public TokenPos(TokenPos copyFrom, int end)
{
this.m_source = copyFrom.m_source;
this.m_type = copyFrom.m_type;
this.m_start = copyFrom.m_start;
this.m_extracted = copyFrom.m_extracted;
m_end = end;
}
public const char escape = '\\'; // can be used only inside quoted strings
public static TokenPos[] tokenize(string src)
{
var count = tokenize(src, null);
if (count < 0)
{
return(null);
}
var tokens = new TokenPos[count];
count = tokenize(src, tokens);
return(tokens);
}
public static ListOrValue parseRoot(TokenPos[] tokens)
{
//ListOrValue retVal = new ListOrValue();
if (tokens == null || tokens.Length < 1)
{
return(new ListOrValue());
}
var tok = new TokenPos(
tokens[0].getSource(),
TokenType.document,
0,
tokens[0].sourceLength()
//retVal.tokenSectionLength()
//retVal.m_value.m_source.Length
);
ListOrValue retVal = new ListOrValue(new List <ListOrValue>(), tok);
//retVal.m_value.m_type = TokenType.document;
//retVal.m_value.m_start = 0;
//retVal.m_value.m_source = tokens[0].m_source;
//retVal.m_value.m_end = retVal.m_value.m_source.Length;
//retVal.m_list = new List<ListOrValue>();
var tokensLength = tokens.Length;
int idx = 0;
while (idx < tokensLength)
{
var item = parseToken(tokens, idx, out idx);
if (item.isDefault())
{
continue;
}
retVal.add(item);
}
if (retVal.count() < 1)
{
return(new ListOrValue());
}
return(retVal);
}
public static int tokenize(string self, TokenPos[] tokens)
{
if (self == null || self == "")
{
return(-1);
}
var curState = ParseState.insideBody;
var lastToken = TokenType.none;
var strLen = self.Length;
var listStartCount = 0;
var listEndCount = 0;
var tokenIdx = 0;
var nextIsScaped = false;
Stack <int> listStartStack = null;
var tokensLen = 0;
if (tokens != null)
{
tokensLen = tokens.Length;
listStartStack = new Stack <int>(tokensLen); // TODO think of an optimized way without heap allocation
}
for (var i = 0; i < strLen; i++)
{
var cur = self[i];
switch (curState)
{
case ParseState.insideBody:
switch (cur)
{
case listStart:
listStartCount++;
if (tokens != null)
{
listStartStack.Push(tokenIdx);
tokens[tokenIdx] = new TokenPos(self, TokenType.list, i, 0);
//tokens[tokenIdx] = new TokenPos() { m_source = self, m_type = TokenType.list, m_start = i };
}
tokenIdx++;
lastToken = TokenType.list;
break;
case listEnd:
listEndCount++;
if (tokens != null)
{
var idx = listStartStack.Pop();
tokens[idx] = new TokenPos(tokens[idx], i + 1);
//tokens[idx].m_end = i + 1;
}
lastToken = TokenType.list;
break;
case stringStart:
curState = ParseState.insideQuotedString;
if (tokens != null)
{
tokens[tokenIdx] = new TokenPos(self, TokenType.quotedString, i, 0);
//tokens[tokenIdx] = new TokenPos() { m_source = self, m_type = TokenType.quotedString, m_start = i };
}
tokenIdx++;
lastToken = TokenType.quotedString;
break;
case escape:
return(-1); // syntax error, can only use scape char inside identifier or string
default:
// it could be entering a neutral section
if (cur < 33) // 33 is the first printable ascii char, the exclamation mark '!'
