/*
* Resolving emphasis tokens is a two part process
*
* 1. Find all valid sequences of * and _ and create `mark` tokens for them
* this is done by CreateEmphasisMarks during the initial character scan
* done by Tokenize
*
* 2. Looks at all these emphasis marks and tries to pair them up
* to make the actual <em> and <strong> tokens
*
* Any unresolved emphasis marks are rendered unaltered as * or _
*/
// Create emphasis mark for sequences of '*' and '_' (part 1)
public Token CreateEmphasisMark(Token previous)
{
// Capture current state
char ch = current;
char altch = ch == '*' ? '_' : '*';
int savepos = position;
// Check for a consecutive sequence of just '_' and '*'
if (bof || char.IsWhiteSpace(CharAtOffset(-1)))
{
while (IsEmphasisChar(current))
SkipForward(1);
if (eof || char.IsWhiteSpace(current))
{
return new Token(TokenType.Html, savepos, position - savepos);
}
// Rewind
position = savepos;
}
// Scan backwards and see if we have space before
while (IsEmphasisChar(CharAtOffset(-1)))
SkipForward(-1);
bool bSpaceBefore = bof || char.IsWhiteSpace(CharAtOffset(-1));
position = savepos;
// Count how many matching emphasis characters
while (current == ch)
{
SkipForward(1);
}
int count=position-savepos;
// Scan forwards and see if we have space after
while (IsEmphasisChar(CharAtOffset(1)))
SkipForward(1);
bool bSpaceAfter = eof || char.IsWhiteSpace(current);
position = savepos + count;
// This should have been stopped by check above
System.Diagnostics.Debug.Assert(!bSpaceBefore || !bSpaceAfter);
if (bSpaceBefore || previous.type == TokenType.closing_mark)
{
return CreateToken(TokenType.opening_mark, savepos, position - savepos);
}
if (bSpaceAfter || previous.type == TokenType.opening_mark)
{
return CreateToken(TokenType.closing_mark, savepos, position - savepos);
}
if (m_Markdown.ExtraMode && ch == '_')
return null;
return CreateToken(TokenType.internal_mark, savepos, position - savepos);
}
// FreeToken - return a token to the spare token pool
internal void FreeToken(Token token)
{
token.data = null;
m_SpareTokens.Push(token);
}
// FreeToken - return a token to the spare token pool
public void FreeToken(Token token)
{
token.data = null;
m_SpareTokens.Push(token);
}
// Split mark token
public Token SplitMarkToken(List<Token> tokens, List<Token> marks, Token token, int position)
{
// Create the new rhs token
Token tokenRhs = CreateToken(token.type, token.startOffset + position, token.length - position);
// Adjust down the length of this token
token.length = position;
// Insert the new token into each of the parent collections
marks.Insert(marks.IndexOf(token) + 1, tokenRhs);
tokens.Insert(tokens.IndexOf(token) + 1, tokenRhs);
// Return the new token
return tokenRhs;
}
// Resolve emphasis marks (part 2)
public void ResolveEmphasisMarks_classic(List<Token> tokens, List<Token> marks)
{
// First pass, do <strong>
for (int i = 0; i < marks.Count; i++)
{
// Get the next opening or public mark
Token opening_mark=marks[i];
if (opening_mark.type!=TokenType.opening_mark && opening_mark.type!=TokenType.internal_mark)
continue;
if (opening_mark.length < 2)
continue;
// Look for a matching closing mark
for (int j = i + 1; j < marks.Count; j++)
{
// Get the next closing or public mark
Token closing_mark = marks[j];
if (closing_mark.type != TokenType.closing_mark && closing_mark.type!=TokenType.internal_mark)
continue;
// Ignore if different type (ie: `*` vs `_`)
if (input[opening_mark.startOffset] != input[closing_mark.startOffset])
continue;
// Must be at least two
if (closing_mark.length < 2)
continue;
// Split the opening mark, keeping the LHS
if (opening_mark.length > 2)
{
SplitMarkToken(tokens, marks, opening_mark, 2);
}
// Split the closing mark, keeping the RHS
if (closing_mark.length > 2)
{
closing_mark=SplitMarkToken(tokens, marks, closing_mark, closing_mark.length-2);
}
// Connect them
opening_mark.type = TokenType.open_strong;
