2 * Copyright (C) 2001, 2002 The Mir-coders group
4 * This file is part of Mir.
6 * Mir is free software; you can redistribute it and/or modify
7 * it under the terms of the GNU General Public License as published by
8 * the Free Software Foundation; either version 2 of the License, or
9 * (at your option) any later version.
11 * Mir is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 * GNU General Public License for more details.
16 * You should have received a copy of the GNU General Public License
17 * along with Mir; if not, write to the Free Software
18 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
20 * In addition, as a special exception, The Mir-coders gives permission to link
21 * the code of this program with any library licensed under the Apache Software License,
22 * The Sun (tm) Java Advanced Imaging library (JAI), The Sun JIMI library
23 * (or with modified versions of the above that use the same license as the above),
24 * and distribute linked combinations including the two. You must obey the
25 * GNU General Public License in all respects for all of the code used other than
26 * the above mentioned libraries. If you modify this file, you may extend this
27 * exception to your version of the file, but you are not obligated to do so.
28 * If you do not wish to do so, delete this exception statement from your version.
33 import gnu.regexp.REException;
35 import java.text.NumberFormat;
36 import java.util.Calendar;
37 import java.util.Date;
38 import java.util.GregorianCalendar;
39 import java.util.TimeZone;
42 * Statische Hilfsmethoden zur Stringbehandlung
44 * @version $Id: StringUtil.java,v 1.35 2004/11/06 19:27:33 idfx Exp $
45 * @author mir-coders group
48 public final class StringUtil {
50 private static RE re_newline2br, re_brbr2p, re_mail, re_url, re_tags,
51 re_tables, re_forbiddenTags;
53 private StringUtil() { } // this avoids contruction
58 re_newline2br = new RE("(\r?\n){1}");
59 re_brbr2p = new RE("(<br>\r?\n<br>){1,}");
60 re_mail = new RE("\\b([a-zA-Z0-9_.-]+)@([a-zA-Z0-9_-]+)\\.([a-zA-Z0-9_.-]+)\\b");
61 re_url = new RE("((https://)|(http://)|(ftp://)){1}([a-zA-Z0-9_-]+).([a-zA-Z0-9_.:-]+)/?([^ \t\r\n<>\\)\\]]+[^ \t\r\n.,<>\\)\\]])");
62 re_tags = new RE("<[^>]*>",RE.REG_ICASE);
63 re_tables = new RE("<[ \t\r\n/]*(table|td|tr)[ \t\r\n]*>",RE.REG_ICASE);
64 re_forbiddenTags = new RE("<[ \t\r\n/]*(html|meta|body|head|script)[ \t\r\n]*>",RE.REG_ICASE);
66 catch (REException e){
67 System.err.println("FATAL: StringUtil: could not precompile REGEX: "+e.toString());
72 * Formats a number with the specified minimum and maximum number of digits.
