AuroraMiddleware/scuffed-code
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
b9ef7bcac4
scuffed-code/icu4c/source/common/ucase.cpp

1383 lines
47 KiB
C++
Raw Blame History

// © 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: UTF-8
* 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"
/* 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 USetAdder *sa, UErrorCode *pErrorCode) {
if(U_FAILURE(*pErrorCode)) {
return;
}
/* add the start code point of each same-value range of the trie */
utrie2_enum(&ucase_props_singleton.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(UChar32 c) {
uint16_t props=UTRIE2_GET16(&ucase_props_singleton.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(&ucase_props_singleton, 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(UChar32 c) {
uint16_t props=UTRIE2_GET16(&ucase_props_singleton.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(&ucase_props_singleton, 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(UChar32 c) {
uint16_t props=UTRIE2_GET16(&ucase_props_singleton.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(&ucase_props_singleton, 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(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(&ucase_props_singleton.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(&ucase_props_singleton, 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 UChar *s, int32_t length, const USetAdder *sa) {
int32_t i, start, limit, result, unfoldRows, unfoldRowWidth, unfoldStringWidth;
if(ucase_props_singleton.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=ucase_props_singleton.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(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(UChar32 c) {
uint16_t props=UTRIE2_GET16(&ucase_props_singleton.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(UChar32 c) {
uint16_t props=UTRIE2_GET16(&ucase_props_singleton.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(UChar32 c) {
uint16_t props=UTRIE2_GET16(&ucase_props_singleton.trie, c);
if(!PROPS_HAS_EXCEPTION(props)) {
return props&UCASE_DOT_MASK;
} else {
const uint16_t *pe=GET_EXCEPTIONS(&ucase_props_singleton, props);
return (*pe>>UCASE_EXC_DOT_SHIFT)&UCASE_DOT_MASK;
}
}
U_CAPI UBool U_EXPORT2
ucase_isSoftDotted(UChar32 c) {
return (UBool)(getDotType(c)==UCASE_SOFT_DOTTED);
}
U_CAPI UBool U_EXPORT2
ucase_isCaseSensitive(UChar32 c) {
uint16_t props=UTRIE2_GET16(&ucase_props_singleton.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_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_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) {
/*
* 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().
*/
char c=*locale++;
// Fastpath for English "en" which is often used for default (=root locale) case mappings,
// and for Chinese "zh": Very common but no special case mapping behavior.
// Then check lowercase vs. uppercase to reduce the number of comparisons
// for other locales without special behavior.
if(c=='e') {
/* el or ell? */
c=*locale++;
if(is_l(c)) {
c=*locale++;
if(is_l(c)) {
c=*locale;
}
if(is_sep(c)) {
return UCASE_LOC_GREEK;
}
}
// en, es, ... -> root
} else if(c=='z') {
return UCASE_LOC_ROOT;
#if U_CHARSET_FAMILY==U_ASCII_FAMILY
} else if(c>='a') { // ASCII a-z = 0x61..0x7a, after A-Z
#elif U_CHARSET_FAMILY==U_EBCDIC_FAMILY
} else if(c<='z') { // EBCDIC a-z = 0x81..0xa9 with two gaps, before A-Z
#else
# error Unknown charset family!
#endif
// lowercase c
if(c=='t') {
/* tr or tur? */
c=*locale++;
if(is_u(c)) {
c=*locale++;
}
if(is_r(c)) {
c=*locale;
if(is_sep(c)) {
return UCASE_LOC_TURKISH;
}
}
} else if(c=='a') {
/* az or aze? */
c=*locale++;
if(is_z(c)) {
c=*locale++;
if(is_e(c)) {
c=*locale;
}
if(is_sep(c)) {
return UCASE_LOC_TURKISH;
}
}
} else if(c=='l') {
/* lt or lit? */
c=*locale++;
if(is_i(c)) {
c=*locale++;
}
if(is_t(c)) {
c=*locale;
if(is_sep(c)) {
return UCASE_LOC_LITHUANIAN;
}
}
} else if(c=='n') {
/* nl or nld? */
c=*locale++;
if(is_l(c)) {
c=*locale++;
if(is_d(c)) {
c=*locale;
}
if(is_sep(c)) {
return UCASE_LOC_DUTCH;
}
}
}
} else {
// uppercase c
// Same code as for lowercase c but also check for 'E'.
