/**
* Parses a regular expression factor. This method handles the
* Fact production in the grammar (see regexp.grammar).
*
* @param start the initial NFA state
*
* @return the terminating NFA state
*
* @throws RegExpException if an error was encountered in the
* pattern string
*/
private NFAState ParseFact(NFAState start)
{
NFAState placeholder = new NFAState();
NFAState end;
end = ParseAtom(placeholder);
switch (PeekChar(0))
{
case '?':
case '*':
case '+':
case '{':
end = ParseAtomModifier(placeholder, end);
break;
}
if (placeholder.incoming.Length > 0 && start.outgoing.Length > 0)
{
start.AddOut(new NFAEpsilonTransition(placeholder));
return(end);
}
else
{
placeholder.MergeInto(start);
return((end == placeholder) ? start : end);
}
}
/**
* Parses a regular expression. This method handles the Expr
* production in the grammar (see regexp.grammar).
*
* @param start the initial NFA state
*
* @return the terminating NFA state
*
* @throws RegExpException if an error was encountered in the
* pattern string
*/
private NFAState ParseExpr(NFAState start)
{
NFAState end = new NFAState();
NFAState subStart;
NFAState subEnd;
do
{
if (PeekChar(0) == '|')
{
ReadChar('|');
}
subStart = new NFAState();
subEnd = ParseTerm(subStart);
if (subStart.incoming.Length == 0)
{
subStart.MergeInto(start);
}
else
{
start.AddOut(new NFAEpsilonTransition(subStart));
}
if (subEnd.outgoing.Length == 0 ||
(!end.HasTransitions() && PeekChar(0) != '|'))
{
subEnd.MergeInto(end);
}
else
{
subEnd.AddOut(new NFAEpsilonTransition(end));
}
} while (PeekChar(0) == '|');
return(end);
}
/// <summary>
/// Parses a regular expression character set. This method handles
/// the contents of the '[...]' construct in a regular expression.
/// </summary>
/// <param name="start">The initial NFA state</param>
/// <returns>The terminating NFA state</returns>
/// <exception cref="RegExpException">
/// If an error was encountered in the pattern string
/// </exception>
private NFAState ParseCharSet(NFAState start)
{
NFAState end = new NFAState();
NFACharRangeTransition range;
char min;
char max;
if (this.PeekChar(0) == '^')
{
this.ReadChar('^');
range = new NFACharRangeTransition(true, this.ignoreCase, end);
}
else
{
range = new NFACharRangeTransition(false, this.ignoreCase, end);
}
start.AddOut(range);
while (this.PeekChar(0) > 0)
{
min = (char)this.PeekChar(0);
switch (min)
{
case ']':
return(end);
case '\\':
range.AddCharacter(this.ReadEscapeChar());
break;
default:
this.ReadChar(min);
if (this.PeekChar(0) == '-' &&
this.PeekChar(1) > 0 &&
this.PeekChar(1) != ']')
{
this.ReadChar('-');
max = this.ReadChar();
range.AddRange(min, max);
}
else
{
range.AddCharacter(min);
}
break;
}
}
return(end);
}
/**
* Adds a string match to this automaton. New states and
* transitions will be added to extend this automaton to support
* the specified string.
*
* @param str the string to match
* @param ignoreCase the case-insensitive match flag
* @param value the match value
*/
public void AddTextMatch(string str, bool ignoreCase, TokenPattern value)
{
NFAState state;
char ch = str[0];
if (ch < 128 && !ignoreCase)
{
state = initialChar[ch];
if (state == null)
{
state = initialChar[ch] = new NFAState();
}
}
else
{
state = initial.AddOut(ch, ignoreCase, null);
}
for (int i = 1; i < str.Length; i++)
{
state = state.AddOut(str[i], ignoreCase, null);
}
state.value = value;
}
/**
* Merges all the transitions in this state into another
* state.
