public TokenObject GetToken()
{
TokenObject t = null;
if (_stack.Count > 0)
{
t = _stack.Pop();
}
else
{
t = GetAnyToken();
}
if (IgnoreComments)
{
while (t is TokenComment)
{
if (_stack.Count > 0)
{
t = _stack.Pop();
}
else
{
t = GetAnyToken();
}
}
}
return(t);
}
private void ThrowOnError(TokenObject t)
{
if (t is TokenError)
{
throw new ApplicationException(t.ToString());
}
}
private void ThrowOnEmptyOrError(TokenObject t)
{
if (t is TokenError)
{
throw new ApplicationException(t.ToString());
}
else if (t is TokenEmpty)
{
throw new ApplicationException("Unexpected end of PDF document.");
}
}
private T ThrowIfNot <T>(TokenObject t) where T : TokenObject
{
if (t is TokenError)
{
throw new ApplicationException(t.ToString());
}
else if (t is TokenEmpty)
{
throw new ApplicationException("Unexpected end of PDF document.");
}
else if (!(t is T))
{
throw new ApplicationException($"Found {t.GetType().Name} instead of {typeof(T).Name}.");
}
return((T)t);
}
public List <TokenXRefEntry> ParseXRef()
{
Tokenizer.IgnoreComments = true;
TokenObject t = Tokenizer.GetToken();
TokenKeyword keyword = t as TokenKeyword;
if ((keyword == null) || (keyword.Value != ParseKeyword.XRef))
{
// Scan entire source creating XRef entries for each indirect object
return(IndirectObjectsToXRef());
}
else
{
List <TokenXRefEntry> entries = new List <TokenXRefEntry>();
ParseXRefSections(entries);
return(entries);
}
}
public List <TokenXRefEntry> IndirectObjectsToXRef()
{
List <TokenXRefEntry> entries = new List <TokenXRefEntry>();
// Start scanning from beginning of the source
Tokenizer.IgnoreComments = false;
Tokenizer.Position = 0;
long lastTrailer = -1;
do
{
long t1Position = Tokenizer.Position;
TokenObject t1 = Tokenizer.GetToken();
if (t1 is TokenInteger)
{
TokenInteger t2 = Tokenizer.GetToken() as TokenInteger;
if (t2 != null)
{
TokenKeyword t3 = Tokenizer.GetToken() as TokenKeyword;
if ((t3 != null) && (t3.Value == ParseKeyword.Obj))
{
TokenInteger id = (TokenInteger)t1;
entries.Add(new TokenXRefEntry(id.Value, t2.Value, t1Position, true));
}
}
}
else if ((t1 is TokenKeyword) && ((TokenKeyword)t1).Value == ParseKeyword.Trailer)
{
lastTrailer = t1Position;
}
} while (Tokenizer.GotoNextLine());
// Leave with the position on the last 'trailer' as caller will then parse it
if (lastTrailer >= 0)
{
Tokenizer.Position = lastTrailer;
}
return(entries);
}
public void ParseXRefSections(List <TokenXRefEntry> entries)
{
while (true)
{
TokenObject t = Tokenizer.GetToken();
ThrowOnError(t);
// Cross-reference table ends when we find a 'trailer' keyword instead of another section
if ((t is TokenKeyword) && (((TokenKeyword)t).Value == ParseKeyword.Trailer))
{
Tokenizer.PushToken(t);
return;
}
// Section starts with an integer object number
TokenInteger start = t as TokenInteger;
if (start == null)
{
throw new ApplicationException($"Cross-reference section number must be an integer.");
}
t = Tokenizer.GetToken();
ThrowOnError(t);
// Section then has an integer length number
TokenInteger length = t as TokenInteger;
if (length == null)
{
throw new ApplicationException($"Cross-reference section length must be an integer.");
}
// Load each line in the section
for (int i = 0, id = start.Value; i < length.Value; i++, id++)
{
TokenObject entry = Tokenizer.GetXRefEntry(id);
ThrowOnError(entry);
entries.Add((TokenXRefEntry)entry);
}
}
}
public ParseDictionary ParseTrailer()
{
Tokenizer.IgnoreComments = true;
TokenObject t = Tokenizer.GetToken();
ThrowOnError(t);
// Cross-reference table ends when we find a 'trailer' keyword instead of another section
if (!(t is TokenKeyword) || (((TokenKeyword)t).Value != ParseKeyword.Trailer))
{
throw new ApplicationException($"Trailer section must start with the 'trailer' keyword.");
