michael@0: /*
michael@0: *******************************************************************************
michael@0: *
michael@0: *   Copyright (C) 2002-2012, International Business Machines
michael@0: *   Corporation and others.  All Rights Reserved.
michael@0: *
michael@0: *******************************************************************************
michael@0: *   file name:  utf_old.h
michael@0: *   encoding:   US-ASCII
michael@0: *   tab size:   8 (not used)
michael@0: *   indentation:4
michael@0: *
michael@0: *   created on: 2002sep21
michael@0: *   created by: Markus W. Scherer
michael@0: */
michael@0: 
michael@0: /**
michael@0:  * \file 
michael@0:  * \brief C API: Deprecated macros for Unicode string handling
michael@0:  */
michael@0: 
michael@0: /**
michael@0:  * 
michael@0:  * The macros in utf_old.h are all deprecated and their use discouraged.
michael@0:  * Some of the design principles behind the set of UTF macros
michael@0:  * have changed or proved impractical.
michael@0:  * Almost all of the old "UTF macros" are at least renamed.
michael@0:  * If you are looking for a new equivalent to an old macro, please see the
michael@0:  * comment at the old one.
michael@0:  *
michael@0:  * Brief summary of reasons for deprecation:
michael@0:  * - Switch on UTF_SIZE (selection of UTF-8/16/32 default string processing)
michael@0:  *   was impractical.
michael@0:  * - Switch on UTF_SAFE etc. (selection of unsafe/safe/strict default string processing)
michael@0:  *   was of little use and impractical.
michael@0:  * - Whole classes of macros became obsolete outside of the UTF_SIZE/UTF_SAFE
michael@0:  *   selection framework: UTF32_ macros (all trivial)
michael@0:  *   and UTF_ default and intermediate macros (all aliases).
michael@0:  * - The selection framework also caused many macro aliases.
michael@0:  * - Change in Unicode standard: "irregular" sequences (3.0) became illegal (3.2).
michael@0:  * - Change of language in Unicode standard:
michael@0:  *   Growing distinction between internal x-bit Unicode strings and external UTF-x
michael@0:  *   forms, with the former more lenient.
michael@0:  *   Suggests renaming of UTF16_ macros to U16_.
michael@0:  * - The prefix "UTF_" without a width number confused some users.
michael@0:  * - "Safe" append macros needed the addition of an error indicator output.
michael@0:  * - "Safe" UTF-8 macros used legitimate (if rarely used) code point values
michael@0:  *   to indicate error conditions.
michael@0:  * - The use of the "_CHAR" infix for code point operations confused some users.
michael@0:  *
michael@0:  * More details:
michael@0:  *
michael@0:  * Until ICU 2.2, utf.h theoretically allowed to choose among UTF-8/16/32
michael@0:  * for string processing, and among unsafe/safe/strict default macros for that.
michael@0:  *
michael@0:  * It proved nearly impossible to write non-trivial, high-performance code
michael@0:  * that is UTF-generic.
michael@0:  * Unsafe default macros would be dangerous for default string processing,
michael@0:  * and the main reason for the "strict" versions disappeared:
michael@0:  * Between Unicode 3.0 and 3.2 all "irregular" UTF-8 sequences became illegal.
michael@0:  * The only other conditions that "strict" checked for were non-characters,
michael@0:  * which are valid during processing. Only during text input/output should they
michael@0:  * be checked, and at that time other well-formedness checks may be
michael@0:  * necessary or useful as well.
michael@0:  * This can still be done by using U16_NEXT and U_IS_UNICODE_NONCHAR
michael@0:  * or U_IS_UNICODE_CHAR.
michael@0:  *
michael@0:  * The old UTF8_..._SAFE macros also used some normal Unicode code points
michael@0:  * to indicate malformed sequences.
michael@0:  * The new UTF8_ macros without suffix use negative values instead.
michael@0:  *
michael@0:  * The entire contents of utf32.h was moved here without replacement
michael@0:  * because all those macros were trivial and
michael@0:  * were meaningful only in the framework of choosing the UTF size.
michael@0:  *
michael@0:  * See Jitterbug 2150 and its discussion on the ICU mailing list
michael@0:  * in September 2002.
michael@0:  *
michael@0:  * <hr>
michael@0:  *
michael@0:  * <em>Obsolete part</em> of pre-ICU 2.4 utf.h file documentation:
michael@0:  *
michael@0:  * <p>The original concept for these files was for ICU to allow
michael@0:  * in principle to set which UTF (UTF-8/16/32) is used internally
michael@0:  * by defining UTF_SIZE to either 8, 16, or 32. utf.h would then define the UChar type
michael@0:  * accordingly. UTF-16 was the default.</p>
michael@0:  *
michael@0:  * <p>This concept has been abandoned.
michael@0:  * A lot of the ICU source code assumes UChar strings are in UTF-16.
michael@0:  * This is especially true for low-level code like
michael@0:  * conversion, normalization, and collation.
michael@0:  * The utf.h header enforces the default of UTF-16.
michael@0:  * The UTF-8 and UTF-32 macros remain for now for completeness and backward compatibility.</p>
michael@0:  *
michael@0:  * <p>Accordingly, utf.h defines UChar to be an unsigned 16-bit integer. If this matches wchar_t, then
michael@0:  * UChar is defined to be exactly wchar_t, otherwise uint16_t.</p>
michael@0:  *
michael@0:  * <p>UChar32 is defined to be a signed 32-bit integer (int32_t), large enough for a 21-bit
michael@0:  * Unicode code point (Unicode scalar value, 0..0x10ffff).
michael@0:  * Before ICU 2.4, the definition of UChar32 was similarly platform-dependent as
michael@0:  * the definition of UChar. For details see the documentation for UChar32 itself.</p>
michael@0:  *
michael@0:  * <p>utf.h also defines a number of C macros for handling single Unicode code points and
michael@0:  * for using UTF Unicode strings. It includes utf8.h, utf16.h, and utf32.h for the actual
michael@0:  * implementations of those macros and then aliases one set of them (for UTF-16) for general use.
michael@0:  * The UTF-specific macros have the UTF size in the macro name prefixes (UTF16_...), while
michael@0:  * the general alias macros always begin with UTF_...</p>
michael@0:  *
michael@0:  * <p>Many string operations can be done with or without error checking.
michael@0:  * Where such a distinction is useful, there are two versions of the macros, "unsafe" and "safe"
michael@0:  * ones with ..._UNSAFE and ..._SAFE suffixes. The unsafe macros are fast but may cause
michael@0:  * program failures if the strings are not well-formed. The safe macros have an additional, boolean
michael@0:  * parameter "strict". If strict is FALSE, then only illegal sequences are detected.
michael@0:  * Otherwise, irregular sequences and non-characters are detected as well (like single surrogates).
michael@0:  * Safe macros return special error code points for illegal/irregular sequences:
michael@0:  * Typically, U+ffff, or values that would result in a code unit sequence of the same length
michael@0:  * as the erroneous input sequence.<br>
michael@0:  * Note that _UNSAFE macros have fewer parameters: They do not have the strictness parameter, and
michael@0:  * they do not have start/length parameters for boundary checking.</p>
michael@0:  *
michael@0:  * <p>Here, the macros are aliased in two steps:
michael@0:  * In the first step, the UTF-specific macros with UTF16_ prefix and _UNSAFE and _SAFE suffixes are
michael@0:  * aliased according to the UTF_SIZE to macros with UTF_ prefix and the same suffixes and signatures.
michael@0:  * Then, in a second step, the default, general alias macros are set to use either the unsafe or
michael@0:  * the safe/not strict (default) or the safe/strict macro;
michael@0:  * these general macros do not have a strictness parameter.</p>
michael@0:  *
michael@0:  * <p>It is possible to change the default choice for the general alias macros to be unsafe, safe/not strict or safe/strict.
