panda.net.http.MultipartStream Maven / Gradle / Ivy
Show all versions of panda-core Show documentation
package panda.net.http;
import static java.lang.String.format;
import java.io.ByteArrayOutputStream;
import java.io.IOException;
import java.io.InputStream;
import java.io.OutputStream;
import java.io.UnsupportedEncodingException;
import panda.io.SizeLimitExceededException;
import panda.io.Streams;
/**
*
* Low level API for processing file uploads.
*
* This class can be used to process data streams conforming to MIME 'multipart' format as defined
* in RFC 1867. Arbitrarily large amounts of data
* in the stream can be processed under constant memory usage.
*
* The format of the stream is defined in the following way:
*
* multipart-body := preamble 1*encapsulation close-delimiter epilogue
* encapsulation := delimiter body CRLF
* delimiter := "--" boundary CRLF
* close-delimiter := "--" boundary "--"
* preamble := <ignore>
* epilogue := <ignore>
* body := header-part CRLF body-part
* header-part := 1*header CRLF
* header := header-name ":" header-value
* header-name := <printable ascii characters except ":">
* header-value := <any ascii characters except CR & LF>
* body-data := <arbitrary data>
*
*
* Note that body-data can contain another mulipart entity. There is limited support for single pass
* processing of such nested streams. The nested stream is required to have a
* boundary token of the same length as the parent stream (see {@link #setBoundary(byte[])}).
*
* Here is an example of usage of this class.
*
*
* try {
* MultipartStream multipartStream = new MultipartStream(input, boundary);
* boolean nextPart = multipartStream.skipPreamble();
* OutputStream output;
* while (nextPart) {
* String header = multipartStream.readHeaders();
* // process headers
* // create some output stream
* multipartStream.readBodyData(output);
* nextPart = multipartStream.readBoundary();
* }
* }
* catch (MultipartStream.MalformedStreamException e) {
* // the stream failed to follow required syntax
* }
* catch (IOException e) {
* // a read or write error occurred
* }
*
*/
public class MultipartStream {
/**
* The {@link ProgressListener} may be used to display a progress bar or do stuff like that.
*/
public static interface ProgressListener {
/**
* Updates the listeners status information.
*
* @param pBytesRead The total number of bytes, which have been read so far.
* @param pContentLength The total number of bytes, which are being read. May be -1, if this
* number is unknown.
* @param pItems The number of the field, which is currently being read. (0 = no item so far, 1
* = first item is being read, ...)
*/
void update(long pBytesRead, long pContentLength, int pItems);
}
/**
* Internal class, which is used to invoke the {@link ProgressListener}.
*/
public static class ProgressNotifier {
/**
* The listener to invoke.
*/
private final ProgressListener listener;
/**
* Number of expected bytes, if known, or -1.
*/
private final long contentLength;
/**
* Number of bytes, which have been read so far.
*/
private long bytesRead;
/**
* Number of items, which have been read so far.
*/
private int items;
/**
* Creates a new instance with the given listener and content length.
*
* @param pListener The listener to invoke.
* @param pContentLength The expected content length.
*/
public ProgressNotifier(ProgressListener pListener, long pContentLength) {
listener = pListener;
contentLength = pContentLength;
}
/**
* Called to indicate that bytes have been read.
*
* @param pBytes Number of bytes, which have been read.
*/
void noteBytesRead(int pBytes) {
/*
* Indicates, that the given number of bytes have been read from the input stream.
*/
bytesRead += pBytes;
notifyListener();
}
/**
* Called to indicate, that a new file item has been detected.
*/
public void noteItem() {
++items;
notifyListener();
}
/**
* Called for notifying the listener.
*/
private void notifyListener() {
if (listener != null) {
listener.update(bytesRead, contentLength, items);
}
}
}
// ----------------------------------------------------- Manifest constants
/**
* The Carriage Return ASCII character value.
