001    /*
002     * Copyright (c) 2005 Marko Karppinen & Co. LLC
003     *
004     * Permission is hereby granted, free of charge, to any person obtaining a 
005     * copy of this software and associated documentation files (the "Software"), 
006     * to deal in the Software without restriction, including without limitation 
007     * the rights to use, copy, modify, merge, publish, distribute, sublicense, 
008     * and/or sell copies of the Software, and to permit persons to whom the 
009     * Software is furnished to do so, subject to the following conditions:
010     *
011     * The above copyright notice and this permission notice shall be included in 
012     * all copies or substantial portions of the Software.
013     *
014     * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR 
015     * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, 
016     * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL 
017     * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER 
018     * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING 
019     * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER 
020     * DEALINGS IN THE SOFTWARE.
021     */
022    
023    package fi.karppinen.xml;
024    
025    import org.xml.sax.ContentHandler;
026    import org.xml.sax.SAXException;
027    
028    /**
029     * Parses plain text with double line break paragraph breaks as XHTML 1.x paragraphs.
030     * 
031     * @version $Id: SimpleParagraphParser.java,v 1.1 2006/10/30 19:57:09 hsivonen Exp $
032     * @author hsivonen
033     */
034    public class SimpleParagraphParser {
035    
036        private char[] content;
037    
038        private XhtmlEmitter emitter;
039    
040        private ContentHandler contentHandler;
041    
042        private String htmlClass = null;
043    
044        private char[] SPACE = { ' ' };
045        
046        private boolean somethingRendered = false;
047    
048        /**
049         *  
050         */
051        public SimpleParagraphParser(String content, ContentHandler contentHandler) {
052            this.content = content.toCharArray();
053            this.emitter = new XhtmlEmitter(contentHandler);
054            this.contentHandler = contentHandler;
055        }
056    
057        public void parse() throws SAXException {
058            int i = 0;
059            int len = content.length;
060            int start = 0;
061            boolean paraOpen = false;
062            boolean prevWasSpace = false;
063            while (i < len) {
064                if (!paraOpen) {
065                    whitespace: while (i < len) {
066                        switch (content[i]) {
067                            case ' ':
068                            case '\n':
069                            case '\r':
070                            case '\t':
071                                i++;
072                                break;
073                            default:
074                                break whitespace;
075                        }
076                    }
077                }
078                if (i < len) {
079                    if (!paraOpen) {
080                        if (htmlClass == null) {
081                            emitter.startElement("p");
082                        } else {
083                            emitter.startElementWithClass("p", htmlClass);
084                        }
085                        somethingRendered = true;
086                        paraOpen = true;
087                        start = i;
088                    }
089                    innerloop: while (i < len) {
090                        switch (content[i]) {
091                            case '\r':
092                                if (i + 1 < len && content[i + 1] == '\n') {
093                                    if (start < i) {
094                                        contentHandler.characters(content, start, i
095                                                - start);
096                                    }
097                                    i++;
098                                    start = i;
099                                    break;
100                                }
101                            // fall thru!
102                            case '\n':
103                                if (prevWasSpace) {
104                                    if (start < i) {
105                                        contentHandler.characters(content, start, i
106                                                - start);
107                                    }
108                                    i++;
109                                    start = i;
110                                    prevWasSpace = false;
111                                    break innerloop;
112                                } else {
113                                    if (start < i) {
114                                        contentHandler.characters(content, start, i
115                                                - start);
116                                    }
117                                    i++;
118                                    start = i;
119                                    emitter.characters(SPACE);
120                                    prevWasSpace = true;
121                                }
122                                break;
123                            default:
124                                prevWasSpace = false;
125                                i++;
126                        }
127                    }
128                }
129                if (paraOpen) {
130                    emitter.endElement("p");
131                    paraOpen = false;
132                }
133            }
134        }
135        /**
136         * Returns the somethingRendered.
137         * 
138         * @return the somethingRendered
139         */
140        public boolean isSomethingRendered() {
141            return somethingRendered;
142        }
143        /**
144         * Returns the htmlClass.
145         * 
146         * @return the htmlClass
147         */
148        public String getHtmlClass() {
149            return htmlClass;
150        }
151        /**
152         * Sets the htmlClass.
153         * 
154         * @param htmlClass the htmlClass to set
155         */
156        public void setHtmlClass(String htmlClass) {
157            this.htmlClass = htmlClass;
158        }
159    }