[kaffe] CVS kaffe (robilad): resynced with gnu classpath: uri fixes

Kaffe CVS cvs-commits at kaffe.org
Sat May 21 08:41:43 PDT 2005


PatchSet 6565 
Date: 2005/05/21 15:26:33
Author: robilad
Branch: HEAD
Tag: (none) 
Log:
resynced with gnu classpath: uri fixes

Members: 
	ChangeLog:1.4092->1.4093 
	libraries/javalib/java/net/URI.java:1.14->1.15 

Index: kaffe/ChangeLog
diff -u kaffe/ChangeLog:1.4092 kaffe/ChangeLog:1.4093
--- kaffe/ChangeLog:1.4092	Sat May 21 15:13:35 2005
+++ kaffe/ChangeLog	Sat May 21 15:26:33 2005
@@ -2,6 +2,47 @@
 
 	Resynced with GNU Classpath.
 
+	2005-05-19  Andrew John Hughes  <gnu_andrew at member.fsf.org>
+
+        * java/net/URI.java:
+        Added more documentation.
+        (RFC2396_MARK): Removed.
+        (RFC2396_UNRESERVED): Changed to RFC3986_UNRESERVED and updated.
+        (RFC2396_REG_NAME): Changed to RFC3986_REG_NAME and updated.
+        (RFC2396_PCHAR): Changed to RFC3986_PCHAR and updated.
+        (RFC2396_SEGMENT): Changed to RFC3986_SEGMENT and updated.
+        (RFC2396_PATH_SEGMENTS): Changed to RFC3986_PATH_SEGMENTS.
+        (RFC3986_UNRESERVED): New field.
+        (RFC3986_SSP): New field.
+        (RFC3986_HOST): New field.
+        (RFC3986_USERINFO): New field.
+        (static): New initializer to initialize patterns with class.
+        (parseURI()): Moved authority parsing to parseServerAuthority().
+        (unquote(String)): Removed invalid exception for non-ASCII chars.
+        (quote(String)): Implemented.
+        (quoteAuthority(String)): Adapted to use new fields.
+        (quote(String,String)): Moved escaping of characters to another
+        method.
+        (quoteHost(String)): Implemented.
+        (quotePath(String)): Adapted to use new fields.
+        (quoteUserInfo(String)): Implemented.
+        (parseServerAuthority()): Implemented.
+        (normalize()): Implemented.
+        (normalizePath(String)): Implemented as part of normalize().
+        (removeLastSegment(StringBuffer)): Likewise.
+        (relativize(java.net.URI)): Implemented.
+        (equals(Object)): Implemented.
+        (hashCode()): Implemented.
+        (compareTo(Object)): Implemented.
+        (compareFragments(java.net.URI)): Implemented.
+        (toString()): Use fields directly.
+        (toASCIIString()): Implemented.
+        (escapeCharacters(String)): Implemented to escape non-ASCII characters.
+
+2005-05-21  Dalibor Topic  <robilad at kaffe.org>
+
+	Resynced with GNU Classpath.
+
 	2005-05-20  Roman Kennke  <roman at kennke.org>
 