*
* @param state the state to merge into
*/
public void MergeInto(NFAState state)
{
for (int i = 0; i < incoming.Length; i++)
{
state.AddIn(incoming[i]);
incoming[i].state = state;
}
incoming = null;
for (int i = 0; i < outgoing.Length; i++)
{
state.AddOut(outgoing[i]);
}
outgoing = null;
}
/// <summary>
/// Parses a regular expression character. This method handles
/// a single normal character in a regular expression.
/// </summary>
/// <param name="start">The initial NFA state</param>
/// <returns>The terminating NFA state</returns>
/// <exception cref="RegExpException">
/// If an error was encountered in the pattern string
/// </exception>
private NFAState ParseChar(NFAState start)
{
switch (this.PeekChar(0))
{
case '\\':
return(this.ParseEscapeChar(start));
case '^':
case '$':
throw new RegExpException(
RegExpException.ErrorType.UnsupportedSpecialCharacter,
this.pos,
this.pattern);
default:
return(start.AddOut(this.ReadChar(), this.ignoreCase, new NFAState()));
}
}
/**
* Parses a regular expression character. This method handles
* a single normal character in a regular expression.
*
* @param start the initial NFA state
*
* @return the terminating NFA state
*
* @throws RegExpException if an error was encountered in the
* pattern string
*/
private NFAState ParseChar(NFAState start)
{
switch (PeekChar(0))
{
case '\\':
return(ParseEscapeChar(start));
case '^':
case '$':
throw new RegExpException(
RegExpException.ErrorType.UNSUPPORTED_SPECIAL_CHARACTER,
pos,
pattern);
default:
return(start.AddOut(ReadChar(), ignoreCase, new NFAState()));
}
}
/// <summary>
/// Parses a regular expression atom. This method handles the
/// Atom production in the grammar (see regexp.grammar).
/// </summary>
/// <param name="start">The initial NFA state</param>
/// <returns>The terminating NFA state</returns>
/// <exception cref="RegExpException">
/// If an error was encountered in the pattern string
/// </exception>
private NFAState ParseAtom(NFAState start)
{
NFAState end;
switch (this.PeekChar(0))
{
case '.':
this.ReadChar('.');
return(start.AddOut(new NFADotTransition(new NFAState())));
case '(':
this.ReadChar('(');
end = this.ParseExpr(start);
this.ReadChar(')');
return(end);
case '[':
this.ReadChar('[');
end = this.ParseCharSet(start);
this.ReadChar(']');
return(end);
case -1:
case ')':
case ']':
case '{':
case '}':
case '?':
case '*':
case '+':
case '|':
throw new RegExpException(
RegExpException.ErrorType.UnexpectedCharacter,
this.pos,
this.pattern);
default:
return(this.ParseChar(start));
}
}
/**
* Parses a regular expression atom. This method handles the
* Atom production in the grammar (see regexp.grammar).
*
* @param start the initial NFA state
*
* @return the terminating NFA state
*
* @throws RegExpException if an error was encountered in the
* pattern string
*/
private NFAState ParseAtom(NFAState start)
{
NFAState end;
switch (PeekChar(0))
{
case '.':
ReadChar('.');
return(start.AddOut(new NFADotTransition(new NFAState())));
case '(':
ReadChar('(');
end = ParseExpr(start);
ReadChar(')');
return(end);
case '[':
ReadChar('[');
end = ParseCharSet(start);
ReadChar(']');
return(end);
case -1:
case ')':
case ']':
case '{':
case '}':
case '?':
case '*':
case '+':
case '|':
throw new RegExpException(
RegExpException.ErrorType.UNEXPECTED_CHARACTER,
pos,
pattern);
default:
return(ParseChar(start));
}
}
/// <summary>
/// Parses a regular expression character escape. This method
/// handles a single character escape in a regular expression.
