/**
* <summary>Moves the pointer to the next token.</summary>
* <remarks>To properly parse the current token, the pointer MUST be just before its starting
* (leading whitespaces are ignored). When this method terminates, the pointer IS
* at the last byte of the current token.</remarks>
* <returns>Whether a new token was found.</returns>
*/
public virtual bool MoveNext(
)
{
StringBuilder buffer = null;
token = null;
int c = 0;
// Skip leading white-space characters.
do
{
c = stream.ReadByte();
if (c == -1)
{
return(false);
}
} while(IsWhitespace(c)); // Keep goin' till there's a white-space character...
// Which character is it?
switch (c)
{
case Symbol.Slash: // Name.
{
tokenType = TokenTypeEnum.Name;
/*
* NOTE: As name objects are simple symbols uniquely defined by sequences of characters,
* the bytes making up the name are never treated as text, so here they are just
* passed through without unescaping.
*/
buffer = new StringBuilder();
while (true)
{
c = stream.ReadByte();
if (c == -1)
{
break; // NOOP.
}
if (IsDelimiter(c) || IsWhitespace(c))
{
break;
}
buffer.Append((char)c);
}
if (c > -1)
{
stream.Skip(-1);
} // Restores the first byte after the current token.
} break;
case '0':
case '1':
case '2':
case '3':
case '4':
case '5':
case '6':
case '7':
case '8':
case '9':
case '.':
case '-':
case '+': // Number.
{
if (c == '.')
{
tokenType = TokenTypeEnum.Real;
}
else // Digit or signum.
{
tokenType = TokenTypeEnum.Integer;
} // By default (it may be real).
// Building the number...
buffer = new StringBuilder();
while (true)
{
buffer.Append((char)c);
c = stream.ReadByte();
if (c == -1)
{
break; // NOOP.
}
else if (c == '.')
{
tokenType = TokenTypeEnum.Real;
}
else if (c < '0' || c > '9')
{
break;
}
}
if (c > -1)
{
stream.Skip(-1);
} // Restores the first byte after the current token.
} break;
case Symbol.OpenSquareBracket: // Array (begin).
tokenType = TokenTypeEnum.ArrayBegin;
break;
case Symbol.CloseSquareBracket: // Array (end).
tokenType = TokenTypeEnum.ArrayEnd;
break;
case Symbol.OpenAngleBracket: // Dictionary (begin) | Hexadecimal string.
{
c = stream.ReadByte();
if (c == -1)
{
throw new ParseException("Unexpected EOF (isolated opening angle-bracket character).");
}
// Is it a dictionary (2nd angle bracket)?
if (c == Symbol.OpenAngleBracket)
{
tokenType = TokenTypeEnum.DictionaryBegin;
break;
}
// Hexadecimal string (single angle bracket).
tokenType = TokenTypeEnum.Hex;
buffer = new StringBuilder();
while (c != Symbol.CloseAngleBracket) // NOT string end.
{
if (!IsWhitespace(c))
{
buffer.Append((char)c);
}
c = stream.ReadByte();
if (c == -1)
{
throw new ParseException("Unexpected EOF (malformed hex string).");
}
}
} break;
case Symbol.CloseAngleBracket: // Dictionary (end).
{
c = stream.ReadByte();
if (c != Symbol.CloseAngleBracket)
{
throw new ParseException("Malformed dictionary.", stream.Position);
}
tokenType = TokenTypeEnum.DictionaryEnd;
} break;
case Symbol.OpenRoundBracket: // Literal string.
{
tokenType = TokenTypeEnum.Literal;
buffer = new StringBuilder();
int level = 0;
while (true)
{
c = stream.ReadByte();
if (c == -1)
{
break;
}
else if (c == Symbol.OpenRoundBracket)
{
level++;
}
else if (c == Symbol.CloseRoundBracket)
{
level--;
}
else if (c == '\\')
{
bool lineBreak = false;
c = stream.ReadByte();
switch (c)
{
case 'n':
c = Symbol.LineFeed;
break;
case 'r':
c = Symbol.CarriageReturn;
break;
case 't':
c = '\t';
break;
case 'b':
c = '\b';
break;
case 'f':
c = '\f';
break;
case Symbol.OpenRoundBracket:
case Symbol.CloseRoundBracket:
case '\\':
break;
case Symbol.CarriageReturn:
lineBreak = true;
c = stream.ReadByte();
if (c != Symbol.LineFeed)
{
stream.Skip(-1);
}
break;
case Symbol.LineFeed:
lineBreak = true;
break;
default:
{
// Is it outside the octal encoding?
