private bool HandleUnits()
{
// Makes a token for the units identifier after a number
int start = CS.Position;
if (CS.CurrentChar == '%')
{
AddToken(CssTokenType.Units, start, 1);
CS.Advance(1);
return(true);
}
if (SkipIdentifier())
{
UnitType unitType = UnitHelpers.GetUnitType(CS.TextProvider, start, CS.Position - start);
if (unitType != UnitType.Unknown)
{
// only known units after whitespace (CSS2)
AddToken(CssTokenType.Units, start, CS.Position - start);
return(true);
}
}
// Not a valid unit identifier, go back
CS.Position = start;
return(false);
}
protected bool SkipName()
{
int start = CS.Position;
while (AtName())
{
if (AtEscape())
{
if (AtSlashNineHack())
{
// Stop the name token at the start of the \9, which allows that CSS workaround to work:
// color: red\9;
// border-width: 50px\9;
break;
}
else
{
SkipEscape();
}
}
else
{
CS.Advance(1);
}
}
// In case the last escaped char had a space after it:
SkipWhitespaceReverse();
return(start != CS.Position);
}
/// <summary>
/// If the user puts \9 after an identifier, then we're going to treat that
/// \9 as a comment. That workaround is used to make different values work in different browsers.
/// Old IE browsers probably ignore the \9 even though it really should be part of the identifier.
/// </summary>
private bool HandleSlashNineHack()
{
if (AtSlashNineHack())
{
int start = CS.Position;
CS.Advance(2);
AddToken(CssTokenType.CommentText, start, CS.Position - start);
return(true);
}
return(false);
}
protected bool SkipUri()
{
// The CSS spec uses this notation for URIs:
// ([!#$%&*-~]|{nonascii}|{escape})*
// http://www.w3.org/TR/CSS21/syndata.html
int start = CS.Position;
while (!CS.IsAtEnd)
{
char ch = CS.CurrentChar;
if (AtComment() || AtScopeBlocker())
{
// Found the end
break;
}
else if (AtEscape())
{
SkipEscape();
}
else if ((ch >= '*' && ch <= '~') ||
ch == '!' ||
ch == '#' ||
ch == '$' ||
ch == '%' ||
ch == '&' ||
IsNonAscii(ch))
{
CS.Advance(1);
}
else
{
// Found the end
break;
}
}
if (CS.Position > start)
{
// In case the last escaped char had a space after it:
SkipWhitespaceReverse();
}
return(start != CS.Position);
}
protected bool SkipUnicodeRange()
{
// Skip U+012???-012FFF
// The rules for a Unicode range in the lexer are more relaxed than in the parser.
// The lexer allows question marks anywhere within the first part of the range,
// while the parser would consider "U+??99" to be invalid.
//
// Lexer: u\+[0-9a-f?]{1,6}(-[0-9a-f]{1,6})?
// Parser: http://www.w3.org/TR/css3-webfonts/#character-range-the-unicode-range-descri
int start = CS.Position;
if (AtUnicodeRange())
{
CS.Advance(3); // Skip "U+X"
for (int count = 1; count < 6; count++, CS.Advance(1))
{
if (!IsHexDigit(CS.CurrentChar) && CS.CurrentChar != '?')
{
break;
}
}
if (CS.CurrentChar == '-' && IsHexDigit(CS.Peek(1)))
{
CS.Advance(2); // Skip "-X"
for (int count = 1; count < 6; count++, CS.Advance(1))
{
if (!IsHexDigit(CS.CurrentChar))
{
break;
}
}
}
}
return(start != CS.Position);
}
protected bool SkipIdentifier()
{
int start = CS.Position;
// Identifiers can start with one dash or two
if (CS.CurrentChar == '-')
{
CS.Advance(1);
if (CS.CurrentChar == '-')
{
CS.Advance(1);
}
}
if (!AtNameStart())
{
// Not a valid name
CS.Position = start;
return(false);
}
else
{
if (AtEscape())
{
SkipEscape();
}
else
{
CS.Advance(1);
}
// The rest is just name chars
SkipName();
return(true);
}
}
private bool HandleHash()
{
// Makes a token for: #foo
int start = CS.Position;
if (CS.CurrentChar == '#')
{
CS.Advance(1);
if (SkipName())
{
AddToken(CssTokenType.HashName, start, CS.Position - start);
}
else
{
// There is no name after the '#'
CS.Position = start;
}
}
return(start != CS.Position);
}
/// <summary>
/// Skips: "progid:Foo.Bar(foo=bar)"
/// </summary>
protected bool SkipProgramIdFunction()
{
int start = CS.Position;
if ((CS.CurrentChar == 'p' || CS.CurrentChar == 'P') &&
TextHelpers.CompareCurrentDecodedString(CS, "progid:", /* ignoreCase = */ true, out _))
{
CS.Advance(7);
while (true)
{
if (SkipIdentifier())
{
// part of the function name
}
else if (CS.CurrentChar == '.')
