001/*
002 * Copyright (C) 2006 The Guava Authors
003 *
004 * Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except
005 * in compliance with the License. You may obtain a copy of the License at
006 *
007 * http://www.apache.org/licenses/LICENSE-2.0
008 *
009 * Unless required by applicable law or agreed to in writing, software distributed under the License
010 * is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express
011 * or implied. See the License for the specific language governing permissions and limitations under
012 * the License.
013 */
014
015package com.google.common.escape;
016
017import static com.google.common.base.Preconditions.checkNotNull;
018
019import com.google.common.annotations.Beta;
020import com.google.common.annotations.GwtCompatible;
021import javax.annotation.CheckForNull;
022
023/**
024 * An object that converts literal text into a format safe for inclusion in a particular context
025 * (such as an XML document). Typically (but not always), the inverse process of "unescaping" the
026 * text is performed automatically by the relevant parser.
027 *
028 * <p>For example, an XML escaper would convert the literal string {@code "Foo<Bar>"} into {@code
029 * "Foo&lt;Bar&gt;"} to prevent {@code "<Bar>"} from being confused with an XML tag. When the
030 * resulting XML document is parsed, the parser API will return this text as the original literal
031 * string {@code "Foo<Bar>"}.
032 *
033 * <p>A {@code CharEscaper} instance is required to be stateless, and safe when used concurrently by
034 * multiple threads.
035 *
036 * <p>Popular escapers are defined as constants in classes like {@link
037 * com.google.common.html.HtmlEscapers} and {@link com.google.common.xml.XmlEscapers}. To create
038 * your own escapers extend this class and implement the {@link #escape(char)} method.
039 *
040 * @author Sven Mawson
041 * @since 15.0
042 */
043@Beta
044@GwtCompatible
045@ElementTypesAreNonnullByDefault
046public abstract class CharEscaper extends Escaper {
047  /** Constructor for use by subclasses. */
048  protected CharEscaper() {}
049
050  /**
051   * Returns the escaped form of a given literal string.
052   *
053   * @param string the literal string to be escaped
054   * @return the escaped form of {@code string}
055   * @throws NullPointerException if {@code string} is null
056   */
057  @Override
058  public String escape(String string) {
059    checkNotNull(string); // GWT specific check (do not optimize)
060    // Inlineable fast-path loop which hands off to escapeSlow() only if needed
061    int length = string.length();
062    for (int index = 0; index < length; index++) {
063      if (escape(string.charAt(index)) != null) {
064        return escapeSlow(string, index);
065      }
066    }
067    return string;
068  }
069
070  /**
071   * Returns the escaped form of the given character, or {@code null} if this character does not
072   * need to be escaped. If an empty array is returned, this effectively strips the input character
073   * from the resulting text.
074   *
075   * <p>If the character does not need to be escaped, this method should return {@code null}, rather
076   * than a one-character array containing the character itself. This enables the escaping algorithm
077   * to perform more efficiently.
078   *
079   * <p>An escaper is expected to be able to deal with any {@code char} value, so this method should
080   * not throw any exceptions.
081   *
082   * @param c the character to escape if necessary
083   * @return the replacement characters, or {@code null} if no escaping was needed
084   */
085  @CheckForNull
086  protected abstract char[] escape(char c);
087
088  /**
089   * Returns the escaped form of a given literal string, starting at the given index. This method is
090   * called by the {@link #escape(String)} method when it discovers that escaping is required. It is
091   * protected to allow subclasses to override the fastpath escaping function to inline their
092   * escaping test. See {@link CharEscaperBuilder} for an example usage.
093   *
094   * @param s the literal string to be escaped
095   * @param index the index to start escaping from
096   * @return the escaped form of {@code string}
097   * @throws NullPointerException if {@code string} is null
098   */
099  protected final String escapeSlow(String s, int index) {
100    int slen = s.length();
101
102    // Get a destination buffer and setup some loop variables.
103    char[] dest = Platform.charBufferFromThreadLocal();
104    int destSize = dest.length;
105    int destIndex = 0;
106    int lastEscape = 0;
107
108    // Loop through the rest of the string, replacing when needed into the
109    // destination buffer, which gets grown as needed as well.
110    for (; index < slen; index++) {
111
112      // Get a replacement for the current character.
113      char[] r = escape(s.charAt(index));
114
115      // If no replacement is needed, just continue.
116      if (r == null) {
117        continue;
118      }
119
120      int rlen = r.length;
121      int charsSkipped = index - lastEscape;
122
123      // This is the size needed to add the replacement, not the full size
124      // needed by the string. We only regrow when we absolutely must, and
125      // when we do grow, grow enough to avoid excessive growing. Grow.
126      int sizeNeeded = destIndex + charsSkipped + rlen;
127      if (destSize < sizeNeeded) {
128        destSize = sizeNeeded + DEST_PAD_MULTIPLIER * (slen - index);
129        dest = growBuffer(dest, destIndex, destSize);
130      }
131
132      // If we have skipped any characters, we need to copy them now.
133      if (charsSkipped > 0) {
134        s.getChars(lastEscape, index, dest, destIndex);
135        destIndex += charsSkipped;
136      }
137
138      // Copy the replacement string into the dest buffer as needed.
139      if (rlen > 0) {
140        System.arraycopy(r, 0, dest, destIndex, rlen);
141        destIndex += rlen;
142      }
143      lastEscape = index + 1;
144    }
145
146    // Copy leftover characters if there are any.
147    int charsLeft = slen - lastEscape;
148    if (charsLeft > 0) {
149      int sizeNeeded = destIndex + charsLeft;
150      if (destSize < sizeNeeded) {
151
152        // Regrow and copy, expensive! No padding as this is the final copy.
153        dest = growBuffer(dest, destIndex, sizeNeeded);
154      }
155      s.getChars(lastEscape, slen, dest, destIndex);
156      destIndex = sizeNeeded;
157    }
158    return new String(dest, 0, destIndex);
159  }
160
161  /**
162   * Helper method to grow the character buffer as needed, this only happens once in a while so it's
163   * ok if it's in a method call. If the index passed in is 0 then no copying will be done.
164   */
165  private static char[] growBuffer(char[] dest, int index, int size) {
166    if (size < 0) { // overflow - should be OutOfMemoryError but GWT/j2cl don't support it
167      throw new AssertionError("Cannot increase internal buffer any further");
168    }
169    char[] copy = new char[size];
170    if (index > 0) {
171      System.arraycopy(dest, 0, copy, 0, index);
172    }
173    return copy;
174  }
175
176  /** The multiplier for padding to use when growing the escape buffer. */
177  private static final int DEST_PAD_MULTIPLIER = 2;
178}