         * java/awt/DefaultKeyboardFocusManager.java:
Index: kaffe/libraries/javalib/java/net/URI.java
diff -u kaffe/libraries/javalib/java/net/URI.java:1.14 kaffe/libraries/javalib/java/net/URI.java:1.15
--- kaffe/libraries/javalib/java/net/URI.java:1.14	Wed May 18 21:34:09 2005
+++ kaffe/libraries/javalib/java/net/URI.java	Sat May 21 15:26:37 2005
@@ -48,7 +48,7 @@
 /**
  * <p>
  * A URI instance represents that defined by 
- * <a href="http://www.ietf.org/rfc/rfc3986.txt">RFC2396</a>,
+ * <a href="http://www.ietf.org/rfc/rfc3986.txt">RFC3986</a>,
  * with some deviations.
  * </p>
  * <p>
@@ -98,7 +98,57 @@
  * this that the path sub-part may also not be undefined, so as to ensure
  * the former.
  * </p>
- * 
+ * <h2>Character Escaping and Quoting</h2>
+ * <p>
+ * The characters that can be used within a valid URI are restricted.
+ * There are two main classes of characters which can't be used as is
+ * within the URI:
+ * </p>
+ * <ol>
+ * <li><strong>Characters outside the US-ASCII character set</strong>.
+ * These have to be <strong>escaped</strong> in order to create
+ * an RFC-compliant URI; this means replacing the character with the
+ * appropriate hexadecimal value, preceded by a `%'.</li>
+ * <li><strong>Illegal characters</strong> (e.g. space characters,
+ * control characters) are quoted, which results in them being encoded
+ * in the same way as non-US-ASCII characters.</li>
+ * </ol>
+ * <p>
+ * The set of valid characters differs depending on the section of the URI:
+ * </p>
+ * <ul>
+ * <li><strong>Scheme</strong>: Must be an alphanumeric, `-', `.' or '+'.</li>
+ * <li><strong>Authority</strong>:Composed of the username, host, port, `@'
+ * and `:'.</li>
+ * <li><strong>Username</strong>: Allows unreserved or percent-encoded
+ * characters, sub-delimiters and `:'.</li>
+ * <li><strong>Host</strong>: Allows unreserved or percent-encoded
+ * characters, sub-delimiters and square brackets (`[' and `]') for IPv6
+ * addresses.</li>
+ * <li><strong>Port</strong>: Digits only.</li>
+ * <li><strong>Path</strong>: Allows the path characters and `/'.
+ * <li><strong>Query</strong>: Allows the path characters, `?' and '/'.
+ * <li><strong>Fragment</strong>: Allows the path characters, `?' and '/'.
+ * </ul>
+ * <p>
+ * These definitions reference the following sets of characters:
+ * </p>
+ * <ul>
+ * <li><strong>Unreserved characters</strong>: The alphanumerics plus
+ * `-', `.', `_', and `~'.</li>
+ * <li><strong>Sub-delimiters</strong>: `!', `$', `&', `(', `)', `*',
+ * `+', `,', `;', `=' and the single-quote itself.</li>
+ * <li><strong>Path characters</strong>: Unreserved and percent-encoded
+ * characters and the sub-delimiters along with `@' and `:'.</li>
+ * </ul>
+ * <p>
+ * The constructors and accessor methods allow the use and retrieval of
+ * URI components which contain non-US-ASCII characters directly.
+ * They are only escaped when the <code>toASCIIString()</code> method
+ * is used.  In contrast, illegal characters are always quoted, with the
+ * exception of the return values of the non-raw accessors.
+ * </p>
+ *
  * @author Ito Kazumitsu (ito.kazumitsu at hitachi-cable.co.jp)
  * @author Dalibor Topic (robilad at kaffe.org)
  * @author Michael Koch (konqueror at gmx.de)
@@ -108,6 +158,9 @@
 public final class URI 
   implements Comparable, Serializable
 {
+  /**
+   * For serialization compatability.
+   */
   static final long serialVersionUID = -6052424284110960213L;
 
   /**
@@ -119,11 +172,14 @@
   private static final String URI_REGEXP =
     "^(([^:/?#]+):)?((//([^/?#]*))?([^?#]*)(\\?([^#]*))?)?(#(.*))?";
 
+  /**
+   * Regular expression for parsing the authority segment.
+   */
   private static final String AUTHORITY_REGEXP =
-    "(([^?#]*)@)?([^?#:]*)(:([^?#]*))?";
+    "(([^?#]*)@)?([^?#:]*)(:([0-9]*))?";
 
   /**
-   * Valid characters (taken from rfc2396)
+   * Valid characters (taken from rfc2396/3986)
    */
   private static final String RFC2396_DIGIT = "0123456789";
   private static final String RFC2396_LOWALPHA = "abcdefghijklmnopqrstuvwxyz";
@@ -131,14 +187,17 @@
   private static final String RFC2396_ALPHA =
     RFC2396_LOWALPHA + RFC2396_UPALPHA;
   private static final String RFC2396_ALPHANUM = RFC2396_DIGIT + RFC2396_ALPHA;
-  private static final String RFC2396_MARK = "-_.!~*'()";
-  private static final String RFC2396_UNRESERVED =
-    RFC2396_ALPHANUM + RFC2396_MARK;
-  private static final String RFC2396_REG_NAME =
-    RFC2396_UNRESERVED + "$,;:@&=+";
-  private static final String RFC2396_PCHAR = RFC2396_UNRESERVED + ":@&=+$,";
-  private static final String RFC2396_SEGMENT = RFC2396_PCHAR + ";";
-  private static final String RFC2396_PATH_SEGMENTS = RFC2396_SEGMENT + "/";
+  private static final String RFC3986_UNRESERVED = RFC2396_ALPHANUM + "-._~";
+  private static final String RFC3986_SUBDELIMS = "!$&'()*+,;=";
+  private static final String RFC3986_REG_NAME =
+    RFC3986_UNRESERVED + RFC3986_SUBDELIMS + "%";
+  private static final String RFC3986_PCHAR = RFC3986_UNRESERVED + 
+    RFC3986_SUBDELIMS + ":@%";
+  private static final String RFC3986_SEGMENT = RFC3986_PCHAR;
+  private static final String RFC3986_PATH_SEGMENTS = RFC3986_SEGMENT + "/";
+  private static final String RFC3986_SSP = RFC3986_PCHAR + "?/";
+  private static final String RFC3986_HOST = RFC3986_REG_NAME + "[]";
+  private static final String RFC3986_USERINFO = RFC3986_REG_NAME + ":";
 