{
// still part of the function name
CS.Advance(1);
}
else if (CS.CurrentChar == '(')
{
// found the end of the function name
CS.Advance(1);
return(true);
}
else
{
// bad function name
CS.Position = start;
break;
}
}
}
return(start != CS.Position);
}
protected bool SkipNumber()
{
// Skips: 123.456
int start = CS.Position;
bool seenDot = false;
bool seenDigit = false;
while (true)
{
if (!char.IsDigit(CS.CurrentChar))
{
if (!seenDot && CS.CurrentChar == '.' && char.IsDigit(CS.Peek(1)))
{
seenDot = true;
}
else
{
break;
}
}
else
{
seenDigit = true;
}
CS.Advance(1);
}
if (!seenDigit)
{
CS.Position = start;
}
return(start != CS.Position);
}
private bool HandleDimension()
{
// Makes tokens for: 100, 100px, +0.7em, -20vw
int start = CS.Position;
if (CS.CurrentChar == '+' || CS.CurrentChar == '-')
{
CS.Advance(1);
}
if (!SkipNumber())
{
CS.Position = start;
return(false);
}
AddToken(CssTokenType.Number, start, CS.Position - start);
// Deal with the units after the number (don't care if it fails)
HandleUnits();
return(true);
}
private bool HandleComment()
{
const CssTokenType startTokenType = CssTokenType.OpenCComment;
const CssTokenType endTokenType = CssTokenType.CloseCComment;
const string commentStart = "/*";
const string commentEnd = "*/";
if (AtScopeBlocker())
{
AddToken(CssTokenType.ScopeBlocker, CS.Position, ScopeBlockerText.Length);
CS.Advance(ScopeBlockerText.Length);
return(true);
}
// Comment: /* foo */
// Makes three tokens for a comment (start, text, and end)
if (!CS.TextProvider.CompareTo(CS.Position, commentStart, ignoreCase: false))
{
return(false);
}
// Skip the start of the comment
AddToken(startTokenType, CS.Position, commentStart.Length);
CS.Advance(commentStart.Length);
// Skip the inner text of the comment
bool endFound = false;
if (!CS.IsAtEnd)
{
SkipWhitespace();
int start = CS.Position;
for (; !CS.IsAtEnd; CS.Advance(1))
{
if (CS.TextProvider.CompareTo(CS.Position, commentEnd, ignoreCase: false))
{
endFound = true;
break;
}
if (AtScopeBlocker())
{
// Catch: /* Foo /* END EXTERNAL SOURCE */
break;
}
}
if (CS.Position >= start)
{
// Create a token for the comment text, don't include trailing whitespace
SkipWhitespaceReverse();
if (CS.Position > start)
{
AddToken(CssTokenType.CommentText, start, CS.Position - start);
}
SkipWhitespace();
}
}
// Skip the end of the comment
if (endFound)
{
AddToken(endTokenType, CS.Position, commentEnd.Length);
CS.Advance(commentEnd.Length);
}
return(true);
}
protected virtual bool HandleToken()
{
CssTokenType tokenType = CssTokenType.Unknown;
int tokenStart = CS.Position;
switch (CS.CurrentChar)
{
case '*':
if (CS.Peek(1) == '=')
{
tokenType = CssTokenType.ContainsString;
CS.Advance(1);
}
else
{
tokenType = CssTokenType.Asterisk;
}
break;
case '&':
tokenType = CssTokenType.Ampersand;
break;
case '.':
tokenType = CssTokenType.Dot;
break;
case '!':
tokenType = CssTokenType.Bang;
break;
case ',':
tokenType = CssTokenType.Comma;
break;
case '^':
if (CS.Peek(1) == '=')
{
tokenType = CssTokenType.BeginsWith;
CS.Advance(1);
}
else
{
tokenType = CssTokenType.Caret;
}
break;
case ':':
if (CS.Peek(1) == ':')
{
tokenType = CssTokenType.DoubleColon;
CS.Advance(1);
}
else
{
tokenType = CssTokenType.Colon;
}
break;
case '$':
if (CS.Peek(1) == '=')
{
tokenType = CssTokenType.EndsWith;
CS.Advance(1);
}
else
{
tokenType = CssTokenType.Dollar;
}
break;
case '=':
