Package org.apache.commons.fileupload2.core

Class MultipartInput

java.lang.Object
org.apache.commons.fileupload2.core.MultipartInput
public final class MultipartInput extends Object
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 setBoundary(byte[])).

Here is an example of usage of this class: 

 try {
     MultipartInput multipartStream = new MultipartInput(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 (MultipartInput.MalformedStreamException e) {
     // the stream failed to follow required syntax
 } catch (IOException e) {
     // a read or write error occurred
 }

  • Field Details

    • CR

      public static final byte CR
      The Carriage Return ASCII character value.
      See Also:

    • LF

      public static final byte LF
      The Line Feed ASCII character value.
      See Also:

    • DASH

      public static final byte DASH
      The dash (-) ASCII character value.
      See Also:

    • HEADER_PART_SIZE_MAX

      public static final int HEADER_PART_SIZE_MAX
      The maximum length of header-part that will be processed (10 kilobytes = 10240 bytes.).
      See Also:

    • DEFAULT_BUFSIZE

      static final int DEFAULT_BUFSIZE
      The default length of the buffer used for processing a request.
      See Also:

    • HEADER_SEPARATOR

      static final byte[] HEADER_SEPARATOR
      A byte sequence that marks the end of header-part (CRLFCRLF).

    • FIELD_SEPARATOR

      static final byte[] FIELD_SEPARATOR
      A byte sequence that that follows a delimiter that will be followed by an encapsulation (CRLF).

    • STREAM_TERMINATOR

      static final byte[] STREAM_TERMINATOR
      A byte sequence that that follows a delimiter of the last encapsulation in the stream (--).

    • BOUNDARY_PREFIX

      static final byte[] BOUNDARY_PREFIX
      A byte sequence that precedes a boundary (CRLF--).

    • input

      private final InputStream input
      The input stream from which data is read.

    • boundaryLength

      private int boundaryLength
      The length of the boundary token plus the leading CRLF--.

    • keepRegion

      private final int keepRegion
      The amount of data, in bytes, that must be kept in the buffer in order to detect delimiters reliably.

    • boundary

      private final byte[] boundary
      The byte sequence that partitions the stream.

    • boundaryTable

      private final int[] boundaryTable
      The table for Knuth-Morris-Pratt search algorithm.

    • bufSize

      private final int bufSize
      The length of the buffer used for processing the request.

    • buffer

      private final byte[] buffer
      The buffer used for processing the request.

    • tail

      private int tail
      The index of last valid character in the buffer + 1.
      0 <= tail <= bufSize

    • headerCharset

      private Charset headerCharset
      The content encoding to use when reading headers.

    • notifier

      private final MultipartInput.ProgressNotifier notifier
      The progress notifier, if any, or null.

  • Constructor Details

    • MultipartInput

      private MultipartInput(InputStream input, byte[] boundary, int bufferSize, MultipartInput.ProgressNotifier notifier)
      Constructs a MultipartInput 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.

      Parameters:
      input - The InputStream to serve as a data source.
      boundary - The token used for dividing the stream into encapsulations.
      bufferSize - The size of the buffer to be used, in bytes.
      notifier - The notifier, which is used for calling the progress listener, if any.
      Throws:
      IllegalArgumentException - If the buffer size is too small.

  • Method Details

    • arrayEquals

      static boolean arrayEquals(byte[] a, byte[] b, int count)
      Compares count first bytes in the arrays a and b.
      Parameters:
      a - The first array to compare.
      b - The second array to compare.
      count - How many bytes should be compared.
      Returns:
      true if count first bytes in arrays a and b are equal.

    • builder

      public static MultipartInput.Builder builder()
      Constructs a new MultipartInput.Builder.
      Returns:
      a new MultipartInput.Builder.

    • computeBoundaryTable

      private void computeBoundaryTable()
      Computes the table used for Knuth-Morris-Pratt search algorithm.

    • discardBodyData

      public long discardBodyData() throws MultipartInput.MalformedStreamException, IOException
      Reads body-data from the current encapsulation and discards it.

      Use this method to skip encapsulations you don't need or don't understand.

      Returns:
      The amount of data discarded.
      Throws:
      MultipartInput.MalformedStreamException - if the stream ends unexpectedly.
      IOException - if an i/o error occurs.

    • findByte

      protected int findByte(byte value, int pos)
      Searches for a byte of specified value in the buffer, starting at the specified position.
      Parameters:
      value - The value to find.
      pos - The starting position for searching.
      Returns:
      The position of byte found, counting from beginning of the buffer, or -1 if not found.

    • findSeparator

      protected int findSeparator()
      Searches for the boundary in the buffer region delimited by head and tail.
      Returns:
      The position of the boundary found, counting from the beginning of the buffer, or -1 if not found.

    • getHeaderCharset

      public Charset getHeaderCharset()
      Gets the character encoding used when reading the headers of an individual part. When not specified, or null, the platform default encoding is used.
      Returns:
      The encoding used to read part headers.

    • newInputStream

      public MultipartInput.ItemInputStream newInputStream()
      Creates a new MultipartInput.ItemInputStream.
      Returns:
      A new instance of MultipartInput.ItemInputStream.

    • readBodyData

      public long readBodyData(OutputStream output) throws MultipartInput.MalformedStreamException, IOException
      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 builder()).

      Parameters:
      output - The Stream to write data into. May be null, in which case this method is equivalent to discardBodyData().
      Returns:
      the amount of data written.
      Throws:
      MultipartInput.MalformedStreamException - if the stream ends unexpectedly.
      IOException - if an i/o error occurs.

    • readBoundary

      public boolean readBoundary() throws FileUploadSizeException, MultipartInput.MalformedStreamException
      Skips a boundary token, and checks whether more encapsulations are contained in the stream.
      Returns:
      true if there are more encapsulations in this stream; false otherwise.
      Throws:
      FileUploadSizeException - if the bytes read from the stream exceeded the size limits
      MultipartInput.MalformedStreamException - if the stream ends unexpectedly or fails to follow required syntax.

    • readByte

      public byte readByte() throws IOException
      Reads a byte from the buffer, and refills it as necessary.
      Returns:
      The next byte from the input stream.
      Throws:
      IOException - if there is no more data available.

    • readHeaders

      public String readHeaders() throws FileUploadSizeException, MultipartInput.MalformedStreamException
      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.

      Returns:
      The header-part of the current encapsulation.
      Throws:
      FileUploadSizeException - if the bytes read from the stream exceeded the size limits.
      MultipartInput.MalformedStreamException - if the stream ends unexpectedly.

    • setBoundary

      public void setBoundary(byte[] boundary) throws MultipartInput.FileUploadBoundaryException
      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.

      Parameters:
      boundary - The boundary to be used for parsing of the nested stream.
      Throws:
      MultipartInput.FileUploadBoundaryException - if the boundary has a different length than the one being currently parsed.

    • setHeaderCharset

      public void setHeaderCharset(Charset headerCharset)
      Sets the character encoding to be used when reading the headers of individual parts. When not specified, or null, the platform default encoding is used.
      Parameters:
      headerCharset - The encoding used to read part headers.

    • skipPreamble

      public boolean skipPreamble() throws IOException
      Finds the beginning of the first encapsulation.
      Returns:
      true if an encapsulation was found in the stream.
      Throws:
      IOException - if an i/o error occurs.