1   /*
2    * Copyright (C) 2009 The Guava Authors
3    *
4    * Licensed under the Apache License, Version 2.0 (the "License");
5    * you may not use this file except in compliance with the License.
6    * You may obtain a copy of the License at
7    *
8    * http://www.apache.org/licenses/LICENSE-2.0
9    *
10   * Unless required by applicable law or agreed to in writing, software
11   * distributed under the License is distributed on an "AS IS" BASIS,
12   * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13   * See the License for the specific language governing permissions and
14   * limitations under the License.
15   */
16  
17  package com.google.common.escape;
18  
19  import static com.google.common.base.Preconditions.checkNotNull;
20  
21  import com.google.common.annotations.Beta;
22  import com.google.common.annotations.GwtCompatible;
23  
24  import java.util.Map;
25  
26  import javax.annotation.Nullable;
27  
28  /**
29   * A {@link UnicodeEscaper} that uses an array to quickly look up replacement
30   * characters for a given code point. An additional safe range is provided that
31   * determines whether code points without specific replacements are to be
32   * considered safe and left unescaped or should be escaped in a general way.
33   *
34   * <p>A good example of usage of this class is for HTML escaping where the
35   * replacement array contains information about the named HTML entities
36   * such as {@code &amp;} and {@code &quot;} while {@link #escapeUnsafe} is
37   * overridden to handle general escaping of the form {@code &#NNNNN;}.
38   *
39   * <p>The size of the data structure used by {@link ArrayBasedUnicodeEscaper} is
40   * proportional to the highest valued code point that requires escaping.
41   * For example a replacement map containing the single character
42   * '{@code \}{@code u1000}' will require approximately 16K of memory. If you
43   * need to create multiple escaper instances that have the same character
44   * replacement mapping consider using {@link ArrayBasedEscaperMap}.
45   *
46   * @author David Beaumont
47   * @since 15.0
48   */
49  @Beta
50  @GwtCompatible
51  public abstract class ArrayBasedUnicodeEscaper extends UnicodeEscaper {
52    // The replacement array (see ArrayBasedEscaperMap).
53    private final char[][] replacements;
54    // The number of elements in the replacement array.
55    private final int replacementsLength;
56    // The first code point in the safe range.
57    private final int safeMin;
58    // The last code point in the safe range.
59    private final int safeMax;
60  
61    // Cropped values used in the fast path range checks.
62    private final char safeMinChar;
63    private final char safeMaxChar;
64  
65    /**
66     * Creates a new ArrayBasedUnicodeEscaper instance with the given replacement
67     * map and specified safe range. If {@code safeMax < safeMin} then no code
68     * points are considered safe.
69     *
70     * <p>If a code point has no mapped replacement then it is checked against the
71     * safe range. If it lies outside that, then {@link #escapeUnsafe} is
72     * called, otherwise no escaping is performed.
73     *
74     * @param replacementMap a map of characters to their escaped representations
75     * @param safeMin the lowest character value in the safe range
76     * @param safeMax the highest character value in the safe range
77     * @param unsafeReplacement the default replacement for unsafe characters or
78     *     null if no default replacement is required
79     */
80    protected ArrayBasedUnicodeEscaper(Map<Character, String> replacementMap,
81        int safeMin, int safeMax, @Nullable String unsafeReplacement) {
82  
83      this(ArrayBasedEscaperMap.create(replacementMap), safeMin, safeMax,
84          unsafeReplacement);
85    }
86  
87    /**
88     * Creates a new ArrayBasedUnicodeEscaper instance with the given replacement
89     * map and specified safe range. If {@code safeMax < safeMin} then no code
90     * points are considered safe. This initializer is useful when explicit
91     * instances of ArrayBasedEscaperMap are used to allow the sharing of large
92     * replacement mappings.
93     *
94     * <p>If a code point has no mapped replacement then it is checked against the
95     * safe range. If it lies outside that, then {@link #escapeUnsafe} is
96     * called, otherwise no escaping is performed.