   /**
    * Index of scheme component in parsed URI.
@@ -170,10 +229,36 @@
    */
   private static final int FRAGMENT_GROUP = 10;
   
+  /**
+   * Index of userinfo component in parsed authority section.
+   */
   private static final int AUTHORITY_USERINFO_GROUP = 2;
+
+  /**
+   * Index of host component in parsed authority section.
+   */
   private static final int AUTHORITY_HOST_GROUP = 3;
+
+  /**
+   * Index of port component in parsed authority section.
+   */
   private static final int AUTHORITY_PORT_GROUP = 5;
-  
+
+  /**
+   * The compiled version of the URI regular expression.
+   */
+  private static final Pattern URI_PATTERN;
+
+  /**
+   * The compiled version of the authority regular expression.
+   */
+  private static final Pattern AUTHORITY_PATTERN;
+
+  /**
+   * The set of valid hexadecimal characters.
+   */
+  private static final String HEX = "0123456789ABCDEF";
+
   private transient String scheme;
   private transient String rawSchemeSpecificPart;
   private transient String schemeSpecificPart;
@@ -192,6 +277,15 @@
   private transient String fragment;
   private String string;
 
+  /**
+   * Static initializer to pre-compile the regular expressions.
+   */
+  static
+  {
+    URI_PATTERN = Pattern.compile(URI_REGEXP);
+    AUTHORITY_PATTERN = Pattern.compile(AUTHORITY_REGEXP);
+  }
+
   private void readObject(ObjectInputStream is)
     throws ClassNotFoundException, IOException
   {
@@ -229,8 +323,7 @@
    */
   private void parseURI(String str) throws URISyntaxException
   {
-    Pattern pattern = Pattern.compile(URI_REGEXP);
-    Matcher matcher = pattern.matcher(str);
+    Matcher matcher = URI_PATTERN.matcher(str);
     
     if (matcher.matches())
       {
@@ -246,37 +339,9 @@
 	rawFragment = getURIGroup(matcher, FRAGMENT_GROUP);
       }
     else
-      throw new URISyntaxException(str, "doesn't match URI regular expression");
-
-    if (rawAuthority != null)
-      {
-	pattern = Pattern.compile(AUTHORITY_REGEXP);
-	matcher = pattern.matcher(rawAuthority);
-
-	if (matcher.matches())
-	  {
-	    rawUserInfo = getURIGroup(matcher, AUTHORITY_USERINFO_GROUP);
-	    rawHost = getURIGroup(matcher, AUTHORITY_HOST_GROUP);
-
-	    String portStr = getURIGroup(matcher, AUTHORITY_PORT_GROUP);
-
-	    if (portStr != null)
-	      try
-		{
-		  port = Integer.parseInt(portStr);
-		}
-	      catch (NumberFormatException e)
-		{
-		  URISyntaxException use =
-		    new URISyntaxException
-		      (str, "doesn't match URI regular expression");
-		  use.initCause(e);
-		  throw use;
-		}
-	  }
-	else
-	  throw new URISyntaxException(str, "doesn't match URI regular expression");
-      }
+      throw new URISyntaxException(str,
+				   "doesn't match URI regular expression");
+    parseServerAuthority();
 
     // We must eagerly unquote the parts, because this is the only time
     // we may throw an exception.
@@ -307,8 +372,6 @@
     for (int i = 0; i < str.length(); i++)
       {
 	char c = str.charAt(i);
-	if (c > 127)
-	  throw new URISyntaxException(str, "Invalid character");
 	if (c == '%')
 	  {
 	    if (i + 2 >= str.length())
@@ -345,8 +408,7 @@
    */
   private static String quote(String str)
   {
-    // FIXME: unimplemented.
-    return str;
+    return quote(str, RFC3986_SSP);
   }
 
   /**
@@ -364,20 +426,17 @@
   {
     // Technically, we should be using RFC2396_AUTHORITY, but
     // it contains no additional characters.
-    return quote(str, RFC2396_REG_NAME);
+    return quote(str, RFC3986_REG_NAME);
   }
 
