/**
* Reads a string that consists of a integer denoting the number of bytes, the bytes (including a
* terminating 0 byte)
*
* @return the string
* @throws IOException if the string could not be read
*/
protected String readString() throws IOException {
// read number of bytes
int bytes = _in.readInt();
if (bytes <= 0) {
throw new IOException("Invalid number of string bytes");
}
String s;
if (bytes > 1) {
s = _in.readUTF(bytes - 1);
} else {
s = "";
}
// read terminating zero
_in.readByte();
return s;
}
/**
* Can be called when a new embedded document is found. Reads the document's header and creates a
* new context on the stack.
*
* @param array true if the document is an embedded array
* @return the json token read
* @throws IOException if an I/O error occurs
*/
protected JsonToken handleNewDocument(boolean array) throws IOException {
if (_in == null) {
// this means Feature.HONOR_DOCUMENT_LENGTH is enabled, and w
// haven't yet started reading. Read the first int to find out the
// length of the document.
byte[] buf = new byte[Integer.SIZE / Byte.SIZE];
int len = 0;
while (len < buf.length) {
int l = _rawInputStream.read(buf, len, buf.length - len);
if (l == -1) {
throw new IOException("Not enough bytes for length of document");
}
len += l;
}
// wrap the input stream by a bounded stream, subtract buf.length from the
// length because the size itself is included in the length
int documentLength = ByteBuffer.wrap(buf).order(ByteOrder.LITTLE_ENDIAN).getInt();
InputStream in = new BoundedInputStream(_rawInputStream, documentLength - buf.length);
// buffer if the raw input stream is not already buffered
if (!(_rawInputStream instanceof BufferedInputStream)) {
in = new StaticBufferedInputStream(in);
}
_counter = new CountingInputStream(in);
_in = new LittleEndianInputStream(_counter);
} else {
// read document header (skip size, we're not interested)
_in.readInt();
}
_contexts.push(new Context(array));
return (array ? JsonToken.START_ARRAY : JsonToken.START_OBJECT);
}
@Override
public void close() throws IOException {
if (isEnabled(JsonParser.Feature.AUTO_CLOSE_SOURCE)) {
_in.close();
}
_closed = true;
}
/**
* Can be called when embedded javascript code with scope is found. Reads the code and the
* embedded document.
*
* @return the json token read
* @throws IOException if an I/O error occurs
*/
protected JsonToken handleJavascriptWithScope() throws IOException {
// skip size
_in.readInt();
String code = readString();
Map<String, Object> doc = readDocument();
getContext().value = new JavaScript(code, doc);
return JsonToken.VALUE_EMBEDDED_OBJECT;
}
/**
* Reads binary data from the input stream
*
* @return the json token read
* @throws IOException if an I/O error occurs
*/
protected JsonToken handleBinary() throws IOException {
int size = _in.readInt();
byte subtype = _in.readByte();
Context ctx = getContext();
switch (subtype) {
case BsonConstants.SUBTYPE_BINARY_OLD:
int size2 = _in.readInt();
byte[] buf2 = new byte[size2];
_in.readFully(buf2);
ctx.value = buf2;
break;
case BsonConstants.SUBTYPE_UUID:
long l1 = _in.readLong();
long l2 = _in.readLong();
ctx.value = new UUID(l1, l2);
break;
default:
byte[] buf = new byte[size];
_in.readFully(buf);
ctx.value = buf;
break;
}
return JsonToken.VALUE_EMBEDDED_OBJECT;
}
/**
* Reads a ObjectID from the input stream
*
* @return the ObjectID
* @throws IOException if the ObjectID could not be read
*/
protected ObjectId readObjectId() throws IOException {
int time = ByteOrderUtil.flip(_in.readInt());
int machine = ByteOrderUtil.flip(_in.readInt());
int inc = ByteOrderUtil.flip(_in.readInt());
return new ObjectId(time, machine, inc);
}
/**
* Reads a timestamp object from the input stream
*
* @return the timestamp
* @throws IOException if the timestamp could not be read
*/
protected Timestamp readTimestamp() throws IOException {
