/**
* 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. 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 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 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;
}
/**
* 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);
}
}
