ucase.cpp - Android社区 - https://www.androidos.net.cn/

// Copyright (C) 2016 and later: Unicode, Inc. and others.
// License & terms of use: http://www.unicode.org/copyright.html
/*
*******************************************************************************
*
*   Copyright (C) 2004-2014, International Business Machines
*   Corporation and others.  All Rights Reserved.
*
*******************************************************************************
*   file name:  ucase.cpp
*   encoding:   US-ASCII
*   tab size:   8 (not used)
*   indentation:4
*
*   created on: 2004aug30
*   created by: Markus W. Scherer
*
*   Low-level Unicode character/string case mapping code.
*   Much code moved here (and modified) from uchar.c.
*/

#include "unicode/utypes.h"
#include "unicode/unistr.h"
#include "unicode/uset.h"
#include "unicode/udata.h" /* UDataInfo */
#include "unicode/utf16.h"
#include "ucmndata.h" /* DataHeader */
#include "udatamem.h"
#include "umutex.h"
#include "uassert.h"
#include "cmemory.h"
#include "utrie2.h"
#include "ucase.h"

struct UCaseProps {
    UDataMemory *mem;
    const int32_t *indexes;
    const uint16_t *exceptions;
    const uint16_t *unfold;

UTrie2 trie;
    uint8_t formatVersion[4];
};

/* ucase_props_data.h is machine-generated by gencase --csource */
#define INCLUDED_FROM_UCASE_CPP
#include "ucase_props_data.h"

/* UCaseProps singleton ----------------------------------------------------- */

U_CAPI const UCaseProps * U_EXPORT2
ucase_getSingleton() {
    return &ucase_props_singleton;
}

/* set of property starts for UnicodeSet ------------------------------------ */

static UBool U_CALLCONV
_enumPropertyStartsRange(const void *context, UChar32 start, UChar32 /*end*/, uint32_t /*value*/) {
    /* add the start code point to the USet */
    const USetAdder *sa=(const USetAdder *)context;
    sa->add(sa->set, start);
    return TRUE;
}

U_CFUNC void U_EXPORT2
ucase_addPropertyStarts(const UCaseProps *csp, const USetAdder *sa, UErrorCode *pErrorCode) {
    if(U_FAILURE(*pErrorCode)) {
        return;
    }

/* add the start code point of each same-value range of the trie */
    utrie2_enum(&csp->trie, NULL, _enumPropertyStartsRange, sa);

/* add code points with hardcoded properties, plus the ones following them */

/* (none right now, see comment below) */

/*
     * Omit code points with hardcoded specialcasing properties
     * because we do not build property UnicodeSets for them right now.
     */
}

/* data access primitives --------------------------------------------------- */

#define GET_EXCEPTIONS(csp, props) ((csp)->exceptions+((props)>>UCASE_EXC_SHIFT))

#define PROPS_HAS_EXCEPTION(props) ((props)&UCASE_EXCEPTION)

/* number of bits in an 8-bit integer value */
static const uint8_t flagsOffset[256]={
    0, 1, 1, 2, 1, 2, 2, 3, 1, 2, 2, 3, 2, 3, 3, 4,
    1, 2, 2, 3, 2, 3, 3, 4, 2, 3, 3, 4, 3, 4, 4, 5,
    1, 2, 2, 3, 2, 3, 3, 4, 2, 3, 3, 4, 3, 4, 4, 5,
    2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6,
    1, 2, 2, 3, 2, 3, 3, 4, 2, 3, 3, 4, 3, 4, 4, 5,
    2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6,
    2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6,
    3, 4, 4, 5, 4, 5, 5, 6, 4, 5, 5, 6, 5, 6, 6, 7,
    1, 2, 2, 3, 2, 3, 3, 4, 2, 3, 3, 4, 3, 4, 4, 5,
    2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6,
    2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6,
    3, 4, 4, 5, 4, 5, 5, 6, 4, 5, 5, 6, 5, 6, 6, 7,
    2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6,
    3, 4, 4, 5, 4, 5, 5, 6, 4, 5, 5, 6, 5, 6, 6, 7,
    3, 4, 4, 5, 4, 5, 5, 6, 4, 5, 5, 6, 5, 6, 6, 7,
    4, 5, 5, 6, 5, 6, 6, 7, 5, 6, 6, 7, 6, 7, 7, 8
};

