/// <summary>
/// Parses an embedded Adobe Type 1 font file.
/// </summary>
/// <param name="inputBytes">The bytes of the font program.</param>
/// <param name="length1">The length in bytes of the clear text portion of the font program.</param>
/// <param name="length2">The length in bytes of the encrypted portion of the font program.</param>
/// <returns>The parsed type 1 font.</returns>
public Type1FontProgram Parse(IInputBytes inputBytes, int length1, int length2)
{
// Sometimes the entire PFB file including the header bytes can be included which prevents parsing in the normal way.
var isEntirePfbFile = inputBytes.Peek() == PfbFileIndicator;
IReadOnlyList <byte> eexecPortion = new byte[0];
if (isEntirePfbFile)
{
var(ascii, binary) = ReadPfbHeader(inputBytes);
eexecPortion = binary;
inputBytes = new ByteArrayInputBytes(ascii);
}
var scanner = new CoreTokenScanner(inputBytes);
if (!scanner.TryReadToken(out CommentToken comment) || !comment.Data.StartsWith("!"))
{
throw new InvalidFontFormatException("The Type1 program did not start with '%!'.");
}
string name;
var parts = comment.Data.Split(new[] { " " }, StringSplitOptions.RemoveEmptyEntries);
if (parts.Length == 3)
{
name = parts[1];
}
else
{
name = "Unknown";
}
var comments = new List <string>();
while (scanner.MoveNext() && scanner.CurrentToken is CommentToken commentToken)
{
comments.Add(commentToken.Data);
}
var dictionaries = new List <DictionaryToken>();
// Override arrays and names since type 1 handles these differently.
var arrayTokenizer = new Type1ArrayTokenizer();
var nameTokenizer = new Type1NameTokenizer();
scanner.RegisterCustomTokenizer((byte)'{', arrayTokenizer);
scanner.RegisterCustomTokenizer((byte)'/', nameTokenizer);
try
{
var tempEexecPortion = new List <byte>();
var tokenSet = new PreviousTokenSet();
tokenSet.Add(scanner.CurrentToken);
while (scanner.MoveNext())
{
if (scanner.CurrentToken is OperatorToken operatorToken)
{
if (Equals(scanner.CurrentToken, OperatorToken.Eexec))
{
int offset = 0;
while (inputBytes.MoveNext())
{
if (inputBytes.CurrentByte == (byte)ClearToMark[offset])
{
offset++;
}
else
{
if (offset > 0)
{
for (int i = 0; i < offset; i++)
{
tempEexecPortion.Add((byte)ClearToMark[i]);
}
}
offset = 0;
}
if (offset == ClearToMark.Length)
{
break;
}
if (offset > 0)
{
continue;
}
tempEexecPortion.Add(inputBytes.CurrentByte);
}
}
else
{
HandleOperator(operatorToken, scanner, tokenSet, dictionaries);
}
}
tokenSet.Add(scanner.CurrentToken);
}
if (!isEntirePfbFile)
{
eexecPortion = tempEexecPortion;
}
}
finally
{
scanner.DeregisterCustomTokenizer(arrayTokenizer);
scanner.DeregisterCustomTokenizer(nameTokenizer);
}
var encoding = GetEncoding(dictionaries);
var matrix = GetFontMatrix(dictionaries);
var boundingBox = GetBoundingBox(dictionaries);
var(privateDictionary, charStrings) = encryptedPortionParser.Parse(eexecPortion, false);
return(new Type1FontProgram(name, encoding, matrix, boundingBox ?? new PdfRectangle(), privateDictionary, charStrings));
}
private void HandleOperator(OperatorToken token, IInputBytes bytes, ISeekableTokenScanner scanner, PreviousTokenSet set, List <DictionaryToken> dictionaries)
{
switch (token.Data)
{
case "dict":
var number = ((NumericToken)set[0]).Int;
var dictionary = ReadDictionary(number, scanner);
dictionaries.Add(dictionary);
break;
case "currentfile":
if (!scanner.MoveNext() || scanner.CurrentToken != OperatorToken.Eexec)
{
return;
}
// For now we will not read this stuff.
SkipEncryptedContent(bytes);
break;
default:
return;
}
}
private static void HandleOperator(OperatorToken token, ISeekableTokenScanner scanner, PreviousTokenSet set, List <DictionaryToken> dictionaries)
{
switch (token.Data)
{
case "dict":
var number = ((NumericToken)set[0]).Int;
var dictionary = ReadDictionary(number, scanner);
dictionaries.Add(dictionary);
break;
default:
return;
}
}
public Type1Font Parse(IInputBytes inputBytes)
{
var scanner = new CoreTokenScanner(inputBytes);
if (!scanner.TryReadToken(out CommentToken comment) || !comment.Data.StartsWith("!"))
{
throw new InvalidFontFormatException("The Type1 program did not start with '%!'.");
}
string name;
var parts = comment.Data.Split(new[] { " " }, StringSplitOptions.RemoveEmptyEntries);
if (parts.Length == 3)
{
name = parts[1];
}
else
{
name = "Unknown";
}
var comments = new List <string>();
while (scanner.MoveNext() && scanner.CurrentToken is CommentToken commentToken)
{
comments.Add(commentToken.Data);
}
var dictionaries = new List <DictionaryToken>();
// Override arrays and names since type 1 handles these differently.
var arrayTokenizer = new Type1ArrayTokenizer();
var nameTokenizer = new Type1NameTokenizer();
scanner.RegisterCustomTokenizer((byte)'{', arrayTokenizer);
scanner.RegisterCustomTokenizer((byte)'/', nameTokenizer);
try
{
var tokenSet = new PreviousTokenSet();
tokenSet.Add(scanner.CurrentToken);
while (scanner.MoveNext())
{
if (scanner.CurrentToken is OperatorToken operatorToken)
{
HandleOperator(operatorToken, inputBytes, scanner, tokenSet, dictionaries);
}
tokenSet.Add(scanner.CurrentToken);
}
}
finally
{
scanner.DeregisterCustomTokenizer(arrayTokenizer);
scanner.DeregisterCustomTokenizer(nameTokenizer);
}
var encoding = GetEncoding(dictionaries);
var matrix = GetFontMatrix(dictionaries);
var boundingBox = GetBoundingBox(dictionaries);
return(new Type1Font(name, encoding, matrix, boundingBox));
}
