private static void ReadProcedure(Type1Tokenizer tokenizer, List <Type1Token> tokens, ref int depth)
{
if (depth == -1)
{
ReadExpected(tokenizer, Type1Token.TokenType.StartProc);
depth = 1;
}
if (depth == 0)
{
return;
}
Type1Token token;
while ((token = tokenizer.GetNext()) != null)
{
if (token.Type == Type1Token.TokenType.StartProc)
{
depth += 1;
ReadProcedure(tokenizer, tokens, ref depth);
}
else if (token.Type == Type1Token.TokenType.EndProc)
{
depth--;
break;
}
else
{
tokens.Add(token);
}
}
}
private static void ReadOtherSubroutines(Type1Tokenizer tokenizer, bool isLenientParsing)
{
var start = tokenizer.GetNext();
if (start.Type == Type1Token.TokenType.StartArray)
{
ReadArrayValues(tokenizer, x => x, true, false);
}
else if (start.Type == Type1Token.TokenType.Integer || start.Type == Type1Token.TokenType.Real)
{
var length = start.AsInt();
ReadExpected(tokenizer, Type1Token.TokenType.Name, "array");
for (var i = 0; i < length; i++)
{
ReadExpected(tokenizer, Type1Token.TokenType.Name, "dup");
ReadNumeric(tokenizer);
ReadTillPut(tokenizer);
}
ReadTillDef(tokenizer);
}
else if (!isLenientParsing)
{
throw new InvalidOperationException($"Failed to read start of /OtherSubrs array. Got start token: {start}.");
}
}
private static IReadOnlyList <Type1Token> ReadProcedure(Type1Tokenizer tokenizer)
{
var tokens = new List <Type1Token>();
var depth = -1;
ReadProcedure(tokenizer, tokens, ref depth);
return(tokens);
}
private static IReadOnlyList <Type1Token> ReadProcedure(Type1Tokenizer tokenizer, bool hasReadStartProc)
{
var tokens = new List <Type1Token>();
var depth = hasReadStartProc ? 1 : -1;
ReadProcedure(tokenizer, tokens, ref depth);
return(tokens);
}
private static decimal ReadNumeric(Type1Tokenizer tokenizer)
{
var token = tokenizer.GetNext();
if (token == null || (token.Type != Type1Token.TokenType.Integer && token.Type != Type1Token.TokenType.Real))
{
throw new InvalidOperationException($"Expected to read a numeric token, instead got: {token}.");
}
return(token.AsDecimal());
}
private static bool ReadBoolean(Type1Tokenizer tokenizer)
{
var token = tokenizer.GetNext();
if (token == null || (!string.Equals(token.Text, "true", StringComparison.OrdinalIgnoreCase) &&
!string.Equals(token.Text, "false", StringComparison.OrdinalIgnoreCase)))
{
throw new InvalidOperationException($"Expected to read a boolean token, instead got: {token}.");
}
return(token.AsBool());
}
private static IReadOnlyList <Type1CharstringDecryptedBytes> ReadCharStrings(Type1Tokenizer tokenizer, int lenIv, bool isLenientParsing)
{
var length = (int)ReadNumeric(tokenizer);
ReadExpected(tokenizer, Type1Token.TokenType.Name, "dict");
// dup begin or CharStrings begin.
ReadExpected(tokenizer, Type1Token.TokenType.Name);
ReadExpected(tokenizer, Type1Token.TokenType.Name, "begin");
var results = new List <Type1CharstringDecryptedBytes>();
for (var i = 0; i < length; i++)
{
var token = tokenizer.GetNext();
if (token.Type == Type1Token.TokenType.Name &&
string.Equals(token.Text, "end", StringComparison.OrdinalIgnoreCase))
{
break;
}
if (token.Type != Type1Token.TokenType.Literal)
{
throw new InvalidOperationException($"Type 1 font error. Expected literal for charstring name, instead got: {token}.");
}
var name = token.Text;
var charstringLength = ReadNumeric(tokenizer);
var charstring = tokenizer.GetNext();
if (!(charstring is Type1DataToken charstringToken))
{
throw new InvalidOperationException($"Got wrong type of token, expected charstring, instead got: {charstring}.");
}
if (!isLenientParsing && charstringToken.Data.Count != charstringLength)
{
throw new InvalidOperationException($"The charstring {charstringToken} did not have the expected length of {charstringLength}.");
}
var data = Decrypt(charstringToken.Data, CharstringEncryptionKey, lenIv);