74 public static synchronized String zeroPaddingNumber(long value, int minDigits,
77 NumberFormat numberFormat = NumberFormat.getInstance();
78 numberFormat.setMinimumIntegerDigits(minDigits);
79 numberFormat.setMaximumIntegerDigits(maxDigits);
80 return numberFormat.format(value);
84 * Wandelt Datum in einen 8-ziffrigen String um (yyyymmdd)
86 * @return 8-ziffriger String (yyyymmdd)
89 public static final String date2webdbDate (GregorianCalendar theDate) {
90 StringBuffer webdbDate = new StringBuffer();
91 webdbDate.append(String.valueOf(theDate.get(Calendar.YEAR)));
92 webdbDate.append(pad2(theDate.get(Calendar.MONTH) + 1));
93 webdbDate.append(pad2(theDate.get(Calendar.DATE)));
94 return webdbDate.toString();
98 * Wandelt Calendar in einen 12-ziffrigen String um (yyyymmddhhmm)
100 * @return 12-ziffriger String (yyyymmdd)
103 public static final String date2webdbDateTime (GregorianCalendar theDate) {
104 StringBuffer webdbDate = new StringBuffer();
105 webdbDate.append(String.valueOf(theDate.get(Calendar.YEAR)));
106 webdbDate.append(pad2(theDate.get(Calendar.MONTH) + 1));
107 webdbDate.append(pad2(theDate.get(Calendar.DATE)));
108 webdbDate.append(pad2(theDate.get(Calendar.HOUR)));
109 webdbDate.append(pad2(theDate.get(Calendar.MINUTE)));
110 return webdbDate.toString();
114 * Return a http://www.w3.org/TR/NOTE-datetime formatted date (yyyy-mm-ddThh:mm:ssTZ)
116 * @return w3approved datetime
119 public static final String date2w3DateTime (GregorianCalendar theDate) {
120 StringBuffer webdbDate = new StringBuffer();
121 webdbDate.append(String.valueOf(theDate.get(Calendar.YEAR)));
122 webdbDate.append("-");
123 webdbDate.append(pad2(theDate.get(Calendar.MONTH) + 1));
124 webdbDate.append("-");
125 webdbDate.append(pad2(theDate.get(Calendar.DATE)));
126 webdbDate.append("T");
127 webdbDate.append(pad2(theDate.get(Calendar.HOUR_OF_DAY)));
128 webdbDate.append(":");
129 webdbDate.append(pad2(theDate.get(Calendar.MINUTE)));
130 webdbDate.append(":");
131 webdbDate.append(pad2(theDate.get(Calendar.SECOND)));
132 //assumes you are an hour-multiple away from UTC....
133 int offset=(theDate.get(Calendar.ZONE_OFFSET)/(60*60*1000));
135 webdbDate.append("-");
138 webdbDate.append("+");
140 webdbDate.append(pad2(Math.abs(offset)));
141 webdbDate.append(":00");
142 return webdbDate.toString();
146 * wandelt Calendar in dd.mm.yyyy / hh.mm um
148 * @return String mit (dd.mm.yyyy / hh.mm um)
150 public static String date2readableDateTime (GregorianCalendar theDate) {
151 String readable = "";
153 readable += pad2(theDate.get(Calendar.DATE));
154 readable += "." + pad2(theDate.get(Calendar.MONTH) + 1);
155 readable += "." + String.valueOf(theDate.get(Calendar.YEAR));
156 hour = theDate.get(Calendar.HOUR);
157 if (theDate.get(Calendar.AM_PM) == Calendar.PM)
159 readable += " / " + pad2(hour);
160 readable += ":" + pad2(theDate.get(Calendar.MINUTE));
165 * deleteForbiddenTags
166 * this method deletes all <script>, <body> and <head>-tags
168 public static final String deleteForbiddenTags(String haystack) {
169 return re_forbiddenTags.substituteAll(haystack,"");
173 * deleteHTMLTableTags
174 * this method deletes all <table>, <tr> and <td>-tags
176 public static final String deleteHTMLTableTags(String haystack) {
177 return re_tables.substituteAll(haystack,"");
181 * wandelt eine Datum in einen 8-buchstabigen String, der durch <code>/</code>
185 * @return String mit <code>/yyyy/mm/dd</code>
187 public static final String webdbDate2path (String webdbDate) {
188 StringBuffer path = new StringBuffer();
189 path.append("/").append(webdbDate.substring(0, 4));
190 path.append("/").append(webdbDate.substring(4, 6));
193 //path.append("/").append(webdbDate.substring(6, 8));
194 return path.toString();