/// </summary>
/// <param name="start">The initial NFA state</param>
/// <returns>The terminating NFA state</returns>
/// <exception cref="RegExpException">
/// If an error was encountered in the pattern string
/// </exception>
private NFAState ParseEscapeChar(NFAState start)
{
NFAState end = new NFAState();
if (this.PeekChar(0) == '\\' && this.PeekChar(1) > 0)
{
switch ((char)this.PeekChar(1))
{
case 'd':
this.ReadChar();
this.ReadChar();
return(start.AddOut(new NFADigitTransition(end)));
case 'D':
this.ReadChar();
this.ReadChar();
return(start.AddOut(new NFANonDigitTransition(end)));
case 's':
this.ReadChar();
this.ReadChar();
return(start.AddOut(new NFAWhitespaceTransition(end)));
case 'S':
this.ReadChar();
this.ReadChar();
return(start.AddOut(new NFANonWhitespaceTransition(end)));
case 'w':
this.ReadChar();
this.ReadChar();
return(start.AddOut(new NFAWordTransition(end)));
case 'W':
this.ReadChar();
this.ReadChar();
return(start.AddOut(new NFANonWordTransition(end)));
}
}
return(start.AddOut(this.ReadEscapeChar(), this.ignoreCase, end));
}
/// <summary>
/// Parses a regular expression. This method handles the <c>Expr</c>
/// production in the grammar (see regexp.grammar).
/// </summary>
/// <param name="start">The initial NFA state</param>
/// <returns>The terminating NFA state</returns>
/// <exception cref="RegExpException">
/// If an error was encountered in the pattern string
/// </exception>
private NFAState ParseExpr(NFAState start)
{
NFAState end = new NFAState();
NFAState subStart;
NFAState subEnd;
do
{
if (this.PeekChar(0) == '|')
{
this.ReadChar('|');
}
subStart = new NFAState();
subEnd = this.ParseTerm(subStart);
if (subStart.Incoming.Count == 0)
{
subStart.MergeInto(start);
}
else
{
start.AddOut(new NFAEpsilonTransition(subStart));
}
if (subEnd.Outgoing.Count == 0 ||
(!end.HasTransitions && this.PeekChar(0) != '|'))
{
subEnd.MergeInto(end);
}
else
{
subEnd.AddOut(new NFAEpsilonTransition(end));
}
}while (this.PeekChar(0) == '|');
return(end);
}
/**
* Parses a regular expression character escape. This method
* handles a single character escape in a regular expression.
*
* @param start the initial NFA state
*
* @return the terminating NFA state
*
* @throws RegExpException if an error was encountered in the
* pattern string
*/
private NFAState ParseEscapeChar(NFAState start) {
NFAState end = new NFAState();
if (PeekChar(0) == '\\' && PeekChar(1) > 0) {
switch ((char) PeekChar(1)) {
case 'd':
ReadChar();
ReadChar();
return start.AddOut(new NFADigitTransition(end));
case 'D':
ReadChar();
ReadChar();
return start.AddOut(new NFANonDigitTransition(end));
case 's':
ReadChar();
ReadChar();
return start.AddOut(new NFAWhitespaceTransition(end));
case 'S':
ReadChar();
ReadChar();
return start.AddOut(new NFANonWhitespaceTransition(end));
case 'w':
ReadChar();
ReadChar();
return start.AddOut(new NFAWordTransition(end));
case 'W':
ReadChar();
ReadChar();
return start.AddOut(new NFANonWordTransition(end));
}
}
return start.AddOut(ReadEscapeChar(), ignoreCase, end);
}
/**
* Parses a regular expression character. This method handles
* a single normal character in a regular expression.
*
* @param start the initial NFA state
*
* @return the terminating NFA state
*
* @throws RegExpException if an error was encountered in the
* pattern string
*/
private NFAState ParseChar(NFAState start) {
switch (PeekChar(0)) {
case '\\':
return ParseEscapeChar(start);
case '^':
case '$':
throw new RegExpException(
RegExpException.ErrorType.UNSUPPORTED_SPECIAL_CHARACTER,
pos,
pattern);
default:
return start.AddOut(ReadChar(), ignoreCase, new NFAState());
}
}
/**
* Parses a regular expression character set. This method handles
* the contents of the '[...]' construct in a regular expression.