   /**
-   * Quote characters in str that are not part of legalCharacters.
-   *
-   * Replace illegal characters by encoding their UTF-8
-   * representation as "%" + hex code for each resulting
-   * UTF-8 character.
+   * Quotes the characters in the supplied string that are not part of
+   * the specified set of legal characters.
    *
-   * @param str The string to quote
-   * @param legalCharacters The set of legal characters
+   * @param str the string to quote
+   * @param legalCharacters the set of legal characters
    *
-   * @return The quoted string.
+   * @return the quoted string.
    */
   private static String quote(String str, String legalCharacters)
   {
@@ -387,23 +446,11 @@
 	char c = str.charAt(i);
 	if (legalCharacters.indexOf(c) == -1)
 	  {
-	    String hex = "0123456789ABCDEF";
 	    if (c <= 127)
-	      sb.append('%').append(hex.charAt(c / 16)).append(hex.charAt(c % 16));
-	    else
 	      {
-		try
-		  {
-		    // this is far from optimal, but it works
-		    byte[] utf8 = str.substring(i, i + 1).getBytes("utf-8");
-		    for (int j = 0; j < utf8.length; j++)
-		      sb.append('%').append(hex.charAt((utf8[j] & 0xff) / 16))
-		        .append(hex.charAt((utf8[j] & 0xff) % 16));
-		  }
-		catch (java.io.UnsupportedEncodingException x)
-		  {
-		    throw (Error) new InternalError().initCause(x);
-		  }
+		sb.append('%');
+		sb.append(HEX.charAt(c / 16));
+		sb.append(HEX.charAt(c % 16));
 	      }
 	  }
 	else
@@ -425,8 +472,7 @@
    */
   private static String quoteHost(String str)
   {
-    // FIXME: unimplemented.
-    return str;
+    return quote(str, RFC3986_HOST);
   }
 
   /**
@@ -444,7 +490,7 @@
   {
     // Technically, we should be using RFC2396_PATH, but
     // it contains no additional characters.
-    return quote(str, RFC2396_PATH_SEGMENTS);
+    return quote(str, RFC3986_PATH_SEGMENTS);
   }
 
   /**
@@ -460,8 +506,7 @@
    */
   private static String quoteUserInfo(String str)
   {
-    // FIXME: unimplemented.
-    return str;
+    return quote(str, RFC3986_USERINFO);
   }
 
   /**
@@ -503,8 +548,6 @@
          + (path == null ? "" : quotePath(path))
          + (query == null ? "" : "?" + quote(query))
          + (fragment == null ? "" : "#" + quote(fragment)));
-
-    parseServerAuthority();
   }
 
   /**
@@ -584,21 +627,198 @@
 
   /**
    * Attempts to parse this URI's authority component, if defined,
-   * into user-information, host, and port components
-   *
-   * @exception URISyntaxException If the given string violates RFC 2396
+   * into user-information, host, and port components.  The purpose
+   * of this method was to disambiguate between some authority sections,
+   * which form invalid server-based authories, but valid registry
+   * based authorities.  In the updated RFC 3986, the authority section
+   * is defined differently, with registry-based authorities part of
+   * the host section.  Thus, this method is now simply an explicit
+   * way of parsing any authority section.
+   *
+   * @return the URI, with the authority section parsed into user
+   *         information, host and port components.
+   * @throws URISyntaxException if the given string violates RFC 2396
    */
   public URI parseServerAuthority() throws URISyntaxException
   {
-    return null;
+    if (rawAuthority != null)
+      {
+	Matcher matcher = AUTHORITY_PATTERN.matcher(rawAuthority);
+
+	if (matcher.matches())
+	  {
+	    rawUserInfo = getURIGroup(matcher, AUTHORITY_USERINFO_GROUP);
+	    rawHost = getURIGroup(matcher, AUTHORITY_HOST_GROUP);
+	    
+	    String portStr = getURIGroup(matcher, AUTHORITY_PORT_GROUP);
+	    
+	    if (portStr != null)
+	      try
+		{
+		  port = Integer.parseInt(portStr);
+		}
+	      catch (NumberFormatException e)
+		{
+		  URISyntaxException use =
+		    new URISyntaxException
+		      (string, "doesn't match URI regular expression");
+		  use.initCause(e);
+		  throw use;
+		}
+	  }
+	else
+	  throw new URISyntaxException(string,
+				       "doesn't match URI regular expression");
+      }
+    return this;
   }
 
