Package com.google.protobuf

Source Code of com.google.protobuf.ByteString$Output

// Protocol Buffers - Google's data interchange format
// Copyright 2008 Google Inc.  All rights reserved.
// http://code.google.com/p/protobuf/
//
// Redistribution and use in source and binary forms, with or without
// modification, are permitted provided that the following conditions are
// met:
//
//     * Redistributions of source code must retain the above copyright
// notice, this list of conditions and the following disclaimer.
//     * Redistributions in binary form must reproduce the above
// copyright notice, this list of conditions and the following disclaimer
// in the documentation and/or other materials provided with the
// distribution.
//     * Neither the name of Google Inc. nor the names of its
// contributors may be used to endorse or promote products derived from
// this software without specific prior written permission.
//
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
// OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

package com.google.protobuf;

import java.io.InputStream;
import java.io.ByteArrayInputStream;
import java.io.ByteArrayOutputStream;
import java.io.FilterOutputStream;
import java.io.UnsupportedEncodingException;
import java.nio.ByteBuffer;

/**
* Immutable array of bytes.
*
* @author crazybob@google.com Bob Lee
* @author kenton@google.com Kenton Varda
*/
public final class ByteString {
  private final byte[] bytes;

  private ByteString(byte[] bytes) {
    this.bytes = bytes;
  }

  /**
   * Gets the byte at the given index.
   *
   * @throws ArrayIndexOutOfBoundsException {@code index} is < 0 or >= size
   */
  public byte byteAt(int index) {
    return bytes[index];
  }

  /**
   * Gets the number of bytes.
   */
  public int size() {
    return this.bytes.length;
  }

  /**
   * Returns {@code true} if the size is {@code 0}, {@code false} otherwise.
   */
  public boolean isEmpty() {
    return this.bytes.length == 0;
  }

  // =================================================================
  // byte[] -> ByteString

  /**
   * Empty ByteString.
   */
  public static final ByteString EMPTY = new ByteString(new byte[0]);

  /**
   * Copies the given bytes into a {@code ByteString}.
   */
  public static ByteString copyFrom(byte[] bytes, int offset, int size) {
    byte[] copy = new byte[size];
    System.arraycopy(bytes, offset, copy, 0, size);
    return new ByteString(copy);
  }

  /**
   * Copies the given bytes into a {@code ByteString}.
   */
  public static ByteString copyFrom(byte[] bytes) {
    return copyFrom(bytes, 0, bytes.length);
  }

  /**
   * Encodes {@code text} into a sequence of bytes using the named charset
   * and returns the result as a {@code ByteString}.
   */
  public static ByteString copyFrom(String text, String charsetName)
      throws UnsupportedEncodingException {
    return new ByteString(text.getBytes(charsetName));
  }

  /**
   * Encodes {@code text} into a sequence of UTF-8 bytes and returns the
   * result as a {@code ByteString}.
   */
  public static ByteString copyFromUtf8(String text) {
    try {
      return new ByteString(text.getBytes("UTF-8"));
    } catch (UnsupportedEncodingException e) {
      throw new RuntimeException("UTF-8 not supported?", e);
    }
  }

  // =================================================================
  // ByteString -> byte[]

  /**
   * Copies bytes into a buffer at the given offset.
   *
   * @param target buffer to copy into
   * @param offset in the target buffer
   */
  public void copyTo(byte[] target, int offset) {
    System.arraycopy(bytes, 0, target, offset, bytes.length);
  }

  /**
   * Copies bytes into a buffer.
   *
   * @param target buffer to copy into
   * @param sourceOffset offset within these bytes
   * @param targetOffset offset within the target buffer
   * @param size number of bytes to copy
   */
  public void copyTo(byte[] target, int sourceOffset, int targetOffset,
      int size) {
    System.arraycopy(bytes, sourceOffset, target, targetOffset, size);
  }

  /**
   * Copies bytes to a {@code byte[]}.
   */
  public byte[] toByteArray() {
    int size = this.bytes.length;
    byte[] copy = new byte[size];
    System.arraycopy(this.bytes, 0, copy, 0, size);
    return copy;
  }

  /**
   * Constructs a new read-only {@code java.nio.ByteBuffer} with the
   * same backing byte array.
   */
  public ByteBuffer asReadOnlyByteBuffer() {
    ByteBuffer byteBuffer = ByteBuffer.wrap(this.bytes);
    return byteBuffer.asReadOnlyBuffer();
  }

