1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/icu/source/common/util.cpp Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,409 @@
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/unimatch.h"
1.15 +#include "unicode/utf16.h"
1.16 +#include "patternprops.h"
1.17 +#include "util.h"
1.18 +
1.19 +// Define UChar constants using hex for EBCDIC compatibility
1.20 +
1.21 +static const UChar BACKSLASH = 0x005C; /*\*/
1.22 +static const UChar UPPER_U = 0x0055; /*U*/
1.23 +static const UChar LOWER_U = 0x0075; /*u*/
1.24 +static const UChar APOSTROPHE = 0x0027; // '\''
1.25 +static const UChar SPACE = 0x0020; // ' '
1.26 +
1.27 +// "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ"
1.28 +static const UChar DIGITS[] = {
1.29 + 48,49,50,51,52,53,54,55,56,57,
1.30 + 65,66,67,68,69,70,71,72,73,74,
1.31 + 75,76,77,78,79,80,81,82,83,84,
1.32 + 85,86,87,88,89,90
1.33 +};
1.34 +
1.35 +U_NAMESPACE_BEGIN
1.36 +
1.37 +UnicodeString& ICU_Utility::appendNumber(UnicodeString& result, int32_t n,
1.38 + int32_t radix, int32_t minDigits) {
1.39 + if (radix < 2 || radix > 36) {
1.40 + // Bogus radix
1.41 + return result.append((UChar)63/*?*/);
1.42 + }
1.43 + // Handle negatives
1.44 + if (n < 0) {
1.45 + n = -n;
1.46 + result.append((UChar)45/*-*/);
1.47 + }
1.48 + // First determine the number of digits
1.49 + int32_t nn = n;
1.50 + int32_t r = 1;
1.51 + while (nn >= radix) {
1.52 + nn /= radix;
1.53 + r *= radix;
1.54 + --minDigits;
1.55 + }
1.56 + // Now generate the digits
1.57 + while (--minDigits > 0) {
1.58 + result.append(DIGITS[0]);
1.59 + }
1.60 + while (r > 0) {
1.61 + int32_t digit = n / r;
1.62 + result.append(DIGITS[digit]);
1.63 + n -= digit * r;
1.64 + r /= radix;
1.65 + }
1.66 + return result;
1.67 +}
1.68 +
1.69 +/**
1.70 + * Return true if the character is NOT printable ASCII.
1.71 + */
1.72 +UBool ICU_Utility::isUnprintable(UChar32 c) {
1.73 + return !(c >= 0x20 && c <= 0x7E);
1.74 +}
1.75 +
1.76 +/**
1.77 + * Escape unprintable characters using \uxxxx notation for U+0000 to
1.78 + * U+FFFF and \Uxxxxxxxx for U+10000 and above. If the character is
1.79 + * printable ASCII, then do nothing and return FALSE. Otherwise,
1.80 + * append the escaped notation and return TRUE.
1.81 + */
1.82 +UBool ICU_Utility::escapeUnprintable(UnicodeString& result, UChar32 c) {
1.83 + if (isUnprintable(c)) {
1.84 + result.append(BACKSLASH);
1.85 + if (c & ~0xFFFF) {
1.86 + result.append(UPPER_U);
1.87 + result.append(DIGITS[0xF&(c>>28)]);
1.88 + result.append(DIGITS[0xF&(c>>24)]);
1.89 + result.append(DIGITS[0xF&(c>>20)]);
1.90 + result.append(DIGITS[0xF&(c>>16)]);
1.91 + } else {
1.92 + result.append(LOWER_U);
1.93 + }
1.94 + result.append(DIGITS[0xF&(c>>12)]);
1.95 + result.append(DIGITS[0xF&(c>>8)]);
1.96 + result.append(DIGITS[0xF&(c>>4)]);
1.97 + result.append(DIGITS[0xF&c]);
1.98 + return TRUE;
1.99 + }
1.100 + return FALSE;
1.101 +}
1.102 +
1.103 +/**
1.104 + * Returns the index of a character, ignoring quoted text.
1.105 + * For example, in the string "abc'hide'h", the 'h' in "hide" will not be
1.106 + * found by a search for 'h'.
1.107 + */
1.108 +// FOR FUTURE USE. DISABLE FOR NOW for coverage reasons.