#define HAS_SLOT(flags, idx) ((flags)&(1<<(idx)))
#define SLOT_OFFSET(flags, idx) flagsOffset[(flags)&((1<<(idx))-1)]

/*
 * Get the value of an optional-value slot where HAS_SLOT(excWord, idx).
 *
 * @param excWord (in) initial exceptions word
 * @param idx (in) desired slot index
 * @param pExc16 (in/out) const uint16_t * after excWord=*pExc16++;
 *               moved to the last uint16_t of the value, use +1 for beginning of next slot
 * @param value (out) int32_t or uint32_t output if hasSlot, otherwise not modified
 */
#define GET_SLOT_VALUE(excWord, idx, pExc16, value) \
    if(((excWord)&UCASE_EXC_DOUBLE_SLOTS)==0) { \
        (pExc16)+=SLOT_OFFSET(excWord, idx); \
        (value)=*pExc16; \
    } else { \
        (pExc16)+=2*SLOT_OFFSET(excWord, idx); \
        (value)=*pExc16++; \
        (value)=((value)<<16)|*pExc16; \
    }

/* simple case mappings ----------------------------------------------------- */

U_CAPI UChar32 U_EXPORT2
ucase_tolower(const UCaseProps *csp, UChar32 c) {
    uint16_t props=UTRIE2_GET16(&csp->trie, c);
    if(!PROPS_HAS_EXCEPTION(props)) {
        if(UCASE_GET_TYPE(props)>=UCASE_UPPER) {
            c+=UCASE_GET_DELTA(props);
        }
    } else {
        const uint16_t *pe=GET_EXCEPTIONS(csp, props);
        uint16_t excWord=*pe++;
        if(HAS_SLOT(excWord, UCASE_EXC_LOWER)) {
            GET_SLOT_VALUE(excWord, UCASE_EXC_LOWER, pe, c);
        }
    }
    return c;
}

U_CAPI UChar32 U_EXPORT2
ucase_toupper(const UCaseProps *csp, UChar32 c) {
    uint16_t props=UTRIE2_GET16(&csp->trie, c);
    if(!PROPS_HAS_EXCEPTION(props)) {
        if(UCASE_GET_TYPE(props)==UCASE_LOWER) {
            c+=UCASE_GET_DELTA(props);
        }
    } else {
        const uint16_t *pe=GET_EXCEPTIONS(csp, props);
        uint16_t excWord=*pe++;
        if(HAS_SLOT(excWord, UCASE_EXC_UPPER)) {
            GET_SLOT_VALUE(excWord, UCASE_EXC_UPPER, pe, c);
        }
    }
    return c;
}

U_CAPI UChar32 U_EXPORT2
ucase_totitle(const UCaseProps *csp, UChar32 c) {
    uint16_t props=UTRIE2_GET16(&csp->trie, c);
    if(!PROPS_HAS_EXCEPTION(props)) {
        if(UCASE_GET_TYPE(props)==UCASE_LOWER) {
            c+=UCASE_GET_DELTA(props);
        }
    } else {
        const uint16_t *pe=GET_EXCEPTIONS(csp, props);
        uint16_t excWord=*pe++;
        int32_t idx;
        if(HAS_SLOT(excWord, UCASE_EXC_TITLE)) {
            idx=UCASE_EXC_TITLE;
        } else if(HAS_SLOT(excWord, UCASE_EXC_UPPER)) {
            idx=UCASE_EXC_UPPER;
        } else {
            return c;
        }
        GET_SLOT_VALUE(excWord, idx, pe, c);
    }
    return c;
}

static const UChar iDot[2] = { 0x69, 0x307 };
static const UChar jDot[2] = { 0x6a, 0x307 };
static const UChar iOgonekDot[3] = { 0x12f, 0x307 };
static const UChar iDotGrave[3] = { 0x69, 0x307, 0x300 };
static const UChar iDotAcute[3] = { 0x69, 0x307, 0x301 };
static const UChar iDotTilde[3] = { 0x69, 0x307, 0x303 };

