com.google.protobuf

Class Internal

java.lang.Object

com.google.protobuf.Internal

public class Internal
extends java.lang.Object

The classes contained within are used internally by the Protocol Buffer library and generated message implementations. They are public only because those generated messages do not reside in the protobuf package. Others should not use this class directly.

Author:

kenton@google.com (Kenton Varda)

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static interface	Internal.EnumLite Interface for an enum value or value descriptor, to be used in FieldSet.
static interface	Internal.EnumLiteMap<T extends Internal.EnumLite> Interface for an object which maps integers to Internal.EnumLites.

Constructor Summary

Constructors

Constructor and Description
Internal()

Method Summary

Methods

Modifier and Type	Method and Description
static ByteString	bytesDefaultValue(java.lang.String bytes) Helper called by generated code to construct default values for bytes fields.
static boolean	isValidUtf8(ByteString byteString) Helper called by generated code to determine if a byte array is a valid UTF-8 encoded string such that the original bytes can be converted to a String object and then back to a byte array round tripping the bytes without loss.
static java.lang.String	stringDefaultValue(java.lang.String bytes) Helper called by generated code to construct default values for string fields.

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

Internal

public Internal()

Method Detail

stringDefaultValue

public static java.lang.String stringDefaultValue(java.lang.String bytes)

Helper called by generated code to construct default values for string fields.

The protocol compiler does not actually contain a UTF-8 decoder -- it just pushes UTF-8-encoded text around without touching it. The one place where this presents a problem is when generating Java string literals. Unicode characters in the string literal would normally need to be encoded using a Unicode escape sequence, which would require decoding them. To get around this, protoc instead embeds the UTF-8 bytes into the generated code and leaves it to the runtime library to decode them.

It gets worse, though. If protoc just generated a byte array, like: new byte[] {0x12, 0x34, 0x56, 0x78} Java actually generates *code* which allocates an array and then fills in each value. This is much less efficient than just embedding the bytes directly into the bytecode. To get around this, we need another work-around. String literals are embedded directly, so protoc actually generates a string literal corresponding to the bytes. The easiest way to do this is to use the ISO-8859-1 character set, which corresponds to the first 256 characters of the Unicode range. Protoc can then use good old CEscape to generate the string.

So we have a string literal which represents a set of bytes which represents another string. This function -- stringDefaultValue -- converts from the generated string to the string we actually want. The generated code calls this automatically.

bytesDefaultValue

public static ByteString bytesDefaultValue(java.lang.String bytes)

Helper called by generated code to construct default values for bytes fields.

This is a lot like stringDefaultValue(java.lang.String), but for bytes fields. In this case we only need the second of the two hacks -- allowing us to embed raw bytes as a string literal with ISO-8859-1 encoding.

isValidUtf8

public static boolean isValidUtf8(ByteString byteString)

Helper called by generated code to determine if a byte array is a valid UTF-8 encoded string such that the original bytes can be converted to a String object and then back to a byte array round tripping the bytes without loss. More precisely, returns true whenever:

   Arrays.equals(byteString.toByteArray(),
     new String(byteString.toByteArray(), "UTF-8").getBytes("UTF-8"))
 

This method rejects "overlong" byte sequences, as well as 3-byte sequences that would map to a surrogate character, in accordance with the restricted definition of UTF-8 introduced in Unicode 3.1. Note that the UTF-8 decoder included in Oracle's JDK has been modified to also reject "overlong" byte sequences, but currently (2011) still accepts 3-byte surrogate character byte sequences.

See the Unicode Standard,
Table 3-6. UTF-8 Bit Distribution,
Table 3-7. Well Formed UTF-8 Byte Sequences.

As of 2011-02, this method simply returns the result of ByteString.isValidUtf8(). Calling that method directly is preferred.

Parameters:

byteString - the string to check

Returns:

whether the byte array is round trippable