   /**
-   * Returns a normalizes versions of the URI
+   * <p>
+   * Returns a normalized version of the URI.  If the URI is opaque,
+   * or its path is already in normal form, then this URI is simply
+   * returned.  Otherwise, the following transformation of the path
+   * element takes place:
+   * </p>
+   * <ol>
+   * <li>All `.' segments are removed.</li>
+   * <li>Each `..' segment which can be paired with a prior non-`..' segment
+   * is removed along with the preceding segment.</li>
+   * <li>A `.' segment is added to the front if the first segment contains
+   * a colon (`:').  This is a deviation from the RFC, which prevents
+   * confusion between the path and the scheme.</li>
+   * </ol>
+   * <p>
+   * The resulting URI will be free of `.' and `..' segments, barring those
+   * that were prepended or which couldn't be paired, respectively.
+   * </p>
+   *
+   * @return the normalized URI.
    */
   public URI normalize()
   {
-    return null;
+    if (isOpaque() || path.indexOf("/./") == -1 && path.indexOf("/../") == -1)
+      return this;
+    try
+      {
+	return new URI(scheme, authority, normalizePath(path), query,
+		       fragment);
+      }
+    catch (URISyntaxException e)
+      {
+	throw (Error) new InternalError("Normalized URI variant could not "+
+					"be constructed").initCause(e);
+      }
+  }
+
+  /**
+   * <p>
+   * Normalize the given path.  The following transformation takes place:
+   * </p>
+   * <ol>
+   * <li>All `.' segments are removed.</li>
+   * <li>Each `..' segment which can be paired with a prior non-`..' segment
+   * is removed along with the preceding segment.</li>
+   * <li>A `.' segment is added to the front if the first segment contains
+   * a colon (`:').  This is a deviation from the RFC, which prevents
+   * confusion between the path and the scheme.</li>
+   * </ol>
+   * <p>
+   * The resulting URI will be free of `.' and `..' segments, barring those
+   * that were prepended or which couldn't be paired, respectively.
+   * </p>
+   * 
+   * @param relativePath the relative path to be normalized.
+   * @return the normalized path.
+   */
+  private String normalizePath(String relativePath)
+  {
+    /* 
+       This follows the algorithm in section 5.2.4. of RFC3986,
+       but doesn't modify the input buffer.
+    */
+    StringBuffer input = new StringBuffer(relativePath);
+    StringBuffer output = new StringBuffer();
+    int start = 0;
+    while (start < input.length())
+      {
+	/* A */
+	if (input.indexOf("../",start) == start)
+	  {
+	    start += 3;
+	    continue;
+	  }
+	if (input.indexOf("./",start) == start)
+	  {
+	    start += 2;
+	    continue;
+	  }
+	/* B */
+	if (input.indexOf("/./",start) == start)
+	  {
+	    start += 2;
+	    continue;
+	  }
+	if (input.indexOf("/.",start) == start
+	    && input.charAt(start + 2) != '.')
+	  {
+	    start += 1;
+	    input.setCharAt(start,'/');
+	    continue;
+	  }
+	/* C */
+	if (input.indexOf("/../",start) == start)
+	  {
+	    start += 3;
+	    removeLastSegment(output);
+	    continue;
+	  }
+	if (input.indexOf("/..",start) == start)
+	  {
+	    start += 2;
+	    input.setCharAt(start,'/');
+	    removeLastSegment(output);
+	    continue;
+	  }
+	/* D */
+	if (start == input.length() - 1 && input.indexOf(".",start) == start)
+	  {
+	    input.delete(0,1);
+	    continue;
+	  }
+	if (start == input.length() - 2 && input.indexOf("..",start) == start)
+	  {
+	    input.delete(0,2);
+	    continue;
+	  }
+	/* E */
+	int indexOfSlash = input.indexOf("/",start);
+	while (indexOfSlash == start)
+	  {
+	    output.append("/");
+	    ++start;
+	    indexOfSlash = input.indexOf("/",start);
+	  }
+	if (indexOfSlash == -1)
+	  indexOfSlash = input.length();
+	output.append(input.substring(start, indexOfSlash));
+        start = indexOfSlash;
+      }
+    return output.toString();
+  }
+
+  /**
+   * Removes the last segment of the path from the specified buffer.
+   *
+   * @param buffer the buffer containing the path.
+   */
+  private void removeLastSegment(StringBuffer buffer)
+  {
+    int lastSlash = buffer.lastIndexOf("/");
+    if (lastSlash == -1)
+      buffer.setLength(0);
+    else
+      buffer.setLength(lastSlash);
   }
 