*
* @param start the initial NFA state
*
* @return the terminating NFA state
*
* @throws RegExpException if an error was encountered in the
* pattern string
*/
private NFAState ParseCharSet(NFAState start) {
NFAState end = new NFAState();
NFACharRangeTransition range;
char min;
char max;
if (PeekChar(0) == '^') {
ReadChar('^');
range = new NFACharRangeTransition(true, ignoreCase, end);
} else {
range = new NFACharRangeTransition(false, ignoreCase, end);
}
start.AddOut(range);
while (PeekChar(0) > 0) {
min = (char) PeekChar(0);
switch (min) {
case ']':
return end;
case '\\':
range.AddCharacter(ReadEscapeChar());
break;
default:
ReadChar(min);
if (PeekChar(0) == '-' &&
PeekChar(1) > 0 &&
PeekChar(1) != ']') {
ReadChar('-');
max = ReadChar();
range.AddRange(min, max);
} else {
range.AddCharacter(min);
}
break;
}
}
return end;
}
/**
* Parses a regular expression atom modifier. This method handles
* the AtomModifier production in the grammar (see regexp.grammar).
*
* @param start the initial NFA state
* @param end the terminal NFA state
*
* @return the terminating NFA state
*
* @throws RegExpException if an error was encountered in the
* pattern string
*/
private NFAState ParseAtomModifier(NFAState start, NFAState end) {
int min = 0;
int max = -1;
int firstPos = pos;
// Read min and max
switch (ReadChar()) {
case '?':
min = 0;
max = 1;
break;
case '*':
min = 0;
max = -1;
break;
case '+':
min = 1;
max = -1;
break;
case '{':
min = ReadNumber();
max = min;
if (PeekChar(0) == ',') {
ReadChar(',');
max = -1;
if (PeekChar(0) != '}') {
max = ReadNumber();
}
}
ReadChar('}');
if (max == 0 || (max > 0 && min > max)) {
throw new RegExpException(
RegExpException.ErrorType.INVALID_REPEAT_COUNT,
firstPos,
pattern);
}
break;
default:
throw new RegExpException(
RegExpException.ErrorType.UNEXPECTED_CHARACTER,
pos - 1,
pattern);
}
// Read possessive or reluctant modifiers
if (PeekChar(0) == '?') {
throw new RegExpException(
RegExpException.ErrorType.UNSUPPORTED_SPECIAL_CHARACTER,
pos,
pattern);
} else if (PeekChar(0) == '+') {
throw new RegExpException(
RegExpException.ErrorType.UNSUPPORTED_SPECIAL_CHARACTER,
pos,
pattern);
}
// Handle supported repeaters
if (min == 0 && max == 1) {
return start.AddOut(new NFAEpsilonTransition(end));
} else if (min == 0 && max == -1) {
if (end.outgoing.Length == 0) {
end.MergeInto(start);
} else {
end.AddOut(new NFAEpsilonTransition(start));
}
return start;
} else if (min == 1 && max == -1) {
if (start.outgoing.Length == 1 &&
end.outgoing.Length == 0 &&
end.incoming.Length == 1 &&
start.outgoing[0] == end.incoming[0]) {
end.AddOut(start.outgoing[0].Copy(end));
} else {
end.AddOut(new NFAEpsilonTransition(start));
}
return end;
} else {
throw new RegExpException(
RegExpException.ErrorType.INVALID_REPEAT_COUNT,
firstPos,
pattern);
}
}
/**
* Parses a regular expression atom. This method handles the
* Atom production in the grammar (see regexp.grammar).
