common/escape/Escapers.java

/*
 * Copyright (C) 2009 The Guava Authors
 *
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 * http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

package com.google.common.escape;

import static com.google.common.base.Preconditions.checkNotNull;

import com.google.common.annotations.Beta;
import com.google.common.annotations.GwtCompatible;

import java.util.HashMap;
import java.util.Map;

import javax.annotation.Nullable;

/**
 * Static utility methods pertaining to {@link Escaper} instances.
 *
 * @author Sven Mawson
 * @author David Beaumont
 * @since 15.0
 */
@Beta
@GwtCompatible
public final class Escapers {
  private Escapers() {}

  /**
   * Returns an {@link Escaper} that does no escaping, passing all character
   * data through unchanged.
   */
  public static Escaper nullEscaper() {
    return NULL_ESCAPER;
  }

  // An Escaper that efficiently performs no escaping.
  // Extending CharEscaper (instead of Escaper) makes Escapers.compose() easier.
  private static final Escaper NULL_ESCAPER = new CharEscaper() {
    @Override public String escape(String string) {
      return checkNotNull(string);
    }

    @Override protected char[] escape(char c) {
      // TODO: Fix tests not to call this directly and make it throw an error.
      return null;
    }
  };

  /**
   * Returns a builder for creating simple, fast escapers. A builder instance
   * can be reused and each escaper that is created will be a snapshot of the
   * current builder state. Builders are not thread safe.
   *
   * <p>The initial state of the builder is such that:
   * <ul>
   * <li>There are no replacement mappings<li>
   * <li>{@code safeMin == Character.MIN_VALUE}</li>
   * <li>{@code safeMax == Character.MAX_VALUE}</li>
   * <li>{@code unsafeReplacement == null}</li>
   * </ul>
   * <p>For performance reasons escapers created by this builder are not
   * Unicode aware and will not validate the well-formedness of their input.
   */
  public static Builder builder() {
    return new Builder();
  }

  /**
   * A builder for simple, fast escapers.
   *
   * <p>Typically an escaper needs to deal with the escaping of high valued
   * characters or code points. In these cases it is necessary to extend either
   * {@link ArrayBasedCharEscaper} or {@link ArrayBasedUnicodeEscaper} to
   * provide the desired behavior. However this builder is suitable for creating
   * escapers that replace a relative small set of characters.
   *
   * @author David Beaumont
   * @since 15.0
   */
  @Beta
  public static final class Builder {
    private final Map<Character, String> replacementMap =
        new HashMap<Character, String>();
    private char safeMin = Character.MIN_VALUE;
    private char safeMax = Character.MAX_VALUE;
    private String unsafeReplacement = null;

    // The constructor is exposed via the builder() method above.
    private Builder() {}

    /**
     * Sets the safe range of characters for the escaper. Characters in this
     * range that have no explicit replacement are considered 'safe' and remain
     * unescaped in the output. If {@code safeMax < safeMin} then the safe range
     * is empty.
     *
     * @param safeMin the lowest 'safe' character
     * @param safeMax the highest 'safe' character
     * @return the builder instance
     */
    public Builder setSafeRange(char safeMin, char safeMax) {
      this.safeMin = safeMin;
      this.safeMax = safeMax;
      return this;
    }

    /**
     * Sets the replacement string for any characters outside the 'safe' range
     * that have no explicit replacement. If {@code unsafeReplacement} is
     * {@code null} then no replacement will occur, if it is {@code ""} then
     * the unsafe characters are removed from the output.
     *
     * @param unsafeReplacement the string to replace unsafe chracters
     * @return the builder instance
     */
    public Builder setUnsafeReplacement(@Nullable String unsafeReplacement) {
      this.unsafeReplacement = unsafeReplacement;
      return this;
    }

    /**
     * Adds a replacement string for the given input character. The specified
     * character will be replaced by the given string whenever it occurs in the
     * input, irrespective of whether it lies inside or outside the 'safe'
     * range.
     *
     * @param c the character to be replaced
     * @param replacement the string to replace the given character
     * @return the builder instance
     * @throws NullPointerException if {@code replacement} is null
     */
    public Builder addEscape(char c, String replacement) {
      checkNotNull(replacement);
      // This can replace an existing character (the builder is re-usable).
      replacementMap.put(c, replacement);
      return this;
    }

    /**
     * Returns a new escaper based on the current state of the builder.
     */
    public Escaper build() {
      return new ArrayBasedCharEscaper(replacementMap, safeMin, safeMax) {
        private final char[] replacementChars =
            unsafeReplacement != null ? unsafeReplacement.toCharArray() : null;
        @Override protected char[] escapeUnsafe(char c) {
          return replacementChars;
        }
      };
    }
  }