   /**
@@ -609,7 +829,7 @@
    * @return The resulting URI, or null when it couldn't be resolved
    * for some reason.
    *
-   * @exception NullPointerException If uri is null
+   * @throws NullPointerException if uri is null
    */
   public URI resolve(URI uri)
   {
@@ -645,16 +865,15 @@
 		  basepath.delete(i + 1, basepath.length());
 
 		basepath.append(path);
-		path = basepath.toString();
-		//  FIXME We must normalize the path here.
-		//  Normalization process omitted.
+		path = normalizePath(basepath.toString());
 	      }
 	  }
 	return new URI(this.scheme, authority, path, query, fragment);
       }
     catch (URISyntaxException e)
       {
-	return null;
+	throw (Error) new InternalError("Resolved URI variant could not "+
+					"be constructed").initCause(e);
       }
   }
 
@@ -665,9 +884,9 @@
    *
    * @return The resulting URI
    *
-   * @exception IllegalArgumentException If the given URI string
+   * @throws IllegalArgumentException If the given URI string
    * violates RFC 2396
-   * @exception NullPointerException If uri is null
+   * @throws NullPointerException If uri is null
    */
   public URI resolve(String str) throws IllegalArgumentException
   {
@@ -675,25 +894,60 @@
   }
 
   /**
-   * Relativizes the given URI against this URI
-   *
-   * @param uri The URI to relativize this URI
-   *
-   * @return The resulting URI
-   *
-   * @exception NullPointerException If uri is null
+   * <p>
+   * Relativizes the given URI against this URI using the following
+   * algorithm:
+   * </p>
+   * <ul>
+   * <li>If either URI is opaque, the given URI is returned.</li>
+   * <li>If the schemes of the URIs differ, the given URI is returned.</li>
+   * <li>If the authority components of the URIs differ, then the given
+   * URI is returned.</li>
+   * <li>If the path of this URI is not a prefix of the supplied URI,
+   * then the given URI is returned.</li>
+   * <li>If all the above conditions hold, a new URI is created using the
+   * query and fragment components of the given URI, along with a path
+   * computed by removing the path of this URI from the start of the path
+   * of the supplied URI.</li>
+   * </ul>
+   *
+   * @param uri the URI to relativize agsint this URI
+   * @return the resulting URI
+   * @throws NullPointerException if the uri is null
    */
   public URI relativize(URI uri)
   {
-    return null;
+    if (isOpaque() || uri.isOpaque())
+      return uri;
+    if (scheme == null && uri.getScheme() != null)
+      return uri;
+    if (scheme != null && !(scheme.equals(uri.getScheme())))
+      return uri;
+    if (rawAuthority == null && uri.getRawAuthority() != null)
+      return uri;
+    if (rawAuthority != null && !(rawAuthority.equals(uri.getRawAuthority())))
+      return uri;
+    if (!(uri.getRawPath().startsWith(rawPath)))
+      return uri;
+    try
+      {
+	return new URI(null, null, 
+		       uri.getRawPath().substring(rawPath.length()),
+		       uri.getRawQuery(), uri.getRawFragment());
+      }
+    catch (URISyntaxException e)
+      {
+	throw (Error) new InternalError("Relativized URI variant could not "+
+					"be constructed").initCause(e);       
+      }
   }
 
   /**
    * Creates an URL from an URI
    *
-   * @exception MalformedURLException If a protocol handler for the URL could
+   * @throws MalformedURLException If a protocol handler for the URL could
    * not be found, or if some other error occurred while constructing the URL
-   * @exception IllegalArgumentException If the URI is not absolute
+   * @throws IllegalArgumentException If the URI is not absolute
    */
   public URL toURL() throws IllegalArgumentException, MalformedURLException
   {
@@ -745,7 +999,7 @@
   }
 
   /**
-   * Returns the rae authority part of this URI
+   * Returns the raw authority part of this URI
    */
   public String getRawAuthority()
   {
@@ -841,33 +1095,225 @@
   }
 