int inc = _in.readInt();
int time = _in.readInt();
return new Timestamp(time, inc);
}
/**
* Skips over a null-terminated string in the input stream
*
* @throws IOException if an I/O error occurs
*/
protected void skipCString() throws IOException {
while (_in.readByte() != 0) ;
}
/**
* @return a null-terminated string read from the input stream
* @throws IOException if the string could not be read
*/
protected String readCString() throws IOException {
return _in.readUTF(-1);
}
@Override
public JsonToken nextToken() throws IOException, JsonParseException {
Context ctx = _contexts.peek();
if (_currToken == null && ctx == null) {
_currToken = handleNewDocument(false);
} else {
_tokenPos = _counter.getPosition();
if (ctx == null) {
if (_currToken == JsonToken.END_OBJECT) {
// end of input
return null;
}
throw new JsonParseException("Found element outside the document", getTokenLocation());
}
if (ctx.state == State.DONE) {
// next field
ctx.reset();
}
boolean readValue = true;
if (ctx.state == State.FIELDNAME) {
readValue = false;
while (true) {
// read field name or end of document
ctx.type = _in.readByte();
if (ctx.type == BsonConstants.TYPE_END) {
// end of document
_currToken = (ctx.array ? JsonToken.END_ARRAY : JsonToken.END_OBJECT);
_contexts.pop();
} else if (ctx.type == BsonConstants.TYPE_UNDEFINED) {
// skip field name and then ignore this token
skipCString();
continue;
} else {
ctx.state = State.VALUE;
_currToken = JsonToken.FIELD_NAME;
if (ctx.array) {
// immediately read value of array element (discard field name)
readValue = true;
skipCString();
ctx.fieldName = null;
} else {
// read field name
ctx.fieldName = readCString();
}
}
break;
}
}
if (readValue) {
// parse element's value
switch (ctx.type) {
case BsonConstants.TYPE_DOUBLE:
ctx.value = _in.readDouble();
_currToken = JsonToken.VALUE_NUMBER_FLOAT;
break;
case BsonConstants.TYPE_STRING:
ctx.value = readString();
_currToken = JsonToken.VALUE_STRING;
break;
case BsonConstants.TYPE_DOCUMENT:
_currToken = handleNewDocument(false);
break;
case BsonConstants.TYPE_ARRAY:
_currToken = handleNewDocument(true);
break;
case BsonConstants.TYPE_BINARY:
_currToken = handleBinary();
break;
case BsonConstants.TYPE_OBJECTID:
ctx.value = readObjectId();
_currToken = JsonToken.VALUE_EMBEDDED_OBJECT;
break;
case BsonConstants.TYPE_BOOLEAN:
boolean b = _in.readBoolean();
ctx.value = b;
_currToken = (b ? JsonToken.VALUE_TRUE : JsonToken.VALUE_FALSE);
break;
case BsonConstants.TYPE_DATETIME:
ctx.value = new Date(_in.readLong());
_currToken = JsonToken.VALUE_EMBEDDED_OBJECT;
break;
case BsonConstants.TYPE_NULL:
_currToken = JsonToken.VALUE_NULL;
break;
case BsonConstants.TYPE_REGEX:
_currToken = handleRegEx();
break;
case BsonConstants.TYPE_DBPOINTER:
_currToken = handleDBPointer();
break;
case BsonConstants.TYPE_JAVASCRIPT:
ctx.value = new JavaScript(readString());
_currToken = JsonToken.VALUE_EMBEDDED_OBJECT;
break;
case BsonConstants.TYPE_SYMBOL:
ctx.value = readSymbol();
_currToken = JsonToken.VALUE_EMBEDDED_OBJECT;
break;
case BsonConstants.TYPE_JAVASCRIPT_WITH_SCOPE:
_currToken = handleJavascriptWithScope();
break;
case BsonConstants.TYPE_INT32:
ctx.value = _in.readInt();
_currToken = JsonToken.VALUE_NUMBER_INT;
break;
case BsonConstants.TYPE_TIMESTAMP:
ctx.value = readTimestamp();
_currToken = JsonToken.VALUE_EMBEDDED_OBJECT;
break;
case BsonConstants.TYPE_INT64:
ctx.value = _in.readLong();
_currToken = JsonToken.VALUE_NUMBER_INT;
break;
case BsonConstants.TYPE_MINKEY:
ctx.value = "MinKey";
_currToken = JsonToken.VALUE_STRING;
break;
case BsonConstants.TYPE_MAXKEY:
ctx.value = "MaxKey";
_currToken = JsonToken.VALUE_STRING;
break;
default:
throw new JsonParseException("Unknown element type " + ctx.type, getTokenLocation());
}
ctx.state = State.DONE;
}
}
return _currToken;
}