*/
public static final byte CR = 0x0D;
/**
* The Line Feed ASCII character value.
*/
public static final byte LF = 0x0A;
/**
* The dash (-) ASCII character value.
*/
public static final byte DASH = 0x2D;
/**
* The maximum length of header-part that will be processed (10 kilobytes = 10240
* bytes.).
*/
public static final int HEADER_PART_SIZE_MAX = 10240;
/**
* The default length of the buffer used for processing a request.
*/
protected static final int DEFAULT_BUFSIZE = 4096;
/**
* A byte sequence that marks the end of header-part (CRLFCRLF).
*/
protected static final byte[] HEADER_SEPARATOR = { CR, LF, CR, LF };
/**
* A byte sequence that that follows a delimiter that will be followed by an encapsulation (
* CRLF).
*/
protected static final byte[] FIELD_SEPARATOR = { CR, LF };
/**
* A byte sequence that that follows a delimiter of the last encapsulation in the stream (
* --).
*/
protected static final byte[] STREAM_TERMINATOR = { DASH, DASH };
/**
* A byte sequence that precedes a boundary (CRLF--).
*/
protected static final byte[] BOUNDARY_PREFIX = { CR, LF, DASH, DASH };
// ----------------------------------------------------------- Data members
/**
* The input stream from which data is read.
*/
private final InputStream input;
/**
* The length of the boundary token plus the leading CRLF--.
*/
private int boundaryLength;
/**
* The amount of data, in bytes, that must be kept in the buffer in order to detect delimiters
* reliably.
*/
private int keepRegion;
/**
* The byte sequence that partitions the stream.
*/
private byte[] boundary;
/**
* The length of the buffer used for processing the request.
*/
private final int bufSize;
/**
* The buffer used for processing the request.
*/
private final byte[] buffer;
/**
* The index of first valid character in the buffer.
* 0 <= head < bufSize
*/
private int head;
/**
* The index of last valid character in the buffer + 1.
* 0 <= tail <= bufSize
*/
private int tail;
/**
* The content encoding to use when reading headers.
*/
private String headerEncoding;
/**
* The progress notifier, if any, or null.
*/
private final ProgressNotifier notifier;
// ----------------------------------------------------------- Constructors
/**
*
* Constructs a MultipartStream with a custom size buffer and no progress notifier.
*
* Note that the buffer must be at least big enough to contain the boundary string, plus 4
* characters for CR/LF and double dash, plus at least one byte of data. Too small a buffer size
* setting will degrade performance.
*
* @param input The InputStream to serve as a data source.
* @param boundary The token used for dividing the stream into encapsulations.
* @param bufSize The size of the buffer to be used, in bytes.
*/
public MultipartStream(InputStream input, byte[] boundary, int bufSize) {
this(input, boundary, bufSize, null);
}
/**
*
* Constructs a MultipartStream with a custom size buffer.
*
* Note that the buffer must be at least big enough to contain the boundary string, plus 4
* characters for CR/LF and double dash, plus at least one byte of data. Too small a buffer size
* setting will degrade performance.
*
* @param input The InputStream to serve as a data source.
* @param boundary The token used for dividing the stream into encapsulations.
* @param bufSize The size of the buffer to be used, in bytes.
* @param pNotifier The notifier, which is used for calling the progress listener, if any.
* @throws IllegalArgumentException If the buffer size is too small
*/
public MultipartStream(InputStream input, byte[] boundary, int bufSize, ProgressNotifier pNotifier) {
if (boundary == null) {
throw new IllegalArgumentException("boundary may not be null");
}
this.input = input;
this.bufSize = bufSize;
this.buffer = new byte[bufSize];
this.notifier = pNotifier;
// We prepend CR/LF to the boundary to chop trailing CR/LF from
// body-data tokens.
this.boundaryLength = boundary.length + BOUNDARY_PREFIX.length;
if (bufSize < this.boundaryLength + 1) {
throw new IllegalArgumentException("The buffer size specified for the MultipartStream is too small");
}
this.boundary = new byte[this.boundaryLength];
this.keepRegion = this.boundary.length;
System.arraycopy(BOUNDARY_PREFIX, 0, this.boundary, 0, BOUNDARY_PREFIX.length);
System.arraycopy(boundary, 0, this.boundary, BOUNDARY_PREFIX.length, boundary.length);
head = 0;
tail = 0;
}
/**
*
* Constructs a MultipartStream with a default size buffer.
