Source code

001/*
002 * Licensed to the Apache Software Foundation (ASF) under one or more
003 * contributor license agreements.  See the NOTICE file distributed with
004 * this work for additional information regarding copyright ownership.
005 * The ASF licenses this file to You under the Apache License, Version 2.0
006 * (the "License"); you may not use this file except in compliance with
007 * the License.  You may obtain a copy of the License at
008 *
009 *      https://www.apache.org/licenses/LICENSE-2.0
010 *
011 * Unless required by applicable law or agreed to in writing, software
012 * distributed under the License is distributed on an "AS IS" BASIS,
013 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
014 * See the License for the specific language governing permissions and
015 * limitations under the License.
016 */
017package org.apache.commons.validator;
018
019import java.io.Serializable;
020import java.util.Arrays;
021import java.util.HashSet;
022import java.util.Set;
023import java.util.regex.Matcher;
024import java.util.regex.Pattern;
025
026import org.apache.commons.validator.routines.InetAddressValidator;
027import org.apache.commons.validator.util.Flags;
028
029/**
030 * <p>Validates URLs.</p>
031 * Behaviour of validation is modified by passing in options:
032 * <ul>
033 * <li>ALLOW_2_SLASHES - [FALSE]  Allows double '/' characters in the path
034 * component.</li>
035 * <li>NO_FRAGMENT- [FALSE]  By default fragments are allowed, if this option is
036 * included then fragments are flagged as illegal.</li>
037 * <li>ALLOW_ALL_SCHEMES - [FALSE] By default only http, https, and ftp are
038 * considered valid schemes.  Enabling this option will let any scheme pass validation.</li>
039 * </ul>
040 *
041 * <p>Originally based in on php script by Debbie Dyer, validation.php v1.2b, Date: 03/07/02,
042 * https://javascript.internet.com. However, this validation now bears little resemblance
043 * to the php original.</p>
044 * <pre>
045 *   Example of usage:
046 *   Construct a UrlValidator with valid schemes of "http", and "https".
047 *
048 *    String[] schemes = {"http","https"}.
049 *    UrlValidator urlValidator = new UrlValidator(schemes);
050 *    if (urlValidator.isValid("ftp://foo.bar.com/")) {
051 *       System.out.println("URL is valid");
052 *    } else {
053 *       System.out.println("URL is invalid");
054 *    }
055 *
056 *    prints "URL is invalid"
057 *   If instead the default constructor is used.
058 *
059 *    UrlValidator urlValidator = new UrlValidator();
060 *    if (urlValidator.isValid("ftp://foo.bar.com/")) {
061 *       System.out.println("URL is valid");
062 *    } else {
063 *       System.out.println("URL is invalid");
064 *    }
065 *
066 *   prints out "URL is valid"
067 *  </pre>
068 *
069 * @see
070 * <a href="https://www.ietf.org/rfc/rfc2396.txt">
071 *  Uniform Resource Identifiers (URI): Generic Syntax
072 * </a>
073 *
074 * @since 1.1
075 * @deprecated Use the new UrlValidator in the routines package. This class
076 * will be removed in a future release.
077 */
078@Deprecated
079public class UrlValidator implements Serializable {
080
081    private static final int TOP_LEVEL_MAX_LEN = 4;
082
083    private static final int TOP_LEVEL_MIN_LEN = 2;
084
085    private static final long serialVersionUID = 24137157400029593L;
086
087    /**
088     * Allows all validly formatted schemes to pass validation instead of
089     * supplying a set of valid schemes.
090     */
091    public static final int ALLOW_ALL_SCHEMES = 1 << 0;
092
093    /**
094     * Allow two slashes in the path component of the URL.
095     */
096    public static final int ALLOW_2_SLASHES = 1 << 1;
097
098    /**
099     * Enabling this options disallows any URL fragments.