{
curState = ParseState.insideNeutral;
if (tokens != null && lastToken == TokenType.unquotedString) // close last unquoted string if any
{
tokens[tokenIdx - 1] = new TokenPos(tokens[tokenIdx - 1], i);
//tokens[tokenIdx - 1].m_end = i;
}
if (tokens != null)
{
tokens[tokenIdx] = new TokenPos(self, TokenType.neutral, i, 0);
//tokens[tokenIdx] = new TokenPos() { m_source = self, m_type = TokenType.neutral, m_start = i };
}
tokenIdx++;
lastToken = TokenType.neutral;
break;
}
// else, it is starting a identifier
if (tokens != null && lastToken == TokenType.neutral) // close last neutral if any
{
tokens[tokenIdx - 1] = new TokenPos(tokens[tokenIdx - 1], i);
//tokens[tokenIdx - 1].m_end = i;
}
if (tokens != null)
{
tokens[tokenIdx] = new TokenPos(self, TokenType.unquotedString, i, 0);
//tokens[tokenIdx] = new TokenPos() { m_source = self, m_type = TokenType.unquotedString, m_start = i };
}
tokenIdx++;
lastToken = TokenType.unquotedString;
curState = ParseState.insideUnquotedString;
break;
}
break;
case ParseState.insideQuotedString:
switch (cur)
{
case escape:
if (nextIsScaped)
{
nextIsScaped = false;
break;
}
nextIsScaped = true;
break;
case stringEnd:
if (nextIsScaped)
{
nextIsScaped = false;
break;
}
if (tokens != null)
{
tokens[tokenIdx - 1] = new TokenPos(tokens[tokenIdx - 1], i + 1);
//tokens[tokenIdx - 1].m_end = i + 1;
}
curState = ParseState.insideBody;
lastToken = TokenType.quotedString;
break;
default:
break; // normal string characters
}
break;
case ParseState.insideUnquotedString:
switch (cur)
{
case listStart:
if (tokens != null) // close last unquoted string
{
tokens[tokenIdx - 1] = new TokenPos(tokens[tokenIdx - 1], i);
//tokens[tokenIdx - 1].m_end = i;
}
tokenIdx++;
i--; // must reevaluate character
curState = ParseState.insideBody;
break;
case listEnd:
if (tokens != null) // close last unquoted string
{
tokens[tokenIdx - 1] = new TokenPos(tokens[tokenIdx - 1], i);
//tokens[tokenIdx - 1].m_end = i;
}
i--; // must reevaluate character
curState = ParseState.insideBody;
break;
case stringStart:
return(-1);
case escape:
return(-1);
default:
if (cur > 32)
{
break;
}
// could be starting a neutral //
if (tokens != null && lastToken == TokenType.unquotedString) // close last unquoted if any
{
tokens[tokenIdx - 1] = new TokenPos(tokens[tokenIdx - 1], i);
//tokens[tokenIdx - 1].m_end = i;
}
if (tokens != null)
{
tokens[tokenIdx] = new TokenPos(self, TokenType.neutral, i, 0);
//tokens[tokenIdx] = new TokenPos() { m_source = self, m_type = TokenType.neutral, m_start = i };
}
tokenIdx++;
curState = ParseState.insideNeutral;
lastToken = TokenType.unquotedString;
break;
}
break;
case ParseState.insideNeutral:
if (cur < 33)
{
break;
}
if (tokens != null) // close last neutral
{
tokens[tokenIdx - 1] = new TokenPos(tokens[tokenIdx - 1], i);
//tokens[tokenIdx - 1].m_end = i;
}
i--; // must reevaluate character
curState = ParseState.insideBody;
lastToken = TokenType.neutral;
break;
}
}
if (curState == ParseState.insideQuotedString)
{
return(-1); // unclosed enquoted string
}
if (listStartCount != listEndCount)
{
return(-1); // unatched open/close list
}
if (tokens != null && tokensLen > 0 && tokens[tokensLen - 1].getEnd() < 1) // close last token
{
tokens[tokensLen - 1] = new TokenPos(tokens[tokensLen - 1], strLen);
//tokens[tokensLen - 1].m_end = strLen;
}
return(tokenIdx);
}
public ListOrValue(List <ListOrValue> list, TokenPos token)
{
m_list = list;
m_value = token;
}