michael@0:  * The default is safe/not strict. It is not recommended to select the unsafe macros as the basis for
michael@0:  * Unicode string handling in ICU! To select this, define UTF_SAFE, UTF_STRICT, or UTF_UNSAFE.</p>
michael@0:  *
michael@0:  * <p>For general use, one should use the default, general macros with UTF_ prefix and no _SAFE/_UNSAFE suffix.
michael@0:  * Only in some cases it may be necessary to control the choice of macro directly and use a less generic alias.
michael@0:  * For example, if it can be assumed that a string is well-formed and the index will stay within the bounds,
michael@0:  * then the _UNSAFE version may be used.
michael@0:  * If a UTF-8 string is to be processed, then the macros with UTF8_ prefixes need to be used.</p>
michael@0:  *
michael@0:  * <hr>
michael@0:  *
michael@0:  * @deprecated ICU 2.4. Use the macros in utf.h, utf16.h, utf8.h instead.
michael@0:  */
michael@0: 
michael@0: #ifndef __UTF_OLD_H__
michael@0: #define __UTF_OLD_H__
michael@0: 
michael@0: #ifndef U_HIDE_DEPRECATED_API
michael@0: 
michael@0: #include "unicode/utf.h"
michael@0: #include "unicode/utf8.h"
michael@0: #include "unicode/utf16.h"
michael@0: 
michael@0: /* Formerly utf.h, part 1 --------------------------------------------------- */
michael@0: 
michael@0: #ifdef U_USE_UTF_DEPRECATES
michael@0: /**
michael@0:  * Unicode string and array offset and index type.
michael@0:  * ICU always counts Unicode code units (UChars) for
michael@0:  * string offsets, indexes, and lengths, not Unicode code points.
michael@0:  *
michael@0:  * @obsolete ICU 2.6. Use int32_t directly instead since this API will be removed in that release.
michael@0:  */
michael@0: typedef int32_t UTextOffset;
michael@0: #endif
michael@0: 
michael@0: /** Number of bits in a Unicode string code unit - ICU uses 16-bit Unicode. @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF_SIZE 16
michael@0: 
michael@0: /**
michael@0:  * The default choice for general Unicode string macros is to use the ..._SAFE macro implementations
michael@0:  * with strict=FALSE.
michael@0:  *
michael@0:  * @deprecated ICU 2.4. Obsolete, see utf_old.h.
michael@0:  */
michael@0: #define UTF_SAFE
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #undef UTF_UNSAFE
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #undef UTF_STRICT
michael@0: 
michael@0: /**
michael@0:  * UTF8_ERROR_VALUE_1 and UTF8_ERROR_VALUE_2 are special error values for UTF-8,
michael@0:  * which need 1 or 2 bytes in UTF-8:
michael@0:  * \code
michael@0:  * U+0015 = NAK = Negative Acknowledge, C0 control character
michael@0:  * U+009f = highest C1 control character
michael@0:  * \endcode
michael@0:  *
michael@0:  * These are used by UTF8_..._SAFE macros so that they can return an error value
michael@0:  * that needs the same number of code units (bytes) as were seen by
michael@0:  * a macro. They should be tested with UTF_IS_ERROR() or UTF_IS_VALID().
michael@0:  *
michael@0:  * @deprecated ICU 2.4. Obsolete, see utf_old.h.
michael@0:  */
michael@0: #define UTF8_ERROR_VALUE_1 0x15
michael@0: 
michael@0: /**
michael@0:  * See documentation on UTF8_ERROR_VALUE_1 for details.
michael@0:  *
michael@0:  * @deprecated ICU 2.4. Obsolete, see utf_old.h.
michael@0:  */
michael@0: #define UTF8_ERROR_VALUE_2 0x9f
michael@0: 
michael@0: /**
michael@0:  * Error value for all UTFs. This code point value will be set by macros with error
michael@0:  * checking if an error is detected.
michael@0:  *
michael@0:  * @deprecated ICU 2.4. Obsolete, see utf_old.h.
michael@0:  */
michael@0: #define UTF_ERROR_VALUE 0xffff
michael@0: 
michael@0: /**
michael@0:  * Is a given 32-bit code an error value
michael@0:  * as returned by one of the macros for any UTF?
michael@0:  *
michael@0:  * @deprecated ICU 2.4. Obsolete, see utf_old.h.
michael@0:  */
michael@0: #define UTF_IS_ERROR(c) \
michael@0:     (((c)&0xfffe)==0xfffe || (c)==UTF8_ERROR_VALUE_1 || (c)==UTF8_ERROR_VALUE_2)
michael@0: 
michael@0: /**
michael@0:  * This is a combined macro: Is c a valid Unicode value _and_ not an error code?
michael@0:  *
michael@0:  * @deprecated ICU 2.4. Obsolete, see utf_old.h.
michael@0:  */
michael@0: #define UTF_IS_VALID(c) \
michael@0:     (UTF_IS_UNICODE_CHAR(c) && \
michael@0:      (c)!=UTF8_ERROR_VALUE_1 && (c)!=UTF8_ERROR_VALUE_2)
michael@0: 
michael@0: /**
michael@0:  * Is this code unit or code point a surrogate (U+d800..U+dfff)?
michael@0:  * @deprecated ICU 2.4. Renamed to U_IS_SURROGATE and U16_IS_SURROGATE, see utf_old.h.
michael@0:  */
michael@0: #define UTF_IS_SURROGATE(uchar) (((uchar)&0xfffff800)==0xd800)
michael@0: 
michael@0: /**
michael@0:  * Is a given 32-bit code point a Unicode noncharacter?
michael@0:  *
michael@0:  * @deprecated ICU 2.4. Renamed to U_IS_UNICODE_NONCHAR, see utf_old.h.
michael@0:  */
michael@0: #define UTF_IS_UNICODE_NONCHAR(c) \
michael@0:     ((c)>=0xfdd0 && \
michael@0:      ((uint32_t)(c)<=0xfdef || ((c)&0xfffe)==0xfffe) && \
michael@0:      (uint32_t)(c)<=0x10ffff)
michael@0: 
michael@0: /**
michael@0:  * Is a given 32-bit value a Unicode code point value (0..U+10ffff)
michael@0:  * that can be assigned a character?
michael@0:  *
michael@0:  * Code points that are not characters include:
michael@0:  * - single surrogate code points (U+d800..U+dfff, 2048 code points)
michael@0:  * - the last two code points on each plane (U+__fffe and U+__ffff, 34 code points)
michael@0:  * - U+fdd0..U+fdef (new with Unicode 3.1, 32 code points)
michael@0:  * - the highest Unicode code point value is U+10ffff
michael@0:  *
michael@0:  * This means that all code points below U+d800 are character code points,
michael@0:  * and that boundary is tested first for performance.
michael@0:  *
michael@0:  * @deprecated ICU 2.4. Renamed to U_IS_UNICODE_CHAR, see utf_old.h.
michael@0:  */
michael@0: #define UTF_IS_UNICODE_CHAR(c) \
michael@0:     ((uint32_t)(c)<0xd800 || \
michael@0:         ((uint32_t)(c)>0xdfff && \
michael@0:          (uint32_t)(c)<=0x10ffff && \
michael@0:          !UTF_IS_UNICODE_NONCHAR(c)))
michael@0: 
michael@0: /* Formerly utf8.h ---------------------------------------------------------- */
michael@0: 
michael@0: /**
michael@0:  * Count the trail bytes for a UTF-8 lead byte.
michael@0:  * @deprecated ICU 2.4. Renamed to U8_COUNT_TRAIL_BYTES, see utf_old.h.