if (c < '0' || c > '7')
{
break;
}
// Octal.
int octal = c - '0';
c = stream.ReadByte();
// Octal end?
if (c < '0' || c > '7')
{
c = octal; stream.Skip(-1); break;
}
octal = (octal << 3) + c - '0';
c = stream.ReadByte();
// Octal end?
if (c < '0' || c > '7')
{
c = octal; stream.Skip(-1); break;
}
octal = (octal << 3) + c - '0';
c = octal & 0xff;
break;
}
}
if (lineBreak)
{
continue;
}
if (c == -1)
{
break;
}
}
else if (c == Symbol.CarriageReturn)
{
c = stream.ReadByte();
if (c == -1)
{
break;
}
else if (c != Symbol.LineFeed)
{
c = Symbol.LineFeed; stream.Skip(-1);
}
}
if (level == -1)
{
break;
}
buffer.Append((char)c);
}
if (c == -1)
{
throw new ParseException("Malformed literal string.");
}
} break;
case Symbol.Percent: // Comment.
{
tokenType = TokenTypeEnum.Comment;
buffer = new StringBuilder();
while (true)
{
c = stream.ReadByte();
if (c == -1 ||
IsEOL(c))
{
break;
}
buffer.Append((char)c);
}
} break;
default: // Keyword.
{
tokenType = TokenTypeEnum.Keyword;
buffer = new StringBuilder();
do
{
buffer.Append((char)c);
c = stream.ReadByte();
if (c == -1)
{
break;
}
} while(!IsDelimiter(c) && !IsWhitespace(c));
if (c > -1)
{
stream.Skip(-1);
} // Restores the first byte after the current token.
} break;
}
if (buffer != null)
{
switch (tokenType)
{
case TokenTypeEnum.Keyword:
{
token = buffer.ToString();
switch ((string)token)
{
case Keyword.False:
case Keyword.True: // Boolean.
tokenType = TokenTypeEnum.Boolean;
token = bool.Parse((string)token);
break;
case Keyword.Null: // Null.
tokenType = TokenTypeEnum.Null;
token = null;
break;
}
} break;
case TokenTypeEnum.Name:
case TokenTypeEnum.Literal:
case TokenTypeEnum.Hex:
case TokenTypeEnum.Comment:
token = buffer.ToString();
break;
case TokenTypeEnum.Integer:
token = Int32.Parse(
buffer.ToString(),
NumberStyles.Integer,
StandardNumberFormatInfo
);
break;
case TokenTypeEnum.Real:
token = Double.Parse(
buffer.ToString(),
NumberStyles.Float,
StandardNumberFormatInfo
);
break;
}
}
return(true);
}
/**
* <summary>Parses the next token [PDF:1.6:3.1].</summary>
* <remarks>To properly parse the current token, the pointer MUST be just before its starting
* (leading whitespaces are ignored). When this method terminates, the pointer IS
* at the last byte of the current token.</remarks>
* <returns>Whether a new token was found.</returns>
*/
public bool MoveNext(
)
{
/*
* NOTE: It'd be interesting to evaluate an alternative regular-expression-based
* implementation...
*/
StringBuilder buffer = null;
token = null;
int c = 0;
// Skip white-space characters [PDF:1.6:3.1.1].
do
{
c = stream.ReadByte();
if (c == -1)
{
return(false);
}
} while(IsWhitespace(c)); // Keep goin' till there's a white-space character...
// Which character is it?
switch (c)
{
case Symbol.Slash: // Name.
tokenType = TokenTypeEnum.Name;
buffer = new StringBuilder();
while (true)
{
c = stream.ReadByte();
if (c == -1)
{
throw new FileFormatException("Unexpected EOF (malformed name object).", stream.Position);
}
if (IsDelimiter(c) || IsWhitespace(c))
{
break;
}
buffer.Append((char)c);
}
stream.Skip(-1); // Recover the first byte after the current token.
break;
case '0':
case '1':
case '2':
case '3':
case '4':
case '5':
case '6':
case '7':
case '8':
case '9':
case '.':
case '-':
case '+': // Number [PDF:1.6:3.2.2] | Indirect reference.
switch (c)
{
case '.': // Decimal point.
tokenType = TokenTypeEnum.Real;
break;
case '-':
case '+': // Signum.
tokenType = TokenTypeEnum.Integer; // By default (it may be real).
break;
default: // Digit.
if (multipleTokenParsing) // Plain number (multiple token parsing -- see indirect reference search).