100     */
101    public static final int NO_FRAGMENTS = 1 << 2;
102
103    private static final String ALPHA_CHARS = "a-zA-Z";
104
105// NOT USED   private static final String ALPHA_NUMERIC_CHARS = ALPHA_CHARS + "\\d";
106
107    private static final String SPECIAL_CHARS = ";/@&=,.?:+$";
108
109    private static final String VALID_CHARS = "[^\\s" + SPECIAL_CHARS + "]";
110
111    // Drop numeric, and  "+-." for now
112    private static final String AUTHORITY_CHARS_REGEX = "\\p{Alnum}\\-\\.";
113
114    private static final String ATOM = VALID_CHARS + '+';
115
116    /**
117     * This expression derived/taken from the BNF for URI (RFC2396).
118     */
119    private static final String URL_REGEX =
120            "^(([^:/?#]+):)?(//([^/?#]*))?([^?#]*)(\\?([^#]*))?(#(.*))?";
121    //                                                                      12            3  4          5       6   7        8 9
122    private static final Pattern URL_PATTERN = Pattern.compile(URL_REGEX);
123
124    /**
125     * Schema/Protocol (ie. http:, ftp:, file:, etc).
126     */
127    private static final int PARSE_URL_SCHEME = 2;
128
129    /**
130     * Includes hostname/ip and port number.
131     */
132    private static final int PARSE_URL_AUTHORITY = 4;
133
134    private static final int PARSE_URL_PATH = 5;
135
136    private static final int PARSE_URL_QUERY = 7;
137
138    private static final int PARSE_URL_FRAGMENT = 9;
139
140    /**
141     * Protocol (for example, http:, ftp:, https:).
142     */
143    private static final Pattern SCHEME_PATTERN = Pattern.compile("^\\p{Alpha}[\\p{Alnum}\\+\\-\\.]*");
144
145    private static final String AUTHORITY_REGEX =
146       "^([" + AUTHORITY_CHARS_REGEX + "]*)(:\\d*)?(.*)?";
147    //                                                                            1                          2  3       4
148    private static final Pattern AUTHORITY_PATTERN = Pattern.compile(AUTHORITY_REGEX);
149
150    private static final int PARSE_AUTHORITY_HOST_IP = 1;
151
152    private static final int PARSE_AUTHORITY_PORT = 2;
153
154    /**
155     * Should always be empty.
156     */
157    private static final int PARSE_AUTHORITY_EXTRA = 3;
158
159    private static final Pattern PATH_PATTERN = Pattern.compile("^(/[-\\w:@&?=+,.!/~*'%$_;]*)?$");
160
161    private static final Pattern QUERY_PATTERN = Pattern.compile("^(.*)$");
162
163    private static final Pattern LEGAL_ASCII_PATTERN = Pattern.compile("^\\p{ASCII}+$");
164
165    private static final Pattern DOMAIN_PATTERN =
166            Pattern.compile("^" + ATOM + "(\\." + ATOM + ")*$");
167
168    private static final Pattern PORT_PATTERN = Pattern.compile("^:(\\d{1,5})$");
169
170    private static final Pattern ATOM_PATTERN = Pattern.compile("^(" + ATOM + ").*?$");
171
172    private static final Pattern ALPHA_PATTERN = Pattern.compile("^[" + ALPHA_CHARS + "]");
173
174    /**
175     * Holds the set of current validation options.
176     */
177    private final Flags options;
178
179    /**
180     * The set of schemes that are allowed to be in a URL.
181     */
182    private final Set<String> allowedSchemes = new HashSet<>();
183
184    /**
185     * If no schemes are provided, default to this set.
186     */
187    protected String[] defaultSchemes = {"http", "https", "ftp"};
188
189    /**
190     * Create a UrlValidator with default properties.
191     */
192    public UrlValidator() {
193        this(null);
194    }
195
196    /**
197     * Initialize a UrlValidator with the given validation options.
198     *
199     * @param options The options should be set using the public constants declared in
200     * this class.  To set multiple options you simply add them together.  For example,
201     * ALLOW_2_SLASHES + NO_FRAGMENTS enables both of those options.