michael@0:  */
michael@0: #define UTF8_COUNT_TRAIL_BYTES(leadByte) (utf8_countTrailBytes[(uint8_t)leadByte])
michael@0: 
michael@0: /**
michael@0:  * Mask a UTF-8 lead byte, leave only the lower bits that form part of the code point value.
michael@0:  * @deprecated ICU 2.4. Renamed to U8_MASK_LEAD_BYTE, see utf_old.h.
michael@0:  */
michael@0: #define UTF8_MASK_LEAD_BYTE(leadByte, countTrailBytes) ((leadByte)&=(1<<(6-(countTrailBytes)))-1)
michael@0: 
michael@0: /** Is this this code point a single code unit (byte)? @deprecated ICU 2.4. Renamed to U8_IS_SINGLE, see utf_old.h. */
michael@0: #define UTF8_IS_SINGLE(uchar) (((uchar)&0x80)==0)
michael@0: /** Is this this code unit the lead code unit (byte) of a code point? @deprecated ICU 2.4. Renamed to U8_IS_LEAD, see utf_old.h. */
michael@0: #define UTF8_IS_LEAD(uchar) ((uint8_t)((uchar)-0xc0)<0x3e)
michael@0: /** Is this this code unit a trailing code unit (byte) of a code point? @deprecated ICU 2.4. Renamed to U8_IS_TRAIL, see utf_old.h. */
michael@0: #define UTF8_IS_TRAIL(uchar) (((uchar)&0xc0)==0x80)
michael@0: 
michael@0: /** Does this scalar Unicode value need multiple code units for storage? @deprecated ICU 2.4. Use U8_LENGTH or test ((uint32_t)(c)>0x7f) instead, see utf_old.h. */
michael@0: #define UTF8_NEED_MULTIPLE_UCHAR(c) ((uint32_t)(c)>0x7f)
michael@0: 
michael@0: /**
michael@0:  * Given the lead character, how many bytes are taken by this code point.
michael@0:  * ICU does not deal with code points >0x10ffff
michael@0:  * unless necessary for advancing in the byte stream.
michael@0:  *
michael@0:  * These length macros take into account that for values >0x10ffff
michael@0:  * the UTF8_APPEND_CHAR_SAFE macros would write the error code point 0xffff
michael@0:  * with 3 bytes.
michael@0:  * Code point comparisons need to be in uint32_t because UChar32
michael@0:  * may be a signed type, and negative values must be recognized.
michael@0:  *
michael@0:  * @deprecated ICU 2.4. Use U8_LENGTH instead, see utf.h.
michael@0:  */
michael@0: #if 1
michael@0: #   define UTF8_CHAR_LENGTH(c) \
michael@0:         ((uint32_t)(c)<=0x7f ? 1 : \
michael@0:             ((uint32_t)(c)<=0x7ff ? 2 : \
michael@0:                 ((uint32_t)((c)-0x10000)>0xfffff ? 3 : 4) \
michael@0:             ) \
michael@0:         )
michael@0: #else
michael@0: #   define UTF8_CHAR_LENGTH(c) \
michael@0:         ((uint32_t)(c)<=0x7f ? 1 : \
michael@0:             ((uint32_t)(c)<=0x7ff ? 2 : \
michael@0:                 ((uint32_t)(c)<=0xffff ? 3 : \
michael@0:                     ((uint32_t)(c)<=0x10ffff ? 4 : \
michael@0:                         ((uint32_t)(c)<=0x3ffffff ? 5 : \
michael@0:                             ((uint32_t)(c)<=0x7fffffff ? 6 : 3) \
michael@0:                         ) \
michael@0:                     ) \
michael@0:                 ) \
michael@0:             ) \
michael@0:         )
michael@0: #endif
michael@0: 
michael@0: /** The maximum number of bytes per code point. @deprecated ICU 2.4. Renamed to U8_MAX_LENGTH, see utf_old.h. */
michael@0: #define UTF8_MAX_CHAR_LENGTH 4
michael@0: 
michael@0: /** Average number of code units compared to UTF-16. @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF8_ARRAY_SIZE(size) ((5*(size))/2)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U8_GET_UNSAFE, see utf_old.h. */
michael@0: #define UTF8_GET_CHAR_UNSAFE(s, i, c) { \
michael@0:     int32_t _utf8_get_char_unsafe_index=(int32_t)(i); \
michael@0:     UTF8_SET_CHAR_START_UNSAFE(s, _utf8_get_char_unsafe_index); \
michael@0:     UTF8_NEXT_CHAR_UNSAFE(s, _utf8_get_char_unsafe_index, c); \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Use U8_GET instead, see utf_old.h. */
michael@0: #define UTF8_GET_CHAR_SAFE(s, start, i, length, c, strict) { \
michael@0:     int32_t _utf8_get_char_safe_index=(int32_t)(i); \
michael@0:     UTF8_SET_CHAR_START_SAFE(s, start, _utf8_get_char_safe_index); \
michael@0:     UTF8_NEXT_CHAR_SAFE(s, _utf8_get_char_safe_index, length, c, strict); \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U8_NEXT_UNSAFE, see utf_old.h. */
michael@0: #define UTF8_NEXT_CHAR_UNSAFE(s, i, c) { \
michael@0:     (c)=(s)[(i)++]; \
michael@0:     if((uint8_t)((c)-0xc0)<0x35) { \
michael@0:         uint8_t __count=UTF8_COUNT_TRAIL_BYTES(c); \
michael@0:         UTF8_MASK_LEAD_BYTE(c, __count); \
michael@0:         switch(__count) { \
michael@0:         /* each following branch falls through to the next one */ \
michael@0:         case 3: \
michael@0:             (c)=((c)<<6)|((s)[(i)++]&0x3f); \
michael@0:         case 2: \
michael@0:             (c)=((c)<<6)|((s)[(i)++]&0x3f); \
michael@0:         case 1: \
michael@0:             (c)=((c)<<6)|((s)[(i)++]&0x3f); \
michael@0:         /* no other branches to optimize switch() */ \
michael@0:             break; \
michael@0:         } \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U8_APPEND_UNSAFE, see utf_old.h. */
michael@0: #define UTF8_APPEND_CHAR_UNSAFE(s, i, c) { \
michael@0:     if((uint32_t)(c)<=0x7f) { \
michael@0:         (s)[(i)++]=(uint8_t)(c); \
michael@0:     } else { \
michael@0:         if((uint32_t)(c)<=0x7ff) { \
michael@0:             (s)[(i)++]=(uint8_t)(((c)>>6)|0xc0); \
michael@0:         } else { \
michael@0:             if((uint32_t)(c)<=0xffff) { \
michael@0:                 (s)[(i)++]=(uint8_t)(((c)>>12)|0xe0); \
michael@0:             } else { \
michael@0:                 (s)[(i)++]=(uint8_t)(((c)>>18)|0xf0); \
michael@0:                 (s)[(i)++]=(uint8_t)((((c)>>12)&0x3f)|0x80); \
michael@0:             } \
michael@0:             (s)[(i)++]=(uint8_t)((((c)>>6)&0x3f)|0x80); \
michael@0:         } \
michael@0:         (s)[(i)++]=(uint8_t)(((c)&0x3f)|0x80); \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U8_FWD_1_UNSAFE, see utf_old.h. */
michael@0: #define UTF8_FWD_1_UNSAFE(s, i) { \
michael@0:     (i)+=1+UTF8_COUNT_TRAIL_BYTES((s)[i]); \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U8_FWD_N_UNSAFE, see utf_old.h. */
michael@0: #define UTF8_FWD_N_UNSAFE(s, i, n) { \
michael@0:     int32_t __N=(n); \
michael@0:     while(__N>0) { \
michael@0:         UTF8_FWD_1_UNSAFE(s, i); \
michael@0:         --__N; \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U8_SET_CP_START_UNSAFE, see utf_old.h. */
michael@0: #define UTF8_SET_CHAR_START_UNSAFE(s, i) { \
michael@0:     while(UTF8_IS_TRAIL((s)[i])) { --(i); } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Use U8_NEXT instead, see utf_old.h. */
michael@0: #define UTF8_NEXT_CHAR_SAFE(s, i, length, c, strict) { \
michael@0:     (c)=(s)[(i)++]; \
michael@0:     if((c)>=0x80) { \
michael@0:         if(UTF8_IS_LEAD(c)) { \
michael@0:             (c)=utf8_nextCharSafeBody(s, &(i), (int32_t)(length), c, strict); \