1.109 +/*
1.110 +int32_t ICU_Utility::quotedIndexOf(const UnicodeString& text,
1.111 + int32_t start, int32_t limit,
1.112 + UChar charToFind) {
1.113 + for (int32_t i=start; i<limit; ++i) {
1.114 + UChar c = text.charAt(i);
1.115 + if (c == BACKSLASH) {
1.116 + ++i;
1.117 + } else if (c == APOSTROPHE) {
1.118 + while (++i < limit
1.119 + && text.charAt(i) != APOSTROPHE) {}
1.120 + } else if (c == charToFind) {
1.121 + return i;
1.122 + }
1.123 + }
1.124 + return -1;
1.125 +}
1.126 +*/
1.127 +
1.128 +/**
1.129 + * Skip over a sequence of zero or more white space characters at pos.
1.130 + * @param advance if true, advance pos to the first non-white-space
1.131 + * character at or after pos, or str.length(), if there is none.
1.132 + * Otherwise leave pos unchanged.
1.133 + * @return the index of the first non-white-space character at or
1.134 + * after pos, or str.length(), if there is none.
1.135 + */
1.136 +int32_t ICU_Utility::skipWhitespace(const UnicodeString& str, int32_t& pos,
1.137 + UBool advance) {
1.138 + int32_t p = pos;
1.139 + const UChar* s = str.getBuffer();
1.140 + p = (int32_t)(PatternProps::skipWhiteSpace(s + p, str.length() - p) - s);
1.141 + if (advance) {
1.142 + pos = p;
1.143 + }
1.144 + return p;
1.145 +}
1.146 +
1.147 +/**
1.148 + * Skip over Pattern_White_Space in a Replaceable.
1.149 + * Skipping may be done in the forward or
1.150 + * reverse direction. In either case, the leftmost index will be
1.151 + * inclusive, and the rightmost index will be exclusive. That is,
1.152 + * given a range defined as [start, limit), the call
1.153 + * skipWhitespace(text, start, limit) will advance start past leading
1.154 + * whitespace, whereas the call skipWhitespace(text, limit, start),
1.155 + * will back up limit past trailing whitespace.
1.156 + * @param text the text to be analyzed
1.157 + * @param pos either the start or limit of a range of 'text', to skip
1.158 + * leading or trailing whitespace, respectively
1.159 + * @param stop either the limit or start of a range of 'text', to skip
1.160 + * leading or trailing whitespace, respectively
1.161 + * @return the new start or limit, depending on what was passed in to
1.162 + * 'pos'
1.163 + */
1.164 +//?FOR FUTURE USE. DISABLE FOR NOW for coverage reasons.
1.165 +//?int32_t ICU_Utility::skipWhitespace(const Replaceable& text,
1.166 +//? int32_t pos, int32_t stop) {
1.167 +//? UChar32 c;
1.168 +//? UBool isForward = (stop >= pos);
1.169 +//?
1.170 +//? if (!isForward) {
1.171 +//? --pos; // pos is a limit, so back up by one
1.172 +//? }
1.173 +//?
1.174 +//? while (pos != stop &&
1.175 +//? PatternProps::isWhiteSpace(c = text.char32At(pos))) {
1.176 +//? if (isForward) {
1.177 +//? pos += U16_LENGTH(c);
1.178 +//? } else {
1.179 +//? pos -= U16_LENGTH(c);
1.180 +//? }
1.181 +//? }
1.182 +//?
1.183 +//? if (!isForward) {
1.184 +//? ++pos; // make pos back into a limit
1.185 +//? }
1.186 +//?
1.187 +//? return pos;
1.188 +//?}
1.189 +
1.190 +/**
1.191 + * Parse a single non-whitespace character 'ch', optionally
1.192 + * preceded by whitespace.
1.193 + * @param id the string to be parsed
1.194 + * @param pos INPUT-OUTPUT parameter. On input, pos[0] is the
1.195 + * offset of the first character to be parsed. On output, pos[0]
1.196 + * is the index after the last parsed character. If the parse
1.197 + * fails, pos[0] will be unchanged.
1.198 + * @param ch the non-whitespace character to be parsed.
1.199 + * @return true if 'ch' is seen preceded by zero or more
1.200 + * whitespace characters.
1.201 + */
1.202 +UBool ICU_Utility::parseChar(const UnicodeString& id, int32_t& pos, UChar ch) {
1.203 + int32_t start = pos;
1.204 + skipWhitespace(id, pos, TRUE);
1.205 + if (pos == id.length() ||
1.206 + id.charAt(pos) != ch) {
1.207 + pos = start;
1.208 + return FALSE;
1.209 + }
1.210 + ++pos;
1.211 + return TRUE;
1.212 +}
1.213 +
1.214 +/**
1.215 + * Parse a pattern string within the given Replaceable and a parsing
1.216 + * pattern. Characters are matched literally and case-sensitively
1.217 + * except for the following special characters:
1.218 + *
1.219 + * ~ zero or more Pattern_White_Space chars
1.220 + *
1.221 + * If end of pattern is reached with all matches along the way,
1.222 + * pos is advanced to the first unparsed index and returned.