}
ParseObjectBase obj = ParseObject();
if ((obj == null) || !(obj is ParseDictionary))
{
throw new ApplicationException($"Trailer section must contain a dictionary.");
}
return((ParseDictionary)obj);
}
public void ParseHeader(out int major, out int minor)
{
// The header is a comment token
Tokenizer.IgnoreComments = false;
TokenObject t = Tokenizer.GetToken();
if (!(t is TokenComment))
{
throw new ApplicationException("Missing PDF header.");
}
TokenComment c = (TokenComment)t;
if (!c.Value.StartsWith("%PDF"))
{
throw new ApplicationException("PDF Header must start with '%PDF'.");
}
string[] splits = c.Value.Substring(5).Split('.');
if (splits.Length != 2)
{
throw new ApplicationException("PDF Header must have a <major>.<minor> version number.");
}
if (!int.TryParse(splits[0].Trim(), out major))
{
throw new ApplicationException("Could not parse the header major version number.");
}
if (!int.TryParse(splits[1].Trim(), out minor))
{
throw new ApplicationException("Could not parse the header minor version number.");
}
Tokenizer.IgnoreComments = true;
}
public ParseObjectBase ParseObject(bool allowEmpty = false)
{
Tokenizer.IgnoreComments = true;
TokenObject t = Tokenizer.GetToken();
if (allowEmpty && (t is TokenEmpty))
{
return(null);
}
else
{
ThrowOnEmptyOrError(t);
}
if (t is TokenName)
{
// Store one instance of each unique name to minimize memory footprint
TokenName tokenName = (TokenName)t;
return(ParseName.GetParse(tokenName.Value));
}
else if (t is TokenInteger)
{
TokenObject t2 = Tokenizer.GetToken();
ThrowOnError(t2);
// An object reference has a second integer, the generation number
if (t2 is TokenInteger)
{
TokenObject t3 = Tokenizer.GetToken();
ThrowOnError(t3);
// An object reference has a third value which is the 'R' keyword
if ((t3 is TokenKeyword) && (((TokenKeyword)t3).Value == ParseKeyword.R))
{
return(new ParseObjectReference(t as TokenInteger, t2 as TokenInteger));
}
Tokenizer.PushToken(t3);
}
Tokenizer.PushToken(t2);
return(new ParseInteger(t as TokenInteger));
}
else if (t is TokenReal)
{
return(new ParseReal(t as TokenReal));
}
else if (t is TokenStringHex)
{
return(new ParseString(t as TokenStringHex));
}
else if (t is TokenStringLiteral)
{
return(new ParseString(t as TokenStringLiteral));
}
else if (t is TokenArrayOpen)
{
List <ParseObjectBase> objects = new List <ParseObjectBase>();
ParseObjectBase entry = null;
while (true)
{
entry = ParseObject();
if (entry == null)
{
break;
}
else
{
ThrowOnEmptyOrError(t);
}
objects.Add(entry);
}
ThrowIfNot <TokenArrayClose>(Tokenizer.GetToken());
return(new ParseArray(objects));
}
else if (t is TokenDictionaryOpen)
{
List <string> names = new List <string>();
List <ParseObjectBase> entries = new List <ParseObjectBase>();
ParseObjectBase value1 = null;
ParseObjectBase value2 = null;
while (true)
{
value1 = ParseObject();
if (value1 == null)
{
break;
}
else
{
ThrowOnEmptyOrError(t);
}
// Key value must be a Name
ParseName name = value1 as ParseName;
if (name == null)
{
throw new ApplicationException($"Dictionary key must be a name instead of {name.GetType().Name}.");
}
value2 = ParseObject();
if (value2 == null)
{
throw new ApplicationException($"Dictionary value missing for key {name.Value}.");
}
else
{
ThrowOnEmptyOrError(t);
}
names.Add(name.Value);
entries.Add(value2);
}
ThrowIfNot <TokenDictionaryClose>(Tokenizer.GetToken());
return(new ParseDictionary(names, entries));
}
else if (t is TokenKeyword)
{
switch ((t as TokenKeyword).Value)
{
case ParseKeyword.True:
return(ParseObjectBase.True);
case ParseKeyword.False:
return(ParseObjectBase.False);
case ParseKeyword.Null:
return(ParseObjectBase.Null);
}
}
else if (t is TokenIdentifier)
{