michael@0:         } else { \
michael@0:             (c)=UTF8_ERROR_VALUE_1; \
michael@0:         } \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Use U8_APPEND instead, see utf_old.h. */
michael@0: #define UTF8_APPEND_CHAR_SAFE(s, i, length, c)  { \
michael@0:     if((uint32_t)(c)<=0x7f) { \
michael@0:         (s)[(i)++]=(uint8_t)(c); \
michael@0:     } else { \
michael@0:         (i)=utf8_appendCharSafeBody(s, (int32_t)(i), (int32_t)(length), c, NULL); \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U8_FWD_1, see utf_old.h. */
michael@0: #define UTF8_FWD_1_SAFE(s, i, length) U8_FWD_1(s, i, length)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U8_FWD_N, see utf_old.h. */
michael@0: #define UTF8_FWD_N_SAFE(s, i, length, n) U8_FWD_N(s, i, length, n)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U8_SET_CP_START, see utf_old.h. */
michael@0: #define UTF8_SET_CHAR_START_SAFE(s, start, i) U8_SET_CP_START(s, start, i)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U8_PREV_UNSAFE, see utf_old.h. */
michael@0: #define UTF8_PREV_CHAR_UNSAFE(s, i, c) { \
michael@0:     (c)=(s)[--(i)]; \
michael@0:     if(UTF8_IS_TRAIL(c)) { \
michael@0:         uint8_t __b, __count=1, __shift=6; \
michael@0: \
michael@0:         /* c is a trail byte */ \
michael@0:         (c)&=0x3f; \
michael@0:         for(;;) { \
michael@0:             __b=(s)[--(i)]; \
michael@0:             if(__b>=0xc0) { \
michael@0:                 UTF8_MASK_LEAD_BYTE(__b, __count); \
michael@0:                 (c)|=(UChar32)__b<<__shift; \
michael@0:                 break; \
michael@0:             } else { \
michael@0:                 (c)|=(UChar32)(__b&0x3f)<<__shift; \
michael@0:                 ++__count; \
michael@0:                 __shift+=6; \
michael@0:             } \
michael@0:         } \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U8_BACK_1_UNSAFE, see utf_old.h. */
michael@0: #define UTF8_BACK_1_UNSAFE(s, i) { \
michael@0:     while(UTF8_IS_TRAIL((s)[--(i)])) {} \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U8_BACK_N_UNSAFE, see utf_old.h. */
michael@0: #define UTF8_BACK_N_UNSAFE(s, i, n) { \
michael@0:     int32_t __N=(n); \
michael@0:     while(__N>0) { \
michael@0:         UTF8_BACK_1_UNSAFE(s, i); \
michael@0:         --__N; \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U8_SET_CP_LIMIT_UNSAFE, see utf_old.h. */
michael@0: #define UTF8_SET_CHAR_LIMIT_UNSAFE(s, i) { \
michael@0:     UTF8_BACK_1_UNSAFE(s, i); \
michael@0:     UTF8_FWD_1_UNSAFE(s, i); \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Use U8_PREV instead, see utf_old.h. */
michael@0: #define UTF8_PREV_CHAR_SAFE(s, start, i, c, strict) { \
michael@0:     (c)=(s)[--(i)]; \
michael@0:     if((c)>=0x80) { \
michael@0:         if((c)<=0xbf) { \
michael@0:             (c)=utf8_prevCharSafeBody(s, start, &(i), c, strict); \
michael@0:         } else { \
michael@0:             (c)=UTF8_ERROR_VALUE_1; \
michael@0:         } \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U8_BACK_1, see utf_old.h. */
michael@0: #define UTF8_BACK_1_SAFE(s, start, i) U8_BACK_1(s, start, i)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U8_BACK_N, see utf_old.h. */
michael@0: #define UTF8_BACK_N_SAFE(s, start, i, n) U8_BACK_N(s, start, i, n)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U8_SET_CP_LIMIT, see utf_old.h. */
michael@0: #define UTF8_SET_CHAR_LIMIT_SAFE(s, start, i, length) U8_SET_CP_LIMIT(s, start, i, length)
michael@0: 
michael@0: /* Formerly utf16.h --------------------------------------------------------- */
michael@0: 
michael@0: /** Is uchar a first/lead surrogate? @deprecated ICU 2.4. Renamed to U_IS_LEAD and U16_IS_LEAD, see utf_old.h. */
michael@0: #define UTF_IS_FIRST_SURROGATE(uchar) (((uchar)&0xfffffc00)==0xd800)
michael@0: 
michael@0: /** Is uchar a second/trail surrogate? @deprecated ICU 2.4. Renamed to U_IS_TRAIL and U16_IS_TRAIL, see utf_old.h. */
michael@0: #define UTF_IS_SECOND_SURROGATE(uchar) (((uchar)&0xfffffc00)==0xdc00)
michael@0: 
michael@0: /** Assuming c is a surrogate, is it a first/lead surrogate? @deprecated ICU 2.4. Renamed to U_IS_SURROGATE_LEAD and U16_IS_SURROGATE_LEAD, see utf_old.h. */
michael@0: #define UTF_IS_SURROGATE_FIRST(c) (((c)&0x400)==0)
michael@0: 
michael@0: /** Helper constant for UTF16_GET_PAIR_VALUE. @deprecated ICU 2.4. Renamed to U16_SURROGATE_OFFSET, see utf_old.h. */
michael@0: #define UTF_SURROGATE_OFFSET ((0xd800<<10UL)+0xdc00-0x10000)
michael@0: 
michael@0: /** Get the UTF-32 value from the surrogate code units. @deprecated ICU 2.4. Renamed to U16_GET_SUPPLEMENTARY, see utf_old.h. */
michael@0: #define UTF16_GET_PAIR_VALUE(first, second) \
michael@0:     (((first)<<10UL)+(second)-UTF_SURROGATE_OFFSET)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_LEAD, see utf_old.h. */
michael@0: #define UTF_FIRST_SURROGATE(supplementary) (UChar)(((supplementary)>>10)+0xd7c0)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_TRAIL, see utf_old.h. */
michael@0: #define UTF_SECOND_SURROGATE(supplementary) (UChar)(((supplementary)&0x3ff)|0xdc00)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_LEAD, see utf_old.h. */
michael@0: #define UTF16_LEAD(supplementary) UTF_FIRST_SURROGATE(supplementary)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_TRAIL, see utf_old.h. */
michael@0: #define UTF16_TRAIL(supplementary) UTF_SECOND_SURROGATE(supplementary)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_IS_SINGLE, see utf_old.h. */
michael@0: #define UTF16_IS_SINGLE(uchar) !UTF_IS_SURROGATE(uchar)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_IS_LEAD, see utf_old.h. */
michael@0: #define UTF16_IS_LEAD(uchar) UTF_IS_FIRST_SURROGATE(uchar)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_IS_TRAIL, see utf_old.h. */
michael@0: #define UTF16_IS_TRAIL(uchar) UTF_IS_SECOND_SURROGATE(uchar)
michael@0: 
michael@0: /** Does this scalar Unicode value need multiple code units for storage? @deprecated ICU 2.4. Use U16_LENGTH or test ((uint32_t)(c)>0xffff) instead, see utf_old.h. */
michael@0: #define UTF16_NEED_MULTIPLE_UCHAR(c) ((uint32_t)(c)>0xffff)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_LENGTH, see utf_old.h. */
michael@0: #define UTF16_CHAR_LENGTH(c) ((uint32_t)(c)<=0xffff ? 1 : 2)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_MAX_LENGTH, see utf_old.h. */
michael@0: #define UTF16_MAX_CHAR_LENGTH 2
michael@0: 
michael@0: /** Average number of code units compared to UTF-16. @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF16_ARRAY_SIZE(size) (size)
michael@0: 
michael@0: /**
michael@0:  * Get a single code point from an offset that points to any
michael@0:  * of the code units that belong to that code point.