if(c=='T') {
/* tr or tur? */
c=*locale++;
if(is_u(c)) {
c=*locale++;
}
if(is_r(c)) {
c=*locale;
if(is_sep(c)) {
return UCASE_LOC_TURKISH;
}
}
} else if(c=='A') {
/* az or aze? */
c=*locale++;
if(is_z(c)) {
c=*locale++;
if(is_e(c)) {
c=*locale;
}
if(is_sep(c)) {
return UCASE_LOC_TURKISH;
}
}
} else if(c=='L') {
/* lt or lit? */
c=*locale++;
if(is_i(c)) {
c=*locale++;
}
if(is_t(c)) {
c=*locale;
if(is_sep(c)) {
return UCASE_LOC_LITHUANIAN;
}
}
} else if(c=='E') {
/* el or ell? */
c=*locale++;
if(is_l(c)) {
c=*locale++;
if(is_l(c)) {
c=*locale;
}
if(is_sep(c)) {
return UCASE_LOC_GREEK;
}
}
} else if(c=='N') {
/* nl or nld? */
c=*locale++;
if(is_l(c)) {
c=*locale++;
if(is_d(c)) {
c=*locale;
}
if(is_sep(c)) {
return UCASE_LOC_DUTCH;
}
}
}
}
return UCASE_LOC_ROOT;
}
/*
* 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(UCaseContextIterator *iter, void *context, int8_t dir) {
UChar32 c;
if(iter==NULL) {
return FALSE;
}
for(/* dir!=0 sets direction */; (c=iter(context, dir))>=0; dir=0) {
int32_t type=ucase_getTypeOrIgnorable(c);
if(type&4) {
/* case-ignorable, continue with the loop */
} else if(type!=UCASE_NONE) {
return TRUE; /* followed by cased letter */
} else {
return FALSE; /* uncased and not case-ignorable */
}
}
return FALSE; /* not followed by cased letter */
}
/* Is preceded by Soft_Dotted character with no intervening cc=230 ? */
static UBool
isPrecededBySoftDotted(UCaseContextIterator *iter, void *context) {
UChar32 c;
int32_t dotType;
int8_t dir;
if(iter==NULL) {
return FALSE;
}
for(dir=-1; (c=iter(context, dir))>=0; dir=0) {
dotType=getDotType(c);
if(dotType==UCASE_SOFT_DOTTED) {
return TRUE; /* preceded by TYPE_i */
} else if(dotType!=UCASE_OTHER_ACCENT) {
return FALSE; /* preceded by different base character (not TYPE_i), or intervening cc==230 */
}
}
return FALSE; /* not preceded by TYPE_i */
}
/*
* See Jitterbug 2344:
* The condition After_I for Turkic-lowercasing of U+0307 combining dot above
* is checked in ICU 2.0, 2.1, 2.6 but was not in 2.2 & 2.4 because
* we made those releases compatible with Unicode 3.2 which had not fixed
* a related bug in SpecialCasing.txt.
*
* From the Jitterbug 2344 text:
* ... this bug is listed as a Unicode erratum
* from 2002-10-31 at http://www.unicode.org/uni2errata/UnicodeErrata.html
* <quote>
* There are two errors in SpecialCasing.txt.
* 1. Missing semicolons on two lines. ... [irrelevant for ICU]
* 2. An incorrect context definition. Correct as follows:
* < 0307; ; 0307; 0307; tr After_Soft_Dotted; # COMBINING DOT ABOVE
* < 0307; ; 0307; 0307; az After_Soft_Dotted; # COMBINING DOT ABOVE
* ---
* > 0307; ; 0307; 0307; tr After_I; # COMBINING DOT ABOVE
* > 0307; ; 0307; 0307; az After_I; # COMBINING DOT ABOVE
* where the context After_I is defined as:
* The last preceding base character was an uppercase I, and there is no
* intervening combining character class 230 (ABOVE).
* </quote>
*
* Note that SpecialCasing.txt even in Unicode 3.2 described the condition as:
*
* # When lowercasing, remove dot_above in the sequence I + dot_above, which will turn into i.
* # This matches the behavior of the canonically equivalent I-dot_above
*
* See also the description in this place in older versions of uchar.c (revision 1.100).
