Java > Open Source Codes > net > myvietnam > mvncore > web > fileupload > MultipartStream


1 /*
2  * Copyright 2001-2005 The Apache Software Foundation
3  *
4  * Licensed under the Apache License, Version 2.0 (the "License");
5  * you may not use this file except in compliance with the License.
6  * You may obtain a copy of the License at
7  *
8  * http://www.apache.org/licenses/LICENSE-2.0
9  *
10  * Unless required by applicable law or agreed to in writing, software
11  * distributed under the License is distributed on an "AS IS" BASIS,
12  * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13  * See the License for the specific language governing permissions and
14  * limitations under the License.
15  */
16 package net.myvietnam.mvncore.web.fileupload;
17
18 import java.io.ByteArrayOutputStream ;
19 import java.io.IOException ;
20 import java.io.InputStream ;
21 import java.io.OutputStream ;
22 import java.io.UnsupportedEncodingException ;
23
24 /**
25  * <p> Low level API for processing file uploads.
26  *
27  * <p> This class can be used to process data streams conforming to MIME
28  * 'multipart' format as defined in
29  * <a HREF="http://www.ietf.org/rfc/rfc1867.txt">RFC 1867</a>. Arbitrarily
30  * large amounts of data in the stream can be processed under constant
31  * memory usage.
32  *
33  * <p> The format of the stream is defined in the following way:<br>
34  *
35  * <code>
36  * multipart-body := preamble 1*encapsulation close-delimiter epilogue<br>
37  * encapsulation := delimiter body CRLF<br>
38  * delimiter := "--" boundary CRLF<br>
39  * close-delimiter := "--" boudary "--"<br>
40  * preamble := &lt;ignore&gt;<br>
41  * epilogue := &lt;ignore&gt;<br>
42  * body := header-part CRLF body-part<br>
43  * header-part := 1*header CRLF<br>
44  * header := header-name ":" header-value<br>
45  * header-name := &lt;printable ascii characters except ":"&gt;<br>
46  * header-value := &lt;any ascii characters except CR & LF&gt;<br>
47  * body-data := &lt;arbitrary data&gt;<br>
48  * </code>
49  *
50  * <p>Note that body-data can contain another mulipart entity. There
51  * is limited support for single pass processing of such nested
52  * streams. The nested stream is <strong>required</strong> to have a
53  * boundary token of the same length as the parent stream (see {@link
54  * #setBoundary(byte[])}).
55  *
56  * <p>Here is an example of usage of this class.<br>
57  *
58  * <pre>
59  * try {
60  * MultipartStream multipartStream = new MultipartStream(input,
61  * boundary);
62  * boolean nextPart = multipartStream.skipPreamble();
63  * OutputStream output;
64  * while(nextPart) {
65  * header = chunks.readHeader();
66  * // process headers
67  * // create some output stream
68  * multipartStream.readBodyPart(output);
69  * nextPart = multipartStream.readBoundary();
70  * }
71  * } catch(MultipartStream.MalformedStreamException e) {
72  * // the stream failed to follow required syntax
73  * } catch(IOException) {
74  * // a read or write error occurred
75  * }
76  *
77  * </pre>
78  *
79  * @author <a HREF="mailto:Rafal.Krzewski@e-point.pl">Rafal Krzewski</a>
80  * @author <a HREF="mailto:martinc@apache.org">Martin Cooper</a>
81  * @author Sean C. Sullivan
82  *
83  * @version $Id: MultipartStream.java,v 1.2 2006/02/12 04:43:11 minhnn Exp $
84  */
85 public class MultipartStream {
86
87     // ----------------------------------------------------- Manifest constants
88
89
90     /**
91      * The Carriage Return ASCII character value.
92      */
93     public static final byte CR = 0x0D;
94
95
96     /**
97      * The Line Feed ASCII character value.
98      */
99     public static final byte LF = 0x0A;
100
101
102     /**
103      * The dash (-) ASCII character value.
104      */
105     public static final byte DASH = 0x2D;
106
107
108     /**
109      * The maximum length of <code>header-part</code> that will be
110      * processed (10 kilobytes = 10240 bytes.).
111      */
112     public static final int HEADER_PART_SIZE_MAX = 10240;
113
114
115     /**
116      * The default length of the buffer used for processing a request.
117      */
118     protected static final int DEFAULT_BUFSIZE = 4096;
119
120
121     /**
122      * A byte sequence that marks the end of <code>header-part</code>
123      * (<code>CRLFCRLF</code>).
124      */
125     protected static final byte[] HEADER_SEPARATOR = {
126             CR, LF, CR, LF };
127
128
129     /**
130      * A byte sequence that that follows a delimiter that will be
131      * followed by an encapsulation (<code>CRLF</code>).
132      */
133     protected static final byte[] FIELD_SEPARATOR = {
134             CR, LF};
135
136
137     /**
138      * A byte sequence that that follows a delimiter of the last
139      * encapsulation in the stream (<code>--</code>).