michael@0:  * Assume 0<=i<length.
michael@0:  *
michael@0:  * This could be used for iteration together with
michael@0:  * UTF16_CHAR_LENGTH() and UTF_IS_ERROR(),
michael@0:  * but the use of UTF16_NEXT_CHAR[_UNSAFE]() and
michael@0:  * UTF16_PREV_CHAR[_UNSAFE]() is more efficient for that.
michael@0:  * @deprecated ICU 2.4. Renamed to U16_GET_UNSAFE, see utf_old.h.
michael@0:  */
michael@0: #define UTF16_GET_CHAR_UNSAFE(s, i, c) { \
michael@0:     (c)=(s)[i]; \
michael@0:     if(UTF_IS_SURROGATE(c)) { \
michael@0:         if(UTF_IS_SURROGATE_FIRST(c)) { \
michael@0:             (c)=UTF16_GET_PAIR_VALUE((c), (s)[(i)+1]); \
michael@0:         } else { \
michael@0:             (c)=UTF16_GET_PAIR_VALUE((s)[(i)-1], (c)); \
michael@0:         } \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Use U16_GET instead, see utf_old.h. */
michael@0: #define UTF16_GET_CHAR_SAFE(s, start, i, length, c, strict) { \
michael@0:     (c)=(s)[i]; \
michael@0:     if(UTF_IS_SURROGATE(c)) { \
michael@0:         uint16_t __c2; \
michael@0:         if(UTF_IS_SURROGATE_FIRST(c)) { \
michael@0:             if((i)+1<(length) && UTF_IS_SECOND_SURROGATE(__c2=(s)[(i)+1])) { \
michael@0:                 (c)=UTF16_GET_PAIR_VALUE((c), __c2); \
michael@0:                 /* strict: ((c)&0xfffe)==0xfffe is caught by UTF_IS_ERROR() and UTF_IS_UNICODE_CHAR() */ \
michael@0:             } else if(strict) {\
michael@0:                 /* unmatched first surrogate */ \
michael@0:                 (c)=UTF_ERROR_VALUE; \
michael@0:             } \
michael@0:         } else { \
michael@0:             if((i)-1>=(start) && UTF_IS_FIRST_SURROGATE(__c2=(s)[(i)-1])) { \
michael@0:                 (c)=UTF16_GET_PAIR_VALUE(__c2, (c)); \
michael@0:                 /* strict: ((c)&0xfffe)==0xfffe is caught by UTF_IS_ERROR() and UTF_IS_UNICODE_CHAR() */ \
michael@0:             } else if(strict) {\
michael@0:                 /* unmatched second surrogate */ \
michael@0:                 (c)=UTF_ERROR_VALUE; \
michael@0:             } \
michael@0:         } \
michael@0:     } else if((strict) && !UTF_IS_UNICODE_CHAR(c)) { \
michael@0:         (c)=UTF_ERROR_VALUE; \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_NEXT_UNSAFE, see utf_old.h. */
michael@0: #define UTF16_NEXT_CHAR_UNSAFE(s, i, c) { \
michael@0:     (c)=(s)[(i)++]; \
michael@0:     if(UTF_IS_FIRST_SURROGATE(c)) { \
michael@0:         (c)=UTF16_GET_PAIR_VALUE((c), (s)[(i)++]); \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_APPEND_UNSAFE, see utf_old.h. */
michael@0: #define UTF16_APPEND_CHAR_UNSAFE(s, i, c) { \
michael@0:     if((uint32_t)(c)<=0xffff) { \
michael@0:         (s)[(i)++]=(uint16_t)(c); \
michael@0:     } else { \
michael@0:         (s)[(i)++]=(uint16_t)(((c)>>10)+0xd7c0); \
michael@0:         (s)[(i)++]=(uint16_t)(((c)&0x3ff)|0xdc00); \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_FWD_1_UNSAFE, see utf_old.h. */
michael@0: #define UTF16_FWD_1_UNSAFE(s, i) { \
michael@0:     if(UTF_IS_FIRST_SURROGATE((s)[(i)++])) { \
michael@0:         ++(i); \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_FWD_N_UNSAFE, see utf_old.h. */
michael@0: #define UTF16_FWD_N_UNSAFE(s, i, n) { \
michael@0:     int32_t __N=(n); \
michael@0:     while(__N>0) { \
michael@0:         UTF16_FWD_1_UNSAFE(s, i); \
michael@0:         --__N; \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_SET_CP_START_UNSAFE, see utf_old.h. */
michael@0: #define UTF16_SET_CHAR_START_UNSAFE(s, i) { \
michael@0:     if(UTF_IS_SECOND_SURROGATE((s)[i])) { \
michael@0:         --(i); \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Use U16_NEXT instead, see utf_old.h. */
michael@0: #define UTF16_NEXT_CHAR_SAFE(s, i, length, c, strict) { \
michael@0:     (c)=(s)[(i)++]; \
michael@0:     if(UTF_IS_FIRST_SURROGATE(c)) { \
michael@0:         uint16_t __c2; \
michael@0:         if((i)<(length) && UTF_IS_SECOND_SURROGATE(__c2=(s)[(i)])) { \
michael@0:             ++(i); \
michael@0:             (c)=UTF16_GET_PAIR_VALUE((c), __c2); \
michael@0:             /* strict: ((c)&0xfffe)==0xfffe is caught by UTF_IS_ERROR() and UTF_IS_UNICODE_CHAR() */ \
michael@0:         } else if(strict) {\
michael@0:             /* unmatched first surrogate */ \
michael@0:             (c)=UTF_ERROR_VALUE; \
michael@0:         } \
michael@0:     } else if((strict) && !UTF_IS_UNICODE_CHAR(c)) { \
michael@0:         /* unmatched second surrogate or other non-character */ \
michael@0:         (c)=UTF_ERROR_VALUE; \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Use U16_APPEND instead, see utf_old.h. */
michael@0: #define UTF16_APPEND_CHAR_SAFE(s, i, length, c) { \
michael@0:     if((uint32_t)(c)<=0xffff) { \
michael@0:         (s)[(i)++]=(uint16_t)(c); \
michael@0:     } else if((uint32_t)(c)<=0x10ffff) { \
michael@0:         if((i)+1<(length)) { \
michael@0:             (s)[(i)++]=(uint16_t)(((c)>>10)+0xd7c0); \
michael@0:             (s)[(i)++]=(uint16_t)(((c)&0x3ff)|0xdc00); \
michael@0:         } else /* not enough space */ { \
michael@0:             (s)[(i)++]=UTF_ERROR_VALUE; \
michael@0:         } \
michael@0:     } else /* c>0x10ffff, write error value */ { \
michael@0:         (s)[(i)++]=UTF_ERROR_VALUE; \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_FWD_1, see utf_old.h. */
michael@0: #define UTF16_FWD_1_SAFE(s, i, length) U16_FWD_1(s, i, length)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_FWD_N, see utf_old.h. */
michael@0: #define UTF16_FWD_N_SAFE(s, i, length, n) U16_FWD_N(s, i, length, n)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_SET_CP_START, see utf_old.h. */
michael@0: #define UTF16_SET_CHAR_START_SAFE(s, start, i) U16_SET_CP_START(s, start, i)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_PREV_UNSAFE, see utf_old.h. */
michael@0: #define UTF16_PREV_CHAR_UNSAFE(s, i, c) { \
michael@0:     (c)=(s)[--(i)]; \
michael@0:     if(UTF_IS_SECOND_SURROGATE(c)) { \
michael@0:         (c)=UTF16_GET_PAIR_VALUE((s)[--(i)], (c)); \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_BACK_1_UNSAFE, see utf_old.h. */
michael@0: #define UTF16_BACK_1_UNSAFE(s, i) { \
michael@0:     if(UTF_IS_SECOND_SURROGATE((s)[--(i)])) { \
michael@0:         --(i); \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_BACK_N_UNSAFE, see utf_old.h. */
michael@0: #define UTF16_BACK_N_UNSAFE(s, i, n) { \
michael@0:     int32_t __N=(n); \
michael@0:     while(__N>0) { \
michael@0:         UTF16_BACK_1_UNSAFE(s, i); \
michael@0:         --__N; \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_SET_CP_LIMIT_UNSAFE, see utf_old.h. */
michael@0: #define UTF16_SET_CHAR_LIMIT_UNSAFE(s, i) { \
michael@0:     if(UTF_IS_FIRST_SURROGATE((s)[(i)-1])) { \
michael@0:         ++(i); \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Use U16_PREV instead, see utf_old.h. */
michael@0: #define UTF16_PREV_CHAR_SAFE(s, start, i, c, strict) { \
