KickJava   Java API By Example, From Geeks To Geeks.

Java > Open Source Codes > org > w3c > tidy > StreamInJavaImpl


1 /*
2  * Java HTML Tidy - JTidy
3  * HTML parser and pretty printer
4  *
5  * Copyright (c) 1998-2000 World Wide Web Consortium (Massachusetts
6  * Institute of Technology, Institut National de Recherche en
7  * Informatique et en Automatique, Keio University). All Rights
8  * Reserved.
9  *
10  * Contributing Author(s):
11  *
12  * Dave Raggett <dsr@w3.org>
13  * Andy Quick <ac.quick@sympatico.ca> (translation to Java)
14  * Gary L Peskin <garyp@firstech.com> (Java development)
15  * Sami Lempinen <sami@lempinen.net> (release management)
16  * Fabrizio Giustina <fgiust at users.sourceforge.net>
17  *
18  * The contributing author(s) would like to thank all those who
19  * helped with testing, bug fixes, and patience. This wouldn't
20  * have been possible without all of you.
21  *
22  * COPYRIGHT NOTICE:
23  *
24  * This software and documentation is provided "as is," and
25  * the copyright holders and contributing author(s) make no
26  * representations or warranties, express or implied, including
27  * but not limited to, warranties of merchantability or fitness
28  * for any particular purpose or that the use of the software or
29  * documentation will not infringe any third party patents,
30  * copyrights, trademarks or other rights.
31  *
32  * The copyright holders and contributing author(s) will not be
33  * liable for any direct, indirect, special or consequential damages
34  * arising out of any use of the software or documentation, even if
35  * advised of the possibility of such damage.
36  *
37  * Permission is hereby granted to use, copy, modify, and distribute
38  * this source code, or portions hereof, documentation and executables,
39  * for any purpose, without fee, subject to the following restrictions:
40  *
41  * 1. The origin of this source code must not be misrepresented.
42  * 2. Altered versions must be plainly marked as such and must
43  * not be misrepresented as being the original source.
44  * 3. This Copyright notice may not be removed or altered from any
45  * source or altered source distribution.
46  *
47  * The copyright holders and contributing author(s) specifically
48  * permit, without fee, and encourage the use of this source code
49  * as a component for supporting the Hypertext Markup Language in
50  * commercial products. If you use this source code in a product,
51  * acknowledgment is not required but would be appreciated.
52  *
53  */
54 package org.w3c.tidy;
55
56 import java.io.IOException JavaDoc;
57 import java.io.InputStream JavaDoc;
58 import java.io.InputStreamReader JavaDoc;
59 import java.io.Reader JavaDoc;
60 import java.io.UnsupportedEncodingException JavaDoc;
61
62
63 /**
64  * StreamIn Implementation using java writers.
65  * @author Fabrizio Giustina
66  * @version $Revision: 1.5 $ ($Author: fgiust $)
67  */
68 public class StreamInJavaImpl implements StreamIn
69 {
70
71     /**
72      * number of characters kept in buffer.
73      */
74     private static final int CHARBUF_SIZE = 5;
75
76     /**
77      * character buffer.
78      */
79     private int[] charbuf = new int[CHARBUF_SIZE];
80
81     /**
82      * actual position in buffer.
83      */
84     private int bufpos;
85
86     /**
87      * Java input stream reader.
88      */
89     private Reader JavaDoc reader;
90
91     /**
92      * has end of stream been reached?
93      */
94     private boolean endOfStream;
95
96     /**
97      * Is char pushed?
98      */
99     private boolean pushed;
100
101     /**
102      * current column number.
103      */
104     private int curcol;
105
106     /**
107      * last column.
108      */
109     private int lastcol;
110
111     /**
112      * current line number.
113      */
114     private int curline;
115
116     /**
117      * tab size in chars.
118      */
119     private int tabsize;
120
121     private int tabs;
122
123     /**
124      * Instantiates a new StreamInJavaImpl.