tokenType = CssTokenType.Equals;
break;
case '>':
tokenType = CssTokenType.Greater;
break;
case '|':
if (CS.Peek(1) == '=')
{
tokenType = CssTokenType.ListBeginsWith;
CS.Advance(1);
}
else if (CS.Peek(1) == '|')
{
tokenType = CssTokenType.DoublePipe;
CS.Advance(1);
}
else
{
tokenType = CssTokenType.Or;
}
break;
case '%':
tokenType = CssTokenType.Percent;
break;
case ';':
tokenType = CssTokenType.Semicolon;
break;
case '/':
if (HandleComment())
{
return(true);
}
else
{
tokenType = CssTokenType.Slash;
}
break;
case '\\':
if (HandleSlashNineHack())
{
return(true);
}
break;
case '<':
if (CS.TextProvider.CompareTo(CS.Position, "<!--", ignoreCase: false))
{
CS.Advance(4);
AddToken(CssTokenType.OpenHtmlComment, tokenStart, CS.Position - tokenStart);
return(true);
}
break;
case '~':
if (CS.Peek(1) == '=')
{
tokenType = CssTokenType.OneOf;
CS.Advance(1);
}
else
{
tokenType = CssTokenType.Tilde;
}
break;
case '(':
tokenType = CssTokenType.OpenFunctionBrace;
break;
case ')':
tokenType = CssTokenType.CloseFunctionBrace;
break;
case '[':
tokenType = CssTokenType.OpenSquareBracket;
break;
case ']':
tokenType = CssTokenType.CloseSquareBracket;
break;
case '{':
tokenType = CssTokenType.OpenCurlyBrace;
break;
case '}':
tokenType = CssTokenType.CloseCurlyBrace;
break;
case '@':
tokenType = CssTokenType.At;
break;
case '\'':
case '\"':
if (HandleString() != CssTokenType.Unknown)
{
return(true);
}
break;
case '#':
if (HandleHash())
{
return(true);
}
else
{
tokenType = CssTokenType.Hash;
}
break;
case '+':
tokenType = CssTokenType.Plus;
break;
case '-':
if (HandleIdentifier())
{
return(true);
}
else if (CS.TextProvider.CompareTo(CS.Position, "-->", ignoreCase: false))
{
CS.Advance(3);
AddToken(CssTokenType.CloseHtmlComment, tokenStart, CS.Position - tokenStart);
return(true);
}
else
{
tokenType = CssTokenType.Minus;
}
break;
case 'u':
case 'U':
if (HandleUnicodeRange())
{
return(true);
}
break;
}
if (tokenType == CssTokenType.Unknown &&
HandleIdentifier())
{
return(true);
}
if (!HandleUnknown())
{
// Some kind of junk in the CSS, just deal with it by making an unknown token
CS.Advance(1);
AddToken(tokenType, tokenStart, CS.Position - tokenStart);
}
return(true);
}
protected CssTokenType SkipString()
{
// This can detect single and double quoted strings, spanning a single or multiple lines.
// Also, unterminated strings are detected. The return value tells you the type:
//
// CssTokenType.Unknown (not a string at all)
// CssTokenType.MultilineString
// CssTokenType.String
// CssTokenType.InvalidString (unterminated)
CssTokenType tokenType = CssTokenType.Unknown;
if (TextHelper.IsQuote(CS.CurrentChar))
{
// Guilty until proven valid
tokenType = CssTokenType.InvalidString;
char quote = CS.CurrentChar;
bool multiLine = false;
CS.Advance(1);
while (!CS.IsAtEnd)
{
if (AtScopeBlocker())
{
break;
}
else if (CS.CurrentChar == quote)
{
// Found matching end quote
tokenType = multiLine ? CssTokenType.MultilineString : CssTokenType.String;
CS.Advance(1);
break;
}
else if (TextHelper.IsNewLine(CS.CurrentChar))
{
break;
}
else if (AtEscape())
{
SkipEscape();
}
else if (CS.CurrentChar == '\\')
{
// must be an escaped line break
CS.Advance(1);
Debug.Assert(TextHelper.IsNewLine(CS.CurrentChar));
SkipNewLine();
multiLine = true;
}
else
{
CS.Advance(1);
}
}
if (tokenType == CssTokenType.InvalidString)
{
SkipWhitespaceReverse();
}
}
return(tokenType);
}
