1   /**
2    * CodeViewer.java
3    * 
4    * Bill Lynch & Matt Tucker
5    * CoolServlets.com, October 1999
6    * 
7    * Please visit CoolServlets.com for high quality, open source Java servlets. 
8    *
9    * Copyright (C) 1999  CoolServlets.com
10   * 
11   * Any errors or suggested improvements to this class can be reported
12   * as instructed on Coolservlets.com. We hope you enjoy
13   * this program... your comments will encourage further development!
14   * 
15   * This software is distributed under the terms of The BSD License.
16   * 
17   * Redistribution and use in source and binary forms, with or without
18   * modification, are permitted provided that the following conditions are met:
19   * * Redistributions of source code must retain the above copyright notice,
20   *   this list of conditions and the following disclaimer.
21   * * Redistributions in binary form must reproduce the above copyright notice,
22   *   this list of conditions and the following disclaimer in the documentation
23   *   and/or other materials provided with the distribution.
24   * Neither name of CoolServlets.com nor the names of its contributors may be
25   * used to endorse or promote products derived from this software without
26   * specific prior written permission.
27   * 
28   * THIS SOFTWARE IS PROVIDED BY COOLSERVLETS.COM AND CONTRIBUTORS ``AS IS'' AND
29   * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
30   * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
31   * DISCLAIMED. IN NO EVENT SHALL COOLSERVLETS.COM OR CONTRIBUTORS BE LIABLE FOR
32   * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
33   * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
34   * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
35   * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
36   * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
37   * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
38   */
39  
40  /*
41   * @(#)CodeViewer.java  1.6 02/06/13
42   */
43  
44  
45  import java.io.*;
46  import java.util.*;
47  
48  /**
49   * A class that syntax highlights Java code by turning it into html.
50   *
51   * <p> A <code>CodeViewer</code> object is created and then keeps state as
52   * lines are passed in. Each line passed in as java text, is returned as syntax
53   * highlighted html text.
54   *
55   * <p> Users of the class can set how the java code will be highlighted with
56   * setter methods.
57   *
58   * <p> Only valid java lines should be passed in since the object maintains
59   * state and may not handle illegal code gracefully.
60   *
61   * <p> The actual system is implemented as a series of filters that deal with
62   * specific portions of the java code. The filters are as follows:
63   *
64   * <pre>
65   *  htmlFilter
66   *     |__
67   *        multiLineCommentFilter
68   *           |___
69   *                inlineCommentFilter
70   *                   |___
71   *                        stringFilter
72   *                           |__
73   *                               keywordFilter
74   * </pre>
75   *
76   * @version 1.6 06/13/02
77   * @author Bill Lynch, Matt Tucker, CoolServlets.com
78   */
79  public class CodeViewer {
80  
81      private static HashMap reservedWords = new HashMap(); // >= Java2 only (also, not thread-safe)
82      //private static Hashtable reservedWords = new Hashtable(); // < Java2 (thread-safe)
83      private boolean inMultiLineComment = false;
84      private String backgroundColor = "#ffffff";
85      private String commentStart = "</font><font size=2 color=\"#0000aa\"><i>";
86      private String commentEnd = "</font></i><font size=2 color=black>";    
87      private String stringStart = "</font><font size=2 color=\"#00bb00\">";
88      private String stringEnd = "</font><font size=2 color=black>";
89      private String reservedWordStart = "<b>";
90      private String reservedWordEnd = "</b>";
91  
92      static {
93          loadHash();
94      }
95  
96      public CodeViewer() {
97      }
98  
99      public void setCommentStart(String commentStart) {
100         this.commentStart = commentStart;
101     }
102     public void setCommentEnd(String commentEnd) {
103         this.commentEnd = commentEnd;
104     }
105     public void setStringStart(String stringStart) {
106         this.stringStart = stringStart;
107     }
108     public void setStringEnd(String stringEnd) {
109         this.stringEnd = stringEnd;
110     }
111     public void setReservedWordStart(String reservedWordStart) {
112         this.reservedWordStart = reservedWordStart;
113     }
114     public void setReservedWordEnd(String reservedWordEnd) {
115         this.reservedWordEnd = reservedWordEnd;
116     }
117 
118     public String getCommentStart() {
119         return commentStart;
120     }
121     public String getCommentEnd() {
122         return commentEnd;
123     }
124     public String getStringStart() {
125         return stringStart;
126     }
127     public String getStringEnd() {
128         return stringEnd;
129     }
130     public String getReservedWordStart() {
131         return reservedWordStart;
132     }
133     public String getReservedWordEnd() {
134         return reservedWordEnd;
135     }
136 
137     /**
138      * Passes off each line to the first filter.
139      * @param   line    The line of Java code to be highlighted.
140      */
141     public String syntaxHighlight( String line ) {
142        return htmlFilter(line);
143     }
144 
145     /*
146      * Filter html tags into more benign text.