198 * Ersetzt in String <code>s</code> das Regexp <code>pattern</code> durch <code>substitute</code>
202 * @return String mit den Ersetzungen
204 public static String regexpReplace(String haystack, String pattern, String substitute) {
206 RE regex = new RE(pattern);
207 return regex.substituteAll(haystack,substitute);
208 } catch(REException ex){
214 * L?scht <code>/</code> am Ende des Strings, falls vorhanden
216 * @return String ohne <code>/</code> am Ende
218 public static final String removeSlash (String path) {
219 return path.length() > 1 && path.endsWith("/") ? path.substring(0, path.length()
224 * formatiert eine Zahl (0-99) zweistellig (z.B. 5 -> 05)
225 * @return zwistellige Zahl
227 public static String pad2 (int number) {
228 return number < 10 ? "0" + number : String.valueOf(number);
232 * formatiert eine Zahl (0-999) dreistellig (z.B. 7 -> 007)
234 * @return 3-stellige Zahl
236 public static String pad3 (int number) {
237 return number < 10 ? "00" + number : number < 100 ? "0" + number : String.valueOf(number);
241 * Liefert Default-Wert def zur?ck, wenn String <code>s</code>
246 * @return geparster int aus s oder def
248 public static int parseInt(String s, int def) {
249 if (s == null) return def;
251 return Integer.parseInt(s);
252 } catch (NumberFormatException e) {
259 * convertNewline2P ist eine regex-routine zum umwandeln von 2 oder mehr newlines (\n)
260 * in den html-tag <p>
261 * nur sinnvoll, wenn text nicht im html-format eingegeben
263 public static String convertNewline2P(String haystack) {
264 return re_brbr2p.substituteAll(haystack,"\n</p><p>");
268 * convertNewline2Break ist eine regex-routine zum umwandeln von 1 newline (\n)
269 * in den html-tag <br>
270 * nur sinnvoll, wenn text nicht im html-format eingegeben
272 public static String convertNewline2Break(String haystack) {
273 return re_newline2br.substituteAll(haystack,"$0<br />");
277 * createMailLinks wandelt text im email-adressenformat
278 * in einen klickbaren link um
279 * nur sinnvoll, wenn text nicht im html-format eingegeben
281 public static String createMailLinks(String haystack) {
282 return re_mail.substituteAll(haystack,"<a href=\"mailto:$0\">$0</a>");
287 * createMailLinks wandelt text im email-adressenformat
288 * in einen klickbaren link um
289 * nur sinnvoll, wenn text nicht im html-format eingegeben
291 public static String createMailLinks(String haystack, String imageRoot, String mailImage) {
292 return re_mail.substituteAll(haystack,"<img src=\""+imageRoot+"/"+mailImage+"\" border=\"0\"/> <a href=\"mailto:$0\">$0</a>");
297 * createURLLinks wandelt text im url-format
298 * in einen klickbaren link um
299 * nur sinnvoll, wenn text nicht im html-format eingegeben
301 public static String createURLLinks(String haystack) {
302 return re_url.substituteAll(haystack,"<a href=\"$0\">$0</a>");
306 * this routine takes text in url format and makes
307 * a clickaeble "<href>" link removing any "illegal" html tags
308 * @param haystack, the url
309 * @param title, the href link text
310 * @param imagRoot, the place to find icons
311 * @param extImage, the url of the icon to show next to the link
312 * @return a String containing the url
314 public static String createURLLinks(String haystack, String title, String imageRoot,String extImage) {
316 return re_url.substituteAll(haystack,"<img src=\""+imageRoot+"/"+extImage+"\" border=\"0\"/> <a href=\"$0\">$0</a>");
318 title = removeHTMLTags(title);
319 return re_url.substituteAll(haystack,"<img src=\""+imageRoot+"/"+extImage+"\" border=\"0\"/> <a href=\"$0\">"+title+"</a>");
324 * this routine takes text in url format and makes
325 * a clickaeble "<href>" link removing any "illegal" html tags
326 * @param haystack, the url
327 * @param imageRoot, the place to find icons