U_CFUNC void U_EXPORT2
ucase_addCaseClosure(const UCaseProps *csp, UChar32 c, const USetAdder *sa) {
    uint16_t props;

/*
     * Hardcode the case closure of i and its relatives and ignore the
     * data file data for these characters.
     * The Turkic dotless i and dotted I with their case mapping conditions
     * and case folding option make the related characters behave specially.
     * This code matches their closure behavior to their case folding behavior.
     */

switch(c) {
    case 0x49:
        /* regular i and I are in one equivalence class */
        sa->add(sa->set, 0x69);
        return;
    case 0x69:
        sa->add(sa->set, 0x49);
        return;
    case 0x130:
        /* dotted I is in a class with <0069 0307> (for canonical equivalence with <0049 0307>) */
        sa->addString(sa->set, iDot, 2);
        return;
    case 0x131:
        /* dotless i is in a class by itself */
        return;
    default:
        /* otherwise use the data file data */
        break;
    }

props=UTRIE2_GET16(&csp->trie, c);
    if(!PROPS_HAS_EXCEPTION(props)) {
        if(UCASE_GET_TYPE(props)!=UCASE_NONE) {
            /* add the one simple case mapping, no matter what type it is */
            int32_t delta=UCASE_GET_DELTA(props);
            if(delta!=0) {
                sa->add(sa->set, c+delta);
            }
        }
    } else {
        /*
         * c has exceptions, so there may be multiple simple and/or
         * full case mappings. Add them all.
         */
        const uint16_t *pe0, *pe=GET_EXCEPTIONS(csp, props);
        const UChar *closure;
        uint16_t excWord=*pe++;
        int32_t idx, closureLength, fullLength, length;

pe0=pe;

/* add all simple case mappings */
        for(idx=UCASE_EXC_LOWER; idx<=UCASE_EXC_TITLE; ++idx) {
            if(HAS_SLOT(excWord, idx)) {
                pe=pe0;
                GET_SLOT_VALUE(excWord, idx, pe, c);
                sa->add(sa->set, c);
            }
        }

/* get the closure string pointer & length */
        if(HAS_SLOT(excWord, UCASE_EXC_CLOSURE)) {
            pe=pe0;
            GET_SLOT_VALUE(excWord, UCASE_EXC_CLOSURE, pe, closureLength);
            closureLength&=UCASE_CLOSURE_MAX_LENGTH; /* higher bits are reserved */
            closure=(const UChar *)pe+1; /* behind this slot, unless there are full case mappings */
        } else {
            closureLength=0;
            closure=NULL;
        }

/* add the full case folding */
        if(HAS_SLOT(excWord, UCASE_EXC_FULL_MAPPINGS)) {
            pe=pe0;
            GET_SLOT_VALUE(excWord, UCASE_EXC_FULL_MAPPINGS, pe, fullLength);

/* start of full case mapping strings */
            ++pe;

fullLength&=0xffff; /* bits 16 and higher are reserved */

/* skip the lowercase result string */
            pe+=fullLength&UCASE_FULL_LOWER;
            fullLength>>=4;

/* add the full case folding string */
            length=fullLength&0xf;
            if(length!=0) {
                sa->addString(sa->set, (const UChar *)pe, length);
                pe+=length;
            }

/* skip the uppercase and titlecase strings */
            fullLength>>=4;
            pe+=fullLength&0xf;
            fullLength>>=4;
            pe+=fullLength;

closure=(const UChar *)pe; /* behind full case mappings */
        }

/* add each code point in the closure string */
        for(idx=0; idx<closureLength;) {
            U16_NEXT_UNSAFE(closure, idx, c);
            sa->add(sa->set, c);
        }
    }
}

