//CLOVER:ON
// protected methods -----------------------------------------------
/// <summary>
/// Parses the byte buffer and stores its trie content into a index and
/// data array.
/// </summary>
/// <param name="bytes">Buffer containing trie data.</param>
protected override sealed void Unserialize(ByteBuffer bytes) // ICU4N TODO: API - make internal and add overload that accepts byte[]
{
int indexDataLength = m_dataOffset_ + m_dataLength_;
m_index_ = ICUBinary.GetChars(bytes, indexDataLength, 0);
m_data_ = m_index_;
m_initialValue_ = m_data_[m_dataOffset_];
}
private void ReadData(ByteBuffer bytes)
{
// read the header
ICUBinary.ReadHeader(bytes, FMT, new IsAcceptable());
// read indexes[]
int count = bytes.GetInt32();
if (count < IX_TOP)
{
throw new IOException("indexes[0] too small in " + DATA_FILE_NAME);
}
indexes = new int[count];
indexes[0] = count;
for (int i = 1; i < count; ++i)
{
indexes[i] = bytes.GetInt32();
}
// read the trie
trie = Trie2_16.CreateFromSerialized(bytes);
int expectedTrieLength = indexes[IX_TRIE_SIZE];
int trieLength = trie.GetSerializedLength();
if (trieLength > expectedTrieLength)
{
throw new IOException(DATA_FILE_NAME + ": not enough bytes for the trie");
}
// skip padding after trie bytes
ICUBinary.SkipBytes(bytes, expectedTrieLength - trieLength);
// read exceptions[]
count = indexes[IX_EXC_LENGTH];
if (count > 0)
{
exceptions = ICUBinary.GetString(bytes, count, 0);
}
// read unfold[]
count = indexes[IX_UNFOLD_LENGTH];
if (count > 0)
{
unfold = ICUBinary.GetChars(bytes, count, 0);
}
}
// protected methods -------------------------------------------------
/// <summary>
/// Read and break up the stream of data passed in as arguments
/// and fills up <see cref="UCharacterName"/>.
/// If unsuccessful false will be returned.
/// </summary>
/// <param name="data">Instance of datablock.</param>
/// <exception cref="IOException">Thrown when there's a data error.</exception>
internal void Read(UCharacterName data)
{
// reading index
m_tokenstringindex_ = m_byteBuffer_.GetInt32();
m_groupindex_ = m_byteBuffer_.GetInt32();
m_groupstringindex_ = m_byteBuffer_.GetInt32();
m_algnamesindex_ = m_byteBuffer_.GetInt32();
// reading tokens
int count = m_byteBuffer_.GetChar();
char[] token = ICUBinary.GetChars(m_byteBuffer_, count, 0);
int size = m_groupindex_ - m_tokenstringindex_;
byte[] tokenstr = new byte[size];
m_byteBuffer_.Get(tokenstr);
data.SetToken(token, tokenstr);
// reading the group information records
count = m_byteBuffer_.GetChar();
data.SetGroupCountSize(count, GROUP_INFO_SIZE_);
count *= GROUP_INFO_SIZE_;
char[] group = ICUBinary.GetChars(m_byteBuffer_, count, 0);
size = m_algnamesindex_ - m_groupstringindex_;
byte[] groupstring = new byte[size];
m_byteBuffer_.Get(groupstring);
data.SetGroup(group, groupstring);
count = m_byteBuffer_.GetInt32();
UCharacterName.AlgorithmName[] alg =
new UCharacterName.AlgorithmName[count];
for (int i = 0; i < count; i++)
{
UCharacterName.AlgorithmName an = ReadAlg();
if (an == null)
{
throw new IOException("unames.icu read error: Algorithmic names creation error");
}
alg[i] = an;
}
data.SetAlgorithm(alg);
}
// private methods ---------------------------------------------------
/// <summary>
/// Reads an individual record of <see cref="UCharacterName.AlgorithmName"/>s
/// </summary>
/// <returns>An instance of <see cref="UCharacterName.AlgorithmName"/>s if read is successful otherwise null.</returns>