140      */
141     protected static final byte[] STREAM_TERMINATOR = {
142             DASH, DASH};
143
144
145     /**
146      * A byte sequence that precedes a boundary (<code>CRLF--</code>).
147      */
148     protected static final byte[] BOUNDARY_PREFIX = {
149             CR, LF, DASH, DASH};
150
151
152     /**
153      * The number of bytes, over and above the boundary size, to use for the
154      * keep region.
155      */
156     private static final int KEEP_REGION_PAD = 3;
157
158
159     // ----------------------------------------------------------- Data members
160
161
162     /**
163      * The input stream from which data is read.
164      */
165     private InputStream input;
166
167
168     /**
169      * The length of the boundary token plus the leading <code>CRLF--</code>.
170      */
171     private int boundaryLength;
172
173
174     /**
175      * The amount of data, in bytes, that must be kept in the buffer in order
176      * to detect delimiters reliably.
177      */
178     private int keepRegion;
179
180
181     /**
182      * The byte sequence that partitions the stream.
183      */
184     private byte[] boundary;
185
186
187     /**
188      * The length of the buffer used for processing the request.
189      */
190     private int bufSize;
191
192
193     /**
194      * The buffer used for processing the request.
195      */
196     private byte[] buffer;
197
198
199     /**
200      * The index of first valid character in the buffer.
201      * <br>
202      * 0 <= head < bufSize
203      */
204     private int head;
205
206
207     /**
208      * The index of last valid characer in the buffer + 1.
209      * <br>
210      * 0 <= tail <= bufSize
211      */
212     private int tail;
213
214
215     /**
216      * The content encoding to use when reading headers.
217      */
218     private String headerEncoding;
219
220
221     // ----------------------------------------------------------- Constructors
222
223
224     /**
225      * Default constructor.
226      *
227      * @see #MultipartStream(InputStream, byte[], int)
228      * @see #MultipartStream(InputStream, byte[])
229      *
230      */
231     public MultipartStream() {
232     }
233
234
235     /**
236      * <p> Constructs a <code>MultipartStream</code> with a custom size buffer.
237      *
238      * <p> Note that the buffer must be at least big enough to contain the
239      * boundary string, plus 4 characters for CR/LF and double dash, plus at
240      * least one byte of data. Too small a buffer size setting will degrade
241      * performance.
242      *
243      * @param input The <code>InputStream</code> to serve as a data source.
244      * @param boundary The token used for dividing the stream into
245      * <code>encapsulations</code>.
246      * @param bufSize The size of the buffer to be used, in bytes.
247      *
248      *
249      * @see #MultipartStream()
250      * @see #MultipartStream(InputStream, byte[])
251      *
252      */
253     public MultipartStream(InputStream input,
254                            byte[] boundary,
255                            int bufSize) {
256         this.input = input;
257         this.bufSize = bufSize;
258         this.buffer = new byte[bufSize];
259
260         // We prepend CR/LF to the boundary to chop trailng CR/LF from
261 // body-data tokens.
262 this.boundary = new byte[boundary.length + BOUNDARY_PREFIX.length];
263         this.boundaryLength = boundary.length + BOUNDARY_PREFIX.length;
264         this.keepRegion = boundary.length + KEEP_REGION_PAD;
265         System.arraycopy(BOUNDARY_PREFIX, 0, this.boundary, 0,
266                 BOUNDARY_PREFIX.length);
267         System.arraycopy(boundary, 0, this.boundary, BOUNDARY_PREFIX.length,
268                 boundary.length);
269
270         head = 0;
271         tail = 0;
272     }
273
274
275     /**
276      * <p> Constructs a <code>MultipartStream</code> with a default size buffer.
277      *
278      * @param input The <code>InputStream</code> to serve as a data source.
279      * @param boundary The token used for dividing the stream into
280      * <code>encapsulations</code>.
281      *
282      * @throws IOException when an error occurs.
283      *
284      * @see #MultipartStream()
285      * @see #MultipartStream(InputStream, byte[], int)
286      *
287      */
288     public MultipartStream(InputStream input,
289                            byte[] boundary)
290         throws IOException {
291         this(input, boundary, DEFAULT_BUFSIZE);
292     }
293
294
295     // --------------------------------------------------------- Public methods
296
297
298     /**
299      * Retrieves the character encoding used when reading the headers of an
300      * individual part. When not specified, or <code>null</code>, the platform