97     *
98     * @param escaperMap the map of replacements
99     * @param safeMin the lowest character value in the safe range
100    * @param safeMax the highest character value in the safe range
101    * @param unsafeReplacement the default replacement for unsafe characters or
102    *     null if no default replacement is required
103    */
104   protected ArrayBasedUnicodeEscaper(ArrayBasedEscaperMap escaperMap,
105       int safeMin, int safeMax, @Nullable String unsafeReplacement) {
106 
107     checkNotNull(escaperMap);  // GWT specific check (do not optimize)
108     this.replacements = escaperMap.getReplacementArray();
109     this.replacementsLength = replacements.length;
110     if (safeMax < safeMin) {
111       // If the safe range is empty, set the range limits to opposite extremes
112       // to ensure the first test of either value will fail.
113       safeMax = -1;
114       safeMin = Integer.MAX_VALUE;
115     }
116     this.safeMin = safeMin;
117     this.safeMax = safeMax;
118 
119     // This is a bit of a hack but lets us do quicker per-character checks in
120     // the fast path code. The safe min/max values are very unlikely to extend
121     // into the range of surrogate characters, but if they do we must not test
122     // any values in that range. To see why, consider the case where:
123     //   safeMin <= {hi,lo} <= safeMax
124     // where {hi,lo} are characters forming a surrogate pair such that:
125     //   codePointOf(hi, lo) > safeMax
126     // which would result in the surrogate pair being (wrongly) considered safe.
127     // If we clip the safe range used during the per-character tests so it is
128     // below the values of characters in surrogate pairs, this cannot occur.
129     // This approach does mean that we break out of the fast path code in cases
130     // where we don't strictly need to, but this situation will almost never
131     // occur in practice.
132     if (safeMin >= Character.MIN_HIGH_SURROGATE) {
133       // The safe range is empty or the all safe code points lie in or above the
134       // surrogate range. Either way the character range is empty.
135       this.safeMinChar = Character.MAX_VALUE;
136       this.safeMaxChar = 0;
137     } else {
138       // The safe range is non empty and contains values below the surrogate
139       // range but may extend above it. We may need to clip the maximum value.
140       this.safeMinChar = (char) safeMin;
141       this.safeMaxChar = (char) Math.min(safeMax,
142                                          Character.MIN_HIGH_SURROGATE - 1);
143     }
144   }
145 
146   /*
147    * This is overridden to improve performance. Rough benchmarking shows that
148    * this almost doubles the speed when processing strings that do not require
149    * any escaping.
150    */
151   @Override
152   public final String escape(String s) {
153     checkNotNull(s);  // GWT specific check (do not optimize)
154     for (int i = 0; i < s.length(); i++) {
155       char c = s.charAt(i);
156       if ((c < replacementsLength && replacements[c] != null) ||
157           c > safeMaxChar || c < safeMinChar) {
158         return escapeSlow(s, i);
159       }
160     }
161     return s;
162   }
163 
164   /* Overridden for performance. */
165   @Override
166   protected final int nextEscapeIndex(CharSequence csq, int index, int end) {
167     while (index < end) {
168       char c = csq.charAt(index);
169       if ((c < replacementsLength && replacements[c] != null) ||
170           c > safeMaxChar || c < safeMinChar) {
171         break;
172       }
173       index++;
174     }
175     return index;
176   }
177 
178   /**
179    * Escapes a single Unicode code point using the replacement array and safe
180    * range values. If the given character does not have an explicit replacement
181    * and lies outside the safe range then {@link #escapeUnsafe} is called.
182    */
183   @Override
184   protected final char[] escape(int cp) {
185     if (cp < replacementsLength) {
186       char[] chars = replacements[cp];
187       if (chars != null) {
188         return chars;
189       }
190     }
191     if (cp >= safeMin && cp <= safeMax) {
192       return null;
193     }
194     return escapeUnsafe(cp);
195   }
196 
197   /**
198    * Escapes a code point that has no direct explicit value in the replacement
199    * array and lies outside the stated safe range. Subclasses should override
200    * this method to provide generalized escaping for code points if required.
201    *
202    * <p>Note that arrays returned by this method must not be modified once they
203    * have been returned. However it is acceptable to return the same array
204    * multiple times (even for different input characters).
205    *
206    * @param cp the Unicode code point to escape
207    * @return the replacement characters, or {@code null} if no escaping was
208    *         required
209    */
210   protected abstract char[] escapeUnsafe(int cp);
211 }