/// <exception cref="IOException">Thrown when file read error occurs or data is corrupted.</exception>
private UCharacterName.AlgorithmName ReadAlg()
{
UCharacterName.AlgorithmName result =
new UCharacterName.AlgorithmName();
int rangestart = m_byteBuffer_.GetInt32();
int rangeend = m_byteBuffer_.GetInt32();
byte type = m_byteBuffer_.Get();
byte variant = m_byteBuffer_.Get();
if (!result.SetInfo(rangestart, rangeend, type, variant))
{
return(null);
}
int size = m_byteBuffer_.GetChar();
if (type == UCharacterName.AlgorithmName.TYPE_1_)
{
char[] factor = ICUBinary.GetChars(m_byteBuffer_, variant, 0);
result.SetFactor(factor);
size -= (variant << 1);
}
StringBuilder prefix = new StringBuilder();
char c = (char)(m_byteBuffer_.Get() & 0x00FF);
while (c != 0)
{
prefix.Append(c);
c = (char)(m_byteBuffer_.Get() & 0x00FF);
}
result.SetPrefix(prefix.ToString());
size -= (ALG_INFO_SIZE_ + prefix.Length + 1);
if (size > 0)
{
byte[] str = new byte[size];
m_byteBuffer_.Get(str);
result.SetFactorString(str);
}
return(result);
}
public char[] Read(int length)
{
//Read the extra data
return(ICUBinary.GetChars(byteBuffer, length, 0));
}
/// <summary>
/// Parses the byte buffer and creates the trie index with it.
/// <para/>
/// The position of the input <see cref="ByteBuffer"/> must be right after the trie header.
/// <para/>
/// This is overwritten by the child classes.
/// </summary>
/// <param name="bytes">Buffer containing trie data.</param>
protected virtual void Unserialize(ByteBuffer bytes)
{
m_index = ICUBinary.GetChars(bytes, m_dataOffset, 0);
}
/// <summary>
/// Create a <see cref="Trie2"/> from its serialized form. Inverse of utrie2_serialize().
/// </summary>
/// <remarks>
/// Reads from the current position and leaves the buffer after the end of the trie.
/// <para/>
/// The serialized format is identical between ICU4C, ICU4J, and ICU4N, so this function
/// will work with serialized <see cref="Trie2"/>s from any.
/// <para/>
/// The actual type of the returned <see cref="Trie2"/> will be either <see cref="Trie2_16"/> or <see cref="Trie2_32"/>, depending
/// on the width of the data.
/// <para/>
/// To obtain the width of the <see cref="Trie2"/>, check the actual class type of the returned <see cref="Trie2"/>.
/// Or use the <see cref="Trie2_16.CreateFromSerialized(ByteBuffer)"/> or <see cref="Trie2_32.CreateFromSerialized(ByteBuffer)"/> method, which will
/// return only <see cref="Trie"/>s of their specific type/size.
/// <para/>
/// The serialized <see cref="Trie2"/> on the stream may be in either little or big endian byte order.
/// This allows using serialized <see cref="Trie"/>s from ICU4C without needing to consider the
/// byte order of the system that created them.
/// </remarks>
/// <param name="bytes">A byte buffer to the serialized form of a UTrie2.</param>
/// <returns>An unserialized <see cref="Trie2"/>, ready for use.</returns>
/// <exception cref="ArgumentException">If the stream does not contain a serialized <see cref="Trie2"/>.</exception>
/// <exception cref="IOException">If a read error occurs in the buffer.</exception>
public static Trie2 CreateFromSerialized(ByteBuffer bytes) // ICU4N TODO: API Create overload that accepts byte[]
{
// From ICU4C utrie2_impl.h
// * Trie2 data structure in serialized form:
// *
// * UTrie2Header header;
// * uint16_t index[header.index2Length];
// * uint16_t data[header.shiftedDataLength<<2]; -- or uint32_t data[...]
