1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/icu/source/common/util_props.cpp Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,214 @@
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 +* 11/19/2001 aliu Creation.
1.11 +**********************************************************************
1.12 +*/
1.13 +
1.14 +#include "unicode/uchar.h"
1.15 +#include "unicode/utf16.h"
1.16 +#include "patternprops.h"
1.17 +#include "util.h"
1.18 +
1.19 +U_NAMESPACE_BEGIN
1.20 +
1.21 +/**
1.22 + * Parse an integer at pos, either of the form \d+ or of the form
1.23 + * 0x[0-9A-Fa-f]+ or 0[0-7]+, that is, in standard decimal, hex,
1.24 + * or octal format.
1.25 + * @param pos INPUT-OUTPUT parameter. On input, the first
1.26 + * character to parse. On output, the character after the last
1.27 + * parsed character.
1.28 + */
1.29 +int32_t ICU_Utility::parseInteger(const UnicodeString& rule, int32_t& pos, int32_t limit) {
1.30 + int32_t count = 0;
1.31 + int32_t value = 0;
1.32 + int32_t p = pos;
1.33 + int8_t radix = 10;
1.34 +
1.35 + if (p < limit && rule.charAt(p) == 48 /*0*/) {
1.36 + if (p+1 < limit && (rule.charAt(p+1) == 0x78 /*x*/ || rule.charAt(p+1) == 0x58 /*X*/)) {
1.37 + p += 2;
1.38 + radix = 16;
1.39 + }
1.40 + else {
1.41 + p++;
1.42 + count = 1;
1.43 + radix = 8;
1.44 + }
1.45 + }
1.46 +
1.47 + while (p < limit) {
1.48 + int32_t d = u_digit(rule.charAt(p++), radix);
1.49 + if (d < 0) {
1.50 + --p;
1.51 + break;
1.52 + }
1.53 + ++count;
1.54 + int32_t v = (value * radix) + d;
1.55 + if (v <= value) {
1.56 + // If there are too many input digits, at some point
1.57 + // the value will go negative, e.g., if we have seen
1.58 + // "0x8000000" already and there is another '0', when
1.59 + // we parse the next 0 the value will go negative.
1.60 + return 0;
1.61 + }
1.62 + value = v;
1.63 + }
1.64 + if (count > 0) {
1.65 + pos = p;
1.66 + }
1.67 + return value;
1.68 +}
1.69 +
1.70 +/**
1.71 + * Parse a pattern string starting at offset pos. Keywords are
1.72 + * matched case-insensitively. Spaces may be skipped and may be
1.73 + * optional or required. Integer values may be parsed, and if
1.74 + * they are, they will be returned in the given array. If
1.75 + * successful, the offset of the next non-space character is
1.76 + * returned. On failure, -1 is returned.
1.77 + * @param pattern must only contain lowercase characters, which
1.78 + * will match their uppercase equivalents as well. A space
1.79 + * character matches one or more required spaces. A '~' character
1.80 + * matches zero or more optional spaces. A '#' character matches
1.81 + * an integer and stores it in parsedInts, which the caller must
1.82 + * ensure has enough capacity.
1.83 + * @param parsedInts array to receive parsed integers. Caller
1.84 + * must ensure that parsedInts.length is >= the number of '#'
1.85 + * signs in 'pattern'.
1.86 + * @return the position after the last character parsed, or -1 if
1.87 + * the parse failed
1.88 + */
1.89 +int32_t ICU_Utility::parsePattern(const UnicodeString& rule, int32_t pos, int32_t limit,
1.90 + const UnicodeString& pattern, int32_t* parsedInts) {
1.91 + // TODO Update this to handle surrogates
1.92 + int32_t p;
1.93 + int32_t intCount = 0; // number of integers parsed
1.94 + for (int32_t i=0; i<pattern.length(); ++i) {
1.95 + UChar cpat = pattern.charAt(i);
1.96 + UChar c;
1.97 + switch (cpat) {
1.98 + case 32 /*' '*/:
1.99 + if (pos >= limit) {
1.100 + return -1;
1.101 + }
1.102 + c = rule.charAt(pos++);
1.103 + if (!PatternProps::isWhiteSpace(c)) {
1.104 + return -1;
1.105 + }
1.106 + // FALL THROUGH to skipWhitespace
1.107 + case 126 /*'~'*/:
1.108 + pos = skipWhitespace(rule, pos);
1.109 + break;
1.110 + case 35 /*'#'*/:
1.111 + p = pos;
1.112 + parsedInts[intCount++] = parseInteger(rule, p, limit);
1.113 + if (p == pos) {
1.114 + // Syntax error; failed to parse integer
1.115 + return -1;
1.116 + }
1.117 + pos = p;
1.118 + break;
1.119 + default:
1.120 + if (pos >= limit) {
1.121 + return -1;
1.122 + }
1.123 + c = (UChar) u_tolower(rule.charAt(pos++));
1.124 + if (c != cpat) {
1.125 + return -1;
1.126 + }
1.127 + break;
1.128 + }
1.129 + }
1.130 + return pos;
1.131 +}
1.132 +
1.133 +/**
1.134 + * Parse a Unicode identifier from the given string at the given
1.135 + * position. Return the identifier, or an empty string if there
1.136 + * is no identifier.
