aoqi@0: /*
aoqi@0:  * Copyright (c) 1997, 2012, Oracle and/or its affiliates. All rights reserved.
aoqi@0:  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
aoqi@0:  *
aoqi@0:  * This code is free software; you can redistribute it and/or modify it
aoqi@0:  * under the terms of the GNU General Public License version 2 only, as
aoqi@0:  * published by the Free Software Foundation.  Oracle designates this
aoqi@0:  * particular file as subject to the "Classpath" exception as provided
aoqi@0:  * by Oracle in the LICENSE file that accompanied this code.
aoqi@0:  *
aoqi@0:  * This code is distributed in the hope that it will be useful, but WITHOUT
aoqi@0:  * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
aoqi@0:  * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
aoqi@0:  * version 2 for more details (a copy is included in the LICENSE file that
aoqi@0:  * accompanied this code).
aoqi@0:  *
aoqi@0:  * You should have received a copy of the GNU General Public License version
aoqi@0:  * 2 along with this work; if not, write to the Free Software Foundation,
aoqi@0:  * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
aoqi@0:  *
aoqi@0:  * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
aoqi@0:  * or visit www.oracle.com if you need additional information or have any
aoqi@0:  * questions.
aoqi@0:  */
aoqi@0: 
aoqi@0: package com.sun.xml.internal.ws.encoding;
aoqi@0: 
aoqi@0: import javax.xml.ws.WebServiceException;
aoqi@0: 
aoqi@0: /**
aoqi@0:  * This class tokenizes RFC822 and MIME headers into the basic
aoqi@0:  * symbols specified by RFC822 and MIME. <p>
aoqi@0:  *
aoqi@0:  * This class handles folded headers (ie headers with embedded
aoqi@0:  * CRLF SPACE sequences). The folds are removed in the returned
aoqi@0:  * tokens.
aoqi@0:  *
aoqi@0:  * @version 1.9, 02/03/27
aoqi@0:  * @author  John Mani
aoqi@0:  */
aoqi@0: 
aoqi@0: class HeaderTokenizer {
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * The Token class represents tokens returned by the
aoqi@0:      * HeaderTokenizer.
aoqi@0:      */
aoqi@0:     static class Token {
aoqi@0: 
aoqi@0:         private int type;
aoqi@0:         private String value;
aoqi@0: 
aoqi@0:         /**
aoqi@0:          * Token type indicating an ATOM.
aoqi@0:          */
aoqi@0:         public static final int ATOM            = -1;
aoqi@0: 
aoqi@0:         /**
aoqi@0:          * Token type indicating a quoted string. The value
aoqi@0:          * field contains the string without the quotes.
aoqi@0:          */
aoqi@0:         public static final int QUOTEDSTRING    = -2;
aoqi@0: 
aoqi@0:         /**
aoqi@0:          * Token type indicating a comment. The value field
aoqi@0:          * contains the comment string without the comment
aoqi@0:          * start and end symbols.
aoqi@0:          */
aoqi@0:         public static final int COMMENT         = -3;
aoqi@0: 
aoqi@0:         /**
aoqi@0:          * Token type indicating end of input.
aoqi@0:          */
aoqi@0:         public static final int  EOF            = -4;
aoqi@0: 
aoqi@0:         /**
aoqi@0:          * Constructor.
aoqi@0:          * @param       type    Token type
aoqi@0:          * @param       value   Token value
aoqi@0:          */
aoqi@0:         public Token(int type, String value) {
aoqi@0:              this.type = type;
aoqi@0:              this.value = value;
aoqi@0:         }
aoqi@0: 
aoqi@0:         /**
aoqi@0:          * Return the type of the token. If the token represents a
aoqi@0:          * delimiter or a control character, the type is that character
aoqi@0:          * itself, converted to an integer. Otherwise, it's value is
aoqi@0:          * one of the following:
aoqi@0:          * <ul>
aoqi@0:          * <li><code>ATOM</code> A sequence of ASCII characters
aoqi@0:          *      delimited by either SPACE, CTL, "(", <"> or the
aoqi@0:          *      specified SPECIALS
aoqi@0:          * <li><code>QUOTEDSTRING</code> A sequence of ASCII characters
aoqi@0:          *      within quotes
aoqi@0:          * <li><code>COMMENT</code> A sequence of ASCII characters
aoqi@0:          *      within "(" and ")".