michael@0:     (c)=(s)[--(i)]; \
michael@0:     if(UTF_IS_SECOND_SURROGATE(c)) { \
michael@0:         uint16_t __c2; \
michael@0:         if((i)>(start) && UTF_IS_FIRST_SURROGATE(__c2=(s)[(i)-1])) { \
michael@0:             --(i); \
michael@0:             (c)=UTF16_GET_PAIR_VALUE(__c2, (c)); \
michael@0:             /* strict: ((c)&0xfffe)==0xfffe is caught by UTF_IS_ERROR() and UTF_IS_UNICODE_CHAR() */ \
michael@0:         } else if(strict) {\
michael@0:             /* unmatched second surrogate */ \
michael@0:             (c)=UTF_ERROR_VALUE; \
michael@0:         } \
michael@0:     } else if((strict) && !UTF_IS_UNICODE_CHAR(c)) { \
michael@0:         /* unmatched first surrogate or other non-character */ \
michael@0:         (c)=UTF_ERROR_VALUE; \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_BACK_1, see utf_old.h. */
michael@0: #define UTF16_BACK_1_SAFE(s, start, i) U16_BACK_1(s, start, i)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_BACK_N, see utf_old.h. */
michael@0: #define UTF16_BACK_N_SAFE(s, start, i, n) U16_BACK_N(s, start, i, n)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_SET_CP_LIMIT, see utf_old.h. */
michael@0: #define UTF16_SET_CHAR_LIMIT_SAFE(s, start, i, length) U16_SET_CP_LIMIT(s, start, i, length)
michael@0: 
michael@0: /* Formerly utf32.h --------------------------------------------------------- */
michael@0: 
michael@0: /*
michael@0: * Old documentation:
michael@0: *
michael@0: *   This file defines macros to deal with UTF-32 code units and code points.
michael@0: *   Signatures and semantics are the same as for the similarly named macros
michael@0: *   in utf16.h.
michael@0: *   utf32.h is included by utf.h after unicode/umachine.h</p>
michael@0: *   and some common definitions.
michael@0: *   <p><b>Usage:</b>  ICU coding guidelines for if() statements should be followed when using these macros.
michael@0: *                  Compound statements (curly braces {}) must be used  for if-else-while...
michael@0: *                  bodies and all macro statements should be terminated with semicolon.</p>
michael@0: */
michael@0: 
michael@0: /* internal definitions ----------------------------------------------------- */
michael@0: 
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_IS_SAFE(c, strict) \
michael@0:     (!(strict) ? \
michael@0:         (uint32_t)(c)<=0x10ffff : \
michael@0:         UTF_IS_UNICODE_CHAR(c))
michael@0: 
michael@0: /*
michael@0:  * For the semantics of all of these macros, see utf16.h.
michael@0:  * The UTF-32 versions are trivial because any code point is
michael@0:  * encoded using exactly one code unit.
michael@0:  */
michael@0: 
michael@0: /* single-code point definitions -------------------------------------------- */
michael@0: 
michael@0: /* classes of code unit values */
michael@0: 
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_IS_SINGLE(uchar) 1
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_IS_LEAD(uchar) 0
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_IS_TRAIL(uchar) 0
michael@0: 
michael@0: /* number of code units per code point */
michael@0: 
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_NEED_MULTIPLE_UCHAR(c) 0
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_CHAR_LENGTH(c) 1
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_MAX_CHAR_LENGTH 1
michael@0: 
michael@0: /* average number of code units compared to UTF-16 */
michael@0: 
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_ARRAY_SIZE(size) (size)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_GET_CHAR_UNSAFE(s, i, c) { \
michael@0:     (c)=(s)[i]; \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_GET_CHAR_SAFE(s, start, i, length, c, strict) { \
michael@0:     (c)=(s)[i]; \
michael@0:     if(!UTF32_IS_SAFE(c, strict)) { \
michael@0:         (c)=UTF_ERROR_VALUE; \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /* definitions with forward iteration --------------------------------------- */
michael@0: 
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_NEXT_CHAR_UNSAFE(s, i, c) { \
michael@0:     (c)=(s)[(i)++]; \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_APPEND_CHAR_UNSAFE(s, i, c) { \
michael@0:     (s)[(i)++]=(c); \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_FWD_1_UNSAFE(s, i) { \
michael@0:     ++(i); \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_FWD_N_UNSAFE(s, i, n) { \
michael@0:     (i)+=(n); \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_SET_CHAR_START_UNSAFE(s, i) { \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_NEXT_CHAR_SAFE(s, i, length, c, strict) { \
michael@0:     (c)=(s)[(i)++]; \
michael@0:     if(!UTF32_IS_SAFE(c, strict)) { \
michael@0:         (c)=UTF_ERROR_VALUE; \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_APPEND_CHAR_SAFE(s, i, length, c) { \
michael@0:     if((uint32_t)(c)<=0x10ffff) { \
michael@0:         (s)[(i)++]=(c); \
michael@0:     } else /* c>0x10ffff, write 0xfffd */ { \
michael@0:         (s)[(i)++]=0xfffd; \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_FWD_1_SAFE(s, i, length) { \
michael@0:     ++(i); \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_FWD_N_SAFE(s, i, length, n) { \
michael@0:     if(((i)+=(n))>(length)) { \
michael@0:         (i)=(length); \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_SET_CHAR_START_SAFE(s, start, i) { \
michael@0: }
michael@0: 
michael@0: /* definitions with backward iteration -------------------------------------- */
michael@0: 
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_PREV_CHAR_UNSAFE(s, i, c) { \
michael@0:     (c)=(s)[--(i)]; \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_BACK_1_UNSAFE(s, i) { \
michael@0:     --(i); \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_BACK_N_UNSAFE(s, i, n) { \
michael@0:     (i)-=(n); \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_SET_CHAR_LIMIT_UNSAFE(s, i) { \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_PREV_CHAR_SAFE(s, start, i, c, strict) { \
michael@0:     (c)=(s)[--(i)]; \
michael@0:     if(!UTF32_IS_SAFE(c, strict)) { \
michael@0:         (c)=UTF_ERROR_VALUE; \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_BACK_1_SAFE(s, start, i) { \
michael@0:     --(i); \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_BACK_N_SAFE(s, start, i, n) { \
michael@0:     (i)-=(n); \
michael@0:     if((i)<(start)) { \
michael@0:         (i)=(start); \
michael@0:     } \
michael@0: }
michael@0: 
michael@0: /** @deprecated ICU 2.4. Obsolete, see utf_old.h. */
michael@0: #define UTF32_SET_CHAR_LIMIT_SAFE(s, i, length) { \
michael@0: }
michael@0: 
michael@0: /* Formerly utf.h, part 2 --------------------------------------------------- */
michael@0: 
michael@0: /**
michael@0:  * Estimate the number of code units for a string based on the number of UTF-16 code units.