1.137 + * @param str the string to parse
1.138 + * @param pos INPUT-OUPUT parameter. On INPUT, pos is the
1.139 + * first character to examine. It must be less than str.length(),
1.140 + * and it must not point to a whitespace character. That is, must
1.141 + * have pos < str.length(). On
1.142 + * OUTPUT, the position after the last parsed character.
1.143 + * @return the Unicode identifier, or an empty string if there is
1.144 + * no valid identifier at pos.
1.145 + */
1.146 +UnicodeString ICU_Utility::parseUnicodeIdentifier(const UnicodeString& str, int32_t& pos) {
1.147 + // assert(pos < str.length());
1.148 + UnicodeString buf;
1.149 + int p = pos;
1.150 + while (p < str.length()) {
1.151 + UChar32 ch = str.char32At(p);
1.152 + if (buf.length() == 0) {
1.153 + if (u_isIDStart(ch)) {
1.154 + buf.append(ch);
1.155 + } else {
1.156 + buf.truncate(0);
1.157 + return buf;
1.158 + }
1.159 + } else {
1.160 + if (u_isIDPart(ch)) {
1.161 + buf.append(ch);
1.162 + } else {
1.163 + break;
1.164 + }
1.165 + }
1.166 + p += U16_LENGTH(ch);
1.167 + }
1.168 + pos = p;
1.169 + return buf;
1.170 +}
1.171 +
1.172 +/**
1.173 + * Parse an unsigned 31-bit integer at the given offset. Use
1.174 + * UCharacter.digit() to parse individual characters into digits.
1.175 + * @param text the text to be parsed
1.176 + * @param pos INPUT-OUTPUT parameter. On entry, pos[0] is the
1.177 + * offset within text at which to start parsing; it should point
1.178 + * to a valid digit. On exit, pos[0] is the offset after the last
1.179 + * parsed character. If the parse failed, it will be unchanged on
1.180 + * exit. Must be >= 0 on entry.
1.181 + * @param radix the radix in which to parse; must be >= 2 and <=
1.182 + * 36.
1.183 + * @return a non-negative parsed number, or -1 upon parse failure.
1.184 + * Parse fails if there are no digits, that is, if pos[0] does not
1.185 + * point to a valid digit on entry, or if the number to be parsed
1.186 + * does not fit into a 31-bit unsigned integer.
1.187 + */
1.188 +int32_t ICU_Utility::parseNumber(const UnicodeString& text,
1.189 + int32_t& pos, int8_t radix) {
1.190 + // assert(pos[0] >= 0);
1.191 + // assert(radix >= 2);
1.192 + // assert(radix <= 36);
1.193 + int32_t n = 0;
1.194 + int32_t p = pos;
1.195 + while (p < text.length()) {
1.196 + UChar32 ch = text.char32At(p);
1.197 + int32_t d = u_digit(ch, radix);
1.198 + if (d < 0) {
1.199 + break;
1.200 + }
1.201 + n = radix*n + d;
1.202 + // ASSUME that when a 32-bit integer overflows it becomes
1.203 + // negative. E.g., 214748364 * 10 + 8 => negative value.
1.204 + if (n < 0) {
1.205 + return -1;
1.206 + }
1.207 + ++p;
1.208 + }
1.209 + if (p == pos) {
1.210 + return -1;
1.211 + }
1.212 + pos = p;
1.213 + return n;
1.214 +}
1.215 +
1.216 +U_NAMESPACE_END
1.217 +