aoqi@0:          * <li><code>EOF</code> End of header
aoqi@0:          * </ul>
aoqi@0:          */
aoqi@0:         public int getType() {
aoqi@0:             return type;
aoqi@0:         }
aoqi@0: 
aoqi@0:         /**
aoqi@0:          * Returns the value of the token just read. When the current
aoqi@0:          * token is a quoted string, this field contains the body of the
aoqi@0:          * string, without the quotes. When the current token is a comment,
aoqi@0:          * this field contains the body of the comment.
aoqi@0:          *
aoqi@0:          * @return      token value
aoqi@0:          */
aoqi@0:         public String getValue() {
aoqi@0:             return value;
aoqi@0:         }
aoqi@0:     }
aoqi@0: 
aoqi@0:     private String string; // the string to be tokenized
aoqi@0:     private boolean skipComments; // should comments be skipped ?
aoqi@0:     private String delimiters; // delimiter string
aoqi@0:     private int currentPos; // current parse position
aoqi@0:     private int maxPos; // string length
aoqi@0:     private int nextPos; // track start of next Token for next()
aoqi@0:     private int peekPos; // track start of next Token for peek()
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * RFC822 specials
aoqi@0:      */
aoqi@0:     private final static String RFC822 = "()<>@,;:\\\"\t .[]";
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * MIME specials
aoqi@0:      */
aoqi@0:     final static String MIME = "()<>@,;:\\\"\t []/?=";
aoqi@0: 
aoqi@0:     // The EOF Token
aoqi@0:     private final static Token EOFToken = new Token(Token.EOF, null);
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * Constructor that takes a rfc822 style header.
aoqi@0:      *
aoqi@0:      * @param   header  The rfc822 header to be tokenized
aoqi@0:      * @param   delimiters      Set of delimiter characters
aoqi@0:      *                          to be used to delimit ATOMS. These
aoqi@0:      *                          are usually <code>RFC822</code> or
aoqi@0:      *                          <code>MIME</code>
aoqi@0:      * @param   skipComments  If true, comments are skipped and
aoqi@0:      *                          not returned as tokens
aoqi@0:      */
aoqi@0:     HeaderTokenizer(String header, String delimiters,
aoqi@0:                            boolean skipComments) {
aoqi@0:         string = (header == null) ? "" : header; // paranoia ?!
aoqi@0:         this.skipComments = skipComments;
aoqi@0:         this.delimiters = delimiters;
aoqi@0:         currentPos = nextPos = peekPos = 0;
aoqi@0:         maxPos = string.length();
aoqi@0:     }
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * Constructor. Comments are ignored and not returned as tokens
aoqi@0:      *
aoqi@0:      * @param   header  The header that is tokenized
aoqi@0:      * @param   delimiters  The delimiters to be used
aoqi@0:      */
aoqi@0:     HeaderTokenizer(String header, String delimiters) {
aoqi@0:             this(header, delimiters, true);
aoqi@0:     }
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * Constructor. The RFC822 defined delimiters - RFC822 - are
aoqi@0:      * used to delimit ATOMS. Also comments are skipped and not
aoqi@0:      * returned as tokens
aoqi@0:      */
aoqi@0:     HeaderTokenizer(String header)  {
aoqi@0:             this(header, RFC822);
aoqi@0:     }
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * Parses the next token from this String. <p>
aoqi@0:      *
aoqi@0:      * Clients sit in a loop calling next() to parse successive
aoqi@0:      * tokens until an EOF Token is returned.