*
* Markus W. Scherer 2003-feb-15
*/
/* Is preceded by base character 'I' with no intervening cc=230 ? */
static UBool
isPrecededBy_I(UCaseContextIterator *iter, void *context) {
UChar32 c;
int32_t dotType;
int8_t dir;
if(iter==NULL) {
return FALSE;
}
for(dir=-1; (c=iter(context, dir))>=0; dir=0) {
if(c==0x49) {
return TRUE; /* preceded by I */
}
dotType=getDotType(c);
if(dotType!=UCASE_OTHER_ACCENT) {
return FALSE; /* preceded by different base character (not I), or intervening cc==230 */
}
}
return FALSE; /* not preceded by I */
}
/* Is followed by one or more cc==230 ? */
static UBool
isFollowedByMoreAbove(UCaseContextIterator *iter, void *context) {
UChar32 c;
int32_t dotType;
int8_t dir;
if(iter==NULL) {
return FALSE;
}
for(dir=1; (c=iter(context, dir))>=0; dir=0) {
dotType=getDotType(c);
if(dotType==UCASE_ABOVE) {
return TRUE; /* at least one cc==230 following */
} else if(dotType!=UCASE_OTHER_ACCENT) {
return FALSE; /* next base character, no more cc==230 following */
}
}
return FALSE; /* no more cc==230 following */
}
/* Is followed by a dot above (without cc==230 in between) ? */
static UBool
isFollowedByDotAbove(UCaseContextIterator *iter, void *context) {
UChar32 c;
int32_t dotType;
int8_t dir;
if(iter==NULL) {
return FALSE;
}
for(dir=1; (c=iter(context, dir))>=0; dir=0) {
if(c==0x307) {
return TRUE;
}
dotType=getDotType(c);
if(dotType!=UCASE_OTHER_ACCENT) {
return FALSE; /* next base character or cc==230 in between */
}
}
return FALSE; /* no dot above following */
}
U_CAPI int32_t U_EXPORT2
ucase_toFullLower(UChar32 c,
UCaseContextIterator *iter, void *context,
const UChar **pString,
int32_t loc) {
// The sign of the result has meaning, input must be non-negative so that it can be returned as is.
U_ASSERT(c >= 0);
UChar32 result=c;
uint16_t props=UTRIE2_GET16(&ucase_props_singleton.trie, c);
if(!PROPS_HAS_EXCEPTION(props)) {
if(UCASE_GET_TYPE(props)>=UCASE_UPPER) {
result=c+UCASE_GET_DELTA(props);
}
} else {
const uint16_t *pe=GET_EXCEPTIONS(&ucase_props_singleton, props), *pe2;
uint16_t excWord=*pe++;
int32_t full;
pe2=pe;
if(excWord&UCASE_EXC_CONDITIONAL_SPECIAL) {
/* use hardcoded conditions and mappings */
/*
* Test for conditional mappings first
* (otherwise the unconditional default mappings are always taken),
* then test for characters that have unconditional mappings in SpecialCasing.txt,
* then get the UnicodeData.txt mappings.
*/
if( loc==UCASE_LOC_LITHUANIAN &&
/* base characters, find accents above */
(((c==0x49 || c==0x4a || c==0x12e) &&
isFollowedByMoreAbove(iter, context)) ||
/* precomposed with accent above, no need to find one */
(c==0xcc || c==0xcd || c==0x128))