202     */
203    public UrlValidator(final int options) {
204        this(null, options);
205    }
206
207    /**
208     * Behavior of validation is modified by passing in several strings options:
209     *
210     * @param schemes Pass in one or more URL schemes to consider valid, passing in
211     *        a null will default to "http,https,ftp" being valid.
212     *        If a non-null schemes is specified then all valid schemes must
213     *        be specified. Setting the ALLOW_ALL_SCHEMES option will
214     *        ignore the contents of schemes.
215     */
216    public UrlValidator(final String[] schemes) {
217        this(schemes, 0);
218    }
219
220    /**
221     * Behaviour of validation is modified by passing in options:
222     *
223     * @param schemes The set of valid schemes.
224     * @param options The options should be set using the public constants declared in
225     * this class.  To set multiple options you simply add them together.  For example,
226     * ALLOW_2_SLASHES + NO_FRAGMENTS enables both of those options.
227     */
228    public UrlValidator(String[] schemes, final int options) {
229        this.options = new Flags(options);
230
231        if (this.options.isOn(ALLOW_ALL_SCHEMES)) {
232            return;
233        }
234
235        if (schemes == null) {
236            schemes = defaultSchemes;
237        }
238
239        allowedSchemes.addAll(Arrays.asList(schemes));
240    }
241
242    /**
243     * Returns the number of times the token appears in the target.
244     *
245     * @param token Token value to be counted.
246     * @param target Target value to count tokens in.
247     * @return the number of tokens.
248     */
249    protected int countToken(final String token, final String target) {
250        int tokenIndex = 0;
251        int count = 0;
252        while (tokenIndex != -1) {
253            tokenIndex = target.indexOf(token, tokenIndex);
254            if (tokenIndex > -1) {
255                tokenIndex++;
256                count++;
257            }
258        }
259        return count;
260    }
261
262    /**
263     * <p>Checks if a field has a valid URL address.</p>
264     *
265     * @param value The value validation is being performed on.  A {@code null}
266     * value is considered invalid.
267     * @return true if the URL is valid.
268     */
269    public boolean isValid(final String value) {
270        if (value == null || !LEGAL_ASCII_PATTERN.matcher(value).matches()) {
271           return false;
272        }
273
274        // Check the whole url address structure
275        final Matcher urlMatcher = URL_PATTERN.matcher(value);
276        if (!urlMatcher.matches() || !isValidScheme(urlMatcher.group(PARSE_URL_SCHEME)) || !isValidAuthority(urlMatcher.group(PARSE_URL_AUTHORITY)) || !isValidPath(urlMatcher.group(PARSE_URL_PATH))) {
277            return false;
278        }
279
280        if (!isValidQuery(urlMatcher.group(PARSE_URL_QUERY))) {
281            return false;
282        }
283
284        if (!isValidFragment(urlMatcher.group(PARSE_URL_FRAGMENT))) {
285            return false;
286        }
287
288        return true;
289    }
290
291    /**
292     * Returns true if the authority is properly formatted.  An authority is the combination
293     * of hostname and port.  A {@code null} authority value is considered invalid.
294     *
295     * @param authority Authority value to validate.
296     * @return true if authority (hostname and port) is valid.
297     */
298    protected boolean isValidAuthority(final String authority) {
299        if (authority == null) {
300            return false;
301        }
302
303        final InetAddressValidator inetAddressValidator =
304                InetAddressValidator.getInstance();
305
306        final Matcher authorityMatcher = AUTHORITY_PATTERN.matcher(authority);
307        if (!authorityMatcher.matches()) {
308            return false;
309        }
310
311        boolean hostname = false;
312        // check if authority is IP address or hostname
313        String hostIP = authorityMatcher.group(PARSE_AUTHORITY_HOST_IP);
314        final boolean ipV4Address = inetAddressValidator.isValid(hostIP);
315
316        if (!ipV4Address) {
317            // Domain is hostname name
318            hostname = DOMAIN_PATTERN.matcher(hostIP).matches();
319        }
320
321        //rightmost hostname will never start with a digit.
