Source code

001/*
002 * SonarQube
003 * Copyright (C) 2009-2017 SonarSource SA
004 * mailto:info AT sonarsource DOT com
005 *
006 * This program is free software; you can redistribute it and/or
007 * modify it under the terms of the GNU Lesser General Public
008 * License as published by the Free Software Foundation; either
009 * version 3 of the License, or (at your option) any later version.
010 *
011 * This program is distributed in the hope that it will be useful,
012 * but WITHOUT ANY WARRANTY; without even the implied warranty of
013 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
014 * Lesser General Public License for more details.
015 *
016 * You should have received a copy of the GNU Lesser General Public License
017 * along with this program; if not, write to the Free Software Foundation,
018 * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
019 */
020package org.sonar.api.utils;
021
022import java.util.HashMap;
023import java.util.Map;
024import java.util.regex.Pattern;
025import javax.annotation.Nullable;
026import org.apache.commons.lang.StringUtils;
027
028/**
029 * Implementation of Ant-style matching patterns.
030 * Contrary to other implementations (like AntPathMatcher from Spring Framework) it is based on {@link Pattern Java Regular Expressions}.
031 * To increase performance it holds an internal cache of all processed patterns.
032 * <p>
033 * Following rules are applied:
034 * <ul>
035 * <li>? matches single character</li>
036 * <li>* matches zero or more characters</li>
037 * <li>** matches zero or more 'directories'</li>
038 * </ul>
039 * <p>
040 * Some examples of patterns:
041 * <ul>
042 * <li><code>org/T?st.java</code> - matches <code>org/Test.java</code> and also <code>org/Tost.java</code></li>
043 * <li><code>org/*.java</code> - matches all <code>.java</code> files in the <code>org</code> directory,
044 * e.g. <code>org/Foo.java</code> or <code>org/Bar.java</code></li>
045 * <li><code>org/**</code> - matches all files underneath the <code>org</code> directory,
046 * e.g. <code>org/Foo.java</code> or <code>org/foo/bar.jsp</code></li>
047 * <li><code>org/&#42;&#42;/Test.java</code> - matches all <code>Test.java</code> files underneath the <code>org</code> directory,
048 * e.g. <code>org/Test.java</code> or <code>org/foo/Test.java</code> or <code>org/foo/bar/Test.java</code></li>
049 * <li><code>org/&#42;&#42;/*.java</code> - matches all <code>.java</code> files underneath the <code>org</code> directory,
050 * e.g. <code>org/Foo.java</code> or <code>org/foo/Bar.java</code> or <code>org/foo/bar/Baz.java</code></li>
051 * </ul>
052 * <p>
053 * Another implementation, which is also based on Java Regular Expressions, can be found in
054 * <a href="https://github.com/JetBrains/intellij-community/blob/idea/107.743/platform/util/src/com/intellij/openapi/util/io/FileUtil.java#L847">FileUtil</a>
055 * from IntelliJ OpenAPI.