// * @internal
// */
// typedef struct UTrie2Header {
// /** "Tri2" in big-endian US-ASCII (0x54726932) */
// uint32_t signature;
// /**
// * options bit field:
// * 15.. 4 reserved (0)
// * 3.. 0 UTrie2ValueBits valueBits
// */
// uint16_t options;
//
// /** UTRIE2_INDEX_1_OFFSET..UTRIE2_MAX_INDEX_LENGTH */
// uint16_t indexLength;
//
// /** (UTRIE2_DATA_START_OFFSET..UTRIE2_MAX_DATA_LENGTH)>>UTRIE2_INDEX_SHIFT */
// uint16_t shiftedDataLength;
//
// /** Null index and data blocks, not shifted. */
// uint16_t index2NullOffset, dataNullOffset;
//
// /**
// * First code point of the single-value range ending with U+10ffff,
// * rounded up and then shifted right by UTRIE2_SHIFT_1.
// */
// uint16_t shiftedHighStart;
// } UTrie2Header;
ByteOrder outerByteOrder = bytes.Order;
try
{
UTrie2Header header = new UTrie2Header();
/* check the signature */
header.signature = bytes.GetInt32();
switch (header.signature)
{
case 0x54726932:
// The buffer is already set to the trie data byte order.
break;
case 0x32697254:
// Temporarily reverse the byte order.
bool isBigEndian = outerByteOrder == ByteOrder.BigEndian;
bytes.Order = isBigEndian ? ByteOrder.LittleEndian : ByteOrder.BigEndian;
header.signature = 0x54726932;
break;
default:
throw new ArgumentException("Buffer does not contain a serialized UTrie2");
}
header.options = bytes.GetChar();
header.indexLength = bytes.GetChar();
header.shiftedDataLength = bytes.GetChar();
header.index2NullOffset = bytes.GetChar();
header.dataNullOffset = bytes.GetChar();
header.shiftedHighStart = bytes.GetChar();
// Trie2 data width - 0: 16 bits
// 1: 32 bits
if ((header.options & UTRIE2_OPTIONS_VALUE_BITS_MASK) > 1)
{
throw new ArgumentException("UTrie2 serialized format error.");
}
ValueWidth width;
Trie2 This;
if ((header.options & UTRIE2_OPTIONS_VALUE_BITS_MASK) == 0)
{
width = ValueWidth.BITS_16;
This = new Trie2_16();
}
else
{
width = ValueWidth.BITS_32;
This = new Trie2_32();
}
This.header = header;
/* get the length values and offsets */
This.indexLength = header.indexLength;
This.dataLength = header.shiftedDataLength << UTRIE2_INDEX_SHIFT;
This.index2NullOffset = header.index2NullOffset;
This.dataNullOffset = header.dataNullOffset;
This.highStart = header.shiftedHighStart << UTRIE2_SHIFT_1;
This.highValueIndex = This.dataLength - UTRIE2_DATA_GRANULARITY;
if (width == ValueWidth.BITS_16)
{
This.highValueIndex += This.indexLength;
}
// Allocate the Trie2 index array. If the data width is 16 bits, the array also
// includes the space for the data.
int indexArraySize = This.indexLength;
if (width == ValueWidth.BITS_16)
{
indexArraySize += This.dataLength;
}
/* Read in the index */
This.index = ICUBinary.GetChars(bytes, indexArraySize, 0);
/* Read in the data. 16 bit data goes in the same array as the index.
* 32 bit data goes in its own separate data array.
*/
if (width == ValueWidth.BITS_16)
{
This.data16 = This.indexLength;
}
else
{
This.data32 = ICUBinary.GetInt32s(bytes, This.dataLength, 0);
}
switch (width)
{
case ValueWidth.BITS_16:
This.data32 = null;
This.initialValue = This.index[This.dataNullOffset];
This.errorValue = This.index[This.data16 + UTRIE2_BAD_UTF8_DATA_OFFSET];
break;
case ValueWidth.BITS_32:
This.data16 = 0;
This.initialValue = This.data32[This.dataNullOffset];
This.errorValue = This.data32[UTRIE2_BAD_UTF8_DATA_OFFSET];
break;
default:
throw new ArgumentException("UTrie2 serialized format error.");
}
return(This);
}
finally
{
bytes.Order = outerByteOrder;
}
}