//Compile \d \D \s \S etc.
internal RENode CompileSimpleMacro(char c)
{
RENode node = null;
RESetNode set = null;
if (@"[]{}()*-+.?\|".Contains(c.ToString()))
{
return new RETextNode(c.ToString());
}
switch (c)
{
case 'd': // [0-9]
node = set = new RESetNode(true);
set.AddRange('0', '9');
break;
case 'D': // [^0-9]
node = set = new RESetNode(false);
set.AddRange('0', '9');
break;
case 's':
node = set = new RESetNode(true);
set.AddChars(" \r\n\f\v\t");
break;
case 'S':
node = set = new RESetNode(false);
set.AddChars(" \r\n\f\v\t");
break;
case 'w': // [a-zA-Z0-9_]
node = set = new RESetNode(true);
set.AddRange('a', 'z');
set.AddRange('A', 'Z');
set.AddRange('0', '9');
set.AddChars("_");
break;
case 'W': // [^a-zA-Z0-9_]
node = set = new RESetNode(false);
set.AddRange('a', 'z');
set.AddRange('A', 'Z');
set.AddRange('0', '9');
set.AddChars("_");
break;
case 'f':
node = new RETextNode("\f");
break;
case 'n':
node = new RETextNode("\n");
break;
case 'r':
node = new RETextNode("\r");
break;
case 't':
node = new RETextNode("\t");
break;
case 'v':
node = new RETextNode("\v");
break;
case 'A':
case 'Z':
case 'z':
node = new RETextNode(String.Empty);
break;
default:
AssertParse(false, "Invalid escape.");
break;
}
return node;
}
//Compile token
internal RENode CompileAtom()
{
RENode atom = null;
RESetNode set = null;
int start = 0;
int end = 0;
AssertParse(!mParseDone, "Reached end of string. No element found.");
AssertParse(!("|)?+*{}".Contains(mCurrent.ToString())), "No element found.");
switch (mCurrent)
{
case '.': //Any single char
atom = set = new RESetNode(true);
set.AddRange(Convert.ToChar(0), Convert.ToChar(127));
NextChar();
break;
case '[': //Positive or negative set
NextChar();
atom = CompileSet();
break;
case '(': //Sub expression
int refIndex = 0; //-2 -> don't capture, -1 -> named capture, 0-> indexed capture
NextChar();
//By default, subexpressions must be captured for future reference,
if (mCurrent == '?')
{
NextChar();
if (mCurrent == ':') //If sub expression begins with ?: it means don't store reference
{
NextChar();
refIndex = -2;
}
else //Named backreference, extract backreference name
{
ExtractBackrefName(ref start, ref end);
refIndex = -1;
}
} //else use indexed backreference
atom = new RESubExprNode(CompileExpr());
AssertParse(mCurrent == ')', "Expected ')'");
NextChar();
if (refIndex == -1) //Named backreference
{
(atom as RESubExprNode).Name = mRegex.ToString().Substring(start, end - start + 1);
mNamedBackRefs.Add(atom);
}
else if (refIndex == 0) //Indexed backreference
{
mBackRefs.Add(atom);
}
break;
case '^':
case '$':
atom = new RETextNode(String.Empty);
NextChar();
break;
case '\\':
NextChar();
if (Char.ToLower(mCurrent, CultureInfo.InvariantCulture) == 'x' || Char.ToLower(mCurrent, CultureInfo.InvariantCulture) == 'u' || mCurrent == '0')
{
atom = new RETextNode(EscapeValue().ToString());
}
else if (Char.IsDigit(mCurrent))
{
atom = GetBackRef((int)EscapeValue());
AssertParse(atom != null, "Couldn't find back reference");
atom = new RESubExprNode(atom);
}
else if (mCurrent == 'k') //referencing a backreference by name
{
NextChar();
ExtractBackrefName(ref start, ref end);
atom = GetBackRef(mRegex.ToString().Substring(start, end - start + 1));
AssertParse(atom != null, "Couldn't find back reference");
atom = new RESubExprNode(atom); //Create a copy of the referenced node
}
else
{
atom = CompileSimpleMacro(mCurrent);
NextChar();
}
break;
default:
int closeIndex = mRegex.ToString().IndexOfAny("-*+?(){}\\[]^$.|".ToCharArray(), mIndex + 1);
if (closeIndex == -1)
{
mParseDone = true;
closeIndex = mRegex.Length - 1;
atom = new RETextNode(mRegex.ToString().Substring(mIndex, closeIndex - mIndex + 1));
}
else
{
atom = new RETextNode(mRegex.ToString().Substring(mIndex, closeIndex - mIndex));
}
mIndex = closeIndex;
mCurrent = mRegex[mIndex];
break;
}
return atom;
}
//Compile a character set (i.e expressions like [abc], [A-Z])
internal RENode CompileSet()
{
RENode atom = null;
char cStart, cEnd;
RESetNode set;
if (mCurrent == ':')
{
NextChar();
int closeIndex = mRegex.ToString().IndexOf(":]", StringComparison.Ordinal);
atom = CompileMacro(mIndex, closeIndex - mIndex);
mIndex = closeIndex;
NextChar();
NextChar();
return atom;
}
if (mCurrent == '^')
{
atom = set = new RESetNode(false);
NextChar();
}
else
{
atom = set = new RESetNode(true);
}
if (mCurrent == '-' || mCurrent == ']') //if - or ] are specified as the first char, escape is not required
{
set.AddChars(mCurrent.ToString());
NextChar();
}
while ((!mParseDone) && (mCurrent != ']'))
{
cStart = CompileSetChar();
if (mCurrent == '-')
{
NextChar();
AssertParse(!mParseDone && mCurrent != ']', "End of range is not specified.");
cEnd = CompileSetChar();
set.AddRange(cStart, cEnd);
}
else
{
set.AddChars(cStart.ToString());
}
}
AssertParse(mCurrent == ']', "Expected ']'.");
NextChar();
return atom;
}
internal override string Generate(Random random)
{
if (this == RECompiler.InvalidNode)
{
//select a character
int pos = random.Next(mNodeText.Length);
//generate any other character using a negative SetNode
RESetNode others = new RESetNode(false);
others.AddChars(mNodeText[pos].ToString());
//replace the character
char backup = mNodeText[pos];
mNodeText[pos] = others.Generate(random)[0];
string result = mNodeText.ToString();
//if this node is repeated it needs to be cleaned up for the next call
mNodeText[pos] = backup;
return result;
}
else
{
return mNodeText.ToString();
}
}