328 * @param extImage, the url of the icon to show next to the link
329 * @param intImage, unused
330 * @return a String containing the url
332 public static String createURLLinks(String haystack, String title, String imageRoot,String extImage,String intImage) {
333 return createURLLinks(haystack, title, imageRoot, extImage);
337 * this method deletes all html tags
339 public static final String removeHTMLTags(String haystack){
340 return re_tags.substituteAll(haystack,"");
344 * this method deletes all but the approved tags html tags
345 * it also deletes approved tags which contain malicious-looking attributes and doesn't work at all
347 public static String approveHTMLTags(String haystack){
349 String approvedTags="a|img|h1|h2|h3|h4|h5|h6|br|b|i|strong|p";
350 String badAttributes="onAbort|onBlur|onChange|onClick|onDblClick|onDragDrop|onError|onFocus|onKeyDown|onKeyPress|onKeyUp|onLoad|onMouseDown|onMouseMove|onMouseOut|onMouseOver|onMouseUp|onMove|onReset|onResize|onSelect|onSubmit|onUnload";
351 String approvedProtocols="rtsp|http|ftp|https|freenet|mailto";
353 // kill all the bad tags that have attributes
354 String s = "<\\s*/?\\s*(?!(("+approvedTags+")\\s))\\w+\\s[^>]*>";
355 RE regex = new RE(s,RE.REG_ICASE);
356 haystack = regex.substituteAll(haystack,"");
358 // kill all the bad tags that are attributeless
359 regex = new RE("<\\s*/?\\s*(?!(("+approvedTags+")\\s*>))\\w+\\s*>",RE.REG_ICASE);
360 haystack = regex.substituteAll(haystack,"");
362 // kill all the tags which have a javascript attribute like onLoad
363 regex = new RE("<[^>]*("+badAttributes+")[^>]*>",RE.REG_ICASE);
364 haystack = regex.substituteAll(haystack,"");
366 // kill all the tags which include a url to an unacceptable protocol
367 regex = new RE("<\\s*a\\s+[^>]*href=(?!(\'|\")?("+approvedProtocols+"))[^>]*>",RE.REG_ICASE);
368 haystack = regex.substituteAll(haystack,"");
371 } catch(REException ex){
372 ex.printStackTrace();
379 * createHTML ruft alle regex-methoden zum unwandeln eines nicht
380 * htmlcodierten string auf und returnt einen htmlcodierten String
382 public static String createHTML(String content){
383 content=convertNewline2Break(content);
384 content=convertNewline2P(content);
385 content=createMailLinks(content);
386 content=createURLLinks(content);
392 * createHTML ruft alle regex-methoden zum unwandeln eines nicht
393 * htmlcodierten string auf und returnt einen htmlcodierten String
395 public static String createHTML(String content,String producerDocRoot,String mailImage,String extImage,String intImage){
396 content=convertNewline2Break(content);
397 content=convertNewline2P(content);
398 content=createMailLinks(content,producerDocRoot,mailImage);
399 content=createURLLinks(content,null,producerDocRoot,extImage,intImage);
404 * Converts mir's horrible internal date format (yyyy-MM-dd HH:mm:ss+zz) into a java Date
406 * @param anInternalDate
409 public static Date convertMirInternalDateToDate(String anInternalDate) {
410 Calendar calendar = new GregorianCalendar();
420 year = Integer.parseInt(anInternalDate.substring(0,4));
421 month = Integer.parseInt(anInternalDate.substring(5,7));
422 day = Integer.parseInt(anInternalDate.substring(8,10));
423 hours = Integer.parseInt(anInternalDate.substring(11,13));
424 minutes = Integer.parseInt(anInternalDate.substring(14,16));
425 seconds = Integer.parseInt(anInternalDate.substring(17,19));
427 timezoneOffset = Integer.parseInt(anInternalDate.substring(20,22));
428 if (anInternalDate.charAt(19) == '-')
429 timezoneOffset = -timezoneOffset;
431 calendar.setTimeZone(TimeZone.getTimeZone("UTC"));
432 calendar.set(year, month-1, day, hours, minutes, seconds);
433 calendar.add(Calendar.HOUR, -timezoneOffset);
435 return calendar.getTime();