/*
 * compare s, which has a length, with t, which has a maximum length or is NUL-terminated
 * must be length>0 and max>0 and length<=max
 */
static inline int32_t
strcmpMax(const UChar *s, int32_t length, const UChar *t, int32_t max) {
    int32_t c1, c2;

max-=length; /* we require length<=max, so no need to decrement max in the loop */
    do {
        c1=*s++;
        c2=*t++;
        if(c2==0) {
            return 1; /* reached the end of t but not of s */
        }
        c1-=c2;
        if(c1!=0) {
            return c1; /* return difference result */
        }
    } while(--length>0);
    /* ends with length==0 */

if(max==0 || *t==0) {
        return 0; /* equal to length of both strings */
    } else {
        return -max; /* return lengh difference */
    }
}

U_CFUNC UBool U_EXPORT2
ucase_addStringCaseClosure(const UCaseProps *csp, const UChar *s, int32_t length, const USetAdder *sa) {
    int32_t i, start, limit, result, unfoldRows, unfoldRowWidth, unfoldStringWidth;

if(csp->unfold==NULL || s==NULL) {
        return FALSE; /* no reverse case folding data, or no string */
    }
    if(length<=1) {
        /* the string is too short to find any match */
        /*
         * more precise would be:
         * if(!u_strHasMoreChar32Than(s, length, 1))
         * but this does not make much practical difference because
         * a single supplementary code point would just not be found
         */
        return FALSE;
    }

const uint16_t *unfold=csp->unfold;
    unfoldRows=unfold[UCASE_UNFOLD_ROWS];
    unfoldRowWidth=unfold[UCASE_UNFOLD_ROW_WIDTH];
    unfoldStringWidth=unfold[UCASE_UNFOLD_STRING_WIDTH];
    unfold+=unfoldRowWidth;

if(length>unfoldStringWidth) {
        /* the string is too long to find any match */
        return FALSE;
    }

/* do a binary search for the string */
    start=0;
    limit=unfoldRows;
    while(start<limit) {
        i=(start+limit)/2;
        const UChar *p=reinterpret_cast<const UChar *>(unfold+(i*unfoldRowWidth));
        result=strcmpMax(s, length, p, unfoldStringWidth);

if(result==0) {
            /* found the string: add each code point, and its case closure */
            UChar32 c;

for(i=unfoldStringWidth; i<unfoldRowWidth && p[i]!=0;) {
                U16_NEXT_UNSAFE(p, i, c);
                sa->add(sa->set, c);
                ucase_addCaseClosure(csp, c, sa);
            }
            return TRUE;
        } else if(result<0) {
            limit=i;
        } else /* result>0 */ {
            start=i+1;
        }
    }

return FALSE; /* string not found */
}

U_NAMESPACE_BEGIN

FullCaseFoldingIterator::FullCaseFoldingIterator()
        : unfold(reinterpret_cast<const UChar *>(ucase_props_singleton.unfold)),
          unfoldRows(unfold[UCASE_UNFOLD_ROWS]),
          unfoldRowWidth(unfold[UCASE_UNFOLD_ROW_WIDTH]),
          unfoldStringWidth(unfold[UCASE_UNFOLD_STRING_WIDTH]),
          currentRow(0),
          rowCpIndex(unfoldStringWidth) {
    unfold+=unfoldRowWidth;
}

UChar32
FullCaseFoldingIterator::next(UnicodeString &full) {
    // Advance past the last-delivered code point.
    const UChar *p=unfold+(currentRow*unfoldRowWidth);
    if(rowCpIndex>=unfoldRowWidth || p[rowCpIndex]==0) {
        ++currentRow;
        p+=unfoldRowWidth;
        rowCpIndex=unfoldStringWidth;
    }
    if(currentRow>=unfoldRows) { return U_SENTINEL; }
    // Set "full" to the NUL-terminated string in the first unfold column.
    int32_t length=unfoldStringWidth;
    while(length>0 && p[length-1]==0) { --length; }
    full.setTo(FALSE, p, length);
    // Return the code point.
    UChar32 c;
    U16_NEXT_UNSAFE(p, rowCpIndex, c);
    return c;
}

U_NAMESPACE_END

/** @return UCASE_NONE, UCASE_LOWER, UCASE_UPPER, UCASE_TITLE */
U_CAPI int32_t U_EXPORT2
ucase_getType(const UCaseProps *csp, UChar32 c) {
    uint16_t props=UTRIE2_GET16(&csp->trie, c);
    return UCASE_GET_TYPE(props);
}