301      * default encoding is used.
302
303      *
304      * @return The encoding used to read part headers.
305      */
306     public String getHeaderEncoding() {
307         return headerEncoding;
308     }
309
310
311     /**
312      * Specifies the character encoding to be used when reading the headers of
313      * individual parts. When not specified, or <code>null</code>, the platform
314      * default encoding is used.
315      *
316      * @param encoding The encoding used to read part headers.
317      */
318     public void setHeaderEncoding(String encoding) {
319         headerEncoding = encoding;
320     }
321
322
323     /**
324      * Reads a byte from the <code>buffer</code>, and refills it as
325      * necessary.
326      *
327      * @return The next byte from the input stream.
328      *
329      * @throws IOException if there is no more data available.
330      */
331     public byte readByte()
332         throws IOException {
333         // Buffer depleted ?
334 if (head == tail) {
335             head = 0;
336             // Refill.
337 tail = input.read(buffer, head, bufSize);
338             if (tail == -1) {
339                 // No more data available.
340 throw new IOException ("No more data is available");
341             }
342         }
343         return buffer[head++];
344     }
345
346
347     /**
348      * Skips a <code>boundary</code> token, and checks whether more
349      * <code>encapsulations</code> are contained in the stream.
350      *
351      * @return <code>true</code> if there are more encapsulations in
352      * this stream; <code>false</code> otherwise.
353      *
354      * @throws MalformedStreamException if the stream ends unexpecetedly or
355      * fails to follow required syntax.
356      */
357     public boolean readBoundary()
358         throws MalformedStreamException {
359         byte[] marker = new byte[2];
360         boolean nextChunk = false;
361
362         head += boundaryLength;
363         try {
364             marker[0] = readByte();
365             if (marker[0] == LF) {
366                 // Work around IE5 Mac bug with input type=image.
367 // Because the boundary delimiter, not including the trailing
368 // CRLF, must not appear within any file (RFC 2046, section
369 // 5.1.1), we know the missing CR is due to a buggy browser
370 // rather than a file containing something similar to a
371 // boundary.
372 return true;
373             }
374
375             marker[1] = readByte();
376             if (arrayequals(marker, STREAM_TERMINATOR, 2)) {
377                 nextChunk = false;
378             } else if (arrayequals(marker, FIELD_SEPARATOR, 2)) {
379                 nextChunk = true;
380             } else {
381                 throw new MalformedStreamException(
382                         "Unexpected characters follow a boundary");
383             }
384         } catch (IOException e) {
385             throw new MalformedStreamException("Stream ended unexpectedly");
386         }
387         return nextChunk;
388     }
389
390
391     /**
392      * <p>Changes the boundary token used for partitioning the stream.
393      *
394      * <p>This method allows single pass processing of nested multipart
395      * streams.
396      *
397      * <p>The boundary token of the nested stream is <code>required</code>
398      * to be of the same length as the boundary token in parent stream.
399      *
400      * <p>Restoring the parent stream boundary token after processing of a
401      * nested stream is left to the application.
402      *
403      * @param boundary The boundary to be used for parsing of the nested
404      * stream.
405      *
406      * @throws IllegalBoundaryException if the <code>boundary</code>
407      * has a different length than the one
408      * being currently parsed.
409      */
410     public void setBoundary(byte[] boundary)
411         throws IllegalBoundaryException {
412         if (boundary.length != boundaryLength - BOUNDARY_PREFIX.length) {
413             throw new IllegalBoundaryException(
414                     "The length of a boundary token can not be changed");
415         }
416         System.arraycopy(boundary, 0, this.boundary, BOUNDARY_PREFIX.length,
417                 boundary.length);
418     }
419
420
421     /**
422      * <p>Reads the <code>header-part</code> of the current
423      * <code>encapsulation</code>.
424      *
425      * <p>Headers are returned verbatim to the input stream, including the
426      * trailing <code>CRLF</code> marker. Parsing is left to the
427      * application.
428      *
429      * <p><strong>TODO</strong> allow limiting maximum header size to
430      * protect against abuse.
431      *
432      * @return The <code>header-part</code> of the current encapsulation.