) {
/*
# Lithuanian
# Lithuanian retains the dot in a lowercase i when followed by accents.
# Introduce an explicit dot above when lowercasing capital I's and J's
# whenever there are more accents above.
# (of the accents used in Lithuanian: grave, acute, tilde above, and ogonek)
0049; 0069 0307; 0049; 0049; lt More_Above; # LATIN CAPITAL LETTER I
004A; 006A 0307; 004A; 004A; lt More_Above; # LATIN CAPITAL LETTER J
012E; 012F 0307; 012E; 012E; lt More_Above; # LATIN CAPITAL LETTER I WITH OGONEK
00CC; 0069 0307 0300; 00CC; 00CC; lt; # LATIN CAPITAL LETTER I WITH GRAVE
00CD; 0069 0307 0301; 00CD; 00CD; lt; # LATIN CAPITAL LETTER I WITH ACUTE
0128; 0069 0307 0303; 0128; 0128; lt; # LATIN CAPITAL LETTER I WITH TILDE
*/
switch(c) {
case 0x49: /* LATIN CAPITAL LETTER I */
*pString=iDot;
return 2;
case 0x4a: /* LATIN CAPITAL LETTER J */
*pString=jDot;
return 2;
case 0x12e: /* LATIN CAPITAL LETTER I WITH OGONEK */
*pString=iOgonekDot;
return 2;
case 0xcc: /* LATIN CAPITAL LETTER I WITH GRAVE */
*pString=iDotGrave;
return 3;
case 0xcd: /* LATIN CAPITAL LETTER I WITH ACUTE */
*pString=iDotAcute;
return 3;
case 0x128: /* LATIN CAPITAL LETTER I WITH TILDE */
*pString=iDotTilde;
return 3;
default:
return 0; /* will not occur */
}
/* # Turkish and Azeri */
} else if(loc==UCASE_LOC_TURKISH && c==0x130) {
/*
# I and i-dotless; I-dot and i are case pairs in Turkish and Azeri
# The following rules handle those cases.
0130; 0069; 0130; 0130; tr # LATIN CAPITAL LETTER I WITH DOT ABOVE
0130; 0069; 0130; 0130; az # LATIN CAPITAL LETTER I WITH DOT ABOVE
*/
return 0x69;
} else if(loc==UCASE_LOC_TURKISH && c==0x307 && isPrecededBy_I(iter, context)) {
/*
# When lowercasing, remove dot_above in the sequence I + dot_above, which will turn into i.
# This matches the behavior of the canonically equivalent I-dot_above
0307; ; 0307; 0307; tr After_I; # COMBINING DOT ABOVE
0307; ; 0307; 0307; az After_I; # COMBINING DOT ABOVE
*/
return 0; /* remove the dot (continue without output) */
} else if(loc==UCASE_LOC_TURKISH && c==0x49 && !isFollowedByDotAbove(iter, context)) {
/*
# When lowercasing, unless an I is before a dot_above, it turns into a dotless i.
0049; 0131; 0049; 0049; tr Not_Before_Dot; # LATIN CAPITAL LETTER I
0049; 0131; 0049; 0049; az Not_Before_Dot; # LATIN CAPITAL LETTER I
*/
return 0x131;
} else if(c==0x130) {
/*
# Preserve canonical equivalence for I with dot. Turkic is handled below.
0130; 0069 0307; 0130; 0130; # LATIN CAPITAL LETTER I WITH DOT ABOVE
*/
*pString=iDot;
return 2;
} else if( c==0x3a3 &&
!isFollowedByCasedLetter(iter, context, 1) &&
isFollowedByCasedLetter(iter, context, -1) /* -1=preceded */
) {
/* greek capital sigma maps depending on surrounding cased letters (see SpecialCasing.txt) */
/*
# Special case for final form of sigma
03A3; 03C2; 03A3; 03A3; Final_Sigma; # GREEK CAPITAL LETTER SIGMA
*/
return 0x3c2; /* greek small final sigma */
} else {
/* no known conditional special case mapping, use a normal mapping */
}
} else if(HAS_SLOT(excWord, UCASE_EXC_FULL_MAPPINGS)) {
GET_SLOT_VALUE(excWord, UCASE_EXC_FULL_MAPPINGS, pe, full);
full&=UCASE_FULL_LOWER;
if(full!=0) {
/* set the output pointer to the lowercase mapping */
*pString=reinterpret_cast<const UChar *>(pe+1);
/* return the string length */
return full;
}
}
if(HAS_SLOT(excWord, UCASE_EXC_LOWER)) {
GET_SLOT_VALUE(excWord, UCASE_EXC_LOWER, pe2, result);
}
}
return (result==c) ? ~result : result;
}
/* internal */
static int32_t
toUpperOrTitle(UChar32 c,
UCaseContextIterator *iter, void *context,
const UChar **pString,
int32_t loc,
UBool upperNotTitle) {
// The sign of the result has meaning, input must be non-negative so that it can be returned as is.