/** @return same as ucase_getType() and set bit 2 if c is case-ignorable */
U_CAPI int32_t U_EXPORT2
ucase_getTypeOrIgnorable(const UCaseProps *csp, UChar32 c) {
    uint16_t props=UTRIE2_GET16(&csp->trie, c);
    return UCASE_GET_TYPE_AND_IGNORABLE(props);
}

/** @return UCASE_NO_DOT, UCASE_SOFT_DOTTED, UCASE_ABOVE, UCASE_OTHER_ACCENT */
static inline int32_t
getDotType(const UCaseProps *csp, UChar32 c) {
    uint16_t props=UTRIE2_GET16(&csp->trie, c);
    if(!PROPS_HAS_EXCEPTION(props)) {
        return props&UCASE_DOT_MASK;
    } else {
        const uint16_t *pe=GET_EXCEPTIONS(csp, props);
        return (*pe>>UCASE_EXC_DOT_SHIFT)&UCASE_DOT_MASK;
    }
}

U_CAPI UBool U_EXPORT2
ucase_isSoftDotted(const UCaseProps *csp, UChar32 c) {
    return (UBool)(getDotType(csp, c)==UCASE_SOFT_DOTTED);
}

U_CAPI UBool U_EXPORT2
ucase_isCaseSensitive(const UCaseProps *csp, UChar32 c) {
    uint16_t props=UTRIE2_GET16(&csp->trie, c);
    return (UBool)((props&UCASE_SENSITIVE)!=0);
}

/* string casing ------------------------------------------------------------ */

/*
 * These internal functions form the core of string case mappings.
 * They map single code points to result code points or strings and take
 * all necessary conditions (context, locale ID, options) into account.
 *
 * They do not iterate over the source or write to the destination
 * so that the same functions are useful for non-standard string storage,
 * such as in a Replaceable (for Transliterator) or UTF-8/32 strings etc.
 * For the same reason, the "surrounding text" context is passed in as a
 * UCaseContextIterator which does not make any assumptions about
 * the underlying storage.
 *
 * This section contains helper functions that check for conditions
 * in the input text surrounding the current code point
 * according to SpecialCasing.txt.
 *
 * Each helper function gets the index
 * - after the current code point if it looks at following text
 * - before the current code point if it looks at preceding text
 *
 * Unicode 3.2 UAX 21 "Case Mappings" defines the conditions as follows:
 *
 * Final_Sigma
 *   C is preceded by a sequence consisting of
 *     a cased letter and a case-ignorable sequence,
 *   and C is not followed by a sequence consisting of
 *     an ignorable sequence and then a cased letter.
 *
 * More_Above
 *   C is followed by one or more characters of combining class 230 (ABOVE)
 *   in the combining character sequence.
 *
 * After_Soft_Dotted
 *   The last preceding character with combining class of zero before C
 *   was Soft_Dotted,
 *   and there is no intervening combining character class 230 (ABOVE).
 *
 * Before_Dot
 *   C is followed by combining dot above (U+0307).
 *   Any sequence of characters with a combining class that is neither 0 nor 230
 *   may intervene between the current character and the combining dot above.
 *
 * The erratum from 2002-10-31 adds the condition
 *
 * After_I
 *   The last preceding base character was an uppercase I, and there is no
 *   intervening combining character class 230 (ABOVE).
 *
 *   (See Jitterbug 2344 and the comments on After_I below.)
 *
 * Helper definitions in Unicode 3.2 UAX 21:
 *
 * D1. A character C is defined to be cased
 *     if it meets any of the following criteria:
 *
 *   - The general category of C is Titlecase Letter (Lt)
 *   - In [CoreProps], C has one of the properties Uppercase, or Lowercase
 *   - Given D = NFD(C), then it is not the case that:
 *     D = UCD_lower(D) = UCD_upper(D) = UCD_title(D)
 *     (This third criterium does not add any characters to the list
 *      for Unicode 3.2. Ignored.)
 *
 * D2. A character C is defined to be case-ignorable
 *     if it meets either of the following criteria:
 *
 *   - The general category of C is
 *     Nonspacing Mark (Mn), or Enclosing Mark (Me), or Format Control (Cf), or
 *     Letter Modifier (Lm), or Symbol Modifier (Sk)
 *   - C is one of the following characters 
 *     U+0027 APOSTROPHE
 *     U+00AD SOFT HYPHEN (SHY)
 *     U+2019 RIGHT SINGLE QUOTATION MARK
 *            (the preferred character for apostrophe)
 *
 * D3. A case-ignorable sequence is a sequence of
 *     zero or more case-ignorable characters.
 */