322        if (hostname) {
323            // LOW-TECH FIX FOR VALIDATOR-202
324            // TODO: Rewrite to use ArrayList and .add semantics: see VALIDATOR-203
325            final char[] chars = hostIP.toCharArray();
326            int size = 1;
327            for (final char element : chars) {
328                if (element == '.') {
329                    size++;
330                }
331            }
332            final String[] domainSegment = new String[size];
333            boolean match = true;
334            int segmentCount = 0;
335            int segmentLength = 0;
336
337            while (match) {
338                final Matcher atomMatcher = ATOM_PATTERN.matcher(hostIP);
339                match = atomMatcher.matches();
340                if (match) {
341                    domainSegment[segmentCount] = atomMatcher.group(1);
342                    segmentLength = domainSegment[segmentCount].length() + 1;
343                    hostIP =
344                            segmentLength >= hostIP.length()
345                            ? ""
346                            : hostIP.substring(segmentLength);
347
348                    segmentCount++;
349                }
350            }
351            final String topLevel = domainSegment[segmentCount - 1];
352
353
354            // First letter of top level must be an alpha
355            // Make sure there's a host name preceding the authority.
356            if (topLevel.length() < TOP_LEVEL_MIN_LEN || topLevel.length() > TOP_LEVEL_MAX_LEN || !ALPHA_PATTERN.matcher(topLevel.substring(0, 1)).matches()
357                    || segmentCount < 2) {
358                return false;
359            }
360        }
361
362        if (!hostname && !ipV4Address) {
363            return false;
364        }
365
366        final String port = authorityMatcher.group(PARSE_AUTHORITY_PORT);
367        if (port != null && !PORT_PATTERN.matcher(port).matches()) {
368            return false;
369        }
370
371        final String extra = authorityMatcher.group(PARSE_AUTHORITY_EXTRA);
372        if (!GenericValidator.isBlankOrNull(extra)) {
373            return false;
374        }
375
376        return true;
377    }
378
379    /**
380     * Returns true if the given fragment is null or fragments are allowed.
381     *
382     * @param fragment Fragment value to validate.
383     * @return true if fragment is valid.
384     */
385    protected boolean isValidFragment(final String fragment) {
386        if (fragment == null) {
387            return true;
388        }
389
390        return options.isOff(NO_FRAGMENTS);
391    }
392
393    /**
394     * Returns true if the path is valid.  A {@code null} value is considered invalid.
395     *
396     * @param path Path value to validate.
397     * @return true if path is valid.
398     */
399    protected boolean isValidPath(final String path) {
400        if (path == null || !PATH_PATTERN.matcher(path).matches()) {
401            return false;
402        }
403
404        final int slash2Count = countToken("//", path);
405        if (options.isOff(ALLOW_2_SLASHES) && slash2Count > 0) {
406            return false;
407        }
408
409        final int slashCount = countToken("/", path);
410        final int dot2Count = countToken("..", path);
411        if (dot2Count > 0 && slashCount - slash2Count - 1 <= dot2Count) {
412            return false;
413        }
414
415        return true;
416    }
417
418    /**
419     * Returns true if the query is null, or it's a properly formatted query string.
420     *
421     * @param query Query value to validate.
422     * @return true if query is valid.
423     */
424    protected boolean isValidQuery(final String query) {
425        if (query == null) {
426            return true;
427        }
428
429        return QUERY_PATTERN.matcher(query).matches();
430    }
431
432    /**
433     * Validate scheme. If schemes[] was initialized to a non-null,
434     * then only those schemes are allowed.  Note this is slightly different
435     * than for the constructor.
436     *
437     * @param scheme The scheme to validate.  A {@code null} value is considered
438     * invalid.
439     * @return true if valid.
440     */
441    protected boolean isValidScheme(final String scheme) {
442        if (scheme == null || !SCHEME_PATTERN.matcher(scheme).matches() || options.isOff(ALLOW_ALL_SCHEMES) && !allowedSchemes.contains(scheme)) {
443            return false;
444        }
445
446        return true;
447    }
448}