  /**
   * Constructs a new {@code String} by decoding the bytes using the
   * specified charset.
   */
  public String toString(String charsetName)
      throws UnsupportedEncodingException {
    return new String(this.bytes, charsetName);
  }

  /**
   * Constructs a new {@code String} by decoding the bytes as UTF-8.
   */
  public String toStringUtf8() {
    try {
      return new String(this.bytes, "UTF-8");
    } catch (UnsupportedEncodingException e) {
      throw new RuntimeException("UTF-8 not supported?", e);
    }
  }

  // =================================================================
  // equals() and hashCode()

  @Override
  public boolean equals(Object o) {
    if (o == this) {
      return true;
    }

    if (!(o instanceof ByteString)) {
      return false;
    }

    ByteString other = (ByteString) o;
    int size = this.bytes.length;
    if (size != other.bytes.length) {
      return false;
    }

    byte[] bytes = this.bytes;
    byte[] otherBytes = other.bytes;
    for (int i = 0; i < size; i++) {
      if (bytes[i] != otherBytes[i]) {
        return false;
      }
    }

    return true;
  }

  volatile int hash = 0;

  @Override
  public int hashCode() {
    int h = this.hash;

    if (h == 0) {
      byte[] bytes = this.bytes;
      int size = this.bytes.length;

      h = size;
      for (int i = 0; i < size; i++) {
        h = h * 31 + bytes[i];
      }
      if (h == 0) {
        h = 1;
      }

      this.hash = h;
    }

    return h;
  }

  // =================================================================
  // Input stream

  /**
   * Creates an {@code InputStream} which can be used to read the bytes.
   */
  public InputStream newInput() {
    return new ByteArrayInputStream(bytes);
  }

  /**
   * Creates a {@link CodedInputStream} which can be used to read the bytes.
   * Using this is more efficient than creating a {@link CodedInputStream}
   * wrapping the result of {@link #newInput()}.
   */
  public CodedInputStream newCodedInput() {
    // We trust CodedInputStream not to modify the bytes, or to give anyone
    // else access to them.
    return CodedInputStream.newInstance(bytes);
  }

  // =================================================================
  // Output stream

  /**
   * Creates a new {@link Output} with the given initial capacity.
   */
  public static Output newOutput(int initialCapacity) {
    return new Output(new ByteArrayOutputStream(initialCapacity));
  }

  /**
   * Creates a new {@link Output}.
   */
  public static Output newOutput() {
    return newOutput(32);
  }

  /**
   * Outputs to a {@code ByteString} instance. Call {@link #toByteString()} to
   * create the {@code ByteString} instance.
   */
  public static final class Output extends FilterOutputStream {
    private final ByteArrayOutputStream bout;

    /**
     * Constructs a new output with the given initial capacity.
     */
    private Output(ByteArrayOutputStream bout) {
      super(bout);
      this.bout = bout;
    }

    /**
     * Creates a {@code ByteString} instance from this {@code Output}.
     */
    public ByteString toByteString() {
      byte[] byteArray = bout.toByteArray();
      return new ByteString(byteArray);
    }
  }

  /**
   * Constructs a new ByteString builder, which allows you to efficiently
   * construct a {@code ByteString} by writing to a {@link CodedOutputSteam}.
   * Using this is much more efficient than calling {@code newOutput()} and
   * wrapping that in a {@code CodedOutputStream}.
   *
   * <p>This is package-private because it's a somewhat confusing interface.
   * Users can call {@link Message#toByteString()} instead of calling this
   * directly.
   *
   * @param size The target byte size of the {@code ByteString}.  You must
   *             write exactly this many bytes before building the result.
   */
  static CodedBuilder newCodedBuilder(int size) {
    return new CodedBuilder(size);
  }

  /** See {@link ByteString#newCodedBuilder(int)}. */
  static final class CodedBuilder {
    private final CodedOutputStream output;
    private final byte[] buffer;

    private CodedBuilder(int size) {
      buffer = new byte[size];
      output = CodedOutputStream.newInstance(buffer);
    }

    public ByteString build() {
      output.checkNoSpaceLeft();

      // We can be confident that the CodedOutputStream will not modify the
      // underlying bytes anymore because it already wrote all of them.  So,
      // no need to make a copy.
      return new ByteString(buffer);
    }

    public CodedOutputStream getCodedOutput() {
      return output;
    }
  }
}
TOP

Related Classes of com.google.protobuf.ByteString$Output

TOP
Copyright © 2018 www.massapi.com. All rights reserved.
All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.