X-Git-Url: https://git.saurik.com/apple/icu.git/blobdiff_plain/374ca955a76ecab1204ca8bfa63ff9238d998416..a01113dcd0f39d5da295ef82785beff9ed86fe38:/icuSources/common/util.cpp?ds=sidebyside diff --git a/icuSources/common/util.cpp b/icuSources/common/util.cpp index 203fde51..a762cc48 100644 --- a/icuSources/common/util.cpp +++ b/icuSources/common/util.cpp @@ -1,6 +1,8 @@ +// © 2016 and later: Unicode, Inc. and others. +// License & terms of use: http://www.unicode.org/copyright.html /* ********************************************************************** -* Copyright (c) 2001-2004, International Business Machines +* Copyright (c) 2001-2011, International Business Machines * Corporation and others. All Rights Reserved. ********************************************************************** * Date Name Description @@ -8,10 +10,10 @@ ********************************************************************** */ -#include "util.h" -#include "unicode/uchar.h" #include "unicode/unimatch.h" -#include "uprops.h" +#include "unicode/utf16.h" +#include "patternprops.h" +#include "util.h" // Define UChar constants using hex for EBCDIC compatibility @@ -29,6 +31,8 @@ static const UChar DIGITS[] = { 85,86,87,88,89,90 }; +U_NAMESPACE_BEGIN + UnicodeString& ICU_Utility::appendNumber(UnicodeString& result, int32_t n, int32_t radix, int32_t minDigits) { if (radix < 2 || radix > 36) { @@ -61,9 +65,6 @@ UnicodeString& ICU_Utility::appendNumber(UnicodeString& result, int32_t n, return result; } -static const UChar HEX[16] = {48,49,50,51,52,53,54,55, // 0-7 - 56,57,65,66,67,68,69,70}; // 8-9 A-F - /** * Return true if the character is NOT printable ASCII. */ @@ -82,17 +83,17 @@ UBool ICU_Utility::escapeUnprintable(UnicodeString& result, UChar32 c) { result.append(BACKSLASH); if (c & ~0xFFFF) { result.append(UPPER_U); - result.append(HEX[0xF&(c>>28)]); - result.append(HEX[0xF&(c>>24)]); - result.append(HEX[0xF&(c>>20)]); - result.append(HEX[0xF&(c>>16)]); + result.append(DIGITS[0xF&(c>>28)]); + result.append(DIGITS[0xF&(c>>24)]); + result.append(DIGITS[0xF&(c>>20)]); + result.append(DIGITS[0xF&(c>>16)]); } else { result.append(LOWER_U); } - result.append(HEX[0xF&(c>>12)]); - result.append(HEX[0xF&(c>>8)]); - result.append(HEX[0xF&(c>>4)]); - result.append(HEX[0xF&c]); + result.append(DIGITS[0xF&(c>>12)]); + result.append(DIGITS[0xF&(c>>8)]); + result.append(DIGITS[0xF&(c>>4)]); + result.append(DIGITS[0xF&c]); return TRUE; } return FALSE; @@ -134,13 +135,8 @@ int32_t ICU_Utility::quotedIndexOf(const UnicodeString& text, int32_t ICU_Utility::skipWhitespace(const UnicodeString& str, int32_t& pos, UBool advance) { int32_t p = pos; - while (p < str.length()) { - UChar32 c = str.char32At(p); - if (!uprv_isRuleWhiteSpace(c)) { - break; - } - p += UTF_CHAR_LENGTH(c); - } + const UChar* s = str.getBuffer(); + p = (int32_t)(PatternProps::skipWhiteSpace(s + p, str.length() - p) - s); if (advance) { pos = p; } @@ -148,8 +144,8 @@ int32_t ICU_Utility::skipWhitespace(const UnicodeString& str, int32_t& pos, } /** - * Skip over whitespace in a Replaceable. Whitespace is defined by - * uprv_isRuleWhiteSpace(). Skipping may be done in the forward or + * Skip over Pattern_White_Space in a Replaceable. + * Skipping may be done in the forward or * reverse direction. In either case, the leftmost index will be * inclusive, and the rightmost index will be exclusive. That is, * given a range defined as [start, limit), the call @@ -175,11 +171,11 @@ int32_t ICU_Utility::skipWhitespace(const UnicodeString& str, int32_t& pos, //? } //? //? while (pos != stop && -//? uprv_isRuleWhiteSpace(c = text.char32At(pos))) { +//? PatternProps::isWhiteSpace(c = text.char32At(pos))) { //? if (isForward) { -//? pos += UTF_CHAR_LENGTH(c); +//? pos += U16_LENGTH(c); //? } else { -//? pos -= UTF_CHAR_LENGTH(c); +//? pos -= U16_LENGTH(c); //? } //? } //? @@ -214,75 +210,12 @@ UBool ICU_Utility::parseChar(const UnicodeString& id, int32_t& pos, UChar ch) { return TRUE; } -/** - * Parse a pattern string starting at offset pos. Keywords are - * matched case-insensitively. Spaces may be skipped and may be - * optional or required. Integer values may be parsed, and if - * they are, they will be returned in the given array. If - * successful, the offset of the next non-space character is - * returned. On failure, -1 is returned. - * @param pattern must only contain lowercase characters, which - * will match their uppercase equivalents as well. A space - * character matches one or more required spaces. A '~' character - * matches zero or more optional spaces. A '#' character matches - * an integer and stores it in parsedInts, which the caller must - * ensure has enough capacity. - * @param parsedInts array to receive parsed integers. Caller - * must ensure that parsedInts.length is >= the number of '#' - * signs in 'pattern'. - * @return the position after the last character parsed, or -1 if - * the parse failed - */ -int32_t ICU_Utility::parsePattern(const