433      *
434      * @throws MalformedStreamException if the stream ends unexpecetedly.
435      */
436     public String readHeaders()
437         throws MalformedStreamException {
438         int i = 0;
439         byte[] b = new byte[1];
440         // to support multi-byte characters
441 ByteArrayOutputStream baos = new ByteArrayOutputStream ();
442         int sizeMax = HEADER_PART_SIZE_MAX;
443         int size = 0;
444         while (i < HEADER_SEPARATOR.length) {
445             try {
446                 b[0] = readByte();
447             } catch (IOException e) {
448                 throw new MalformedStreamException("Stream ended unexpectedly");
449             }
450             size++;
451             if (b[0] == HEADER_SEPARATOR[i]) {
452                 i++;
453             } else {
454                 i = 0;
455             }
456             if (size <= sizeMax) {
457                 baos.write(b[0]);
458             }
459         }
460
461         String headers = null;
462         if (headerEncoding != null) {
463             try {
464                 headers = baos.toString(headerEncoding);
465             } catch (UnsupportedEncodingException e) {
466                 // Fall back to platform default if specified encoding is not
467 // supported.
468 headers = baos.toString();
469             }
470         } else {
471             headers = baos.toString();
472         }
473
474         return headers;
475     }
476
477
478     /**
479      * <p>Reads <code>body-data</code> from the current
480      * <code>encapsulation</code> and writes its contents into the
481      * output <code>Stream</code>.
482      *
483      * <p>Arbitrary large amounts of data can be processed by this
484      * method using a constant size buffer. (see {@link
485      * #MultipartStream(InputStream,byte[],int) constructor}).
486      *
487      * @param output The <code>Stream</code> to write data into.
488      *
489      * @return the amount of data written.
490      *
491      * @throws MalformedStreamException if the stream ends unexpectedly.
492      * @throws IOException if an i/o error occurs.
493      */
494     public int readBodyData(OutputStream output)
495         throws MalformedStreamException,
496                IOException {
497         boolean done = false;
498         int pad;
499         int pos;
500         int bytesRead;
501         int total = 0;
502         while (!done) {
503             // Is boundary token present somewere in the buffer?
504 pos = findSeparator();
505             if (pos != -1) {
506                 // Write the rest of the data before the boundary.
507 output.write(buffer, head, pos - head);
508                 total += pos - head;
509                 head = pos;
510                 done = true;
511             } else {
512                 // Determine how much data should be kept in the
513 // buffer.
514 if (tail - head > keepRegion) {
515                     pad = keepRegion;
516                 } else {
517                     pad = tail - head;
518                 }
519                 // Write out the data belonging to the body-data.
520 output.write(buffer, head, tail - head - pad);
521
522                 // Move the data to the beginning of the buffer.
523 total += tail - head - pad;
524                 System.arraycopy(buffer, tail - pad, buffer, 0, pad);
525
526                 // Refill buffer with new data.
527 head = 0;
528                 bytesRead = input.read(buffer, pad, bufSize - pad);
529
530                 // [pprrrrrrr]
531 if (bytesRead != -1) {
532                     tail = pad + bytesRead;
533                 } else {
534                     // The last pad amount is left in the buffer.
535 // Boundary can't be in there so write out the
536 // data you have and signal an error condition.
537 output.write(buffer, 0, pad);
538                     output.flush();
539                     total += pad;
540                     throw new MalformedStreamException(
541                             "Stream ended unexpectedly");
542                 }
543             }
544         }
545         output.flush();
546         return total;
547     }
548
549
550     /**
551      * <p> Reads <code>body-data</code> from the current
552      * <code>encapsulation</code> and discards it.
553      *
554      * <p>Use this method to skip encapsulations you don't need or don't
555      * understand.
556      *
557      * @return The amount of data discarded.
558      *
559      * @throws MalformedStreamException if the stream ends unexpectedly.
560      * @throws IOException if an i/o error occurs.
561      */
562     public int discardBodyData()
563         throws MalformedStreamException,
564                IOException {
565         boolean done = false;
566         int pad;
567         int pos;