147      */ 
148     private String htmlFilter( String line ) {
149         if( line == null || line.equals("") ) {
150             return "";
151         }
152 
153         // replace ampersands with HTML escape sequence for ampersand;
154         line = replace(line, "&", "&#38;");
155 
156         // replace the \\ with HTML escape sequences. fixes a problem when
157         // backslashes preceed quotes.
158         line = replace(line, "\\\\", "&#92;&#92;" );
159 
160         // replace \" sequences with HTML escape sequences;
161         line = replace(line, "" + (char)92 + (char)34, "&#92;&#34");
162 
163         // replace less-than signs which might be confused
164         // by HTML as tag angle-brackets;
165         line = replace(line, "<", "&#60;");
166         // replace greater-than signs which might be confused
167         // by HTML as tag angle-brackets;
168         line = replace(line, ">", "&#62;");
169         
170         return multiLineCommentFilter(line);
171     }
172 
173     /*
174      * Filter out multiLine comments. State is kept with a private boolean
175      * variable.
176      */     
177     private String multiLineCommentFilter(String line) {
178         if (line == null || line.equals("")) {
179             return "";
180         }
181         StringBuffer buf = new StringBuffer();
182         int index;
183         //First, check for the end of a multi-line comment.
184         if (inMultiLineComment && (index = line.indexOf("*/")) > -1 && !isInsideString(line,index)) {
185             inMultiLineComment = false;               
186             buf.append(line.substring(0,index));
187             buf.append("*/").append(commentEnd);
188             if (line.length() > index+2) {
189                 buf.append(inlineCommentFilter(line.substring(index+2)));
190             }
191             return buf.toString();
192         }
193         //If there was no end detected and we're currently in a multi-line
194         //comment, we don't want to do anymore work, so return line.
195         else if (inMultiLineComment) {
196             return line;
197         }
198         //We're not currently in a comment, so check to see if the start
199         //of a multi-line comment is in this line.
200         else if ((index = line.indexOf("/*")) > -1 && !isInsideString(line,index)) {
201             inMultiLineComment = true;
202             //Return result of other filters + everything after the start
203             //of the multiline comment. We need to pass the through the
204             //to the multiLineComment filter again in case the comment ends
205             //on the same line.
206             buf.append(inlineCommentFilter(line.substring(0,index)));
207             buf.append(commentStart).append("/*");
208             buf.append(multiLineCommentFilter(line.substring(index+2)));
209             return buf.toString();
210         }
211         //Otherwise, no useful multi-line comment information was found so
212         //pass the line down to the next filter for processesing.
213         else {
214             return inlineCommentFilter(line);
215         }
216     }      
217 
218     /*
219      * Filter inline comments from a line and formats them properly.
220      */
221     private String inlineCommentFilter(String line) {
222         if (line == null || line.equals("")) {
223             return "";
224         }
225         StringBuffer buf = new StringBuffer();
226         int index;
227         if ((index = line.indexOf("//")) > -1 && !isInsideString(line,index)) {
228             buf.append(stringFilter(line.substring(0,index)));
229             buf.append(commentStart);
230             buf.append(line.substring(index));
231             buf.append(commentEnd);
232         }
233         else {
234             buf.append(stringFilter(line));
235         }
236         return buf.toString();
237     } 
238 
239     /*
240      * Filters strings from a line of text and formats them properly.
241      */
242     private String stringFilter(String line) {
243         if (line == null || line.equals("")) {
244             return "";
245         }
246         StringBuffer buf = new StringBuffer();
247         if (line.indexOf("\"") <= -1) {
248             return keywordFilter(line);
249         }
250         int start = 0;
251         int startStringIndex = -1;
252         int endStringIndex = -1;
253         int tempIndex;
254         //Keep moving through String characters until we want to stop...
255         while ((tempIndex = line.indexOf("\"")) > -1) {
256             //We found the beginning of a string
257             if (startStringIndex == -1) {
258                 startStringIndex = 0;
259                 buf.append( stringFilter(line.substring(start,tempIndex)) );
260                 buf.append(stringStart).append("\"");
261                 line = line.substring(tempIndex+1);
262             }
263             //Must be at the end
264             else {
265                 startStringIndex = -1;
266                 endStringIndex = tempIndex;
267                 buf.append(line.substring(0,endStringIndex+1));
268                 buf.append(stringEnd);
269                 line = line.substring(endStringIndex+1);
270             }
271         }
272 
273         buf.append( keywordFilter(line) );
274 
275         return buf.toString();
276     }
277 
278     /*
279      * Filters keywords from a line of text and formats them properly.