1.223 + * Otherwise -1 is returned.
1.224 + * @param pat pattern that controls parsing
1.225 + * @param text text to be parsed, starting at index
1.226 + * @param index offset to first character to parse
1.227 + * @param limit offset after last character to parse
1.228 + * @return index after last parsed character, or -1 on parse failure.
1.229 + */
1.230 +int32_t ICU_Utility::parsePattern(const UnicodeString& pat,
1.231 + const Replaceable& text,
1.232 + int32_t index,
1.233 + int32_t limit) {
1.234 + int32_t ipat = 0;
1.235 +
1.236 + // empty pattern matches immediately
1.237 + if (ipat == pat.length()) {
1.238 + return index;
1.239 + }
1.240 +
1.241 + UChar32 cpat = pat.char32At(ipat);
1.242 +
1.243 + while (index < limit) {
1.244 + UChar32 c = text.char32At(index);
1.245 +
1.246 + // parse \s*
1.247 + if (cpat == 126 /*~*/) {
1.248 + if (PatternProps::isWhiteSpace(c)) {
1.249 + index += U16_LENGTH(c);
1.250 + continue;
1.251 + } else {
1.252 + if (++ipat == pat.length()) {
1.253 + return index; // success; c unparsed
1.254 + }
1.255 + // fall thru; process c again with next cpat
1.256 + }
1.257 + }
1.258 +
1.259 + // parse literal
1.260 + else if (c == cpat) {
1.261 + index += U16_LENGTH(c);
1.262 + ipat += U16_LENGTH(cpat);
1.263 + if (ipat == pat.length()) {
1.264 + return index; // success; c parsed
1.265 + }
1.266 + // fall thru; get next cpat
1.267 + }
1.268 +
1.269 + // match failure of literal
1.270 + else {
1.271 + return -1;
1.272 + }
1.273 +
1.274 + cpat = pat.char32At(ipat);
1.275 + }
1.276 +
1.277 + return -1; // text ended before end of pat
1.278 +}
1.279 +
1.280 +/**
1.281 + * Append a character to a rule that is being built up. To flush
1.282 + * the quoteBuf to rule, make one final call with isLiteral == TRUE.
1.283 + * If there is no final character, pass in (UChar32)-1 as c.
1.284 + * @param rule the string to append the character to
1.285 + * @param c the character to append, or (UChar32)-1 if none.
1.286 + * @param isLiteral if true, then the given character should not be
1.287 + * quoted or escaped. Usually this means it is a syntactic element
1.288 + * such as > or $
1.289 + * @param escapeUnprintable if true, then unprintable characters
1.290 + * should be escaped using \uxxxx or \Uxxxxxxxx. These escapes will
1.291 + * appear outside of quotes.
1.292 + * @param quoteBuf a buffer which is used to build up quoted
1.293 + * substrings. The caller should initially supply an empty buffer,
1.294 + * and thereafter should not modify the buffer. The buffer should be
1.295 + * cleared out by, at the end, calling this method with a literal
1.296 + * character.
1.297 + */
1.298 +void ICU_Utility::appendToRule(UnicodeString& rule,
1.299 + UChar32 c,
1.300 + UBool isLiteral,
1.301 + UBool escapeUnprintable,
1.302 + UnicodeString& quoteBuf) {
1.303 + // If we are escaping unprintables, then escape them outside
1.304 + // quotes. \u and \U are not recognized within quotes. The same
1.305 + // logic applies to literals, but literals are never escaped.
1.306 + if (isLiteral ||
1.307 + (escapeUnprintable && ICU_Utility::isUnprintable(c))) {
1.308 + if (quoteBuf.length() > 0) {
1.309 + // We prefer backslash APOSTROPHE to double APOSTROPHE
1.310 + // (more readable, less similar to ") so if there are
1.311 + // double APOSTROPHEs at the ends, we pull them outside
1.312 + // of the quote.
1.313 +
1.314 + // If the first thing in the quoteBuf is APOSTROPHE
1.315 + // (doubled) then pull it out.