results.Add(new Type1CharstringDecryptedBytes(name, data, i));
ReadTillDef(tokenizer);
}
ReadExpected(tokenizer, Type1Token.TokenType.Name, "end");
return(results);
}
private static void ReadExpected(Type1Tokenizer tokenizer, Type1Token.TokenType type, string text = null)
{
var token = tokenizer.GetNext();
if (token == null)
{
throw new InvalidOperationException($"Type 1 Encrypted portion ended when a token with text '{text}' was expected.");
}
if (token.Type != type || (text != null && !string.Equals(token.Text, text, StringComparison.OrdinalIgnoreCase)))
{
throw new InvalidOperationException($"Found invalid token {token} when type {type} with text {text} was expected.");
}
}
private static void ReadTillDef(Type1Tokenizer tokenizer, bool skip = false)
{
if (string.Equals(tokenizer.CurrentToken.Text, "def", StringComparison.OrdinalIgnoreCase))
{
return;
}
Type1Token token;
while ((token = tokenizer.GetNext()) != null)
{
if (token.Type == Type1Token.TokenType.Name)
{
if (string.Equals(token.Text, "ND", StringComparison.OrdinalIgnoreCase) || string.Equals(token.Text, "|-", StringComparison.OrdinalIgnoreCase))
{
return;
}
if (string.Equals(token.Text, "def", StringComparison.OrdinalIgnoreCase))
{
break;
}
// PostScript wrapper (ignored for now)
if (string.Equals(token.Text, "systemdict"))
{
ReadExpected(tokenizer, Type1Token.TokenType.Literal, "internaldict");
ReadExpected(tokenizer, Type1Token.TokenType.Name, "known");
ReadExpected(tokenizer, Type1Token.TokenType.StartProc);
// TODO: read values from the wrapper, see line 396 of Type1Parser.java
// Skips the entire contents
while ((token = tokenizer.GetNext()) != null)
{
if (token.Type == Type1Token.TokenType.Name && (token.Text == "ND" || token.Text == "def"))
{
return;
}
}
}
}
else if (!skip)
{
throw new InvalidOperationException($"Encountered unexpected non-name token while reading till 'def' token: {token}");
}
}
}
private static IReadOnlyList <T> ReadArrayValues <T>(Type1Tokenizer tokenizer, Func <Type1Token, T> converter, bool hasReadStart = false,
bool includeDef = true)
{
if (!hasReadStart)
{
ReadExpected(tokenizer, Type1Token.TokenType.StartArray);
}
var results = new List <T>();
Type1Token token;
while ((token = tokenizer.GetNext()) != null)
{
if (token.Type == Type1Token.TokenType.EndArray)
{
break;
}
if (token.Type == Type1Token.TokenType.StartArray)
{
var array = ReadArrayValues <T>(tokenizer, converter, true, false);
results.AddRange(array);
}
try
{
var result = converter(token);
results.Add(result);
}
catch (Exception ex)
{
throw new InvalidOperationException($"Conversion of token '{token}' to value of type {typeof(T).Name} failed.", ex);
}
}
if (includeDef)
{
ReadTillDef(tokenizer);
}
return(results);
}
private static void ReadTillPut(Type1Tokenizer tokenizer)
{
Type1Token token;
while ((token = tokenizer.GetNext()) != null)
{
if (string.Equals(token.Text, "put", StringComparison.OrdinalIgnoreCase))
{
break;
}
switch (token.Text)
{
case "NP":
case "|":
return;
}
}
}
private static IReadOnlyList <Type1CharstringDecryptedBytes> ReadSubroutines(Type1Tokenizer tokenizer, int lenIv, bool isLenientParsing)
{
var length = (int)ReadNumeric(tokenizer);
var subroutines = new List <Type1CharstringDecryptedBytes>(length);
ReadExpected(tokenizer, Type1Token.TokenType.Name, "array");
for (var i = 0; i < length; i++)
{
var current = tokenizer.GetNext();
if (current.Type != Type1Token.TokenType.Name || !string.Equals(current.Text, "dup"))
{
break;
}
var index = (int)ReadNumeric(tokenizer);
var byteLength = (int)ReadNumeric(tokenizer);
var charstring = tokenizer.GetNext();
if (!(charstring is Type1DataToken charstringToken))
{
throw new InvalidOperationException($"Found an unexpected token instead of subroutine charstring: {charstring}.");
}
if (!isLenientParsing && charstringToken.Data.Count != byteLength)
{