#define is_a(c) ((c)=='a' || (c)=='A')
#define is_d(c) ((c)=='d' || (c)=='D')
#define is_e(c) ((c)=='e' || (c)=='E')
#define is_i(c) ((c)=='i' || (c)=='I')
#define is_l(c) ((c)=='l' || (c)=='L')
#define is_n(c) ((c)=='n' || (c)=='N')
#define is_r(c) ((c)=='r' || (c)=='R')
#define is_t(c) ((c)=='t' || (c)=='T')
#define is_u(c) ((c)=='u' || (c)=='U')
#define is_z(c) ((c)=='z' || (c)=='Z')

/* separator? */
#define is_sep(c) ((c)=='_' || (c)=='-' || (c)==0)

/**
 * Requires non-NULL locale ID but otherwise does the equivalent of
 * checking for language codes as if uloc_getLanguage() were called:
 * Accepts both 2- and 3-letter codes and accepts case variants.
 */
U_CFUNC int32_t
ucase_getCaseLocale(const char *locale, int32_t *locCache) {
    int32_t result;
    char c;

if(locCache!=NULL && (result=*locCache)!=UCASE_LOC_UNKNOWN) {
        return result;
    }

result=UCASE_LOC_ROOT;

/*
     * This function used to use uloc_getLanguage(), but the current code
     * removes the dependency of this low-level code on uloc implementation code
     * and is faster because not the whole locale ID has to be
     * examined and copied/transformed.
     *
     * Because this code does not want to depend on uloc, the caller must
     * pass in a non-NULL locale, i.e., may need to call uloc_getDefault().
     */
    c=*locale++;
    if(is_t(c)) {
        /* tr or tur? */
        c=*locale++;
        if(is_u(c)) {
            c=*locale++;
        }
        if(is_r(c)) {
            c=*locale;
            if(is_sep(c)) {
                result=UCASE_LOC_TURKISH;
            }
        }
    } else if(is_a(c)) {
        /* az or aze? */
        c=*locale++;
        if(is_z(c)) {
            c=*locale++;
            if(is_e(c)) {
                c=*locale;
            }
            if(is_sep(c)) {
                result=UCASE_LOC_TURKISH;
            }
        }
    } else if(is_l(c)) {
        /* lt or lit? */
        c=*locale++;
        if(is_i(c)) {
            c=*locale++;
        }
        if(is_t(c)) {
            c=*locale;
            if(is_sep(c)) {
                result=UCASE_LOC_LITHUANIAN;
            }
        }
    } else if(is_e(c)) {
        /* el or ell? */
        c=*locale++;
        if(is_l(c)) {
            c=*locale++;
            if(is_l(c)) {
                c=*locale;
            }
            if(is_sep(c)) {
                result=UCASE_LOC_GREEK;
            }
        }
    } else if(is_n(c)) {
        /* nl or nld? */
        c=*locale++;
        if(is_l(c)) {
            c=*locale++;
            if(is_d(c)) {
                c=*locale;
            }
            if(is_sep(c)) {
                result=UCASE_LOC_DUTCH;
            }
        }
    }

if(locCache!=NULL) {
        *locCache=result;
    }
    return result;
}

/*
 * Is followed by
 *   {case-ignorable}* cased
 * ?
 * (dir determines looking forward/backward)
 * If a character is case-ignorable, it is skipped regardless of whether
 * it is also cased or not.
 */
static UBool
isFollowedByCasedLetter(const UCaseProps *csp, UCaseContextIterator *iter, void *context, int8_t dir) {
    UChar32 c;