056 * 
057 * 
058 * @since 1.10
059 */
060public class WildcardPattern {
061
062  private static final Map<String, WildcardPattern> CACHE = new HashMap<>();
063  private static final String SPECIAL_CHARS = "()[]^$.{}+|";
064
065  private Pattern pattern;
066  private String stringRepresentation;
067
068  protected WildcardPattern(String pattern, String directorySeparator) {
069    this.stringRepresentation = pattern;
070    this.pattern = Pattern.compile(toRegexp(pattern, directorySeparator));
071  }
072
073  private static String toRegexp(String antPattern, String directorySeparator) {
074    final String escapedDirectorySeparator = '\\' + directorySeparator;
075
076    final StringBuilder sb = new StringBuilder(antPattern.length());
077
078    sb.append('^');
079
080    int i = antPattern.startsWith("/") || antPattern.startsWith("\\") ? 1 : 0;
081    while (i < antPattern.length()) {
082      final char ch = antPattern.charAt(i);
083
084      if (SPECIAL_CHARS.indexOf(ch) != -1) {
085        // Escape regexp-specific characters
086        sb.append('\\').append(ch);
087      } else if (ch == '*') {
088        if (i + 1 < antPattern.length() && antPattern.charAt(i + 1) == '*') {
089          // Double asterisk
090          // Zero or more directories
091          if (i + 2 < antPattern.length() && isSlash(antPattern.charAt(i + 2))) {
092            sb.append("(?:.*").append(escapedDirectorySeparator).append("|)");
093            i += 2;
094          } else {
095            sb.append(".*");
096            i += 1;
097          }
098        } else {
099          // Single asterisk
100          // Zero or more characters excluding directory separator
101          sb.append("[^").append(escapedDirectorySeparator).append("]*?");
102        }
103      } else if (ch == '?') {
104        // Any single character excluding directory separator
105        sb.append("[^").append(escapedDirectorySeparator).append("]");
106      } else if (isSlash(ch)) {
107        // Directory separator
108        sb.append(escapedDirectorySeparator);
109      } else {
110        // Single character
111        sb.append(ch);
112      }
113
114      i++;
115    }
116
117    sb.append('$');
118
119    return sb.toString();
120  }
121
122  private static boolean isSlash(char ch) {
123    return ch == '/' || ch == '\\';
124  }
125
126  /**
127   * Returns string representation of this pattern.
128   * 
129   * @since 2.5
130   */
131  @Override
132  public String toString() {
133    return stringRepresentation;
134  }
135
136  /**
137   * Returns true if specified value matches this pattern.
138   */
139  public boolean match(String value) {
140    value = StringUtils.removeStart(value, "/");
141    value = StringUtils.removeEnd(value, "/");
142    return pattern.matcher(value).matches();
143  }
144
145  /**
146   * Returns true if specified value matches one of specified patterns.
147   * 
148   * @since 2.4
149   */
150  public static boolean match(WildcardPattern[] patterns, String value) {
151    for (WildcardPattern pattern : patterns) {
152      if (pattern.match(value)) {
153        return true;
154      }
155    }
156    return false;
157  }
158
159  /**
160   * Creates pattern with "/" as a directory separator.
161   * 
162   * @see #create(String, String)
163   */
164  public static WildcardPattern create(String pattern) {
165    return create(pattern, "/");
166  }
167
168  /**
169   * Creates array of patterns with "/" as a directory separator.
170   * 
171   * @see #create(String, String)
172   */
173  public static WildcardPattern[] create(@Nullable String[] patterns) {
174    if (patterns == null) {
175      return new WildcardPattern[0];
176    }
177    WildcardPattern[] exclusionPAtterns = new WildcardPattern[patterns.length];
178    for (int i = 0; i < patterns.length; i++) {
179      exclusionPAtterns[i] = create(patterns[i]);
180    }
181    return exclusionPAtterns;
182  }
183
184  /**
185   * Creates pattern with specified separator for directories.
186   * <p>
187   * This is used to match Java-classes, i.e. <code>org.foo.Bar</code> against <code>org/**</code>.
188   * <b>However usage of character other than "/" as a directory separator is misleading and should be avoided,
189   * so method {@link #create(String)} is preferred over this one.</b>
190   * 
191   * <p>
192   * Also note that no matter whether forward or backward slashes were used in the <code>antPattern</code>
193   * the returned pattern will use <code>directorySeparator</code>.
194   * Thus to match Windows-style path "dir\file.ext" against pattern "dir/file.ext" normalization should be performed.
195   * 
196   */
197  public static WildcardPattern create(String pattern, String directorySeparator) {
198    String key = pattern + directorySeparator;
199    WildcardPattern wildcardPattern = CACHE.get(key);
200    if (wildcardPattern == null) {
201      wildcardPattern = new WildcardPattern(pattern, directorySeparator);
202      CACHE.put(key, wildcardPattern);
203    }
204    return wildcardPattern;
205  }
206}