*
* @param input The InputStream to serve as a data source.
* @param boundary The token used for dividing the stream into encapsulations.
* @param pNotifier An object for calling the progress listener, if any.
* @see #MultipartStream(InputStream, byte[], int, ProgressNotifier)
*/
public MultipartStream(InputStream input, byte[] boundary, ProgressNotifier pNotifier) {
this(input, boundary, DEFAULT_BUFSIZE, pNotifier);
}
/**
*
* Constructs a MultipartStream with a default size buffer.
*
* @param input The InputStream to serve as a data source.
* @param boundary The token used for dividing the stream into encapsulations.
*/
public MultipartStream(InputStream input, byte[] boundary) {
this(input, boundary, DEFAULT_BUFSIZE, null);
}
// --------------------------------------------------------- Public methods
/**
* Retrieves the character encoding used when reading the headers of an individual part. When
* not specified, or null, the platform default encoding is used.
*
* @return The encoding used to read part headers.
*/
public String getHeaderEncoding() {
return headerEncoding;
}
/**
* Specifies the character encoding to be used when reading the headers of individual parts.
* When not specified, or null, the platform default encoding is used.
*
* @param encoding The encoding used to read part headers.
*/
public void setHeaderEncoding(String encoding) {
headerEncoding = encoding;
}
/**
* Reads a byte from the buffer, and refills it as necessary.
*
* @return The next byte from the input stream.
* @throws IOException if there is no more data available.
*/
public byte readByte() throws IOException {
// Buffer depleted ?
if (head == tail) {
head = 0;
// Refill.
tail = input.read(buffer, head, bufSize);
if (tail == -1) {
// No more data available.
throw new IOException("No more data is available");
}
if (notifier != null) {
notifier.noteBytesRead(tail);
}
}
return buffer[head++];
}
/**
* Skips a boundary token, and checks whether more encapsulations are
* contained in the stream.
*
* @return true if there are more encapsulations in this stream; false
* otherwise.
* @throws SizeLimitExceededException if the bytes read from the stream exceeded the size limits
* @throws MalformedStreamException if the stream ends unexpectedly or fails to follow required
* syntax.
*/
public boolean readBoundary() throws SizeLimitExceededException, MalformedStreamException {
byte[] marker = new byte[2];
boolean nextChunk = false;
head += boundaryLength;
try {
marker[0] = readByte();
if (marker[0] == LF) {
// Work around IE5 Mac bug with input type=image.
// Because the boundary delimiter, not including the trailing
// CRLF, must not appear within any file (RFC 2046, section
// 5.1.1), we know the missing CR is due to a buggy browser
// rather than a file containing something similar to a
// boundary.
return true;
}
marker[1] = readByte();
if (arrayequals(marker, STREAM_TERMINATOR, 2)) {
nextChunk = false;
}
else if (arrayequals(marker, FIELD_SEPARATOR, 2)) {
nextChunk = true;
}
else {
throw new MalformedStreamException("Unexpected characters follow a boundary");
}
}
catch (SizeLimitExceededException e) {
// wraps a SizeException, re-throw as it will be unwrapped later
throw e;
}
catch (IOException e) {
throw new MalformedStreamException("Stream ended unexpectedly", e);
}
return nextChunk;
}
/**
*
* Changes the boundary token used for partitioning the stream.
*
* This method allows single pass processing of nested multipart streams.
*
* The boundary token of the nested stream is required to be of the same length as
* the boundary token in parent stream.