U_ASSERT(c >= 0);
UChar32 result=c;
uint16_t props=UTRIE2_GET16(&ucase_props_singleton.trie, c);
if(!PROPS_HAS_EXCEPTION(props)) {
if(UCASE_GET_TYPE(props)==UCASE_LOWER) {
result=c+UCASE_GET_DELTA(props);
}
} else {
const uint16_t *pe=GET_EXCEPTIONS(&ucase_props_singleton, props), *pe2;
uint16_t excWord=*pe++;
int32_t full, idx;
pe2=pe;
if(excWord&UCASE_EXC_CONDITIONAL_SPECIAL) {
/* use hardcoded conditions and mappings */
if(loc==UCASE_LOC_TURKISH && c==0x69) {
/*
# Turkish and Azeri
# I and i-dotless; I-dot and i are case pairs in Turkish and Azeri
# The following rules handle those cases.
# When uppercasing, i turns into a dotted capital I
0069; 0069; 0130; 0130; tr; # LATIN SMALL LETTER I
0069; 0069; 0130; 0130; az; # LATIN SMALL LETTER I
*/
return 0x130;
} else if(loc==UCASE_LOC_LITHUANIAN && c==0x307 && isPrecededBySoftDotted(iter, context)) {
/*
# Lithuanian
# Lithuanian retains the dot in a lowercase i when followed by accents.
# Remove DOT ABOVE after "i" with upper or titlecase
0307; 0307; ; ; lt After_Soft_Dotted; # COMBINING DOT ABOVE
*/
return 0; /* remove the dot (continue without output) */
} else {
/* no known conditional special case mapping, use a normal mapping */
}
} else if(HAS_SLOT(excWord, UCASE_EXC_FULL_MAPPINGS)) {
GET_SLOT_VALUE(excWord, UCASE_EXC_FULL_MAPPINGS, pe, full);
/* start of full case mapping strings */
++pe;
/* skip the lowercase and case-folding result strings */
pe+=full&UCASE_FULL_LOWER;
full>>=4;
pe+=full&0xf;
full>>=4;
if(upperNotTitle) {
full&=0xf;
} else {
/* skip the uppercase result string */
pe+=full&0xf;
full=(full>>4)&0xf;
}
if(full!=0) {
/* set the output pointer to the result string */
*pString=reinterpret_cast<const UChar *>(pe);
/* return the string length */
return full;
}
}
if(!upperNotTitle && HAS_SLOT(excWord, UCASE_EXC_TITLE)) {
idx=UCASE_EXC_TITLE;
} else if(HAS_SLOT(excWord, UCASE_EXC_UPPER)) {
/* here, titlecase is same as uppercase */
idx=UCASE_EXC_UPPER;
} else {
return ~c;
}
GET_SLOT_VALUE(excWord, idx, pe2, result);
}
return (result==c) ? ~result : result;
}
U_CAPI int32_t U_EXPORT2
ucase_toFullUpper(UChar32 c,
UCaseContextIterator *iter, void *context,
const UChar **pString,
int32_t caseLocale) {
return toUpperOrTitle(c, iter, context, pString, caseLocale, TRUE);
}
U_CAPI int32_t U_EXPORT2
ucase_toFullTitle(UChar32 c,
UCaseContextIterator *iter, void *context,
const UChar **pString,
int32_t caseLocale) {
return toUpperOrTitle(c, iter, context, pString, caseLocale, FALSE);
}
/* case folding ------------------------------------------------------------- */
/*
* Case folding is similar to lowercasing.
* The result may be a simple mapping, i.e., a single code point, or
* a full mapping, i.e., a string.
* If the case folding for a code point is the same as its simple (1:1) lowercase mapping,
* then only the lowercase mapping is stored.
*
* Some special cases are hardcoded because their conditions cannot be
* parsed and processed from CaseFolding.txt.
*
* Unicode 3.2 CaseFolding.txt specifies for its status field:
# C: common case folding, common mappings shared by both simple and full mappings.
# F: full case folding, mappings that cause strings to grow in length. Multiple characters are separated by spaces.
# S: simple case folding, mappings to single characters where different from F.
# T: special case for uppercase I and dotted uppercase I
# - For non-Turkic languages, this mapping is normally not used.
# - For Turkic languages (tr, az), this mapping can be used instead of the normal mapping for these characters.
#
# Usage:
# A. To do a simple case folding, use the mappings with status C + S.
# B. To do a full case folding, use the mappings with status C + F.
#
# The mappings with status T can be used or omitted depending on the desired case-folding
# behavior. (The default option is to exclude them.)
* Unicode 3.2 has 'T' mappings as follows:
0049; T; 0131; # LATIN CAPITAL LETTER I
0130; T; 0069; # LATIN CAPITAL LETTER I WITH DOT ABOVE
* while the default mappings for these code points are:
0049; C; 0069; # LATIN CAPITAL LETTER I
0130; F; 0069 0307; # LATIN CAPITAL LETTER I WITH DOT ABOVE
* U+0130 has no simple case folding (simple-case-folds to itself).