1.316 + while (quoteBuf.length() >= 2 &&
1.317 + quoteBuf.charAt(0) == APOSTROPHE &&
1.318 + quoteBuf.charAt(1) == APOSTROPHE) {
1.319 + rule.append(BACKSLASH).append(APOSTROPHE);
1.320 + quoteBuf.remove(0, 2);
1.321 + }
1.322 + // If the last thing in the quoteBuf is APOSTROPHE
1.323 + // (doubled) then remove and count it and add it after.
1.324 + int32_t trailingCount = 0;
1.325 + while (quoteBuf.length() >= 2 &&
1.326 + quoteBuf.charAt(quoteBuf.length()-2) == APOSTROPHE &&
1.327 + quoteBuf.charAt(quoteBuf.length()-1) == APOSTROPHE) {
1.328 + quoteBuf.truncate(quoteBuf.length()-2);
1.329 + ++trailingCount;
1.330 + }
1.331 + if (quoteBuf.length() > 0) {
1.332 + rule.append(APOSTROPHE);
1.333 + rule.append(quoteBuf);
1.334 + rule.append(APOSTROPHE);
1.335 + quoteBuf.truncate(0);
1.336 + }
1.337 + while (trailingCount-- > 0) {
1.338 + rule.append(BACKSLASH).append(APOSTROPHE);
1.339 + }
1.340 + }
1.341 + if (c != (UChar32)-1) {
1.342 + /* Since spaces are ignored during parsing, they are
1.343 + * emitted only for readability. We emit one here
1.344 + * only if there isn't already one at the end of the
1.345 + * rule.
1.346 + */
1.347 + if (c == SPACE) {
1.348 + int32_t len = rule.length();
1.349 + if (len > 0 && rule.charAt(len-1) != c) {
1.350 + rule.append(c);
1.351 + }
1.352 + } else if (!escapeUnprintable || !ICU_Utility::escapeUnprintable(rule, c)) {
1.353 + rule.append(c);
1.354 + }
1.355 + }
1.356 + }
1.357 +
1.358 + // Escape ' and '\' and don't begin a quote just for them
1.359 + else if (quoteBuf.length() == 0 &&
1.360 + (c == APOSTROPHE || c == BACKSLASH)) {
1.361 + rule.append(BACKSLASH);
1.362 + rule.append(c);
1.363 + }
1.364 +
1.365 + // Specials (printable ascii that isn't [0-9a-zA-Z]) and
1.366 + // whitespace need quoting. Also append stuff to quotes if we are
1.367 + // building up a quoted substring already.
1.368 + else if (quoteBuf.length() > 0 ||
1.369 + (c >= 0x0021 && c <= 0x007E &&
1.370 + !((c >= 0x0030/*'0'*/ && c <= 0x0039/*'9'*/) ||
1.371 + (c >= 0x0041/*'A'*/ && c <= 0x005A/*'Z'*/) ||
1.372 + (c >= 0x0061/*'a'*/ && c <= 0x007A/*'z'*/))) ||
1.373 + PatternProps::isWhiteSpace(c)) {
1.374 + quoteBuf.append(c);
1.375 + // Double ' within a quote
1.376 + if (c == APOSTROPHE) {
1.377 + quoteBuf.append(c);
1.378 + }
1.379 + }
1.380 +
1.381 + // Otherwise just append
1.382 + else {
1.383 + rule.append(c);
1.384 + }
1.385 +}
1.386 +
1.387 +void ICU_Utility::appendToRule(UnicodeString& rule,
1.388 + const UnicodeString& text,
1.389 + UBool isLiteral,
1.390 + UBool escapeUnprintable,
1.391 + UnicodeString& quoteBuf) {
1.392 + for (int32_t i=0; i<text.length(); ++i) {
1.393 + appendToRule(rule, text[i], isLiteral, escapeUnprintable, quoteBuf);
1.394 + }
1.395 +}
1.396 +
1.397 +/**
1.398 + * Given a matcher reference, which may be null, append its
1.399 + * pattern as a literal to the given rule.
1.400 + */
1.401 +void ICU_Utility::appendToRule(UnicodeString& rule,
1.402 + const UnicodeMatcher* matcher,
1.403 + UBool escapeUnprintable,
1.404 + UnicodeString& quoteBuf) {
1.405 + if (matcher != NULL) {
1.406 + UnicodeString pat;
1.407 + appendToRule(rule, matcher->toPattern(pat, escapeUnprintable),
1.408 + TRUE, escapeUnprintable, quoteBuf);
1.409 + }
1.410 +}
1.411 +
1.412 +U_NAMESPACE_END