568         int bytesRead;
569         int total = 0;
570         while (!done) {
571             // Is boundary token present somewere in the buffer?
572 pos = findSeparator();
573             if (pos != -1) {
574                 // Write the rest of the data before the boundary.
575 total += pos - head;
576                 head = pos;
577                 done = true;
578             } else {
579                 // Determine how much data should be kept in the
580 // buffer.
581 if (tail - head > keepRegion) {
582                     pad = keepRegion;
583                 } else {
584                     pad = tail - head;
585                 }
586                 total += tail - head - pad;
587
588                 // Move the data to the beginning of the buffer.
589 System.arraycopy(buffer, tail - pad, buffer, 0, pad);
590
591                 // Refill buffer with new data.
592 head = 0;
593                 bytesRead = input.read(buffer, pad, bufSize - pad);
594
595                 // [pprrrrrrr]
596 if (bytesRead != -1) {
597                     tail = pad + bytesRead;
598                 } else {
599                     // The last pad amount is left in the buffer.
600 // Boundary can't be in there so signal an error
601 // condition.
602 total += pad;
603                     throw new MalformedStreamException(
604                             "Stream ended unexpectedly");
605                 }
606             }
607         }
608         return total;
609     }
610
611
612     /**
613      * Finds the beginning of the first <code>encapsulation</code>.
614      *
615      * @return <code>true</code> if an <code>encapsulation</code> was found in
616      * the stream.
617      *
618      * @throws IOException if an i/o error occurs.
619      */
620     public boolean skipPreamble()
621         throws IOException {
622         // First delimiter may be not preceeded with a CRLF.
623 System.arraycopy(boundary, 2, boundary, 0, boundary.length - 2);
624         boundaryLength = boundary.length - 2;
625         try {
626             // Discard all data up to the delimiter.
627 discardBodyData();
628
629             // Read boundary - if succeded, the stream contains an
630 // encapsulation.
631 return readBoundary();
632         } catch (MalformedStreamException e) {
633             return false;
634         } finally {
635             // Restore delimiter.
636 System.arraycopy(boundary, 0, boundary, 2, boundary.length - 2);
637             boundaryLength = boundary.length;
638             boundary[0] = CR;
639             boundary[1] = LF;
640         }
641     }
642
643
644     /**
645      * Compares <code>count</code> first bytes in the arrays
646      * <code>a</code> and <code>b</code>.
647      *
648      * @param a The first array to compare.
649      * @param b The second array to compare.
650      * @param count How many bytes should be compared.
651      *
652      * @return <code>true</code> if <code>count</code> first bytes in arrays
653      * <code>a</code> and <code>b</code> are equal.
654      */
655     public static boolean arrayequals(byte[] a,
656                                       byte[] b,
657                                       int count) {
658         for (int i = 0; i < count; i++) {
659             if (a[i] != b[i]) {
660                 return false;
661             }
662         }
663         return true;
664     }
665
666
667     /**
668      * Searches for a byte of specified value in the <code>buffer</code>,
669      * starting at the specified <code>position</code>.
670      *
671      * @param value The value to find.
672      * @param pos The starting position for searching.
673      *
674      * @return The position of byte found, counting from beginning of the
675      * <code>buffer</code>, or <code>-1</code> if not found.
676      */
677     protected int findByte(byte value,
678                            int pos) {
679         for (int i = pos; i < tail; i++) {
680             if (buffer[i] == value) {
681                 return i;
682             }
683         }
684
685         return -1;
686     }
687
688
689     /**
690      * Searches for the <code>boundary</code> in the <code>buffer</code>
691      * region delimited by <code>head</code> and <code>tail</code>.
692      *
693      * @return The position of the boundary found, counting from the
694      * beginning of the <code>buffer</code>, or <code>-1</code> if
695      * not found.
696      */
697     protected int findSeparator() {
698         int first;
699         int match = 0;
700         int maxpos = tail - boundaryLength;
701         for (first = head;
702              (first <= maxpos) && (match != boundaryLength);
703              first++) {
704             first = findByte(boundary[0], first);
705             if (first == -1 || (first > maxpos)) {