*
* @param start the initial NFA state
*
* @return the terminating NFA state
*
* @throws RegExpException if an error was encountered in the
* pattern string
*/
private NFAState ParseAtom(NFAState start) {
NFAState end;
switch (PeekChar(0)) {
case '.':
ReadChar('.');
return start.AddOut(new NFADotTransition(new NFAState()));
case '(':
ReadChar('(');
end = ParseExpr(start);
ReadChar(')');
return end;
case '[':
ReadChar('[');
end = ParseCharSet(start);
ReadChar(']');
return end;
case -1:
case ')':
case ']':
case '{':
case '}':
case '?':
case '*':
case '+':
case '|':
throw new RegExpException(
RegExpException.ErrorType.UNEXPECTED_CHARACTER,
pos,
pattern);
default:
return ParseChar(start);
}
}
/**
* Parses a regular expression factor. This method handles the
* Fact production in the grammar (see regexp.grammar).
*
* @param start the initial NFA state
*
* @return the terminating NFA state
*
* @throws RegExpException if an error was encountered in the
* pattern string
*/
private NFAState ParseFact(NFAState start) {
NFAState placeholder = new NFAState();
NFAState end;
end = ParseAtom(placeholder);
switch (PeekChar(0)) {
case '?':
case '*':
case '+':
case '{':
end = ParseAtomModifier(placeholder, end);
break;
}
if (placeholder.incoming.Length > 0 && start.outgoing.Length > 0) {
start.AddOut(new NFAEpsilonTransition(placeholder));
return end;
} else {
placeholder.MergeInto(start);
return (end == placeholder) ? start : end;
}
}
/**
* Parses a regular expression. This method handles the Expr
* production in the grammar (see regexp.grammar).
*
* @param start the initial NFA state
*
* @return the terminating NFA state
*
* @throws RegExpException if an error was encountered in the
* pattern string
*/
private NFAState ParseExpr(NFAState start) {
NFAState end = new NFAState();
NFAState subStart;
NFAState subEnd;
do {
if (PeekChar(0) == '|') {
ReadChar('|');
}
subStart = new NFAState();
subEnd = ParseTerm(subStart);
if (subStart.incoming.Length == 0) {
subStart.MergeInto(start);
} else {
start.AddOut(new NFAEpsilonTransition(subStart));
}
if (subEnd.outgoing.Length == 0 ||
(!end.HasTransitions() && PeekChar(0) != '|')) {
subEnd.MergeInto(end);
} else {
subEnd.AddOut(new NFAEpsilonTransition(end));
}
} while (PeekChar(0) == '|');
return end;
}
/**
* Merges all the transitions in this state into another
* state.
*
* @param state the state to merge into
*/
public void MergeInto(NFAState state)
{
for (int i = 0; i < incoming.Length; i++) {
state.AddIn(incoming[i]);
incoming[i].state = state;
}
incoming = null;
for (int i = 0; i < outgoing.Length; i++) {
state.AddOut(outgoing[i]);
}
outgoing = null;
}
/**
* Parses a regular expression atom modifier. This method handles
* the AtomModifier production in the grammar (see regexp.grammar).
*
* @param start the initial NFA state
* @param end the terminal NFA state
*
* @return the terminating NFA state
*
* @throws RegExpException if an error was encountered in the
* pattern string
*/
private NFAState ParseAtomModifier(NFAState start, NFAState end)
{
int min = 0;
int max = -1;
int firstPos = pos;
// Read min and max
switch (ReadChar())
{
case '?':
min = 0;
max = 1;
break;
case '*':
min = 0;
max = -1;
break;
case '+':
min = 1;
max = -1;
break;
case '{':
min = ReadNumber();
max = min;
if (PeekChar(0) == ',')
{
ReadChar(',');
max = -1;
if (PeekChar(0) != '}')
{
max = ReadNumber();
}
}
ReadChar('}');
if (max == 0 || (max > 0 && min > max))
{
throw new RegExpException(
RegExpException.ErrorType.INVALID_REPEAT_COUNT,
firstPos,
pattern);
}
break;
default:
throw new RegExpException(
RegExpException.ErrorType.UNEXPECTED_CHARACTER,
pos - 1,
pattern);
}
// Read possessive or reluctant modifiers
if (PeekChar(0) == '?')