*/
/* return the simple case folding mapping for c */
U_CAPI UChar32 U_EXPORT2
ucase_fold(UChar32 c, uint32_t options) {
uint16_t props=UTRIE2_GET16(&ucase_props_singleton.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(&ucase_props_singleton, props);
uint16_t excWord=*pe++;
int32_t idx;
if(excWord&UCASE_EXC_CONDITIONAL_FOLD) {
/* special case folding mappings, hardcoded */
if((options&_FOLD_CASE_OPTIONS_MASK)==U_FOLD_CASE_DEFAULT) {
/* default mappings */
if(c==0x49) {
/* 0049; C; 0069; # LATIN CAPITAL LETTER I */
return 0x69;
} else if(c==0x130) {
/* no simple case folding for U+0130 */
return c;
}
} else {
/* Turkic mappings */
if(c==0x49) {
/* 0049; T; 0131; # LATIN CAPITAL LETTER I */
return 0x131;
} else if(c==0x130) {
/* 0130; T; 0069; # LATIN CAPITAL LETTER I WITH DOT ABOVE */
return 0x69;
}
}
}
if(HAS_SLOT(excWord, UCASE_EXC_FOLD)) {
idx=UCASE_EXC_FOLD;
} else if(HAS_SLOT(excWord, UCASE_EXC_LOWER)) {
idx=UCASE_EXC_LOWER;
} else {
return c;
}
GET_SLOT_VALUE(excWord, idx, pe, c);
}
return c;
}
/*
* Issue for canonical caseless match (UAX #21):
* Turkic casefolding (using "T" mappings in CaseFolding.txt) does not preserve
* canonical equivalence, unlike default-option casefolding.
* For example, I-grave and I + grave fold to strings that are not canonically
* equivalent.
* For more details, see the comment in unorm_compare() in unorm.cpp
* and the intermediate prototype changes for Jitterbug 2021.
* (For example, revision 1.104 of uchar.c and 1.4 of CaseFolding.txt.)
*
* This did not get fixed because it appears that it is not possible to fix
* it for uppercase and lowercase characters (I-grave vs. i-grave)
* together in a way that they still fold to common result strings.
*/
U_CAPI int32_t U_EXPORT2
ucase_toFullFolding(UChar32 c,
const UChar **pString,
uint32_t options) {
// The sign of the result has meaning, input must be non-negative so that it can be returned as is.
U_ASSERT(c >= 0);
UChar32 result=c;
uint16_t props=UTRIE2_GET16(&ucase_props_singleton.trie, c);
if(!PROPS_HAS_EXCEPTION(props)) {
if(UCASE_GET_TYPE(props)>=UCASE_UPPER) {
result=c+UCASE_GET_DELTA(props);
}
} else {
const uint16_t *pe=GET_EXCEPTIONS(&ucase_props_singleton, props), *pe2;
uint16_t excWord=*pe++;
int32_t full, idx;
pe2=pe;
if(excWord&UCASE_EXC_CONDITIONAL_FOLD) {
/* use hardcoded conditions and mappings */
if((options&_FOLD_CASE_OPTIONS_MASK)==U_FOLD_CASE_DEFAULT) {
/* default mappings */
if(c==0x49) {
/* 0049; C; 0069; # LATIN CAPITAL LETTER I */
return 0x69;
} else if(c==0x130) {
/* 0130; F; 0069 0307; # LATIN CAPITAL LETTER I WITH DOT ABOVE */
*pString=iDot;
return 2;
}
} else {
/* Turkic mappings */
if(c==0x49) {
/* 0049; T; 0131; # LATIN CAPITAL LETTER I */
return 0x131;
} else if(c==0x130) {
/* 0130; T; 0069; # LATIN CAPITAL LETTER I WITH DOT ABOVE */
return 0x69;
}
}
} else if(HAS_SLOT(excWord, UCASE_EXC_FULL_MAPPINGS)) {
GET_SLOT_VALUE(excWord, UCASE_EXC_FULL_MAPPINGS, pe, full);
/* start of full case mapping strings */
++pe;
/* skip the lowercase result string */
pe+=full&UCASE_FULL_LOWER;
full=(full>>4)&0xf;
if(full!=0) {
/* set the output pointer to the result string */
*pString=reinterpret_cast<const UChar *>(pe);
/* return the string length */
return full;
}
}
if(HAS_SLOT(excWord, UCASE_EXC_FOLD)) {
idx=UCASE_EXC_FOLD;
} else if(HAS_SLOT(excWord, UCASE_EXC_LOWER)) {
idx=UCASE_EXC_LOWER;
} else {
return ~c;
}
GET_SLOT_VALUE(excWord, idx, pe2, result);
}
return (result==c) ? ~result : result;
}
/* case mapping properties API ---------------------------------------------- */
/* public API (see uchar.h) */
U_CAPI UBool U_EXPORT2
u_isULowercase(UChar32 c) {
return (UBool)(UCASE_LOWER==ucase_getType(c));
}
U_CAPI UBool U_EXPORT2
u_isUUppercase(UChar32 c) {
return (UBool)(UCASE_UPPER==ucase_getType(c));
}
/* Transforms the Unicode character to its lower case equivalent.*/
U_CAPI UChar32 U_EXPORT2
u_tolower(UChar32 c) {
return ucase_tolower(c);
}
/* Transforms the Unicode character to its upper case equivalent.*/
U_CAPI UChar32 U_EXPORT2
u_toupper(UChar32 c) {
return ucase_toupper(c);
}
/* Transforms the Unicode character to its title case equivalent.*/
U_CAPI UChar32 U_EXPORT2
u_totitle(UChar32 c) {
return ucase_totitle(c);
}
/* return the simple case folding mapping for c */
U_CAPI UChar32 U_EXPORT2
u_foldCase(UChar32 c, uint32_t options) {
return ucase_fold(c, options);
}
U_CFUNC int32_t U_EXPORT2
ucase_hasBinaryProperty(UChar32 c, UProperty which) {
/* case mapping properties */
const UChar *resultString;
switch(which) {
case UCHAR_LOWERCASE:
return (UBool)(UCASE_LOWER==ucase_getType(c));
case UCHAR_UPPERCASE:
return (UBool)(UCASE_UPPER==ucase_getType(c));
case UCHAR_SOFT_DOTTED:
return ucase_isSoftDotted(c);
case UCHAR_CASE_SENSITIVE:
return ucase_isCaseSensitive(c);
case UCHAR_CASED:
return (UBool)(UCASE_NONE!=ucase_getType(c));
case UCHAR_CASE_IGNORABLE:
return (UBool)(ucase_getTypeOrIgnorable(c)>>2);
/*
* Note: The following Changes_When_Xyz are defined as testing whether
* the NFD form of the input changes when Xyz-case-mapped.
* However, this simpler implementation of these properties,
* ignoring NFD, passes the tests.
* The implementation needs to be changed if the tests start failing.
* When that happens, optimizations should be used to work with the
* per-single-code point ucase_toFullXyz() functions unless
* the NFD form has more than one code point,
* and the property starts set needs to be the union of the
* start sets for normalization and case mappings.
*/
case UCHAR_CHANGES_WHEN_LOWERCASED:
return (UBool)(ucase_toFullLower(c, NULL, NULL, &resultString, UCASE_LOC_ROOT)>=0);
case UCHAR_CHANGES_WHEN_UPPERCASED:
return (UBool)(ucase_toFullUpper(c, NULL, NULL, &resultString, UCASE_LOC_ROOT)>=0);
case UCHAR_CHANGES_WHEN_TITLECASED:
return (UBool)(ucase_toFullTitle(c, NULL, NULL, &resultString, UCASE_LOC_ROOT)>=0);
/* case UCHAR_CHANGES_WHEN_CASEFOLDED: -- in uprops.c */
case UCHAR_CHANGES_WHEN_CASEMAPPED:
return (UBool)(
ucase_toFullLower(c, NULL, NULL, &resultString, UCASE_LOC_ROOT)>=0 ||
ucase_toFullUpper(c, NULL, NULL, &resultString, UCASE_LOC_ROOT)>=0 ||
ucase_toFullTitle(c, NULL, NULL, &resultString, UCASE_LOC_ROOT)>=0);
default:
return FALSE;
}
}
Reference in New Issue View Git Blame Copy Permalink