125      * @param stream
126      * @param encoding
127      * @param tabsize
128      * @throws UnsupportedEncodingException
129      */
130     public StreamInJavaImpl(InputStream JavaDoc stream, String JavaDoc encoding, int tabsize) throws UnsupportedEncodingException JavaDoc
131     {
132         reader = new InputStreamReader JavaDoc(stream, encoding);
133         this.pushed = false;
134         this.tabsize = tabsize;
135         this.curline = 1;
136         this.curcol = 1;
137         this.endOfStream = false;
138     }
139
140     /**
141      * @see org.w3c.tidy.StreamIn#readCharFromStream()
142      */
143     public int readCharFromStream()
144     {
145         int c;
146         try
147         {
148             c = reader.read();
149             if (c < 0)
150             {
151                 endOfStream = true;
152             }
153
154         }
155         catch (IOException JavaDoc e)
156         {
157             // @todo how to handle?
158 endOfStream = true;
159             return END_OF_STREAM;
160         }
161
162         return c;
163     }
164
165     /**
166      * @see org.w3c.tidy.StreamIn#readChar()
167      */
168     public int readChar()
169     {
170         int c;
171
172         if (this.pushed)
173         {
174             c = this.charbuf[--(this.bufpos)];
175             if ((this.bufpos) == 0)
176             {
177                 this.pushed = false;
178             }
179
180             if (c == '\n')
181             {
182                 this.curcol = 1;
183                 this.curline++;
184                 return c;
185             }
186
187             this.curcol++;
188             return c;
189         }
190
191         this.lastcol = this.curcol;
192
193         if (this.tabs > 0)
194         {
195             this.curcol++;
196             this.tabs--;
197             return ' ';
198         }
199
200         c = readCharFromStream();
201
202         if (c < 0)
203         {
204             endOfStream = true;
205             return END_OF_STREAM;
206         }
207
208         if (c == '\n')
209         {
210             this.curcol = 1;
211             this.curline++;
212             return c;
213         }
214         else if (c == '\r') // \r\n
215 {
216             c = readCharFromStream();
217             if (c != '\n')
218             {
219                 if (c != END_OF_STREAM)
220                 {
221                     ungetChar(c);
222                 }
223                 c = '\n';
224             }
225             this.curcol = 1;
226             this.curline++;
227             return c;
228         }
229
230         if (c == '\t')
231         {
232             this.tabs = this.tabsize - ((this.curcol - 1) % this.tabsize) - 1;
233             this.curcol++;
234             c = ' ';
235             return c;
236         }
237
238         this.curcol++;
239
240         return c;
241     }
242
243     /**
244      * @see org.w3c.tidy.StreamIn#ungetChar(int)
245      */
246     public void ungetChar(int c)
247     {
248         this.pushed = true;
249         if (this.bufpos >= CHARBUF_SIZE)
250         {
251             // pop last element
252 System.arraycopy(this.charbuf, 0, this.charbuf, 1, CHARBUF_SIZE - 1);
253             this.bufpos--;
254         }
255         this.charbuf[(this.bufpos)++] = c;
256
257         if (c == '\n')
258         {
259             --this.curline;
260         }
261
262         this.curcol = this.lastcol;
263     }
264
265     /**
266      * @see org.w3c.tidy.StreamIn#isEndOfStream()
267      */
268     public boolean isEndOfStream()
269     {
270         return endOfStream;
271     }
272
273     /**
274      * Getter for <code>curcol</code>.
275      * @return Returns the curcol.
276      */
277     public int getCurcol()
278     {
279         return this.curcol;
280     }
281
282     /**
283      * Getter for <code>curline</code>.
284      * @return Returns the curline.
285      */
286     public int getCurline()
287     {
288         return this.curline;
289     }
290
291     /**
292      * @see org.w3c.tidy.StreamIn#setLexer(org.w3c.tidy.Lexer)
293      */
294     public void setLexer(Lexer lexer)
295     {
296         // unused in the java implementation
297 }
298
299 }
Popular Tags