  /**
   * Returns a {@link UnicodeEscaper} equivalent to the given escaper instance.
   * If the escaper is already a UnicodeEscaper then it is simply returned,
   * otherwise it is wrapped in a UnicodeEscaper.
   *
   * <p>When a {@link CharEscaper} escaper is wrapped by this method it acquires
   * extra behavior with respect to the well-formedness of Unicode character
   * sequences and will throw {@link IllegalArgumentException} when given bad
   * input.
   *
   * @param escaper the instance to be wrapped
   * @return a UnicodeEscaper with the same behavior as the given instance
   * @throws NullPointerException if escaper is null
   * @throws IllegalArgumentException if escaper is not a UnicodeEscaper or a
   *         CharEscaper
   */
  static UnicodeEscaper asUnicodeEscaper(Escaper escaper) {
    checkNotNull(escaper);
    if (escaper instanceof UnicodeEscaper) {
      return (UnicodeEscaper) escaper;
    } else if (escaper instanceof CharEscaper) {
      return wrap((CharEscaper) escaper);
    }
    // In practice this shouldn't happen because it would be very odd not to
    // extend either CharEscaper or UnicodeEscaper for non trivial cases.
    throw new IllegalArgumentException("Cannot create a UnicodeEscaper from: " +
        escaper.getClass().getName());
  }

  /**
   * Returns a string that would replace the given character in the specified
   * escaper, or {@code null} if no replacement should be made. This method is
   * intended for use in tests through the {@code EscaperAsserts} class;
   * production users of {@link CharEscaper} should limit themselves to its
   * public interface.
   *
   * @param c the character to escape if necessary
   * @return the replacement string, or {@code null} if no escaping was needed
   */
  public static String computeReplacement(CharEscaper escaper, char c) {
    return stringOrNull(escaper.escape(c));
  }

  /**
   * Returns a string that would replace the given character in the specified
   * escaper, or {@code null} if no replacement should be made. This method is
   * intended for use in tests through the {@code EscaperAsserts} class;
   * production users of {@link UnicodeEscaper} should limit themselves to its
   * public interface.
   *
   * @param cp the Unicode code point to escape if necessary
   * @return the replacement string, or {@code null} if no escaping was needed
   */
  public static String computeReplacement(UnicodeEscaper escaper, int cp) {
    return stringOrNull(escaper.escape(cp));
  }

  private static String stringOrNull(char[] in) {
    return (in == null) ? null : new String(in);
  }

  /** Private helper to wrap a CharEscaper as a UnicodeEscaper. */
  private static UnicodeEscaper wrap(final CharEscaper escaper) {
    return new UnicodeEscaper() {
      @Override protected char[] escape(int cp) {
        // If a code point maps to a single character, just escape that.
        if (cp < Character.MIN_SUPPLEMENTARY_CODE_POINT) {
          return escaper.escape((char) cp);
        }
        // Convert the code point to a surrogate pair and escape them both.
        // Note: This code path is horribly slow and typically allocates 4 new
        // char[] each time it is invoked. However this avoids any
        // synchronization issues and makes the escaper thread safe.
        char[] surrogateChars = new char[2];
        Character.toChars(cp, surrogateChars, 0);
        char[] hiChars = escaper.escape(surrogateChars[0]);
        char[] loChars = escaper.escape(surrogateChars[1]);

        // If either hiChars or lowChars are non-null, the CharEscaper is trying
        // to escape the characters of a surrogate pair separately. This is
        // uncommon and applies only to escapers that assume UCS-2 rather than
        // UTF-16. See: http://en.wikipedia.org/wiki/UTF-16/UCS-2
        if (hiChars == null && loChars == null) {
          // We expect this to be the common code path for most escapers.
          return null;
        }
        // Combine the characters and/or escaped sequences into a single array.
        int hiCount = hiChars != null ? hiChars.length : 1;
        int loCount = loChars != null ? loChars.length : 1;
        char[] output = new char[hiCount + loCount];
        if (hiChars != null) {
          // TODO: Is this faster than System.arraycopy() for small arrays?
          for (int n = 0; n < hiChars.length; ++n) {
            output[n] = hiChars[n];
          }
        } else {
          output[0] = surrogateChars[0];
        }
        if (loChars != null) {
          for (int n = 0; n < loChars.length; ++n) {
            output[hiCount + n] = loChars[n];
          }
        } else {
          output[hiCount] = surrogateChars[1];
        }
        return output;
      }
    };
  }
}