throw new InvalidOperationException($"The subroutine charstring {charstringToken} did not have the expected length of {byteLength}.");
}
var subroutine = Decrypt(charstringToken.Data, CharstringEncryptionKey, lenIv);
subroutines.Add(new Type1CharstringDecryptedBytes(subroutine, index));
ReadTillPut(tokenizer);
}
ReadTillDef(tokenizer);
return(subroutines);
}
private static void ReadExpectedAfterOptional(Type1Tokenizer tokenizer, Type1Token.TokenType optionalType, string optionalText,
Type1Token.TokenType type, string text)
{
var token = tokenizer.GetNext();
if (token == null)
{
throw new InvalidOperationException($"Type 1 Encrypted portion ended when a token with text '{optionalText}' or '{text}' was expected.");
}
if (token.Type == type && string.Equals(token.Text, text, StringComparison.OrdinalIgnoreCase))
{
return;
}
if (token.Type == optionalType && string.Equals(token.Text, optionalText, StringComparison.OrdinalIgnoreCase))
{
ReadExpected(tokenizer, type, text);
return;
}
throw new InvalidOperationException($"Found invalid token {token} when type {type} with text {text} was expected.");
}
private static void ReadTillDef(Type1Tokenizer tokenizer, bool skip = false)
{
Type1Token token;
while ((token = tokenizer.GetNext()) != null)
{
if (token.Type == Type1Token.TokenType.Name)
{
if (string.Equals(token.Text, "ND", StringComparison.OrdinalIgnoreCase) || string.Equals(token.Text, "|-", StringComparison.OrdinalIgnoreCase))
{
return;
}
if (string.Equals(token.Text, "def", StringComparison.OrdinalIgnoreCase))
{
break;
}
}
else if (!skip)
{
throw new InvalidOperationException($"Encountered unexpected non-name token while reading till 'def' token: {token}");
}
}
}
public (Type1PrivateDictionary, Type1CharStrings) Parse(IReadOnlyList <byte> bytes, bool isLenientParsing)
{
if (!IsBinary(bytes))
{
bytes = ConvertHexToBinary(bytes);
}
var decrypted = Decrypt(bytes, EexecEncryptionKey, EexecRandomBytes);
if (decrypted.Count == 0)
{
var defaultPrivateDictionary = new Type1PrivateDictionary(new Type1PrivateDictionary.Builder());
var defaultCharstrings = new Type1CharStrings(new Dictionary <string, Type1CharStrings.CommandSequence>(),
new Dictionary <int, string>(),
new Dictionary <int, Type1CharStrings.CommandSequence>());
return(defaultPrivateDictionary, defaultCharstrings);
}
var tokenizer = new Type1Tokenizer(new ByteArrayInputBytes(decrypted));
/*
* After 4 random characters follows the /Private dictionary and the /CharString dictionary.
* The first defines a number of technical terms involving character construction, and contains also an array of subroutines used in character paths.
* The second contains the character descriptions themselves.
* Both the subroutines and the character descriptions are yet again encrypted in a fashion similar to the entire binary segment, but now with an initial value of R = 4330 instead of 55665.
*/
while (!tokenizer.CurrentToken.IsPrivateDictionary)
{
tokenizer.GetNext();
if (tokenizer.CurrentToken == null)
{
throw new InvalidOperationException("Did not find the private dictionary start token.");
}
}
var next = tokenizer.GetNext();
if (next?.Type != Type1Token.TokenType.Integer)
{
throw new InvalidOperationException($"No length token was present in the stream following the private dictionary start, instead got {next}.");
}
var length = next.AsInt();
ReadExpected(tokenizer, Type1Token.TokenType.Name, "dict");
// Could also be "/Private 10 dict def Private begin" instead of the "dup"
ReadExpectedAfterOptional(tokenizer, Type1Token.TokenType.Name, "def", Type1Token.TokenType.Name, "dup");
ReadExpected(tokenizer, Type1Token.TokenType.Name, "begin");
/*
* The lenIV entry is an integer specifying the number of random bytes at the beginning of charstrings for charstring encryption.
* The default value of lenIV is 4.