closing_mark.type = TokenType.close_strong;
// Continue after the closing mark
i = marks.IndexOf(closing_mark);
break;
}
}
// Second pass, do <em>
for (int i = 0; i < marks.Count; i++)
{
// Get the next opening or public mark
Token opening_mark = marks[i];
if (opening_mark.type != TokenType.opening_mark && opening_mark.type != TokenType.internal_mark)
continue;
// Look for a matching closing mark
for (int j = i + 1; j < marks.Count; j++)
{
// Get the next closing or public mark
Token closing_mark = marks[j];
if (closing_mark.type != TokenType.closing_mark && closing_mark.type != TokenType.internal_mark)
continue;
// Ignore if different type (ie: `*` vs `_`)
if (input[opening_mark.startOffset] != input[closing_mark.startOffset])
continue;
// Split the opening mark, keeping the LHS
if (opening_mark.length > 1)
{
SplitMarkToken(tokens, marks, opening_mark, 1);
}
// Split the closing mark, keeping the RHS
if (closing_mark.length > 1)
{
closing_mark = SplitMarkToken(tokens, marks, closing_mark, closing_mark.length - 1);
}
// Connect them
opening_mark.type = TokenType.open_em;
closing_mark.type = TokenType.close_em;
// Continue after the closing mark
i = marks.IndexOf(closing_mark);
break;
}
}
}
// Resolve emphasis marks (part 2)
public void ResolveEmphasisMarks(List<Token> tokens, List<Token> marks)
{
bool bContinue = true;
while (bContinue)
{
bContinue = false;
for (int i = 0; i < marks.Count; i++)
{
// Get the next opening or public mark
Token opening_mark = marks[i];
if (opening_mark.type != TokenType.opening_mark && opening_mark.type != TokenType.internal_mark)
continue;
// Look for a matching closing mark
for (int j = i + 1; j < marks.Count; j++)
{
// Get the next closing or public mark
Token closing_mark = marks[j];
if (closing_mark.type != TokenType.closing_mark && closing_mark.type != TokenType.internal_mark)
break;
// Ignore if different type (ie: `*` vs `_`)
if (input[opening_mark.startOffset] != input[closing_mark.startOffset])
continue;
// strong or em?
int style = Math.Min(opening_mark.length, closing_mark.length);
// Triple or more on both ends?
if (style >= 3)
{
style = (style % 2)==1 ? 1 : 2;
}
// Split the opening mark, keeping the RHS
if (opening_mark.length > style)
{
opening_mark = SplitMarkToken(tokens, marks, opening_mark, opening_mark.length - style);
i--;
}
// Split the closing mark, keeping the LHS
if (closing_mark.length > style)
{
SplitMarkToken(tokens, marks, closing_mark, style);
}
// Connect them
opening_mark.type = style == 1 ? TokenType.open_em : TokenType.open_strong;
closing_mark.type = style == 1 ? TokenType.close_em : TokenType.close_strong;
// Remove the matched marks
marks.Remove(opening_mark);
marks.Remove(closing_mark);
bContinue = true;
break;
}
}
}
}
// Scan the input string, creating tokens for anything special
public void Tokenize(string str, int start, int len)
{
// Prepare
base.Reset(str, start, len);
m_Tokens.Clear();
List<Token> emphasis_marks = null;
List<Abbreviation> Abbreviations=m_Markdown.GetAbbreviations();
bool ExtraMode = m_Markdown.ExtraMode;
// Scan string
int start_text_token = position;
while (!eof)
{
int end_text_token=position;
// Work out token
Token token = null;
switch (current)
{
case '*':
case '_':
// Create emphasis mark
token = CreateEmphasisMark();
if (token != null)
{
// Store marks in a separate list the we'll resolve later
switch (token.type)
{
case TokenType.internal_mark:
case TokenType.opening_mark:
case TokenType.closing_mark:
if (emphasis_marks == null)
{
emphasis_marks = new List<Token>();
}
emphasis_marks.Add(token);
break;
}
}
break;
case '`':
token = ProcessCodeSpan();
break;
case '[':
case '!':
{
// Process link reference
int linkpos = position;
token = ProcessLinkOrImageOrFootnote();
// Rewind if invalid syntax
// (the '[' or '!' will be treated as a regular character and processed below)
if (token == null)
position = linkpos;
break;
}
case '<':
{
// Is it a valid html tag?
