_token.m_symbol = m_grammar.StartSymbol;
start.m_state = m_lrState;
m_lrStack[m_lrStackIndex] = start;
m_reductionCount = Undefined; // there are no reductions yet.
}
public void SetSourceCode(string sourceCode)
{
m_buffer = sourceCode;
m_charIndex = 0;
m_lineStart = 0;
m_lineNumber = 1;
m_lineLength = Undefined;
// Put grammar start symbol into LR parsing stack.
m_lrState = m_grammar.InitialLRState;
m_lrStack[m_lrStackIndex] = new LRStackItem { m_token = { m_symbol = m_grammar.StartSymbol }};
m_reductionCount = Undefined; // there are no reductions yet.
}
private void RsesizeLRStackIfNeed(int index)
{
if (index > m_lrStack.Length - 1)
{
LRStackItem[] tmp = new LRStackItem[m_lrStack.Length * 2];
m_lrStack.CopyTo(tmp, 0);
m_lrStack = tmp;
}
}
