private static BinaryHelpFileHeader ReadFileHeader(Stream stream)
{
using (BinaryReader reader = new BinaryReader(new StreamView(stream, 0x46, 2)))
{
BinaryHelpFileHeader header = new BinaryHelpFileHeader();
header.Version = reader.ReadUInt16();
header.Attributes = (HelpFileAttributes)reader.ReadUInt16();
header.ControlCharacter = reader.ReadByte();
header.Padding1 = reader.ReadByte();
header.TopicCount = reader.ReadUInt16();
header.ContextCount = reader.ReadUInt16();
header.DisplayWidth = reader.ReadByte();
header.Padding2 = reader.ReadByte();
header.Padding3 = reader.ReadUInt16();
byte[] stringData = reader.ReadBytes(14);
header.DatabaseName = Encoding.ASCII.GetString(stringData);
int k = header.DatabaseName.IndexOf('\0');
if (k >= 0)
{
header.DatabaseName = header.DatabaseName.Substring(0, k);
}
header.Reserved1 = reader.ReadInt32();
header.TopicIndexOffset = reader.ReadInt32();
header.ContextStringsOffset = reader.ReadInt32();
header.ContextMapOffset = reader.ReadInt32();
header.KeywordsOffset = reader.ReadInt32();
header.HuffmanTreeOffset = reader.ReadInt32();
header.TopicTextOffset = reader.ReadInt32();
header.Reserved2 = reader.ReadInt32();
header.Reserved3 = reader.ReadInt32();
header.DatabaseSize = reader.ReadInt32();
// TODO: make them warnings instead of errors.
// COBOL.HLP contains "JCK\r" in Reserved1.
#if false
// Validate unknown fields.
if (header.Version != 2)
{
throw new NotSupportedException("Unexpected Version field.");
}
if (header.Padding1 != 0 || header.Padding2 != 0 || header.Padding3 != 0)
{
throw new NotSupportedException("Unexpected Padding field.");
}
if (header.Reserved1 != 0 || header.Reserved2 != 0 || header.Reserved3 != 0)
{
throw new NotSupportedException("Unexpected Reserved field.");
}
#endif
return(header);
}
}
private static UInt16[] ReadContextMap(BinaryReader reader, BinaryHelpFileHeader header)
{
if (reader.BaseStream.Position != header.ContextMapOffset)
{
throw new InvalidDataException("Incorrect Context Map section position.");
}
UInt16[] contextMap = new UInt16[header.ContextCount];
for (int i = 0; i < header.ContextCount; i++)
{
contextMap[i] = reader.ReadUInt16();
}
return(contextMap);
}
private static string[] ReadContextStrings(BinaryReader reader, BinaryHelpFileHeader header)
{
if (reader.BaseStream.Position != header.ContextStringsOffset)
{
throw new InvalidDataException("Incorrect Context Strings section position.");
}
// TODO: the NULL at the very end produces an extra, empty context string.
// TODO: check exact number of context strings.
int size = header.ContextMapOffset - header.ContextStringsOffset;
string all = Encoding.ASCII.GetString(reader.ReadBytes(size));
return(all.Split('\0'));
}
private static int[] ReadTopicIndex(BinaryReader reader, BinaryHelpFileHeader header)
{
if (reader.BaseStream.Position != header.TopicIndexOffset)
{
throw new InvalidDataException("Incorrect Topic Index section position.");
}
int[] topicOffsets = new int[header.TopicCount + 1];
for (int i = 0; i <= header.TopicCount; i++)
{
topicOffsets[i] = reader.ReadInt32();
}
return(topicOffsets);
}
private static HuffmanTree ReadHuffmanTree(BinaryReader reader, BinaryHelpFileHeader header)
{
if (reader.BaseStream.Position != header.HuffmanTreeOffset)
{
throw new InvalidDataException("Incorrect Huffman Tree section position.");
}
//int sectionSize = file.Header.TopicTextOffset - file.Header.HuffmanTreeOffset;
HuffmanTree tree = HuffmanTree.Deserialize(reader);
if (tree.IsEmpty || tree.IsSingular)
{
throw new InvalidDataException("Invalid huffman tree.");
}
return(tree);
}
// TODO: this section doesn't terminate by itself?!