// Store one instance of each unique identifier to minimize memory footprint
TokenIdentifier tokenIdentifier = (TokenIdentifier)t;
return(ParseIdentifier.GetParse(tokenIdentifier.Value));
}
// Token is not one that starts an object, so put the token back
Tokenizer.PushToken(t);
return(null);
}
public ParseIndirectObject ParseIndirectObject()
{
Tokenizer.IgnoreComments = true;
TokenObject t = Tokenizer.GetToken();
ThrowOnEmptyOrError(t);
// Indirect object starts with an integer, the object identifier
if (!(t is TokenInteger))
{
Tokenizer.PushToken(t);
return(null);
}
// Second is another integer, the generation number
TokenObject u = Tokenizer.GetToken();
ThrowOnEmptyOrError(u);
if (!(u is TokenInteger))
{
Tokenizer.PushToken(t);
Tokenizer.PushToken(u);
return(null);
}
// This is the keyword 'obj'
TokenObject v = Tokenizer.GetToken();
ThrowOnEmptyOrError(v);
if (!(v is TokenKeyword) || ((v as TokenKeyword).Value != ParseKeyword.Obj))
{
Tokenizer.PushToken(t);
Tokenizer.PushToken(u);
Tokenizer.PushToken(v);
return(null);
}
// Get actual object that is the content
ParseObjectBase obj = ParseObject();
if (obj == null)
{
throw new ApplicationException($"Indirect object has missing content.");
}
// Must be followed by either 'endobj' or 'stream'
v = Tokenizer.GetToken();
ThrowOnEmptyOrError(v);
TokenKeyword keyword = v as TokenKeyword;
if (keyword == null)
{
// PH: FIXME:
//throw new ApplicationException($"Indirect object has missing 'endobj or 'stream'.");
return(new ParseIndirectObject(t as TokenInteger, u as TokenInteger, new ParseInteger(0)));
}
if (keyword.Value == ParseKeyword.EndObj)
{
return(new ParseIndirectObject(t as TokenInteger, u as TokenInteger, obj));
}
else if (keyword.Value == ParseKeyword.Stream)
{
ParseDictionary dictionary = obj as ParseDictionary;
if (dictionary == null)
{
throw new ApplicationException($"Stream must be preceded by a dictionary.");
}
if (!dictionary.ContainsName("Length"))
{
throw new ApplicationException($"Stream dictionary must contain a 'Length' entry.");
}
ParseObjectBase lengthObj = dictionary["Length"];
// Resolve any object reference
ParseObjectReference reference = lengthObj as ParseObjectReference;
if (reference != null)
{
lengthObj = OnResolveReference(reference);
}
ParseInteger length = lengthObj as ParseInteger;
if (length == null)
{
throw new ApplicationException($"Stream dictionary has a 'Length' entry that is not an integer entry.");
}
if (length.Value < 0)
{
throw new ApplicationException($"Stream dictionary has a 'Length' less than 0.");
}
byte[] bytes = Tokenizer.GetBytes(length.Value);
if (bytes == null)
{
throw new ApplicationException($"Cannot read in expected {length.Value} bytes from stream.");
}
// Stream contents must be followed by 'endstream'
v = Tokenizer.GetToken();
ThrowOnEmptyOrError(v);
keyword = v as TokenKeyword;
if (keyword == null)
{
throw new ApplicationException($"Stream has missing 'endstream' after content.");
}
if (keyword.Value != ParseKeyword.EndStream)
{
throw new ApplicationException($"Stream has unexpected keyword {keyword.Value} instead of 'endstream'.");
}
// Stream contents must be followed by 'endobj'
v = Tokenizer.GetToken();
ThrowOnEmptyOrError(v);
keyword = v as TokenKeyword;
if (keyword == null)
{
throw new ApplicationException($"Indirect object has missing 'endobj'.");
}
if (keyword.Value != ParseKeyword.EndObj)
{
throw new ApplicationException($"Indirect object has unexpected keyword {keyword.Value} instead of 'endobj'.");
}
return(new ParseIndirectObject(t as TokenInteger, u as TokenInteger, new ParseStream(dictionary, bytes)));
}
else
{
throw new ApplicationException($"Indirect object has unexpected keyword {keyword.Value}.");
}
}
public void PushToken(TokenObject token)
{
_stack.Push(token);
}