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; 021 022/** 023 * An object that converts literal text into a format safe for inclusion in a particular context 024 * (such as an XML document). Typically (but not always), the inverse process of "unescaping" the 025 * text is performed automatically by the relevant parser. 026 * 027 * <p>For example, an XML escaper would convert the literal string {@code "Foo<Bar>"} into {@code 028 * "Foo<Bar>"} to prevent {@code "<Bar>"} from being confused with an XML tag. When the 029 * resulting XML document is parsed, the parser API will return this text as the original literal 030 * string {@code "Foo<Bar>"}. 031 * 032 * <p>A {@code CharEscaper} instance is required to be stateless, and safe when used concurrently by 033 * multiple threads. 034 * 035 * <p>Popular escapers are defined as constants in classes like {@link 036 * com.google.common.html.HtmlEscapers} and {@link com.google.common.xml.XmlEscapers}. To create 037 * your own escapers extend this class and implement the {@link #escape(char)} method. 038 * 039 * @author Sven Mawson 040 * @since 15.0 041 */ 042@Beta 043@GwtCompatible 044public abstract class CharEscaper extends Escaper { 045 /** Constructor for use by subclasses. */ 046 protected CharEscaper() {} 047 048 /** 049 * Returns the escaped form of a given literal string. 050 * 051 * @param string the literal string to be escaped 052 * @return the escaped form of {@code string} 053 * @throws NullPointerException if {@code string} is null 054 */ 055 @Override 056 public String escape(String string) { 057 checkNotNull(string); // GWT specific check (do not optimize) 058 // Inlineable fast-path loop which hands off to escapeSlow() only if needed 059 int length = string.length(); 060 for (int index = 0; index < length; index++) { 061 if (escape(string.charAt(index)) != null) { 062 return escapeSlow(string, index); 063 } 064 } 065 return string; 066 } 067 068 /** 069 * Returns the escaped form of the given character, or {@code null} if this character does not 070 * need to be escaped. If an empty array is returned, this effectively strips the input character 071 * from the resulting text. 072 * 073 * <p>If the character does not need to be escaped, this method should return {@code null}, rather 074 * than a one-character array containing the character itself. This enables the escaping algorithm 075 * to perform more efficiently. 076 * 077 * <p>An escaper is expected to be able to deal with any {@code char} value, so this method should 078 * not throw any exceptions. 079 * 080 * @param c the character to escape if necessary 081 * @return the replacement characters, or {@code null} if no escaping was needed 082 */ 083 protected abstract char[] escape(char c); 084 085 /** 086 * Returns the escaped form of a given literal string, starting at the given index. This method is 087 * called by the {@link #escape(String)} method when it discovers that escaping is required. It is 088 * protected to allow subclasses to override the fastpath escaping function to inline their 089 * escaping test. See {@link CharEscaperBuilder} for an example usage. 090 * 091 * @param s the literal string to be escaped 092 * @param index the index to start escaping from 093 * @return the escaped form of {@code string} 094 * @throws NullPointerException if {@code string} is null 095 */ 096 protected final String escapeSlow(String s, int index) { 097 int slen = s.length(); 098 099 // Get a destination buffer and setup some loop variables. 100 char[] dest = Platform.charBufferFromThreadLocal(); 101 int destSize = dest.length; 102 int destIndex = 0; 103 int lastEscape = 0; 104 105 // Loop through the rest of the string, replacing when needed into the 106 // destination buffer, which gets grown as needed as well. 107 for (; index < slen; index++) { 108 109 // Get a replacement for the current character. 110 char[] r = escape(s.charAt(index)); 111 112 // If no replacement is needed, just continue. 113 if (r == null) { 114 continue; 115 } 116 117 int rlen = r.length; 118 int charsSkipped = index - lastEscape; 119 120 // This is the size needed to add the replacement, not the full size 121 // needed by the string. We only regrow when we absolutely must, and 122 // when we do grow, grow enough to avoid excessive growing. Grow. 123 int sizeNeeded = destIndex + charsSkipped + rlen; 124 if (destSize < sizeNeeded) { 125 destSize = sizeNeeded + DEST_PAD_MULTIPLIER * (slen - index); 126 dest = growBuffer(dest, destIndex, destSize); 127 } 128 129 // If we have skipped any characters, we need to copy them now. 130 if (charsSkipped > 0) { 131 s.getChars(lastEscape, index, dest, destIndex); 132 destIndex += charsSkipped; 133 } 134 135 // Copy the replacement string into the dest buffer as needed. 136 if (rlen > 0) { 137 System.arraycopy(r, 0, dest, destIndex, rlen); 138 destIndex += rlen; 139 } 140 lastEscape = index + 1; 141 } 142 143 // Copy leftover characters if there are any. 144 int charsLeft = slen - lastEscape; 145 if (charsLeft > 0) { 146 int sizeNeeded = destIndex + charsLeft; 147 if (destSize < sizeNeeded) { 148 149 // Regrow and copy, expensive! No padding as this is the final copy. 150 dest = growBuffer(dest, destIndex, sizeNeeded); 151 } 152 s.getChars(lastEscape, slen, dest, destIndex); 153 destIndex = sizeNeeded; 154 } 155 return new String(dest, 0, destIndex); 156 } 157 158 /** 159 * Helper method to grow the character buffer as needed, this only happens once in a while so it's 160 * ok if it's in a method call. If the index passed in is 0 then no copying will be done. 161 */ 162 private static char[] growBuffer(char[] dest, int index, int size) { 163 if (size < 0) { // overflow - should be OutOfMemoryError but GWT/j2cl don't support it 164 throw new AssertionError("Cannot increase internal buffer any further"); 165 } 166 char[] copy = new char[size]; 167 if (index > 0) { 168 System.arraycopy(dest, 0, copy, 0, index); 169 } 170 return copy; 171 } 172 173 /** The multiplier for padding to use when growing the escape buffer. */ 174 private static final int DEST_PAD_MULTIPLIER = 2; 175}