[Archive-access-cvs] SF.net SVN: archive-access:[3034] trunk/archive-access/projects/wayback/ wayback-core/src/main/java/org/archive/wayback/core/Resource.java

[<bra...@us...>]

Revision: 3034
          http://archive-access.svn.sourceforge.net/archive-access/?rev=3034&view=rev
Author:   bradtofel
Date:     2010-04-14 18:44:53 +0000 (Wed, 14 Apr 2010)

Log Message:
-----------
BUGFIX(unreported): chunked peek-ahead was looking for 10-13 line-ending, not 13-10. Now it allows either, as well as a lone 10.
COMMENT/JAVADOC update

Modified Paths:
--------------
    trunk/archive-access/projects/wayback/wayback-core/src/main/java/org/archive/wayback/core/Resource.java

Modified: trunk/archive-access/projects/wayback/wayback-core/src/main/java/org/archive/wayback/core/Resource.java
===================================================================
--- trunk/archive-access/projects/wayback/wayback-core/src/main/java/org/archive/wayback/core/Resource.java	2010-04-14 02:33:08 UTC (rev 3033)
+++ trunk/archive-access/projects/wayback/wayback-core/src/main/java/org/archive/wayback/core/Resource.java	2010-04-14 18:44:53 UTC (rev 3034)
@@ -34,6 +34,13 @@
  * Abstraction on top of a document stored in a WaybackCollection. Currently
  * implemented subclasses include ArcResource and WarcResource.
  * 
+ * This implementation needs some pretty drastic refactoring.. May have to wait
+ * for 2.0. This should be a byte-oriented record, and allow wrapping the 
+ * interior byte-stream in on the more full featured HTTP libraries 
+ * (jetty/apache-http-client/w3c-http-reference).
+ * 
+ * For now, it is a system-wide assumption that all resources are HTTP based.
+ * 
  * @author Brad Tofel
  * @version $Date$, $Revision$
  */
@@ -42,8 +49,20 @@
 	private InputStream is;
 
     public abstract void close() throws IOException;
+	/**
+	 * Assumes an HTTP resource - return the HTTP response code
+	 * @return the HTTP response code from the HTTP message
+	 */
 	public abstract int getStatusCode();
+	/**
+	 * @return the size in bytes of the record payload, including HTTP header
+	 */
 	public abstract long getRecordLength();
+	/**
+	 * Assumes an HTTP response - return the HTTP headers, not including the
+	 * HTTP Message header
+	 * @return key-value Map of HTTP headers 
+	 */
 	public abstract Map<String,String> getHttpHeaders();
 
 	private void validate() throws IOException {
@@ -59,10 +78,13 @@
 			this.is = new BufferedInputStream(is);
 		}
 	}
+
 	/**
 	 * indicate that there is a Transfer-Encoding: chunked header, so the input
-	 *   data should be dechunked as it is read.
-	 * @throws IOException
+	 *   data should be dechunked as it is read. This method actually peeks
+	 *   ahead to verify that there is a hex-encoded chunk length before
+	 *   assuming the data is chunked.
+	 * @throws IOException for usual reasons
 	 */
 	public void setChunkedEncoding() throws IOException {
 		validate();
@@ -70,21 +92,31 @@
 		int max = 50;
 		is.mark(max+2);
 		int cur = 0;
+		int hexFound = 0;
 		boolean isChunked = false;
 		while(cur < max) {
 			int nextC = is.read();
-			if(nextC == 10) {
+			// allow CRLF and plain ole LF:
+			if((nextC == 13) || (nextC == 10)) {
 				// must have read at least 1 hex char:
-				if(cur > 0) {
+				if(hexFound > 0) {
+					if(nextC == 10) {
+						isChunked = true;
+						break;
+					}
 					nextC = is.read();
-					if(nextC == 13) {
+					if(nextC == 10) {
 						isChunked = true;
 						break;
 					}
 				}
+				// keep looking to allow some blank lines. 
 			} else {
 				// better be a hex character:
-				if(!isHex(nextC)) {
+				if(isHex(nextC)) {
+					hexFound++;
+				} else {
+					// not a hex digit: not a chunked stream.
 					break;
 				}
 			}
@@ -108,80 +140,45 @@
 		}
 		return false;
 	}
-	
-	/**
-	 * @return
-	 * @throws IOException
-	 * @see java.io.BufferedInputStream#available()
-	 */
+
 	public int available() throws IOException {
 		validate();
 		return is.available();
 	}
-	/**
-	 * @param readlimit
-	 * @see java.io.BufferedInputStream#mark(int)
-	 */
+
 	public void mark(int readlimit) {
 		if(is != null) {
 			is.mark(readlimit);
 		}
 	}
-	/**
-	 * @return
-	 * @see java.io.BufferedInputStream#markSupported()
-	 */
+
 	public boolean markSupported() {
 		if(is == null) {
 			return false;
 		}
 		return is.markSupported();
 	}
-	/**
-	 * @return
-	 * @throws IOException
-	 * @see java.io.BufferedInputStream#read()
-	 */
+
 	public int read() throws IOException {
 		validate();
 		return is.read();
 	}
-	/**
-	 * @param b
-	 * @param off
-	 * @param len
-	 * @return
-	 * @throws IOException
-	 * @see java.io.BufferedInputStream#read(byte[], int, int)
-	 */
+
 	public int read(byte[] b, int off, int len) throws IOException {
 		validate();
 		return is.read(b, off, len);
 	}
-	/**
-	 * @param b
-	 * @return
-	 * @throws IOException
-	 * @see java.io.FilterInputStream#read(byte[])
-	 */
+
 	public int read(byte[] b) throws IOException {
 		validate();
 		return is.read(b);
 	}
-	/**
-	 * @throws IOException
-	 * @see java.io.BufferedInputStream#reset()
-	 */
+
 	public void reset() throws IOException {
 		validate();
 		is.reset();
 	}
-	/**
-	 * @param n
-	 * @return
-	 * @throws IOException
-	 * @see java.io.BufferedInputStream#skip(long)
-	 */
+
 	public long skip(long n) throws IOException {
 		validate();
 		return is.skip(n);


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.