{
tokenType = TokenTypeEnum.Integer; // By default (it may be real).
}
else // Maybe an indirect reference (postfix notation [PDF:1.6:3.2.9]).
{
/*
* NOTE: We need to identify this pattern:
* ref := { int int 'R' }
*/
// Enable multiple token parsing!
// NOTE: This state MUST be disabled before returning.
multipleTokenParsing = true;
// 1. Object number.
// Try the possible object number!
stream.Skip(-1); MoveNext();
// Isn't it a valid object number?
if (tokenType != TokenTypeEnum.Integer)
{
// Disable multiple token parsing!
multipleTokenParsing = false;
return(true);
}
// Assign object number!
int objectNumber = (int)token;
// Backup the recovery position!
long oldOffset = stream.Position;
// 2. Generation number.
// Try the possible generation number!
MoveNext();
// Isn't it a valid generation number?
if (tokenType != TokenTypeEnum.Integer)
{
// Rollback!
stream.Seek(oldOffset);
token = objectNumber; tokenType = TokenTypeEnum.Integer;
// Disable multiple token parsing!
multipleTokenParsing = false;
return(true);
}
// Assign generation number!
int generationNumber = (int)token;
// 3. Reference keyword.
// Try the possible reference keyword!
MoveNext();
// Isn't it a valid reference keyword?
if (tokenType != TokenTypeEnum.Reference)
{
// Rollback!
stream.Seek(oldOffset);
token = objectNumber; tokenType = TokenTypeEnum.Integer;
// Disable multiple token parsing!
multipleTokenParsing = false;
return(true);
}
token = new Reference(objectNumber, generationNumber);
// Disable multiple token parsing!
multipleTokenParsing = false;
return(true);
}
break;
}
// Building the number...
buffer = new StringBuilder();
do
{
buffer.Append((char)c);
c = stream.ReadByte();
if (c == -1)
{
throw new FileFormatException("Unexpected EOF (malformed number object).", stream.Position);
}
if (c == '.')
{
tokenType = TokenTypeEnum.Real;
}
else if (c < '0' || c > '9')
{
break;
}
} while(true);
stream.Skip(-1); // Recover the first byte after the current token.
break;
case Symbol.OpenSquareBracket: // Array (begin).
tokenType = TokenTypeEnum.ArrayBegin;
break;
case Symbol.CloseSquareBracket: // Array (end).
tokenType = TokenTypeEnum.ArrayEnd;
break;
case Symbol.OpenAngleBracket: // Dictionary (begin) | Hexadecimal string.
c = stream.ReadByte();
if (c == -1)
{
throw new FileFormatException("Unexpected EOF (isolated opening angle-bracket character).", stream.Position);
}
// Is it a dictionary (2nd angle bracket [PDF:1.6:3.2.6])?
if (c == Symbol.OpenAngleBracket)
{
tokenType = TokenTypeEnum.DictionaryBegin;
break;
}
// Hexadecimal string (single angle bracket [PDF:1.6:3.2.3]).
tokenType = TokenTypeEnum.Hex;
// [FIX:0.0.4:4] It skipped after the first hexadecimal character, missing it.
buffer = new StringBuilder();
while (c != Symbol.CloseAngleBracket) // NOT string end.