{
throw new RegExpException(
RegExpException.ErrorType.UNSUPPORTED_SPECIAL_CHARACTER,
pos,
pattern);
}
else if (PeekChar(0) == '+')
{
throw new RegExpException(
RegExpException.ErrorType.UNSUPPORTED_SPECIAL_CHARACTER,
pos,
pattern);
}
// Handle supported repeaters
if (min == 0 && max == 1)
{
return(start.AddOut(new NFAEpsilonTransition(end)));
}
else if (min == 0 && max == -1)
{
if (end.outgoing.Length == 0)
{
end.MergeInto(start);
}
else
{
end.AddOut(new NFAEpsilonTransition(start));
}
return(start);
}
else if (min == 1 && max == -1)
{
if (start.outgoing.Length == 1 &&
end.outgoing.Length == 0 &&
end.incoming.Length == 1 &&
start.outgoing[0] == end.incoming[0])
{
end.AddOut(start.outgoing[0].Copy(end));
}
else
{
end.AddOut(new NFAEpsilonTransition(start));
}
return(end);
}
else
{
throw new RegExpException(
RegExpException.ErrorType.INVALID_REPEAT_COUNT,
firstPos,
pattern);
}
}
/// <summary>
/// Parses a regular expression atom modifier. This method handles
/// the AtomModifier production in the grammar (see regexp.grammar).
/// </summary>
/// <param name="start">The initial NFA state</param>
/// <param name="end">The terminal NFA state</param>
/// <returns>The terminating NFA state</returns>
/// <exception cref="RegExpException">
/// If an error was encountered in the pattern string
/// </exception>
private NFAState ParseAtomModifier(
NFAState start,
NFAState end)
{
int min;
int max;
int firstPos = this.pos;
// Read min and max
switch (this.ReadChar())
{
case '?':
min = 0;
max = 1;
break;
case '*':
min = 0;
max = -1;
break;
case '+':
min = 1;
max = -1;
break;
case '{':
min = this.ReadNumber();
max = min;
if (this.PeekChar(0) == ',')
{
this.ReadChar(',');
max = -1;
if (this.PeekChar(0) != '}')
{
max = this.ReadNumber();
}
}
this.ReadChar('}');
if (max == 0 || (max > 0 && min > max))
{
throw new RegExpException(
RegExpException.ErrorType.InvalidRepeatCount,
firstPos,
this.pattern);
}
break;
default:
throw new RegExpException(
RegExpException.ErrorType.UnexpectedCharacter,
this.pos - 1,
this.pattern);
}
// Read possessive or reluctant modifiers
if (this.PeekChar(0) == '?')
{
throw new RegExpException(
RegExpException.ErrorType.UnsupportedSpecialCharacter,
this.pos,
this.pattern);
}
else if (this.PeekChar(0) == '+')
{
throw new RegExpException(
RegExpException.ErrorType.UnsupportedSpecialCharacter,
this.pos,
this.pattern);
}
// Handle supported repeaters
if (min == 0 && max == 1)
{
return(start.AddOut(new NFAEpsilonTransition(end)));
}
else if (min == 0 && max == -1)
{
if (end.Outgoing.Count == 0)
{
end.MergeInto(start);
}
else
{
end.AddOut(new NFAEpsilonTransition(start));
}
return(start);
}
else if (min == 1 && max == -1)
{
if (start.Outgoing.Count == 1 &&
end.Outgoing.Count == 0 &&
end.Incoming.Count == 1 &&
start.Outgoing[0] == end.Incoming[0])
{
end.AddOut(start.Outgoing[0].Copy(end));
}
else
{
end.AddOut(new NFAEpsilonTransition(start));
}
return(end);
}
else
{
throw new RegExpException(
RegExpException.ErrorType.InvalidRepeatCount,
firstPos,
this.pattern);
}
}