*
* Restoring the parent stream boundary token after processing of a nested stream is left to the
* application.
*
* @param boundary The boundary to be used for parsing of the nested stream.
* @throws IllegalBoundaryException if the boundary has a different length than the
* one being currently parsed.
*/
public void setBoundary(byte[] boundary) throws IllegalBoundaryException {
if (boundary.length != boundaryLength - BOUNDARY_PREFIX.length) {
throw new IllegalBoundaryException("The length of a boundary token can not be changed");
}
System.arraycopy(boundary, 0, this.boundary, BOUNDARY_PREFIX.length, boundary.length);
}
/**
*
* Reads the header-part of the current encapsulation.
*
* Headers are returned verbatim to the input stream, including the trailing CRLF
* marker. Parsing is left to the application.
*
* TODO allow limiting maximum header size to protect against abuse.
*
* @return The header-part of the current encapsulation.
* @throws SizeLimitExceededException if the bytes read from the stream exceeded the size limits.
* @throws MalformedStreamException if the stream ends unexpectedly.
*/
public String readHeaders() throws SizeLimitExceededException, MalformedStreamException {
int i = 0;
byte b;
// to support multi-byte characters
ByteArrayOutputStream baos = new ByteArrayOutputStream();
int size = 0;
while (i < HEADER_SEPARATOR.length) {
try {
b = readByte();
}
catch (SizeLimitExceededException e) {
// wraps a SizeException, re-throw as it will be unwrapped later
throw e;
}
catch (IOException e) {
throw new MalformedStreamException("Stream ended unexpectedly");
}
if (++size > HEADER_PART_SIZE_MAX) {
throw new MalformedStreamException(format(
"Header section has more than %s bytes (maybe it is not properly terminated)",
Integer.valueOf(HEADER_PART_SIZE_MAX)));
}
if (b == HEADER_SEPARATOR[i]) {
i++;
}
else {
i = 0;
}
baos.write(b);
}
String headers = null;
if (headerEncoding != null) {
try {
headers = baos.toString(headerEncoding);
}
catch (UnsupportedEncodingException e) {
// Fall back to platform default if specified encoding is not
// supported.
headers = baos.toString();
}
}
else {
headers = baos.toString();
}
return headers;
}
/**
*
* Reads body-data from the current encapsulation and writes its
* contents into the output Stream.
*
* Arbitrary large amounts of data can be processed by this method using a constant size buffer.
* (see {@link #MultipartStream(InputStream,byte[],int, MultipartStream.ProgressNotifier)
* constructor}).
*
* @param output The Stream to write data into. May be null, in which case this
* method is equivalent to {@link #discardBodyData()}.
* @return the amount of data written.
* @throws MalformedStreamException if the stream ends unexpectedly.
* @throws IOException if an i/o error occurs.
*/
public int readBodyData(OutputStream output) throws MalformedStreamException, IOException {
final InputStream istream = newInputStream();
return (int)Streams.copy(istream, output);
}
/**
* Creates a new {@link ItemInputStream}.
*
* @return A new instance of {@link ItemInputStream}.
*/
public ItemInputStream newInputStream() {
return new ItemInputStream();
}
/**
*
* Reads body-data from the current encapsulation and discards it.
*
* Use this method to skip encapsulations you don't need or don't understand.
*
* @return The amount of data discarded.
* @throws MalformedStreamException if the stream ends unexpectedly.
* @throws IOException if an i/o error occurs.
*/
public int discardBodyData() throws MalformedStreamException, IOException {
return readBodyData(Streams.nullOutputStream());
}
/**
* Finds the beginning of the first encapsulation.
*
* @return true if an encapsulation was found in the stream.
* @throws IOException if an i/o error occurs.
*/
public boolean skipPreamble() throws IOException {
// First delimiter may be not preceeded with a CRLF.