{
buffer.Append((char)c);
c = stream.ReadByte();
if (c == -1)
{
throw new FileFormatException("Unexpected EOF (malformed hex string).", stream.Position);
}
}
break;
case Symbol.CloseAngleBracket: // Dictionary (end).
c = stream.ReadByte();
if (c != Symbol.CloseAngleBracket)
{
throw new FileFormatException("Malformed dictionary.", stream.Position);
}
tokenType = TokenTypeEnum.DictionaryEnd;
break;
case Symbol.OpenRoundBracket: // Literal string.
tokenType = TokenTypeEnum.Literal;
buffer = new StringBuilder();
int level = 0;
while (true)
{
c = stream.ReadByte();
if (c == -1)
{
break;
}
if (c == Symbol.OpenRoundBracket)
{
level++;
}
else if (c == Symbol.CloseRoundBracket)
{
level--;
}
else if (c == '\\')
{
bool lineBreak = false;
c = stream.ReadByte();
switch (c)
{
case 'n':
c = Symbol.LineFeed;
break;
case 'r':
c = Symbol.CarriageReturn;
break;
case 't':
c = '\t';
break;
case 'b':
c = '\b';
break;
case 'f':
c = '\f';
break;
case Symbol.OpenRoundBracket:
case Symbol.CloseRoundBracket:
case '\\':
break;
case Symbol.CarriageReturn:
lineBreak = true;
c = stream.ReadByte();
if (c != Symbol.LineFeed)
{
stream.Skip(-1);
}
break;
case Symbol.LineFeed:
lineBreak = true;
break;
default:
{
// Is it outside the octal encoding?
if (c < '0' || c > '7')
{
break;
}
// Octal [PDF:1.6:3.2.3].
int octal = c - '0';
c = stream.ReadByte();
// Octal end?
if (c < '0' || c > '7')
{
c = octal; stream.Skip(-1); break;
}
octal = (octal << 3) + c - '0';
c = stream.ReadByte();
// Octal end?
if (c < '0' || c > '7')
{
c = octal; stream.Skip(-1); break;
}
octal = (octal << 3) + c - '0';
c = octal & 0xff;
break;
}
}
if (lineBreak)
{
continue;
}
if (c == -1)
{
break;
}
}
else if (c == Symbol.CarriageReturn)
{
c = stream.ReadByte();
if (c == -1)
{
break;
}
if (c != Symbol.LineFeed)
{
c = Symbol.LineFeed; stream.Skip(-1);
}
}
if (level == -1)
{
break;
}
buffer.Append((char)c);
}
if (c == -1)
{
throw new FileFormatException("Malformed literal string.", stream.Position);
}
break;
case Symbol.CapitalR: // Indirect reference.
tokenType = TokenTypeEnum.Reference;
break;
case Symbol.Percent: // Comment [PDF:1.6:3.1.2].
tokenType = TokenTypeEnum.Comment;
buffer = new StringBuilder();
while (true)
{
c = stream.ReadByte();
if (c == -1 ||
IsEOL(c))
{
break;
}
buffer.Append((char)c);
}
break;
default: // Keyword object.
tokenType = TokenTypeEnum.Keyword;
buffer = new StringBuilder();
do
{
buffer.Append((char)c);
c = stream.ReadByte();
if (c == -1)
{
break;
}
} while(!IsDelimiter(c) && !IsWhitespace(c));
stream.Skip(-1); // Recover the first byte after the current token.
break;
}
if (buffer != null)
{
/*
* Current token initialization.
*/
// Wich token type?
switch (tokenType)
{
case TokenTypeEnum.Keyword:
token = buffer.ToString();
// Late recognition.
switch ((string)token)
{
case Keyword.False:
case Keyword.True: // Boolean.
tokenType = TokenTypeEnum.Boolean;
token = bool.Parse((string)token);
break;
case Keyword.Null: // Null.
tokenType = TokenTypeEnum.Null;
token = null;
break;
}
break;
case TokenTypeEnum.Comment:
case TokenTypeEnum.Hex:
case TokenTypeEnum.Name:
token = buffer.ToString();
break;
case TokenTypeEnum.Literal:
token = buffer.ToString();
// Late recognition.
if (((string)token).StartsWith(Keyword.DatePrefix)) // Date.
{
tokenType = TokenTypeEnum.Date;
token = PdfDate.ToDate((string)token);
}
break;
case TokenTypeEnum.Integer:
token = Int32.Parse(
buffer.ToString(),
NumberStyles.Integer,
StandardNumberFormatInfo
);
break;
case TokenTypeEnum.Real:
// [FIX:1668410] Parsing of float numbers was buggy (localized default number format).
token = Single.Parse(
buffer.ToString(),
NumberStyles.Float,
StandardNumberFormatInfo
);
break;
}
}
return(true);
}