aoqi@0:      *
aoqi@0:      * @return          the next Token
aoqi@0:      * @exception WebServiceException if the parse fails
aoqi@0:      */
aoqi@0:     Token next() throws WebServiceException {
aoqi@0:         Token tk;
aoqi@0: 
aoqi@0:         currentPos = nextPos; // setup currentPos
aoqi@0:         tk = getNext();
aoqi@0:         nextPos = peekPos = currentPos; // update currentPos and peekPos
aoqi@0:         return tk;
aoqi@0:     }
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * Peek at the next token, without actually removing the token
aoqi@0:      * from the parse stream. Invoking this method multiple times
aoqi@0:      * will return successive tokens, until <code>next()</code> is
aoqi@0:      * called. <p>
aoqi@0:      *
aoqi@0:      * @return          the next Token
aoqi@0:      * @exception       WebServiceException if the parse fails
aoqi@0:      */
aoqi@0:     Token peek() throws WebServiceException {
aoqi@0:         Token tk;
aoqi@0: 
aoqi@0:         currentPos = peekPos; // setup currentPos
aoqi@0:         tk = getNext();
aoqi@0:         peekPos = currentPos; // update peekPos
aoqi@0:         return tk;
aoqi@0:     }
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * Return the rest of the Header.
aoqi@0:      *
aoqi@0:      * @return String   rest of header. null is returned if we are
aoqi@0:      *                  already at end of header
aoqi@0:      */
aoqi@0:     String getRemainder() {
aoqi@0:             return string.substring(nextPos);
aoqi@0:     }
aoqi@0: 
aoqi@0:     /*
aoqi@0:      * Return the next token starting from 'currentPos'. After the
aoqi@0:      * parse, 'currentPos' is updated to point to the start of the
aoqi@0:      * next token.
aoqi@0:      */
aoqi@0:     private Token getNext() throws WebServiceException {
aoqi@0:         // If we're already at end of string, return EOF
aoqi@0:         if (currentPos >= maxPos)
aoqi@0:             return EOFToken;
aoqi@0: 
aoqi@0:         // Skip white-space, position currentPos beyond the space
aoqi@0:         if (skipWhiteSpace() == Token.EOF)
aoqi@0:             return EOFToken;
aoqi@0: 
aoqi@0:         char c;
aoqi@0:         int start;
aoqi@0:         boolean filter = false;
aoqi@0: 
aoqi@0:         c = string.charAt(currentPos);
aoqi@0: 
aoqi@0:         // Check or Skip comments and position currentPos
aoqi@0:         // beyond the comment
aoqi@0:         while (c == '(') {
aoqi@0:             // Parsing comment ..
aoqi@0:             int nesting;
aoqi@0:             for (start = ++currentPos, nesting = 1;
aoqi@0:              nesting > 0 && currentPos < maxPos;
aoqi@0:              currentPos++) {
aoqi@0:             c = string.charAt(currentPos);
aoqi@0:             if (c == '\\') {  // Escape sequence
aoqi@0:                 currentPos++; // skip the escaped character
aoqi@0:                 filter = true;
aoqi@0:             } else if (c == '\r')
aoqi@0:                 filter = true;
aoqi@0:             else if (c == '(')
aoqi@0:                 nesting++;
aoqi@0:             else if (c == ')')
aoqi@0:                 nesting--;
aoqi@0:             }
aoqi@0:             if (nesting != 0)
aoqi@0:             throw new WebServiceException("Unbalanced comments");
aoqi@0: 
aoqi@0:             if (!skipComments) {
aoqi@0:             // Return the comment, if we are asked to.
aoqi@0:             // Note that the comment start & end markers are ignored.
aoqi@0:             String s;
aoqi@0:             if (filter) // need to go thru the token again.
aoqi@0:                 s = filterToken(string, start, currentPos-1);
aoqi@0:             else
aoqi@0:                 s = string.substring(start,currentPos-1);
aoqi@0: 
aoqi@0:             return new Token(Token.COMMENT, s);
aoqi@0:             }
aoqi@0: 
aoqi@0:             // Skip any whitespace after the comment.
