public List<TokenMatch> Parse(string grammar)
{
// $month:@starttoken.value $day:@number $dayend:( st | nd | rd | th )? ( ','? $year:@number )?"
_pos = 0;
_grammar = grammar;
_errors = new List<string>();
END = _grammar.Length;
_matches = new List<TokenMatch>();
while (_pos < END)
{
var c = _grammar[_pos];
// Case 1: Begin a token match.
if (c == '$' || c == '\'')
{
var match = ReadToken();
AddMatch(match);
}
// Case 2: Move past white space.
else if (c == '\t' || c == ' ')
{
_pos++;
}
// Case 3: Last token was optional
else if (c == '?')
{
_pos++;
var last = this.Last();
if(_groupCount > 0 && last.IsGroup)
{
var group = ((TokenGroup) last);
last = group.Matches[group.Matches.Count - 1];
}
last.IsRequired = false;
}
// Case 4: "(" starts a grouping of token matches
else if( c == '(')
{
_pos++;
this.BeginGroup();
}
else if (c == ')')
{
_pos++;
CloseGroup();
}
// Case 5: word
else if (char.IsLetter(c))
{
var word = ReadWord();
var match = new TokenMatch();
match.Text = word;
match.IsRequired = true;
AddMatch(match);
}
else
_pos++;
}
return _matches;
}
private void AddMatch(TokenMatch match)
{
var last = Last();
if(_groupCount > 0 && last != null && last.IsGroup)
{
((TokenGroup) last).Matches.Add(match);
}
else
{
_matches.Add(match);
}
}
/// <summary>
/// Reads a token value : usually after ":" and includes type e.g. @number @bool
/// </summary>
/// <param name="match"></param>
private void ReadTokenValue(TokenMatch match)
{
var c = _grammar[_pos];
if (c == '@')
{
_pos++;
var tokenType = ReadWord();
match.TokenType = "@" + tokenType;
if (_pos < END)
{
// Check for ".value"
c = _grammar[_pos];
if (c == '.')
{
_pos++;
var val = ReadWord();
if (val == "value")
{
match.TokenPropEnabled = true;
match.TokenPropValue = val;
}
}
}
}
else if (c == '(')
{
_pos++;
var words = ReadList();
match.Values = words;
}
else if (c == ':')
{
_pos++;
match.Text = ":";
}
}
/// <summary>
/// Reads a token e.g. usually begins with "$"
/// </summary>
/// <returns></returns>
private TokenMatch ReadToken()
{
var start = _pos;
var match = new TokenMatch();
match.IsRequired = true;
while (_pos < END)
{
var c = _grammar[_pos];
// Case 1: Begin a token match.
if (c == '$')
{
_pos++;
var word = ReadWord();
match.Name = word;
// Expect ":" followed by tokenvalue
this.Expect(':');
this.ReadTokenValue(match);
break;
}
// Case 2: Move past white space.
else if (c == '\t' || c == ' ')
{
_pos++;
}
else if ( c == '\'')
{
_pos++;
match.Text = ReadString();
break;
}
else
break;
}
return match;
}
public void AddMatch(TokenMatch m)
{
this.Matches.Add(m);
}
private void DoParse()
{
while (_pos < END)
{
var c = _grammar[_pos];
// Case 1: Begin a token match.
if (c == '$' || c == '\'')
{
var match = ReadToken();
AddMatch(match);
}
// Case 2: '@' specific type of word/identifier
else if (c == '@')
{
_pos++;
var tokenType = ReadWord();
var match = new TokenMatch();
match.TokenType = "@" + tokenType;
match.IsRequired = true;
AddMatch(match);
}
// Case 3: Move past white space.
else if (c == '\t' || c == ' ')
{
_pos++;
}
// Case 4: Last token was optional
else if (c == '?')
{
_pos++;
var last = Last();
if (_groupCount > 0 && last.IsGroup)
{
var group = ((TokenGroup)last);
last = group.Matches[group.Matches.Count - 1];
}
last.IsRequired = false;
}
// Case 5: "(" starts a grouping of token matches
else if (c == '(')
{
_pos++;
BeginGroup();
}
else if (c == ')')
{
_pos++;
CloseGroup();
}
// Case 6: word
else if (char.IsLetter(c))
{
var word = ReadWord();
var match = new TokenMatch();
match.Text = word;
match.IsRequired = true;
AddMatch(match);
}
// Case 7: |
else if (c == '|')
{
_orCount++;
_operatorStack.Add("|");
_pos++;
}
// Case 8: escape \
else if (c == '\\')
{
_pos++;
var escapedChar = _grammar[_pos];
var match = new TokenMatch();
match.Text = escapedChar.ToString();
match.IsRequired = true;
AddMatch(match);
// Move past escaped char
_pos++;
}
// Case 9: #reference
else if (c == '#')
{
var match = ReadToken();
match.Ref = match.Name;
match.Name = null;
AddMatch(match);
}
else
{
_pos++;
}
}
}