System.arraycopy(boundary, 2, boundary, 0, boundary.length - 2);
boundaryLength = boundary.length - 2;
try {
// Discard all data up to the delimiter.
discardBodyData();
// Read boundary - if succeeded, the stream contains an
// encapsulation.
return readBoundary();
}
catch (MalformedStreamException e) {
return false;
}
finally {
// Restore delimiter.
System.arraycopy(boundary, 0, boundary, 2, boundary.length - 2);
boundaryLength = boundary.length;
boundary[0] = CR;
boundary[1] = LF;
}
}
/**
* Compares count first bytes in the arrays a and b.
*
* @param a The first array to compare.
* @param b The second array to compare.
* @param count How many bytes should be compared.
* @return true if count first bytes in arrays a and
* b are equal.
*/
public static boolean arrayequals(byte[] a, byte[] b, int count) {
for (int i = 0; i < count; i++) {
if (a[i] != b[i]) {
return false;
}
}
return true;
}
/**
* Searches for a byte of specified value in the buffer, starting at the specified
* position.
*
* @param value The value to find.
* @param pos The starting position for searching.
* @return The position of byte found, counting from beginning of the buffer, or
* -1 if not found.
*/
protected int findByte(byte value, int pos) {
for (int i = pos; i < tail; i++) {
if (buffer[i] == value) {
return i;
}
}
return -1;
}
/**
* Searches for the boundary in the buffer region delimited by
* head and tail.
*
* @return The position of the boundary found, counting from the beginning of the
* buffer, or -1 if not found.
*/
protected int findSeparator() {
int first;
int match = 0;
int maxpos = tail - boundaryLength;
for (first = head; first <= maxpos && match != boundaryLength; first++) {
first = findByte(boundary[0], first);
if (first == -1 || first > maxpos) {
return -1;
}
for (match = 1; match < boundaryLength; match++) {
if (buffer[first + match] != boundary[match]) {
break;
}
}
}
if (match == boundaryLength) {
return first - 1;
}
return -1;
}
/**
* Thrown to indicate that the input stream fails to follow the required syntax.
*/
public static class MalformedStreamException extends IOException {
/**
* The UID to use when serializing this instance.
*/
private static final long serialVersionUID = 6466926458059796677L;
/**
* Constructs a MalformedStreamException with no detail message.
*/
public MalformedStreamException() {
super();
}
/**
* Constructs an MalformedStreamException with the specified detail message.
*
* @param message The detail message.
*/
public MalformedStreamException(String message) {
super(message);
}
/**
* @param message the exception message
* @param cause the cause exception
*/
public MalformedStreamException(String message, Throwable cause) {
super(message, cause);
}
}
/**
* Thrown upon attempt of setting an invalid boundary token.
*/
public static class IllegalBoundaryException extends IOException {
/**
* The UID to use when serializing this instance.
*/
private static final long serialVersionUID = -161533165102632918L;
/**
* Constructs an IllegalBoundaryException with no detail message.
*/
public IllegalBoundaryException() {
super();
}
/**
* Constructs an IllegalBoundaryException with the specified detail message.
*
* @param message The detail message.
*/
public IllegalBoundaryException(String message) {
super(message);
}
}
/**
* An {@link InputStream} for reading an items contents.
*/
public class ItemInputStream extends InputStream {
/**
* The number of bytes, which have been read so far.
*/
private long total;
/**
* The number of bytes, which must be hold, because they might be a part of the boundary.
*/
private int pad;
/**
* The current offset in the buffer.
*/
private int pos;
/**
* Whether the stream is already closed.
*/
private boolean closed;
/**
* Creates a new instance.
*/
ItemInputStream() {
findSeparator();
}
/**
* Called for finding the separator.
*/
private void findSeparator() {
pos = MultipartStream.this.findSeparator();
if (pos == -1) {
if (tail - head > keepRegion) {
pad = keepRegion;
}
else {
pad = tail - head;
}
}
}
/**
* Returns the number of bytes, which have been read by the stream.