*/
var lenIv = Len4Bytes;
var builder = new Type1PrivateDictionary.Builder();
for (var i = 0; i < length; i++)
{
var token = tokenizer.GetNext();
// premature end
if (token == null || token.Type != Type1Token.TokenType.Literal)
{
break;
}
var key = token.Text;
switch (key)
{
case Type1Symbols.RdProcedure:
case Type1Symbols.RdProcedureAlt:
{
var procedureTokens = ReadProcedure(tokenizer);
builder.Rd = procedureTokens;
ReadTillDef(tokenizer);
break;
}
case Type1Symbols.NoAccessDef:
case Type1Symbols.NoAccessDefAlt:
{
var procedureTokens = ReadProcedure(tokenizer);
builder.NoAccessDef = procedureTokens;
ReadTillDef(tokenizer);
break;
}
case Type1Symbols.NoAccessPut:
case Type1Symbols.NoAccessPutAlt:
{
var procedureTokens = ReadProcedure(tokenizer);
builder.NoAccessPut = procedureTokens;
ReadTillDef(tokenizer);
break;
}
case Type1Symbols.BlueValues:
{
var blueValues = ReadArrayValues(tokenizer, x => x.AsInt());
builder.BlueValues = blueValues;
break;
}
case Type1Symbols.OtherBlues:
{
var otherBlues = ReadArrayValues(tokenizer, x => x.AsInt());
builder.OtherBlues = otherBlues;
break;
}
case Type1Symbols.StdHorizontalStemWidth:
{
var widths = ReadArrayValues(tokenizer, x => x.AsDecimal());
var width = widths[0];
builder.StandardHorizontalWidth = width;
break;
}
case Type1Symbols.StdVerticalStemWidth:
{
var widths = ReadArrayValues(tokenizer, x => x.AsDecimal());
var width = widths[0];
builder.StandardVerticalWidth = width;
break;
}
case Type1Symbols.StemSnapHorizontalWidths:
{
var widths = ReadArrayValues(tokenizer, x => x.AsDecimal());
builder.StemSnapHorizontalWidths = widths;
break;
}
case Type1Symbols.StemSnapVerticalWidths:
{
var widths = ReadArrayValues(tokenizer, x => x.AsDecimal());
builder.StemSnapVerticalWidths = widths;
break;
}
case Type1Symbols.BlueScale:
{
builder.BlueScale = ReadNumeric(tokenizer);
ReadTillDef(tokenizer);
break;
}
case Type1Symbols.ForceBold:
{
builder.ForceBold = ReadBoolean(tokenizer);
ReadTillDef(tokenizer);
break;
}
case Type1Symbols.MinFeature:
{
var procedureTokens = ReadProcedure(tokenizer);
builder.MinFeature = new MinFeature(procedureTokens[0].AsInt(), procedureTokens[1].AsInt());
ReadTillDef(tokenizer);
break;
}
case Type1Symbols.Password:
{
var password = (int)ReadNumeric(tokenizer);
if (password != Password && !isLenientParsing)
{
throw new InvalidOperationException($"Type 1 font had the wrong password: {password}");
}
builder.Password = password;
ReadTillDef(tokenizer);
break;
}
case Type1Symbols.UniqueId:
{
var id = (int)ReadNumeric(tokenizer);
builder.UniqueId = id;
ReadTillDef(tokenizer);
break;
}
case Type1Symbols.Len4:
{
lenIv = (int)ReadNumeric(tokenizer);
ReadTillDef(tokenizer);
break;
}
case Type1Symbols.BlueShift:
{
builder.BlueShift = (int)ReadNumeric(tokenizer);
ReadTillDef(tokenizer);
break;
}
case Type1Symbols.BlueFuzz:
{
builder.BlueFuzz = (int)ReadNumeric(tokenizer);
ReadTillDef(tokenizer);
break;
}
case Type1Symbols.FamilyBlues:
{
builder.FamilyBlues = ReadArrayValues(tokenizer, x => x.AsInt());
break;
}
case Type1Symbols.FamilyOtherBlues:
{
builder.FamilyOtherBlues = ReadArrayValues(tokenizer, x => x.AsInt());
break;
}
case Type1Symbols.LanguageGroup:
{
builder.LanguageGroup = (int)ReadNumeric(tokenizer);
ReadTillDef(tokenizer);
break;
}
case Type1Symbols.RndStemUp:
{
builder.RoundStemUp = ReadBoolean(tokenizer);
ReadTillDef(tokenizer);
break;
}
case Type1Symbols.Subroutines:
{
builder.Subroutines = ReadSubroutines(tokenizer, lenIv, isLenientParsing);
break;
}
case Type1Symbols.OtherSubroutines:
{
ReadOtherSubroutines(tokenizer, isLenientParsing);
ReadTillDef(tokenizer);
break;
}
case Type1Symbols.ExpansionFactor:
{
builder.ExpansionFactor = ReadNumeric(tokenizer);
ReadTillDef(tokenizer);
break;
}
case Type1Symbols.Erode:
{
ReadTillDef(tokenizer, true);
break;
}
default:
{
ReadTillDef(tokenizer, true);
break;
}
}
}
var currentToken = tokenizer.CurrentToken;
IReadOnlyList <Type1CharstringDecryptedBytes> charStrings;
if (currentToken != null)
{
while (currentToken != null && currentToken.Type != Type1Token.TokenType.Literal &&
!string.Equals(currentToken.Text, "CharStrings", StringComparison.OrdinalIgnoreCase))
{
currentToken = tokenizer.GetNext();
}
if (currentToken != null)
{
charStrings = ReadCharStrings(tokenizer, lenIv, isLenientParsing);
}
else
{
charStrings = new Type1CharstringDecryptedBytes[0];
}
}
else
{
charStrings = new Type1CharstringDecryptedBytes[0];
}
var privateDictionary = builder.Build();
var instructions = Type1CharStringParser.Parse(charStrings, builder.Subroutines ?? new Type1CharstringDecryptedBytes[0]);
return(privateDictionary, instructions);
}