SyntaxNode _newTokenNode(char token, int position) {
if (!__inCharacterClass && !__caseSensitive && (_isUpperCase(token) || _isLowerCase(token))) {
CharacterClassNode node = new CharacterClassNode(position);
node._addToken(token);
node._addToken(_toggleCase(token));
return node;
}
return new TokenNode(token, position);
}
private SyntaxNode __atom() throws MalformedPatternException {
SyntaxNode current;
if (__lookahead == '(') {
__match('(');
++__openParen;
current = __regex();
__match(')');
++__closeParen;
} else if (__lookahead == '[') current = __characterClass();
else if (__lookahead == '.') {
CharacterClassNode characterSet;
__match('.');
characterSet = new NegativeCharacterClassNode(__position++);
if (__multiline) characterSet._addToken('\n');
current = characterSet;
} else if (__lookahead == '\\') {
current = __backslashToken();
} /*else if(__lookahead == '^') {
current =
new TokenNode((char)LeafNode._BEGIN_LINE_MARKER_TOKEN, __position++);
__match('^');
} else if(__lookahead == '$') {
current =
new TokenNode((char)LeafNode._END_LINE_MARKER_TOKEN, __position++);
__match('$');
} */ else if (!__isMetachar(__lookahead)) {
current = _newTokenNode(__lookahead, __position++);
__match(__lookahead);
} else
throw new MalformedPatternException(
"Parse error: unexpected character " + __lookahead + " at position " + __bytesRead);
return current;
}
private SyntaxNode __characterClass() throws MalformedPatternException {
char lastToken, token;
SyntaxNode node;
CharacterClassNode current;
__match('[');
__inCharacterClass = true;
if (__lookahead == '^') {
__match('^');
current = new NegativeCharacterClassNode(__position++);
} else current = new CharacterClassNode(__position++);
while (__lookahead != ']' && __lookahead != _END_OF_INPUT) {
if (__lookahead == '\\') {
node = __backslashToken();
--__position;
// __backslashToken() (actually newTokenNode()) does not take care of
// case insensitivity when __inCharacterClass is true.
if (node instanceof TokenNode) {
lastToken = ((TokenNode) node)._token;
current._addToken(lastToken);
if (!__caseSensitive) current._addToken(_toggleCase(lastToken));
} else {
CharacterClassNode slash;
slash = (CharacterClassNode) node;
// This could be made more efficient by manipulating the
// characterSet elements of the CharacterClassNodes but
// for the moment, this is more clear.
for (token = 0; token < LeafNode._NUM_TOKENS; token++) {
if (slash._matches(token)) current._addToken(token);
}
// A byproduct of this act is that when a '-' occurs after
// a \d, \w, etc. it is not interpreted as a range and no
// parse exception is thrown.
// This is considered a feature and not a bug for now.
continue;
}
} else {
lastToken = __lookahead;
current._addToken(__lookahead);
if (!__caseSensitive) current._addToken(_toggleCase(__lookahead));
__match(__lookahead);
}
// In Perl, a - is a token if it occurs at the beginning
// or end of the character class. Anywhere else, it indicates
// a range.
// A byproduct of this implementation is that if a '-' occurs
// after the end of a range, it is interpreted as a '-' and no
// exception is thrown. e.g., the second dash in [a-z-x]
// This is considered a feature and not a bug for now.