aoqi@0:             if (skipWhiteSpace() == Token.EOF)
aoqi@0:             return EOFToken;
aoqi@0:             c = string.charAt(currentPos);
aoqi@0:         }
aoqi@0: 
aoqi@0:         // Check for quoted-string and position currentPos
aoqi@0:         //  beyond the terminating quote
aoqi@0:         if (c == '"') {
aoqi@0:             for (start = ++currentPos; currentPos < maxPos; currentPos++) {
aoqi@0:             c = string.charAt(currentPos);
aoqi@0:             if (c == '\\') { // Escape sequence
aoqi@0:                 currentPos++;
aoqi@0:                 filter = true;
aoqi@0:             } else if (c == '\r')
aoqi@0:                 filter = true;
aoqi@0:             else if (c == '"') {
aoqi@0:                 currentPos++;
aoqi@0:                 String s;
aoqi@0: 
aoqi@0:                 if (filter)
aoqi@0:                 s = filterToken(string, start, currentPos-1);
aoqi@0:                 else
aoqi@0:                 s = string.substring(start,currentPos-1);
aoqi@0: 
aoqi@0:                 return new Token(Token.QUOTEDSTRING, s);
aoqi@0:             }
aoqi@0:             }
aoqi@0:             throw new WebServiceException("Unbalanced quoted string");
aoqi@0:         }
aoqi@0: 
aoqi@0:         // Check for SPECIAL or CTL
aoqi@0:         if (c < 040 || c >= 0177 || delimiters.indexOf(c) >= 0) {
aoqi@0:             currentPos++; // re-position currentPos
aoqi@0:             char ch[] = new char[1];
aoqi@0:             ch[0] = c;
aoqi@0:             return new Token((int)c, new String(ch));
aoqi@0:         }
aoqi@0: 
aoqi@0:         // Check for ATOM
aoqi@0:         for (start = currentPos; currentPos < maxPos; currentPos++) {
aoqi@0:             c = string.charAt(currentPos);
aoqi@0:             // ATOM is delimited by either SPACE, CTL, "(", <">
aoqi@0:             // or the specified SPECIALS
aoqi@0:             if (c < 040 || c >= 0177 || c == '(' || c == ' ' ||
aoqi@0:             c == '"' || delimiters.indexOf(c) >= 0)
aoqi@0:             break;
aoqi@0:         }
aoqi@0:         return new Token(Token.ATOM, string.substring(start, currentPos));
aoqi@0:         }
aoqi@0: 
aoqi@0:         // Skip SPACE, HT, CR and NL
aoqi@0:         private int skipWhiteSpace() {
aoqi@0:         char c;
aoqi@0:         for (; currentPos < maxPos; currentPos++)
aoqi@0:             if (((c = string.charAt(currentPos)) != ' ') &&
aoqi@0:             (c != '\t') && (c != '\r') && (c != '\n'))
aoqi@0:             return currentPos;
aoqi@0:         return Token.EOF;
aoqi@0:     }
aoqi@0: 
aoqi@0:     /* Process escape sequences and embedded LWSPs from a comment or
aoqi@0:      * quoted string.
aoqi@0:      */
aoqi@0:     private static String filterToken(String s, int start, int end) {
aoqi@0:         StringBuffer sb = new StringBuffer();
aoqi@0:         char c;
aoqi@0:         boolean gotEscape = false;
aoqi@0:         boolean gotCR = false;
aoqi@0: 
aoqi@0:         for (int i = start; i < end; i++) {
aoqi@0:             c = s.charAt(i);
aoqi@0:             if (c == '\n' && gotCR) {
aoqi@0:             // This LF is part of an unescaped
aoqi@0:             // CRLF sequence (i.e, LWSP). Skip it.
aoqi@0:             gotCR = false;
aoqi@0:             continue;
aoqi@0:             }
aoqi@0: 
aoqi@0:             gotCR = false;
aoqi@0:             if (!gotEscape) {
aoqi@0:             // Previous character was NOT '\'
aoqi@0:             if (c == '\\') // skip this character
aoqi@0:                 gotEscape = true;
aoqi@0:             else if (c == '\r') // skip this character
aoqi@0:                 gotCR = true;
aoqi@0:             else // append this character
aoqi@0:                 sb.append(c);
aoqi@0:             } else {
aoqi@0:             // Previous character was '\'. So no need to
aoqi@0:             // bother with any special processing, just
aoqi@0:             // append this character
aoqi@0:             sb.append(c);
aoqi@0:             gotEscape = false;
aoqi@0:             }
aoqi@0:         }
aoqi@0:         return sb.toString();
aoqi@0:     }
aoqi@0: }