   /**
-   * Compares the URI with a given object
-   *
-   * @param obj The obj to compare the URI with
+   * <p> 
+   * Compares the URI with the given object for equality.  If the
+   * object is not a <code>URI</code>, then the method returns false.
+   * Otherwise, the following criteria are observed:
+   * </p>
+   * <ul>
+   * <li>The scheme of the URIs must either be null (undefined) in both cases,
+   * or equal, ignorant of case.</li>
+   * <li>The raw fragment of the URIs must either be null (undefined) in both
+   * cases, or equal, ignorant of case.</li>
+   * <li>Both URIs must be of the same type (opaque or hierarchial)</li>
+   * <li><strong>For opaque URIs:</strong></li>
+   * <ul>
+   * <li>The raw scheme-specific parts must be equal.</li>
+   * </ul>
+   * <li>For hierarchical URIs:</li>
+   * <ul>
+   * <li>The raw paths must be equal, ignorant of case.</li>
+   * <li>The raw queries are either both undefined or both equal, ignorant
+   * of case.</li>
+   * <li>The raw authority sections are either both undefined or:</li>
+   * <li><strong>For registry-based authorities:</strong></li>
+   * <ul><li>they are equal.</li></ul>
+   * <li><strong>For server-based authorities:</strong></li>
+   * <ul>
+   * <li>the hosts are equal, ignoring case</li>
+   * <li>the ports are equal</li>
+   * <li>the user information components are equal</li>
+   * </ul>
+   * </ul>
+   * </ul>
+   *
+   * @param obj the obj to compare the URI with.
+   * @return <code>true</code> if the objects are equal, according to
+   *         the specification above.
    */
   public boolean equals(Object obj)
   {
-    return false;
+    if (!(obj instanceof URI))
+      return false;
+    URI uriObj = (URI) obj;
+    if (scheme == null)
+      {
+	if (uriObj.getScheme() != null)
+	  return false;
+      }
+    else
+      if (!(scheme.equalsIgnoreCase(uriObj.getScheme())))
+	return false;
+    if (rawFragment == null)
+      {
+	if (uriObj.getRawFragment() != null)
+	  return false;
+      }
+    else
+      if (!(rawFragment.equalsIgnoreCase(uriObj.getRawFragment())))
+	return false;
+    boolean opaqueThis = isOpaque();
+    boolean opaqueObj = uriObj.isOpaque();
+    if (opaqueThis && opaqueObj)
+      return rawSchemeSpecificPart.equals(uriObj.getRawSchemeSpecificPart());
+    else if (!opaqueThis && !opaqueObj)
+      {
+	boolean common = rawPath.equalsIgnoreCase(uriObj.getRawPath())
+	  && ((rawQuery == null && uriObj.getRawQuery() == null)
+	      || rawQuery.equalsIgnoreCase(uriObj.getRawQuery()));
+	if (rawAuthority == null && uriObj.getRawAuthority() == null)
+	  return common;
+	if (host == null)
+	  return common 
+	    && rawAuthority.equalsIgnoreCase(uriObj.getRawAuthority());
+	return common 
+	  && host.equalsIgnoreCase(uriObj.getHost())
+	  && port == uriObj.getPort()
+	  && (rawUserInfo == null ?
+	      uriObj.getRawUserInfo() == null :
+	      rawUserInfo.equalsIgnoreCase(uriObj.getRawUserInfo()));
+      }
+    else
+      return false;
   }
 
   /**
-   * Computes the hascode of the URI
+   * Computes the hashcode of the URI
    */
   public int hashCode()
   {
-    return 0;
+    return (getScheme() == null ? 0 : 13 * getScheme().hashCode())
+      + 17 * getRawSchemeSpecificPart().hashCode()
+      + (getRawFragment() == null ? 0 : 21 + getRawFragment().hashCode());
   }
 