if (__lookahead == '-') {
__match('-');
if (__lookahead == ']') {
current._addToken('-');
break;
} else if (__lookahead == '\\') {
node = __backslashToken();
--__position;
if (node instanceof TokenNode) token = ((TokenNode) node)._token;
else
throw new MalformedPatternException(
"Parse error: invalid range specified at position " + __bytesRead);
} else {
token = __lookahead;
__match(__lookahead);
}
if (token < lastToken)
throw new MalformedPatternException(
"Parse error: invalid range specified at position " + __bytesRead);
current._addTokenRange(lastToken + 1, token);
if (!__caseSensitive)
current._addTokenRange(_toggleCase((char) (lastToken + 1)), _toggleCase(token));
}
}
__match(']');
__inCharacterClass = false;
return current;
}
private SyntaxNode __backslashToken() throws MalformedPatternException {
SyntaxNode current;
char token;
int number;
__match('\\');
if (__lookahead == 'x') {
__match('x');
// Parse a hexadecimal number
current = _newTokenNode((char) __parseUnsignedInteger(16, 2, 2), __position++);
} else if (__lookahead == 'c') {
__match('c');
// Create a control character
token = Character.toUpperCase(__lookahead);
token = (char) (token > 63 ? token - 64 : token + 64);
current = new TokenNode(token, __position++);
__match(__lookahead);
} else if (__lookahead >= '0' && __lookahead <= '9') {
__match(__lookahead);
if (__lookahead >= '0' && __lookahead <= '9') {
// We have an octal character or a multi-digit backreference.
// Assume octal character for now.
__putback();
number = __parseUnsignedInteger(10, 2, 3);
number = Integer.parseInt(Integer.toString(number), 8);
current = _newTokenNode((char) number, __position++);
} else {
// We have either \0, an escaped digit, or a backreference.
__putback();
if (__lookahead == '0') {
// \0 matches the null character
__match('0');
current = new TokenNode('\0', __position++);
} else {
// Either an escaped digit or backreference.
number = Character.digit(__lookahead, 10);
current = _newTokenNode(__lookahead, __position++);
__match(__lookahead);
}
}
} else if (__lookahead == 'b') {
// Inside of a character class the \b means backspace, otherwise
// it means a word boundary
// if(__inCharacterClass)
// \b always means backspace
current = new TokenNode('\b', __position++);
/*
else
current = new TokenNode((char)LeafNode._WORD_BOUNDARY_MARKER_TOKEN,
position++);
*/
__match('b');
} /*else if(__lookahead == 'B' && !__inCharacterClass){
current = new TokenNode((char)LeafNode._NONWORD_BOUNDARY_MARKER_TOKEN,
position++);
__match('B');
} */ else {
CharacterClassNode characterSet;
token = __lookahead;
switch (__lookahead) {
case 'n':
token = '\n';
break;
case 'r':
token = '\r';
break;
case 't':
token = '\t';
break;
case 'f':
token = '\f';
break;
}
switch (token) {
case 'd':
characterSet = new CharacterClassNode(__position++);
characterSet._addTokenRange('0', '9');
current = characterSet;
break;
case 'D':
characterSet = new NegativeCharacterClassNode(__position++);
characterSet._addTokenRange('0', '9');
current = characterSet;
break;
case 'w':
characterSet = new CharacterClassNode(__position++);
characterSet._addTokenRange('0', '9');
characterSet._addTokenRange('a', 'z');
characterSet._addTokenRange('A', 'Z');
characterSet._addToken('_');
current = characterSet;
break;
case 'W':
characterSet = new NegativeCharacterClassNode(__position++);
characterSet._addTokenRange('0', '9');
characterSet._addTokenRange('a', 'z');
characterSet._addTokenRange('A', 'Z');
characterSet._addToken('_');
current = characterSet;
break;
case 's':
characterSet = new CharacterClassNode(__position++);
characterSet._addToken(' ');
characterSet._addToken('\f');
characterSet._addToken('\n');
characterSet._addToken('\r');
characterSet._addToken('\t');
current = characterSet;
break;
case 'S':
characterSet = new NegativeCharacterClassNode(__position++);
characterSet._addToken(' ');
characterSet._addToken('\f');
characterSet._addToken('\n');
characterSet._addToken('\r');
characterSet._addToken('\t');
current = characterSet;
break;
default:
current = _newTokenNode(token, __position++);
break;
}
__match(__lookahead);
}
return current;
}