michael@0:  *
michael@0:  * @deprecated ICU 2.4. Obsolete, see utf_old.h.
michael@0:  */
michael@0: #define UTF_ARRAY_SIZE(size) UTF16_ARRAY_SIZE(size)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_GET_UNSAFE, see utf_old.h. */
michael@0: #define UTF_GET_CHAR_UNSAFE(s, i, c)                 UTF16_GET_CHAR_UNSAFE(s, i, c)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Use U16_GET instead, see utf_old.h. */
michael@0: #define UTF_GET_CHAR_SAFE(s, start, i, length, c, strict) UTF16_GET_CHAR_SAFE(s, start, i, length, c, strict)
michael@0: 
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_NEXT_UNSAFE, see utf_old.h. */
michael@0: #define UTF_NEXT_CHAR_UNSAFE(s, i, c)                UTF16_NEXT_CHAR_UNSAFE(s, i, c)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Use U16_NEXT instead, see utf_old.h. */
michael@0: #define UTF_NEXT_CHAR_SAFE(s, i, length, c, strict)  UTF16_NEXT_CHAR_SAFE(s, i, length, c, strict)
michael@0: 
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_APPEND_UNSAFE, see utf_old.h. */
michael@0: #define UTF_APPEND_CHAR_UNSAFE(s, i, c)              UTF16_APPEND_CHAR_UNSAFE(s, i, c)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Use U16_APPEND instead, see utf_old.h. */
michael@0: #define UTF_APPEND_CHAR_SAFE(s, i, length, c)        UTF16_APPEND_CHAR_SAFE(s, i, length, c)
michael@0: 
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_FWD_1_UNSAFE, see utf_old.h. */
michael@0: #define UTF_FWD_1_UNSAFE(s, i)                       UTF16_FWD_1_UNSAFE(s, i)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_FWD_1, see utf_old.h. */
michael@0: #define UTF_FWD_1_SAFE(s, i, length)                 UTF16_FWD_1_SAFE(s, i, length)
michael@0: 
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_FWD_N_UNSAFE, see utf_old.h. */
michael@0: #define UTF_FWD_N_UNSAFE(s, i, n)                    UTF16_FWD_N_UNSAFE(s, i, n)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_FWD_N, see utf_old.h. */
michael@0: #define UTF_FWD_N_SAFE(s, i, length, n)              UTF16_FWD_N_SAFE(s, i, length, n)
michael@0: 
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_SET_CP_START_UNSAFE, see utf_old.h. */
michael@0: #define UTF_SET_CHAR_START_UNSAFE(s, i)              UTF16_SET_CHAR_START_UNSAFE(s, i)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_SET_CP_START, see utf_old.h. */
michael@0: #define UTF_SET_CHAR_START_SAFE(s, start, i)         UTF16_SET_CHAR_START_SAFE(s, start, i)
michael@0: 
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_PREV_UNSAFE, see utf_old.h. */
michael@0: #define UTF_PREV_CHAR_UNSAFE(s, i, c)                UTF16_PREV_CHAR_UNSAFE(s, i, c)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Use U16_PREV instead, see utf_old.h. */
michael@0: #define UTF_PREV_CHAR_SAFE(s, start, i, c, strict)   UTF16_PREV_CHAR_SAFE(s, start, i, c, strict)
michael@0: 
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_BACK_1_UNSAFE, see utf_old.h. */
michael@0: #define UTF_BACK_1_UNSAFE(s, i)                      UTF16_BACK_1_UNSAFE(s, i)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_BACK_1, see utf_old.h. */
michael@0: #define UTF_BACK_1_SAFE(s, start, i)                 UTF16_BACK_1_SAFE(s, start, i)
michael@0: 
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_BACK_N_UNSAFE, see utf_old.h. */
michael@0: #define UTF_BACK_N_UNSAFE(s, i, n)                   UTF16_BACK_N_UNSAFE(s, i, n)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_BACK_N, see utf_old.h. */
michael@0: #define UTF_BACK_N_SAFE(s, start, i, n)              UTF16_BACK_N_SAFE(s, start, i, n)
michael@0: 
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_SET_CP_LIMIT_UNSAFE, see utf_old.h. */
michael@0: #define UTF_SET_CHAR_LIMIT_UNSAFE(s, i)              UTF16_SET_CHAR_LIMIT_UNSAFE(s, i)
michael@0: 
michael@0: /** @deprecated ICU 2.4. Renamed to U16_SET_CP_LIMIT, see utf_old.h. */
michael@0: #define UTF_SET_CHAR_LIMIT_SAFE(s, start, i, length) UTF16_SET_CHAR_LIMIT_SAFE(s, start, i, length)
michael@0: 
michael@0: /* Define default macros (UTF-16 "safe") ------------------------------------ */
michael@0: 
michael@0: /**
michael@0:  * Does this code unit alone encode a code point (BMP, not a surrogate)?
michael@0:  * Same as UTF16_IS_SINGLE.
michael@0:  * @deprecated ICU 2.4. Renamed to U_IS_SINGLE and U16_IS_SINGLE, see utf_old.h.
michael@0:  */
michael@0: #define UTF_IS_SINGLE(uchar) U16_IS_SINGLE(uchar)
michael@0: 
michael@0: /**
michael@0:  * Is this code unit the first one of several (a lead surrogate)?
michael@0:  * Same as UTF16_IS_LEAD.
michael@0:  * @deprecated ICU 2.4. Renamed to U_IS_LEAD and U16_IS_LEAD, see utf_old.h.
michael@0:  */
michael@0: #define UTF_IS_LEAD(uchar) U16_IS_LEAD(uchar)
michael@0: 
michael@0: /**
michael@0:  * Is this code unit one of several but not the first one (a trail surrogate)?
michael@0:  * Same as UTF16_IS_TRAIL.
michael@0:  * @deprecated ICU 2.4. Renamed to U_IS_TRAIL and U16_IS_TRAIL, see utf_old.h.
michael@0:  */
michael@0: #define UTF_IS_TRAIL(uchar) U16_IS_TRAIL(uchar)
michael@0: 
michael@0: /**
michael@0:  * Does this code point require multiple code units (is it a supplementary code point)?
michael@0:  * Same as UTF16_NEED_MULTIPLE_UCHAR.
michael@0:  * @deprecated ICU 2.4. Use U16_LENGTH or test ((uint32_t)(c)>0xffff) instead.
michael@0:  */
michael@0: #define UTF_NEED_MULTIPLE_UCHAR(c) UTF16_NEED_MULTIPLE_UCHAR(c)
michael@0: 
michael@0: /**
michael@0:  * How many code units are used to encode this code point (1 or 2)?
michael@0:  * Same as UTF16_CHAR_LENGTH.
michael@0:  * @deprecated ICU 2.4. Renamed to U16_LENGTH, see utf_old.h.