   /**
-   * Compare the URI with another object that must be an URI too
+   * Compare the URI with another object that must also be a URI.
+   * Undefined components are taken to be less than any other component.
+   * The following criteria are observed:
+   * </p>
+   * <ul>
+   * <li>Two URIs with different schemes are compared according to their
+   * scheme, regardless of case.</li>
+   * <li>A hierarchical URI is less than an opaque URI with the same
+   * scheme.</li>
+   * <li><strong>For opaque URIs:</strong></li>
+   * <ul>
+   * <li>URIs with differing scheme-specific parts are ordered according
+   * to the ordering of the scheme-specific part.</li>
+   * <li>URIs with the same scheme-specific part are ordered by the
+   * raw fragment.</li>
+   * </ul>
+   * <li>For hierarchical URIs:</li>
+   * <ul>
+   * <li>URIs are ordered according to their raw authority sections,
+   * if they are unequal.</li>
+   * <li><strong>For registry-based authorities:</strong></li>
+   * <ul><li>they are ordered according to the ordering of the authority
+   * component.</li></ul>
+   * <li><strong>For server-based authorities:</strong></li>
+   * <ul>
+   * <li>URIs are ordered according to the raw user information.</li>
+   * <li>URIs with the same user information are ordered by the host,
+   * ignoring case.</li>
+   * <lI>URIs with the same host are ordered by the port.</li>
+   * </ul>
+   * <li>URIs with the same authority section are ordered by the raw path.</li>
+   * <li>URIs with the same path are ordered by their raw query.</li>
+   * <li>URIs with the same query are ordered by their raw fragments.</li>
+   * </ul>
+   * </ul>
    *
    * @param obj This object to compare this URI with
+   * @return a negative integer, zero or a positive integer depending
+   *         on whether this URI is less than, equal to or greater
+   *         than that supplied, respectively.
+   * @throws ClassCastException if the given object is not a URI
+   */
+  public int compareTo(Object obj) 
+    throws ClassCastException
+  {
+    URI uri = (URI) obj;
+    if (scheme == null && uri.getScheme() != null)
+      return -1;
+    if (scheme != null)
+      {
+	int sCompare = scheme.compareToIgnoreCase(uri.getScheme()); 
+	if (sCompare != 0)
+	  return sCompare;
+      }
+    boolean opaqueThis = isOpaque();
+    boolean opaqueObj = uri.isOpaque();
+    if (opaqueThis && !opaqueObj)
+      return 1;
+    if (!opaqueThis && opaqueObj)
+      return -1;
+    if (opaqueThis)
+      {
+	int ssCompare = 
+	  rawSchemeSpecificPart.compareTo(uri.getRawSchemeSpecificPart());
+	if (ssCompare == 0)
+	  return compareFragments(uri);
+	else
+	  return ssCompare;
+      }
+    if (rawAuthority == null && uri.getRawAuthority() != null)
+      return -1;
+    if (rawAuthority != null)
+      {
+	int aCompare = rawAuthority.compareTo(uri.getRawAuthority());
+	if (aCompare != 0)
+	  {
+	    if (host == null)
+	      return aCompare;
+	    if (rawUserInfo == null && uri.getRawUserInfo() != null)
+	      return -1;
+	    int uCompare = rawUserInfo.compareTo(uri.getRawUserInfo());
+	    if (uCompare != 0)
+	      return uCompare;
+	    if (host == null && uri.getHost() != null)
+	      return -1;
+	    int hCompare = host.compareTo(uri.getHost());
+	    if (hCompare != 0)
+	      return hCompare;
+	    return new Integer(port).compareTo(new Integer(uri.getPort()));
+	  }
+      }
+    if (rawPath == null && uri.getRawPath() != null)
+      return -1;
+    if (rawPath != null)
+      {
+	int pCompare = rawPath.compareTo(uri.getRawPath()); 
+	if (pCompare != 0)
+	  return pCompare;
+      }
+    if (rawQuery == null && uri.getRawQuery() != null)
+      return -1;
+    if (rawQuery != null)
+      {
+	int qCompare = rawQuery.compareTo(uri.getRawQuery());
+	if (qCompare != 0)
+	  return qCompare;
+      }
+    return compareFragments(uri);
+  }
+
+  /**
+   * Compares the fragment of this URI with that of the supplied URI.
    *
-   * @exception ClassCastException If given object ist not an URI
+   * @param uri the URI to compare with this one.
+   * @return a negative integer, zero or a positive integer depending
+   *         on whether this uri's fragment is less than, equal to
+   *         or greater than the fragment of the uri supplied, respectively.
    */
-  public int compareTo(Object obj) throws ClassCastException
+  private int compareFragments(URI uri)
   {
-    return 0;
+    if (rawFragment == null && uri.getRawFragment() != null)
+      return -1;
+    else if (rawFragment == null)
+      return 0;
+    else
+      return rawFragment.compareTo(uri.getRawFragment());
   }
 
   /**
@@ -878,16 +1324,79 @@
    */
   public String toString()
   {
-    return (getScheme() == null ? "" : getScheme() + ":")
-           + getRawSchemeSpecificPart()
-           + (getRawFragment() == null ? "" : "#" + getRawFragment());
+    return (scheme == null ? "" : scheme + ":")
+      + rawSchemeSpecificPart
+      + (rawFragment == null ? "" : "#" + rawFragment);
   }
 
   /**
-   * Returns the URI as US-ASCII string
+   * Returns the URI as US-ASCII string.  This is the same as the result
+   * from <code>toString()</code> for URIs that don't contain any non-US-ASCII
+   * characters.  Otherwise, the non-US-ASCII characters are replaced
+   * by their percent-encoded representations.
+   *
+   * @return a string representation of the URI, containing only US-ASCII
+   *         characters.
    */
   public String toASCIIString()
   {
-    return "";
+    String strRep = toString();
+    boolean inNonAsciiBlock = false;
+    StringBuffer buffer = new StringBuffer();
+    StringBuffer encBuffer = null;
+    for (int i = 0; i < strRep.length(); i++)
+      {
+	char c = strRep.charAt(i);
+	if (c <= 127)

*** Patch too long, truncated ***




More information about the kaffe mailing list