1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/icu/source/common/ruleiter.cpp Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,160 @@
1.4 +/*
1.5 +**********************************************************************
1.6 +* Copyright (c) 2003-2011, International Business Machines
1.7 +* Corporation and others. All Rights Reserved.
1.8 +**********************************************************************
1.9 +* Author: Alan Liu
1.10 +* Created: September 24 2003
1.11 +* Since: ICU 2.8
1.12 +**********************************************************************
1.13 +*/
1.14 +#include "ruleiter.h"
1.15 +#include "unicode/parsepos.h"
1.16 +#include "unicode/symtable.h"
1.17 +#include "unicode/unistr.h"
1.18 +#include "unicode/utf16.h"
1.19 +#include "patternprops.h"
1.20 +
1.21 +/* \U87654321 or \ud800\udc00 */
1.22 +#define MAX_U_NOTATION_LEN 12
1.23 +
1.24 +U_NAMESPACE_BEGIN
1.25 +
1.26 +RuleCharacterIterator::RuleCharacterIterator(const UnicodeString& theText, const SymbolTable* theSym,
1.27 + ParsePosition& thePos) :
1.28 + text(theText),
1.29 + pos(thePos),
1.30 + sym(theSym),
1.31 + buf(0),
1.32 + bufPos(0)
1.33 +{}
1.34 +
1.35 +UBool RuleCharacterIterator::atEnd() const {
1.36 + return buf == 0 && pos.getIndex() == text.length();
1.37 +}
1.38 +
1.39 +UChar32 RuleCharacterIterator::next(int32_t options, UBool& isEscaped, UErrorCode& ec) {
1.40 + if (U_FAILURE(ec)) return DONE;
1.41 +
1.42 + UChar32 c = DONE;
1.43 + isEscaped = FALSE;
1.44 +
1.45 + for (;;) {
1.46 + c = _current();
1.47 + _advance(U16_LENGTH(c));
1.48 +
1.49 + if (c == SymbolTable::SYMBOL_REF && buf == 0 &&
1.50 + (options & PARSE_VARIABLES) != 0 && sym != 0) {
1.51 + UnicodeString name = sym->parseReference(text, pos, text.length());
1.52 + // If name is empty there was an isolated SYMBOL_REF;
1.53 + // return it. Caller must be prepared for this.
1.54 + if (name.length() == 0) {
1.55 + break;
1.56 + }
1.57 + bufPos = 0;
1.58 + buf = sym->lookup(name);
1.59 + if (buf == 0) {
1.60 + ec = U_UNDEFINED_VARIABLE;
1.61 + return DONE;
1.62 + }
1.63 + // Handle empty variable value
1.64 + if (buf->length() == 0) {
1.65 + buf = 0;
1.66 + }
1.67 + continue;
1.68 + }
1.69 +
1.70 + if ((options & SKIP_WHITESPACE) != 0 && PatternProps::isWhiteSpace(c)) {
1.71 + continue;
1.72 + }
1.73 +
1.74 + if (c == 0x5C /*'\\'*/ && (options & PARSE_ESCAPES) != 0) {
1.75 + UnicodeString tempEscape;
1.76 + int32_t offset = 0;
1.77 + c = lookahead(tempEscape, MAX_U_NOTATION_LEN).unescapeAt(offset);
1.78 + jumpahead(offset);
1.79 + isEscaped = TRUE;
1.80 + if (c < 0) {
1.81 + ec = U_MALFORMED_UNICODE_ESCAPE;
1.82 + return DONE;
1.83 + }
1.84 + }
1.85 +
1.86 + break;
1.87 + }
1.88 +
1.89 + return c;
1.90 +}
1.91 +
1.92 +void RuleCharacterIterator::getPos(RuleCharacterIterator::Pos& p) const {
1.93 + p.buf = buf;
1.94 + p.pos = pos.getIndex();
1.95 + p.bufPos = bufPos;
1.96 +}
1.97 +
1.98 +void RuleCharacterIterator::setPos(const RuleCharacterIterator::Pos& p) {
1.99 + buf = p.buf;
1.100 + pos.setIndex(p.pos);
1.101 + bufPos = p.bufPos;
1.102 +}
1.103 +
1.104 +void RuleCharacterIterator::skipIgnored(int32_t options) {
1.105 + if ((options & SKIP_WHITESPACE) != 0) {
1.106 + for (;;) {
1.107 + UChar32 a = _current();
1.108 + if (!PatternProps::isWhiteSpace(a)) break;
1.109 + _advance(U16_LENGTH(a));
1.110 + }
1.111 + }
1.112 +}
1.113 +
1.114 +UnicodeString& RuleCharacterIterator::lookahead(UnicodeString& result, int32_t maxLookAhead) const {
1.115 + if (maxLookAhead < 0) {
1.116 + maxLookAhead = 0x7FFFFFFF;
1.117 + }
1.118 + if (buf != 0) {
1.119 + buf->extract(bufPos, maxLookAhead, result);
1.120 + } else {
1.121 + text.extract(pos.getIndex(), maxLookAhead, result);
1.122 + }
1.123 + return result;
1.124 +}
1.125 +
1.126 +void RuleCharacterIterator::jumpahead(int32_t count) {
1.127 + _advance(count);
1.128 +}
1.129 +
1.130 +/*
1.131 +UnicodeString& RuleCharacterIterator::toString(UnicodeString& result) const {
1.132 + int32_t b = pos.getIndex();
1.133 + text.extract(0, b, result);
1.134 + return result.append((UChar) 0x7C).append(text, b, 0x7FFFFFFF); // Insert '|' at index
1.135 +}
1.136 +*/
1.137 +
1.138 +UChar32 RuleCharacterIterator::_current() const {
1.139 + if (buf != 0) {
1.140 + return buf->char32At(bufPos);
1.141 + } else {
1.142 + int i = pos.getIndex();
1.143 + return (i < text.length()) ? text.char32At(i) : (UChar32)DONE;
1.144 + }
1.145 +}
1.146 +
1.147 +void RuleCharacterIterator::_advance(int32_t count) {
1.148 + if (buf != 0) {
1.149 + bufPos += count;
1.150 + if (bufPos == buf->length()) {
1.151 + buf = 0;
1.152 + }
1.153 + } else {
1.154 + pos.setIndex(pos.getIndex() + count);
1.155 + if (pos.getIndex() > text.length()) {
1.156 + pos.setIndex(text.length());
1.157 + }
1.158 + }
1.159 +}
1.160 +
1.161 +U_NAMESPACE_END
1.162 +
1.163 +//eof