private static byte[][] ReadKeywords(
BinaryReader reader, BinaryHelpFileHeader header)
{
if (reader.BaseStream.Position != header.KeywordsOffset)
{
throw new InvalidDataException("Incorrect Keywords section position.");
}
int sectionSize = (header.HuffmanTreeOffset > 0 ? header.HuffmanTreeOffset : header.TopicTextOffset)
- header.KeywordsOffset;
byte[] section = reader.ReadBytes(sectionSize);
if (section.Length != sectionSize)
{
throw new InvalidDataException("Cannot fully read dictionary section.");
}
return(KeywordListSerializer.Deserialize(section).ToArray());
}
/// <summary>
/// Deserializes the next help database from a binary reader.
/// </summary>
/// <remarks>
/// This method throws an exception if it encounters an irrecoverable
/// error in the input, such as an IO error or malformed input in the
/// meta data. It raises an <c>InvalidTopicData</c> event for each
/// format error it encounters during topic deserialization.
/// </remarks>
public HelpDatabase DeserializeDatabase(
Stream stream, SerializationOptions options)
{
if (stream == null)
{
throw new ArgumentNullException(nameof(stream));
}
if (options == null)
{
options = new SerializationOptions();
}
CheckSignature(stream);
BinaryHelpFileHeader header = ReadFileHeader(stream);
bool isCaseSensitive = (header.Attributes & HelpFileAttributes.CaseSensitive) != 0;
HelpDatabase database = new HelpDatabase(header.DatabaseName, isCaseSensitive);
options.ControlCharacter = Graphic437.GetChars(new byte[] { header.ControlCharacter })[0];
using (Stream streamView = new StreamView(stream, header.DatabaseSize, 0x46))
using (BinaryReader reader = new BinaryReader(streamView))
{
int[] topicOffsets = ReadTopicIndex(reader, header);
// Read Context Strings and Context Map sections.
if (true)
{
string[] contextStrings = ReadContextStrings(reader, header);
UInt16[] contextMap = ReadContextMap(reader, header);
for (int i = 0; i < header.ContextCount; i++)
{
database.AddContext(contextStrings[i], contextMap[i]);
}
}
// Read Keywords section.
if (header.KeywordsOffset > 0)
{
options.Keywords = ReadKeywords(reader, header);
options.Compression |= CompressionFlags.Keyword;
options.Compression |= CompressionFlags.ExtendedKeyword;
}
else
{
options.Keywords = null;
}
// Read Huffman Tree section.
if (header.HuffmanTreeOffset > 0)
{
options.HuffmanTree = ReadHuffmanTree(reader, header);
// file.HuffmanTree.Dump();
options.Compression |= CompressionFlags.Huffman;
}
else
{
options.HuffmanTree = null;
}
// Read topic data.
if (reader.BaseStream.Position != header.TopicTextOffset)
{
throw new InvalidDataException("Incorrect topic position.");
}
for (int i = 0; i < header.TopicCount; i++)
{
if (reader.BaseStream.Position != topicOffsets[i])
{
throw new InvalidDataException("Incorrect topic position.");
}
int inputLength = topicOffsets[i + 1] - topicOffsets[i];
byte[] inputData = reader.ReadBytes(inputLength);
HelpTopic topic = DeserializeTopic(inputData, options);
database.Topics.Add(topic);
}
// TODO: check position
if (reader.BaseStream.Position != topicOffsets[header.TopicCount])
{
throw new InvalidDataException("Incorrect topic end position.");
}
if (reader.BaseStream.Position != header.DatabaseSize)
{
throw new InvalidDataException("Incorrect database size.");
}
#if DEBUG
System.Diagnostics.Debug.WriteLine(string.Format(
"Decoded database {0} of {1} bytes.",
database.Name, header.DatabaseSize));
#endif
}
return(database);
}