michael@0: /*
michael@0: *******************************************************************************
michael@0: *
michael@0: *   Copyright (C) 2000, International Business Machines
michael@0: *   Corporation and others.  All Rights Reserved.
michael@0: *
michael@0: *******************************************************************************
michael@0: *
michael@0: * File writejava.c
michael@0: *
michael@0: * Modification History:
michael@0: *
michael@0: *   Date        Name        Description
michael@0: *   01/11/02    Ram        Creation.
michael@0: *******************************************************************************
michael@0: */
michael@0: 
michael@0: #ifndef RLE_H
michael@0: #define RLE_H 1
michael@0: 
michael@0: #include "unicode/utypes.h"
michael@0: #include "unicode/ustring.h"
michael@0: 
michael@0: U_CDECL_BEGIN
michael@0: /**
michael@0:  * Construct a string representing a byte array.  Use run-length encoding.
michael@0:  * Two bytes are packed into a single char, with a single extra zero byte at
michael@0:  * the end if needed.  A byte represents itself, unless it is the
michael@0:  * ESCAPE_BYTE.  Then the following notations are possible:
michael@0:  *   ESCAPE_BYTE ESCAPE_BYTE   ESCAPE_BYTE literal
michael@0:  *   ESCAPE_BYTE n b           n instances of byte b
michael@0:  * Since an encoded run occupies 3 bytes, we only encode runs of 4 or
michael@0:  * more bytes.  Thus we have n > 0 and n != ESCAPE_BYTE and n <= 0xFF.
michael@0:  * If we encounter a run where n == ESCAPE_BYTE, we represent this as:
michael@0:  *   b ESCAPE_BYTE n-1 b
michael@0:  * The ESCAPE_BYTE value is chosen so as not to collide with commonly
michael@0:  * seen values.
michael@0:  */
michael@0: int32_t
michael@0: byteArrayToRLEString(const uint8_t* src,int32_t srcLen, uint16_t* buffer,int32_t bufLen, UErrorCode* status);
michael@0: 
michael@0: 
michael@0: /**
michael@0:  * Construct a string representing a char array.  Use run-length encoding.
michael@0:  * A character represents itself, unless it is the ESCAPE character.  Then
michael@0:  * the following notations are possible:
michael@0:  *   ESCAPE ESCAPE   ESCAPE literal
michael@0:  *   ESCAPE n c      n instances of character c
michael@0:  * Since an encoded run occupies 3 characters, we only encode runs of 4 or
michael@0:  * more characters.  Thus we have n > 0 and n != ESCAPE and n <= 0xFFFF.
michael@0:  * If we encounter a run where n == ESCAPE, we represent this as:
michael@0:  *   c ESCAPE n-1 c
michael@0:  * The ESCAPE value is chosen so as not to collide with commonly
michael@0:  * seen values.
michael@0:  */
michael@0: int32_t
michael@0: usArrayToRLEString(const uint16_t* src,int32_t srcLen,uint16_t* buffer, int32_t bufLen,UErrorCode* status);
michael@0: 
michael@0: /**
michael@0:  * Construct an array of bytes from a run-length encoded string.
michael@0:  */
michael@0: int32_t
michael@0: rleStringToByteArray(uint16_t* src, int32_t srcLen, uint8_t* target, int32_t tgtLen, UErrorCode* status); 
michael@0: /**
michael@0:  * Construct an array of shorts from a run-length encoded string.
michael@0:  */
michael@0: int32_t
michael@0: rleStringToUCharArray(uint16_t* src, int32_t srcLen, uint16_t* target, int32_t tgtLen, UErrorCode* status);
michael@0: 
michael@0: U_CDECL_END
michael@0: 
michael@0: #endif