706                 return -1;
707             }
708             for (match = 1; match < boundaryLength; match++) {
709                 if (buffer[first + match] != boundary[match]) {
710                     break;
711                 }
712             }
713         }
714         if (match == boundaryLength) {
715             return first - 1;
716         }
717         return -1;
718     }
719
720     /**
721      * Returns a string representation of this object.
722      *
723      * @return The string representation of this object.
724      */
725     public String toString() {
726         StringBuffer sbTemp = new StringBuffer ();
727         sbTemp.append("boundary='");
728         sbTemp.append(String.valueOf(boundary));
729         sbTemp.append("'\nbufSize=");
730         sbTemp.append(bufSize);
731         return sbTemp.toString();
732     }
733
734     /**
735      * Thrown to indicate that the input stream fails to follow the
736      * required syntax.
737      */
738     public static class MalformedStreamException
739         extends IOException {
740         /**
741          * Constructs a <code>MalformedStreamException</code> with no
742          * detail message.
743          */
744         public MalformedStreamException() {
745             super();
746         }
747
748         /**
749          * Constructs an <code>MalformedStreamException</code> with
750          * the specified detail message.
751          *
752          * @param message The detail message.
753          */
754         public MalformedStreamException(String message) {
755             super(message);
756         }
757     }
758
759
760     /**
761      * Thrown upon attempt of setting an invalid boundary token.
762      */
763     public static class IllegalBoundaryException
764         extends IOException {
765         /**
766          * Constructs an <code>IllegalBoundaryException</code> with no
767          * detail message.
768          */
769         public IllegalBoundaryException() {
770             super();
771         }
772
773         /**
774          * Constructs an <code>IllegalBoundaryException</code> with
775          * the specified detail message.
776          *
777          * @param message The detail message.
778          */
779         public IllegalBoundaryException(String message) {
780             super(message);
781         }
782     }
783
784
785     // ------------------------------------------------------ Debugging methods
786
787
788     // These are the methods that were used to debug this stuff.
789 /*
790
791     // Dump data.
792     protected void dump()
793     {
794         System.out.println("01234567890");
795         byte[] temp = new byte[buffer.length];
796         for(int i=0; i<buffer.length; i++)
797         {
798             if (buffer[i] == 0x0D || buffer[i] == 0x0A)
799             {
800                 temp[i] = 0x21;
801             }
802             else
803             {
804                 temp[i] = buffer[i];
805             }
806         }
807         System.out.println(new String(temp));
808         int i;
809         for (i=0; i<head; i++)
810             System.out.print(" ");
811         System.out.println("h");
812         for (i=0; i<tail; i++)
813             System.out.print(" ");
814         System.out.println("t");
815         System.out.flush();
816     }
817
818     // Main routine, for testing purposes only.
819     //
820     // @param args A String[] with the command line arguments.
821     // @throws Exception, a generic exception.
822     public static void main( String[] args )
823         throws Exception
824     {
825         File boundaryFile = new File("boundary.dat");
826         int boundarySize = (int)boundaryFile.length();
827         byte[] boundary = new byte[boundarySize];
828         FileInputStream input = new FileInputStream(boundaryFile);
829         input.read(boundary,0,boundarySize);
830
831         input = new FileInputStream("multipart.dat");
832         MultipartStream chunks = new MultipartStream(input, boundary);
833
834         int i = 0;
835         String header;
836         OutputStream output;
837         boolean nextChunk = chunks.skipPreamble();
838         while (nextChunk)
839         {
840             header = chunks.readHeaders();
841             System.out.println("!"+header+"!");
842             System.out.println("wrote part"+i+".dat");
843             output = new FileOutputStream("part"+(i++)+".dat");
844             chunks.readBodyData(output);
845             nextChunk = chunks.readBoundary();
846         }
847     }
848
849     */
850 }
851

A to Z: JavaDoc & Examples

---

Daily Java News & Articles

---

Open Source Projects

---

Open Source Codes

---

Free Computer Books

---

Remove Frame

---

Popular Tags