280      */
281     private String keywordFilter( String line ) {
282         if( line == null || line.equals("") ) {
283             return "";
284         }
285         StringBuffer buf = new StringBuffer();
286         HashMap usedReservedWords = new HashMap(); // >= Java2 only (not thread-safe)
287         //Hashtable usedReservedWords = new Hashtable(); // < Java2 (thread-safe)
288         int i=0, startAt=0;
289         char ch;
290         StringBuffer temp = new StringBuffer();
291         while( i < line.length() ) {
292             temp.setLength(0);
293             ch = line.charAt(i);
294             startAt = i;
295             // 65-90, uppercase letters
296             // 97-122, lowercase letters
297             while( i<line.length() && ( ( ch >= 65 && ch <= 90 )
298                     || ( ch >= 97 && ch <= 122 ) ) ) {
299                 temp.append(ch);
300                 i++;
301                 if( i < line.length() ) {
302                     ch = line.charAt(i);
303                 }
304             }
305             String tempString = temp.toString();
306             if( reservedWords.containsKey(tempString) && !usedReservedWords.containsKey(tempString)) {
307                 usedReservedWords.put(tempString,tempString);
308                 line = replace( line, tempString, (reservedWordStart+tempString+reservedWordEnd) );
309                 i += (reservedWordStart.length() + reservedWordEnd.length());
310             }
311             else {
312                 i++;
313             }            
314         }
315         buf.append(line);
316         return buf.toString();
317     }
318 
319     /*
320      * All important replace method. Replaces all occurences of oldString in
321      * line with newString.
322      */
323     private String replace( String line, String oldString, String newString ) {
324         int i=0;
325         while( ( i=line.indexOf( oldString, i ) ) >= 0 ) {
326             line = (new StringBuffer().append(line.substring(0,i)).append(newString).append(line.substring(i+oldString.length()))).toString();
327             i += newString.length();
328         }
329         return line;
330     }
331 
332     /*
333      * Checks to see if some position in a line is between String start and
334      * ending characters. Not yet used in code or fully working :)
335      */
336     private boolean isInsideString(String line, int position) {
337         if (line.indexOf("\"") < 0) {
338             return false;
339         }
340         int index;
341         String left = line.substring(0,position);
342         String right = line.substring(position);
343         int leftCount = 0;
344         int rightCount = 0;
345         while ((index = left.indexOf("\"")) > -1) {
346             leftCount ++;
347             left = left.substring(index+1); 
348         }
349         while ((index = right.indexOf("\"")) > -1) {
350             rightCount ++;
351             right = right.substring(index+1); 
352         }
353         if (rightCount % 2 != 0 && leftCount % 2 != 0) {
354             return true;
355         }
356         else {
357             return false;
358         }        
359     }
360 
361     /*
362      * Load Hashtable (or HashMap) with Java reserved words.
363      */
364     private static void loadHash() {
365         reservedWords.put( "abstract", "abstract" );
366         reservedWords.put( "do", "do" );
367         reservedWords.put( "inner", "inner" );
368         reservedWords.put( "public", "public" );
369         reservedWords.put( "var", "var" );
370         reservedWords.put( "boolean", "boolean" );
371         reservedWords.put( "continue", "continue" );
372         reservedWords.put( "int", "int" );
373         reservedWords.put( "return", "return" );
374         reservedWords.put( "void", "void" );
375         reservedWords.put( "break", "break" );
376         reservedWords.put( "else", "else" );
377         reservedWords.put( "interface", "interface" );
378         reservedWords.put( "short", "short" );
379         reservedWords.put( "volatile", "volatile" );
380         reservedWords.put( "byvalue", "byvalue" );
381         reservedWords.put( "extends", "extends" );
382         reservedWords.put( "long", "long" );
383         reservedWords.put( "static", "static" );
384         reservedWords.put( "while", "while" );
385         reservedWords.put( "case", "case" );
386         reservedWords.put( "final", "final" );
387         reservedWords.put( "naive", "naive" );
388         reservedWords.put( "super", "super" );
389         reservedWords.put( "transient", "transient" );
390         reservedWords.put( "cast", "cast" );
391         reservedWords.put( "float", "float" );
392         reservedWords.put( "new", "new" );
393         reservedWords.put( "rest", "rest" );
394         reservedWords.put( "catch", "catch" );
395         reservedWords.put( "for", "for" );
396         reservedWords.put( "null", "null" );
397         reservedWords.put( "synchronized", "synchronized" );
398         reservedWords.put( "char", "char" );
399         reservedWords.put( "finally", "finally" );
400         reservedWords.put( "operator", "operator" );
401         reservedWords.put( "this", "this" );
402         reservedWords.put( "class", "class" );
403         reservedWords.put( "generic", "generic" );
404         reservedWords.put( "outer", "outer" );
405         reservedWords.put( "switch", "switch" );
406         reservedWords.put( "const", "const" );
407         reservedWords.put( "goto", "goto" );
408         reservedWords.put( "package", "package" );
409         reservedWords.put( "throw", "throw" );
410         reservedWords.put( "double", "double" );
411         reservedWords.put( "if", "if" );
412         reservedWords.put( "private", "private" );
413         reservedWords.put( "true", "true" );
414         reservedWords.put( "default", "default" );
415         reservedWords.put( "import", "import" );
416         reservedWords.put( "protected", "protected" );
417         reservedWords.put( "try", "try" );
418     }
419 }
420