*
* @return Number of bytes, which have been read so far.
*/
public long getBytesRead() {
return total;
}
/**
* Returns the number of bytes, which are currently available, without blocking.
*
* @throws IOException An I/O error occurs.
* @return Number of bytes in the buffer.
*/
@Override
public int available() throws IOException {
if (pos == -1) {
return tail - head - pad;
}
return pos - head;
}
/**
* Offset when converting negative bytes to integers.
*/
private static final int BYTE_POSITIVE_OFFSET = 256;
protected void assertClosed() throws IOException {
if (closed) {
throw new IOException("Cannot read a closed input stream!");
}
}
/**
* Returns the next byte in the stream.
*
* @return The next byte in the stream, as a non-negative integer, or -1 for EOF.
* @throws IOException An I/O error occurred.
*/
@Override
public int read() throws IOException {
assertClosed();
if (available() == 0 && makeAvailable() == 0) {
return -1;
}
++total;
int b = buffer[head++];
if (b >= 0) {
return b;
}
return b + BYTE_POSITIVE_OFFSET;
}
/**
* Reads bytes into the given buffer.
*
* @param b The destination buffer, where to write to.
* @param off Offset of the first byte in the buffer.
* @param len Maximum number of bytes to read.
* @return Number of bytes, which have been actually read, or -1 for EOF.
* @throws IOException An I/O error occurred.
*/
@Override
public int read(byte[] b, int off, int len) throws IOException {
assertClosed();
if (len == 0) {
return 0;
}
int res = available();
if (res == 0) {
res = makeAvailable();
if (res == 0) {
return -1;
}
}
res = Math.min(res, len);
System.arraycopy(buffer, head, b, off, res);
head += res;
total += res;
return res;
}
/**
* Closes the input stream.
*
* @throws IOException An I/O error occurred.
*/
@Override
public void close() throws IOException {
close(false);
}
/**
* Closes the input stream.
*
* @param pCloseUnderlying Whether to close the underlying stream (hard close)
* @throws IOException An I/O error occurred.
*/
public void close(boolean pCloseUnderlying) throws IOException {
if (closed) {
return;
}
if (pCloseUnderlying) {
closed = true;
input.close();
}
else {
for (;;) {
int av = available();
if (av == 0) {
av = makeAvailable();
if (av == 0) {
break;
}
}
skip(av);
}
}
closed = true;
}
/**
* Skips the given number of bytes.
*
* @param bytes Number of bytes to skip.
* @return The number of bytes, which have actually been skipped.
* @throws IOException An I/O error occurred.
*/
@Override
public long skip(long bytes) throws IOException {
assertClosed();
int av = available();
if (av == 0) {
av = makeAvailable();
if (av == 0) {
return 0;
}
}
long res = Math.min(av, bytes);
head += res;
return res;
}
/**
* Attempts to read more data.
*
* @return Number of available bytes
* @throws IOException An I/O error occurred.
*/
private int makeAvailable() throws IOException {
if (pos != -1) {
return 0;
}
// Move the data to the beginning of the buffer.
total += tail - head - pad;
System.arraycopy(buffer, tail - pad, buffer, 0, pad);
// Refill buffer with new data.
head = 0;
tail = pad;
for (;;) {
int bytesRead = input.read(buffer, tail, bufSize - tail);
if (bytesRead == -1) {
// The last pad amount is left in the buffer.
// Boundary can't be in there so signal an error
// condition.
final String msg = "Stream ended unexpectedly";
throw new MalformedStreamException(msg);
}
if (notifier != null) {
notifier.noteBytesRead(bytesRead);
}
tail += bytesRead;
findSeparator();
int av = available();
if (av > 0 || pos != -1) {
return av;
}
}
}
/**
* Returns, whether the stream is closed.
*
* @return True, if the stream is closed, otherwise false.
*/
public boolean isClosed() {
return closed;
}
}
}