michael@0:  */
michael@0: #define UTF_CHAR_LENGTH(c) U16_LENGTH(c)
michael@0: 
michael@0: /**
michael@0:  * How many code units are used at most for any Unicode code point (2)?
michael@0:  * Same as UTF16_MAX_CHAR_LENGTH.
michael@0:  * @deprecated ICU 2.4. Renamed to U16_MAX_LENGTH, see utf_old.h.
michael@0:  */
michael@0: #define UTF_MAX_CHAR_LENGTH U16_MAX_LENGTH
michael@0: 
michael@0: /**
michael@0:  * Set c to the code point that contains the code unit i.
michael@0:  * i could point to the lead or the trail surrogate for the code point.
michael@0:  * i is not modified.
michael@0:  * Same as UTF16_GET_CHAR.
michael@0:  * \pre 0<=i<length
michael@0:  *
michael@0:  * @deprecated ICU 2.4. Renamed to U16_GET, see utf_old.h.
michael@0:  */
michael@0: #define UTF_GET_CHAR(s, start, i, length, c) U16_GET(s, start, i, length, c)
michael@0: 
michael@0: /**
michael@0:  * Set c to the code point that starts at code unit i
michael@0:  * and advance i to beyond the code units of this code point (post-increment).
michael@0:  * i must point to the first code unit of a code point.
michael@0:  * Otherwise c is set to the trail unit (surrogate) itself.
michael@0:  * Same as UTF16_NEXT_CHAR.
michael@0:  * \pre 0<=i<length
michael@0:  * \post 0<i<=length
michael@0:  *
michael@0:  * @deprecated ICU 2.4. Renamed to U16_NEXT, see utf_old.h.
michael@0:  */
michael@0: #define UTF_NEXT_CHAR(s, i, length, c) U16_NEXT(s, i, length, c)
michael@0: 
michael@0: /**
michael@0:  * Append the code units of code point c to the string at index i
michael@0:  * and advance i to beyond the new code units (post-increment).
michael@0:  * The code units beginning at index i will be overwritten.
michael@0:  * Same as UTF16_APPEND_CHAR.
michael@0:  * \pre 0<=c<=0x10ffff
michael@0:  * \pre 0<=i<length
michael@0:  * \post 0<i<=length
michael@0:  *
michael@0:  * @deprecated ICU 2.4. Use U16_APPEND instead, see utf_old.h.
michael@0:  */
michael@0: #define UTF_APPEND_CHAR(s, i, length, c) UTF16_APPEND_CHAR_SAFE(s, i, length, c)
michael@0: 
michael@0: /**
michael@0:  * Advance i to beyond the code units of the code point that begins at i.
michael@0:  * I.e., advance i by one code point.
michael@0:  * Same as UTF16_FWD_1.
michael@0:  * \pre 0<=i<length
michael@0:  * \post 0<i<=length
michael@0:  *
michael@0:  * @deprecated ICU 2.4. Renamed to U16_FWD_1, see utf_old.h.
michael@0:  */
michael@0: #define UTF_FWD_1(s, i, length) U16_FWD_1(s, i, length)
michael@0: 
michael@0: /**
michael@0:  * Advance i to beyond the code units of the n code points where the first one begins at i.
michael@0:  * I.e., advance i by n code points.
michael@0:  * Same as UT16_FWD_N.
michael@0:  * \pre 0<=i<length
michael@0:  * \post 0<i<=length
michael@0:  *
michael@0:  * @deprecated ICU 2.4. Renamed to U16_FWD_N, see utf_old.h.
michael@0:  */
michael@0: #define UTF_FWD_N(s, i, length, n) U16_FWD_N(s, i, length, n)
michael@0: 
michael@0: /**
michael@0:  * Take the random-access index i and adjust it so that it points to the beginning
michael@0:  * of a code point.
michael@0:  * The input index points to any code unit of a code point and is moved to point to
michael@0:  * the first code unit of the same code point. i is never incremented.
michael@0:  * In other words, if i points to a trail surrogate that is preceded by a matching
michael@0:  * lead surrogate, then i is decremented. Otherwise it is not modified.
michael@0:  * This can be used to start an iteration with UTF_NEXT_CHAR() from a random index.
michael@0:  * Same as UTF16_SET_CHAR_START.
michael@0:  * \pre start<=i<length
michael@0:  * \post start<=i<length
michael@0:  *
michael@0:  * @deprecated ICU 2.4. Renamed to U16_SET_CP_START, see utf_old.h.
michael@0:  */
michael@0: #define UTF_SET_CHAR_START(s, start, i) U16_SET_CP_START(s, start, i)
michael@0: 
michael@0: /**
michael@0:  * Set c to the code point that has code units before i
michael@0:  * and move i backward (towards the beginning of the string)
michael@0:  * to the first code unit of this code point (pre-increment).
michael@0:  * i must point to the first code unit after the last unit of a code point (i==length is allowed).
michael@0:  * Same as UTF16_PREV_CHAR.
michael@0:  * \pre start<i<=length
michael@0:  * \post start<=i<length
michael@0:  *
michael@0:  * @deprecated ICU 2.4. Renamed to U16_PREV, see utf_old.h.
michael@0:  */
michael@0: #define UTF_PREV_CHAR(s, start, i, c) U16_PREV(s, start, i, c)
michael@0: 
michael@0: /**
michael@0:  * Move i backward (towards the beginning of the string)
michael@0:  * to the first code unit of the code point that has code units before i.
michael@0:  * I.e., move i backward by one code point.
michael@0:  * i must point to the first code unit after the last unit of a code point (i==length is allowed).
michael@0:  * Same as UTF16_BACK_1.
michael@0:  * \pre start<i<=length
michael@0:  * \post start<=i<length
michael@0:  *
michael@0:  * @deprecated ICU 2.4. Renamed to U16_BACK_1, see utf_old.h.
michael@0:  */
michael@0: #define UTF_BACK_1(s, start, i) U16_BACK_1(s, start, i)
michael@0: 
michael@0: /**
michael@0:  * Move i backward (towards the beginning of the string)
michael@0:  * to the first code unit of the n code points that have code units before i.
michael@0:  * I.e., move i backward by n code points.
michael@0:  * i must point to the first code unit after the last unit of a code point (i==length is allowed).
michael@0:  * Same as UTF16_BACK_N.
michael@0:  * \pre start<i<=length
michael@0:  * \post start<=i<length
michael@0:  *
michael@0:  * @deprecated ICU 2.4. Renamed to U16_BACK_N, see utf_old.h.
michael@0:  */
michael@0: #define UTF_BACK_N(s, start, i, n) U16_BACK_N(s, start, i, n)
michael@0: 
michael@0: /**
michael@0:  * Take the random-access index i and adjust it so that it points beyond
michael@0:  * a code point. The input index points beyond any code unit
michael@0:  * of a code point and is moved to point beyond the last code unit of the same
michael@0:  * code point. i is never decremented.
michael@0:  * In other words, if i points to a trail surrogate that is preceded by a matching
michael@0:  * lead surrogate, then i is incremented. Otherwise it is not modified.
michael@0:  * This can be used to start an iteration with UTF_PREV_CHAR() from a random index.
michael@0:  * Same as UTF16_SET_CHAR_LIMIT.
michael@0:  * \pre start<i<=length
michael@0:  * \post start<i<=length
michael@0:  *
michael@0:  * @deprecated ICU 2.4. Renamed to U16_SET_CP_LIMIT, see utf_old.h.
michael@0:  */
michael@0: #define UTF_SET_CHAR_LIMIT(s, start, i, length) U16_SET_CP_LIMIT(s, start, i, length)
michael@0: 
michael@0: #endif /* U_HIDE_DEPRECATED_API */
michael@0: 
michael@0: #endif
michael@0: