1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/icu/source/i18n/name2uni.cpp Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,256 @@
1.4 +/*
1.5 +**********************************************************************
1.6 +* Copyright (C) 2001-2011, International Business Machines
1.7 +* Corporation and others. All Rights Reserved.
1.8 +**********************************************************************
1.9 +* Date Name Description
1.10 +* 06/07/01 aliu Creation.
1.11 +**********************************************************************
1.12 +*/
1.13 +
1.14 +#include "unicode/utypes.h"
1.15 +
1.16 +#if !UCONFIG_NO_TRANSLITERATION
1.17 +
1.18 +#include "unicode/unifilt.h"
1.19 +#include "unicode/uchar.h"
1.20 +#include "unicode/uniset.h"
1.21 +#include "unicode/utf16.h"
1.22 +#include "cmemory.h"
1.23 +#include "name2uni.h"
1.24 +#include "patternprops.h"
1.25 +#include "uprops.h"
1.26 +#include "uinvchar.h"
1.27 +#include "util.h"
1.28 +
1.29 +U_NAMESPACE_BEGIN
1.30 +
1.31 +UOBJECT_DEFINE_RTTI_IMPLEMENTATION(NameUnicodeTransliterator)
1.32 +
1.33 +static const UChar OPEN[] = {92,78,126,123,126,0}; // "\N~{~"
1.34 +static const UChar OPEN_DELIM = 92; // '\\' first char of OPEN
1.35 +static const UChar CLOSE_DELIM = 125; // '}'
1.36 +static const UChar SPACE = 32; // ' '
1.37 +
1.38 +U_CDECL_BEGIN
1.39 +
1.40 +// USetAdder implementation
1.41 +// Does not use uset.h to reduce code dependencies
1.42 +static void U_CALLCONV
1.43 +_set_add(USet *set, UChar32 c) {
1.44 + uset_add(set, c);
1.45 +}
1.46 +
1.47 +// These functions aren't used.
1.48 +/*static void U_CALLCONV
1.49 +_set_addRange(USet *set, UChar32 start, UChar32 end) {
1.50 + ((UnicodeSet *)set)->add(start, end);
1.51 +}
1.52 +
1.53 +static void U_CALLCONV
1.54 +_set_addString(USet *set, const UChar *str, int32_t length) {
1.55 + ((UnicodeSet *)set)->add(UnicodeString((UBool)(length<0), str, length));
1.56 +}*/
1.57 +
1.58 +U_CDECL_END
1.59 +
1.60 +/**
1.61 + * Constructs a transliterator with the default delimiters '{' and
1.62 + * '}'.
1.63 + */
1.64 +NameUnicodeTransliterator::NameUnicodeTransliterator(UnicodeFilter* adoptedFilter) :
1.65 + Transliterator(UNICODE_STRING("Name-Any", 8), adoptedFilter) {
1.66 +
1.67 + UnicodeSet *legalPtr = &legal;
1.68 + // Get the legal character set
1.69 + USetAdder sa = {
1.70 + (USet *)legalPtr, // USet* == UnicodeSet*
1.71 + _set_add,
1.72 + NULL, // Don't need _set_addRange
1.73 + NULL, // Don't need _set_addString
1.74 + NULL, // Don't need remove()
1.75 + NULL
1.76 + };
1.77 + uprv_getCharNameCharacters(&sa);
1.78 +}
1.79 +
1.80 +/**
1.81 + * Destructor.
1.82 + */
1.83 +NameUnicodeTransliterator::~NameUnicodeTransliterator() {}
1.84 +
1.85 +/**
1.86 + * Copy constructor.
1.87 + */
1.88 +NameUnicodeTransliterator::NameUnicodeTransliterator(const NameUnicodeTransliterator& o) :
1.89 + Transliterator(o), legal(o.legal) {}
1.90 +
1.91 +/**
1.92 + * Assignment operator.
1.93 + */
1.94 +/*NameUnicodeTransliterator& NameUnicodeTransliterator::operator=(
1.95 + const NameUnicodeTransliterator& o) {
1.96 + Transliterator::operator=(o);
1.97 + // not necessary: the legal sets should all be the same -- legal=o.legal;
1.98 + return *this;
1.99 +}*/
1.100 +
1.101 +/**
1.102 + * Transliterator API.
1.103 + */
1.104 +Transliterator* NameUnicodeTransliterator::clone(void) const {
1.105 + return new NameUnicodeTransliterator(*this);
1.106 +}
1.107 +
1.108 +/**
1.109 + * Implements {@link Transliterator#handleTransliterate}.
1.110 + */
1.111 +void NameUnicodeTransliterator::handleTransliterate(Replaceable& text, UTransPosition& offsets,
1.112 + UBool isIncremental) const {
1.113 + // The failure mode, here and below, is to behave like Any-Null,
1.114 + // if either there is no name data (max len == 0) or there is no
1.115 + // memory (malloc() => NULL).
1.116 +
1.117 + int32_t maxLen = uprv_getMaxCharNameLength();
1.118 + if (maxLen == 0) {
1.119 + offsets.start = offsets.limit;
1.120 + return;
1.121 + }
1.122 +
1.123 + // Accomodate the longest possible name
1.124 + ++maxLen; // allow for temporary trailing space
1.125 + char* cbuf = (char*) uprv_malloc(maxLen);
1.126 + if (cbuf == NULL) {
1.127 + offsets.start = offsets.limit;
1.128 + return;
1.129 + }
1.130 +
1.131 + UnicodeString openPat(TRUE, OPEN, -1);
1.132 + UnicodeString str, name;
1.133 +
1.134 + int32_t cursor = offsets.start;
1.135 + int32_t limit = offsets.limit;
1.136 +
1.137 + // Modes:
1.138 + // 0 - looking for open delimiter
1.139 + // 1 - after open delimiter
1.140 + int32_t mode = 0;
1.141 + int32_t openPos = -1; // open delim candidate pos
1.142 +
1.143 + UChar32 c;
1.144 + while (cursor < limit) {
1.145 + c = text.char32At(cursor);
1.146 +
1.147 + switch (mode) {
1.148 + case 0: // looking for open delimiter
1.149 + if (c == OPEN_DELIM) { // quick check first
1.150 + openPos = cursor;
1.151 + int32_t i =
1.152 + ICU_Utility::parsePattern(openPat, text, cursor, limit);
1.153 + if (i >= 0 && i < limit) {
1.154 + mode = 1;
1.155 + name.truncate(0);
1.156 + cursor = i;
1.157 + continue; // *** reprocess char32At(cursor)
1.158 + }
1.159 + }
1.160 + break;
1.161 +
1.162 + case 1: // after open delimiter
1.163 + // Look for legal chars. If \s+ is found, convert it
1.164 + // to a single space. If closeDelimiter is found, exit
1.165 + // the loop. If any other character is found, exit the
1.166 + // loop. If the limit is reached, exit the loop.
1.167 +
1.168 + // Convert \s+ => SPACE. This assumes there are no
1.169 + // runs of >1 space characters in names.
1.170 + if (PatternProps::isWhiteSpace(c)) {
1.171 + // Ignore leading whitespace
1.172 + if (name.length() > 0 &&
1.173 + name.charAt(name.length()-1) != SPACE) {
1.174 + name.append(SPACE);
1.175 + // If we are too long then abort. maxLen includes
1.176 + // temporary trailing space, so use '>'.
1.177 + if (name.length() > maxLen) {
1.178 + mode = 0;
1.179 + }
1.180 + }
1.181 + break;
1.182 + }
1.183 +
1.184 + if (c == CLOSE_DELIM) {
1.185 + int32_t len = name.length();
1.186 +
1.187 + // Delete trailing space, if any
1.188 + if (len > 0 &&
1.189 + name.charAt(len-1) == SPACE) {
1.190 + --len;
1.191 + }
1.192 +
1.193 + if (uprv_isInvariantUString(name.getBuffer(), len)) {
1.194 + name.extract(0, len, cbuf, maxLen, US_INV);
1.195 +
1.196 + UErrorCode status = U_ZERO_ERROR;
1.197 + c = u_charFromName(U_EXTENDED_CHAR_NAME, cbuf, &status);
1.198 + if (U_SUCCESS(status)) {
1.199 + // Lookup succeeded
1.200 +
1.201 + // assert(U16_LENGTH(CLOSE_DELIM) == 1);
1.202 + cursor++; // advance over CLOSE_DELIM
1.203 +
1.204 + str.truncate(0);
1.205 + str.append(c);
1.206 + text.handleReplaceBetween(openPos, cursor, str);
1.207 +
1.208 + // Adjust indices for the change in the length of
1.209 + // the string. Do not assume that str.length() ==
1.210 + // 1, in case of surrogates.
1.211 + int32_t delta = cursor - openPos - str.length();
1.212 + cursor -= delta;
1.213 + limit -= delta;
1.214 + // assert(cursor == openPos + str.length());
1.215 + }
1.216 + }
1.217 + // If the lookup failed, we leave things as-is and
1.218 + // still switch to mode 0 and continue.
1.219 + mode = 0;
1.220 + openPos = -1; // close off candidate
1.221 + continue; // *** reprocess char32At(cursor)
1.222 + }
1.223 +
1.224 + // Check if c is a legal char. We assume here that
1.225 + // legal.contains(OPEN_DELIM) is FALSE, so when we abort a
1.226 + // name, we don't have to go back to openPos+1.
1.227 + if (legal.contains(c)) {
1.228 + name.append(c);
1.229 + // If we go past the longest possible name then abort.
1.230 + // maxLen includes temporary trailing space, so use '>='.
1.231 + if (name.length() >= maxLen) {
1.232 + mode = 0;
1.233 + }
1.234 + }
1.235 +
1.236 + // Invalid character
1.237 + else {
1.238 + --cursor; // Backup and reprocess this character
1.239 + mode = 0;
1.240 + }
1.241 +
1.242 + break;
1.243 + }
1.244 +
1.245 + cursor += U16_LENGTH(c);
1.246 + }
1.247 +
1.248 + offsets.contextLimit += limit - offsets.limit;
1.249 + offsets.limit = limit;
1.250 + // In incremental mode, only advance the cursor up to the last
1.251 + // open delimiter candidate.
1.252 + offsets.start = (isIncremental && openPos >= 0) ? openPos : cursor;
1.253 +
1.254 + uprv_free(cbuf);
1.255 +}
1.256 +
1.257 +U_NAMESPACE_END
1.258 +
1.259 +#endif /* #if !UCONFIG_NO_TRANSLITERATION */