int save = position;
HtmlTag tag = HtmlTag.Parse(this);
if (tag != null)
{
if (!m_Markdown.SafeMode || tag.IsSafe())
{
// Yes, create a token for it
token = CreateToken(TokenType.HtmlTag, save, position - save);
}
else
{
// No, rewrite and encode it
position = save;
}
}
else
{
// No, rewind and check if it's a valid autolink eg: <google.com>
position = save;
token = ProcessAutoLink();
if (token == null)
position = save;
}
break;
}
case '&':
{
// Is it a valid html entity
int save=position;
string unused=null;
if (SkipHtmlEntity(ref unused))
{
// Yes, create a token for it
token = CreateToken(TokenType.Html, save, position - save);
}
break;
}
case ' ':
{
// Check for double space at end of a line
if (CharAtOffset(1)==' ' && IsLineEnd(CharAtOffset(2)))
{
// Yes, skip it
SkipForward(2);
// Don't put br's at the end of a paragraph
if (!eof)
{
SkipEol();
token = CreateToken(TokenType.br, end_text_token, 0);
}
}
break;
}
case '\\':
{
// Special handling for escaping <autolinks>
/*
if (CharAtOffset(1) == '<')
{
// Is it an autolink?
int savepos = position;
SkipForward(1);
bool AutoLink = ProcessAutoLink() != null;
position = savepos;
if (AutoLink)
{
token = CreateToken(TokenType.Text, position + 1, 1);
SkipForward(2);
}
}
else
*/
{
// Check followed by an escapable character
if (Utils.IsEscapableChar(CharAtOffset(1), ExtraMode))
{
token = CreateToken(TokenType.Text, position + 1, 1);
SkipForward(2);
}
}
break;
}
}
// Look for abbreviations.
if (token == null && Abbreviations!=null && !Char.IsLetterOrDigit(CharAtOffset(-1)))
{
var savepos = position;
foreach (var abbr in Abbreviations)
{
if (SkipString(abbr.Abbr) && !Char.IsLetterOrDigit(current))
{
token = CreateToken(TokenType.abbreviation, abbr);
break;
}
position = savepos;
}
}
// If token found, append any preceeding text and the new token to the token list
if (token!=null)
{
// Create a token for everything up to the special character
if (end_text_token > start_text_token)
{
m_Tokens.Add(CreateToken(TokenType.Text, start_text_token, end_text_token-start_text_token));
}
// Add the new token
m_Tokens.Add(token);
// Remember where the next text token starts
start_text_token=position;
}
else
{
// Skip a single character and keep looking
SkipForward(1);
}
}
// Append a token for any trailing text after the last token.
if (position > start_text_token)
{
m_Tokens.Add(CreateToken(TokenType.Text, start_text_token, position-start_text_token));
}
// Do we need to resolve and emphasis marks?
if (emphasis_marks != null)
{
ResolveEmphasisMarks(m_Tokens, emphasis_marks);
}
// Done!
return;
}
// FreeToken - return a token to the spare token pool
public void FreeToken(Token token)
{
token.data = null;
m_SpareTokens.Push(token);
}
// FreeToken - return a token to the spare token pool
internal void FreeToken(Token token)
{
token.data = null;
m_SpareTokens.Push(token);
}