UnicodeString& rule, int32_t pos, int32_t limit, - const UnicodeString& pattern, int32_t* parsedInts) { - // TODO Update this to handle surrogates - int32_t p; - int32_t intCount = 0; // number of integers parsed - for (int32_t i=0; i= limit) { - return -1; - } - c = rule.charAt(pos++); - if (!uprv_isRuleWhiteSpace(c)) { - return -1; - } - // FALL THROUGH to skipWhitespace - case 126 /*'~'*/: - pos = skipWhitespace(rule, pos); - break; - case 35 /*'#'*/: - p = pos; - parsedInts[intCount++] = parseInteger(rule, p, limit); - if (p == pos) { - // Syntax error; failed to parse integer - return -1; - } - pos = p; - break; - default: - if (pos >= limit) { - return -1; - } - c = (UChar) u_tolower(rule.charAt(pos++)); - if (c != cpat) { - return -1; - } - break; - } - } - return pos; -} - /** * Parse a pattern string within the given Replaceable and a parsing * pattern. Characters are matched literally and case-sensitively * except for the following special characters: * - * ~ zero or more uprv_isRuleWhiteSpace chars + * ~ zero or more Pattern_White_Space chars * * If end of pattern is reached with all matches along the way, * pos is advanced to the first unparsed index and returned. @@ -311,8 +244,8 @@ int32_t ICU_Utility::parsePattern(const UnicodeString& pat, // parse \s* if (cpat == 126 /*~*/) { - if (uprv_isRuleWhiteSpace(c)) { - index += UTF_CHAR_LENGTH(c); + if (PatternProps::isWhiteSpace(c)) { + index += U16_LENGTH(c); continue; } else { if (++ipat == pat.length()) { @@ -324,8 +257,8 @@ int32_t ICU_Utility::parsePattern(const UnicodeString& pat, // parse literal else if (c == cpat) { - index += UTF_CHAR_LENGTH(c); - ipat += UTF_CHAR_LENGTH(cpat); + index += U16_LENGTH(c); + ipat += U16_LENGTH(cpat); if (ipat == pat.length()) { return index; // success; c parsed } @@ -343,140 +276,6 @@ int32_t ICU_Utility::parsePattern(const UnicodeString& pat, return -1; // text ended before end of pat } -/** - * Parse an integer at pos, either of the form \d+ or of the form - * 0x[0-9A-Fa-f]+ or 0[0-7]+, that is, in standard decimal, hex, - * or octal format. - * @param pos INPUT-OUTPUT parameter. On input, the first - * character to parse. On output, the character after the last - * parsed character. - */ -int32_t ICU_Utility::parseInteger(const UnicodeString& rule, int32_t& pos, int32_t limit) { - int32_t count = 0; - int32_t value = 0; - int32_t p = pos; - int8_t radix = 10; - - if (p < limit && rule.charAt(p) == 48 /*0*/) { - if (p+1 < limit && (rule.charAt(p+1) == 0x78 /*x*/ || rule.charAt(p+1) == 0x58 /*X*/)) { - p += 2; - radix = 16; - } - else { - p++; - count = 1; - radix = 8; - } - } - - while (p < limit) { - int32_t d = u_digit(rule.charAt(p++), radix); - if (d < 0) { - --p; - break; - } - ++count; - int32_t v = (value * radix) + d; - if (v <= value) { - // If there are too many input digits, at some point - // the value will go negative, e.g., if we have seen - // "0x8000000" already and there is another '0', when - // we parse the next 0 the value will go negative. - return 0; - } - value = v; - } - if (count > 0) { - pos = p; - } - return value; -} - -/** - * Parse a Unicode identifier from the given string at the given - * position. Return the identifier, or an empty string if there - * is no identifier. - * @param str the string to parse - * @param pos INPUT-OUPUT parameter. On INPUT, pos is the - * first character to examine. It must be less than str.length(), - * and it must not point to a whitespace character. That is, must - * have pos < str.length() and - * !uprv_isRuleWhiteSpace(str.char32At(pos)). On - * OUTPUT, the position after the last parsed character. - * @return the Unicode identifier, or an empty string if there is - * no valid identifier at pos. - */ -UnicodeString ICU_Utility::parseUnicodeIdentifier(const UnicodeString& str, int32_t& pos) { - // assert(pos < str.length()); - // assert(!uprv_isRuleWhiteSpace(str.char32At(pos))); - UnicodeString buf; - int p = pos; - while (p < str.length()) { - UChar32 ch = str.char32At(p); - if (buf.length() == 0) { - if (u_isIDStart(ch)) { - buf.append(ch); - } else { - buf.truncate(0); - return buf; - } - } else { - if (u_isIDPart(ch)) { - buf.append(ch); - } else { - break; - } - } - p += UTF_CHAR_LENGTH(ch); - } - pos = p; - return buf; -} - -/** - * Parse an unsigned 31-bit integer at the given offset. Use - * UCharacter.digit() to parse individual characters into digits. - * @param text the text to be parsed - * @param pos INPUT-OUTPUT parameter. On entry, pos[0] is the - * offset within text at which to start parsing; it should point - * to a valid digit. On exit, pos[0] is the offset after the last - * parsed character. If the parse failed, it will be unchanged on - * exit. Must be >= 0 on entry. - * @param radix the radix in which to parse; must be >= 2 and <= - * 36. - * @return a non-negative parsed number, or -1 upon parse failure. - * Parse fails if there are no digits, that is, if pos[0] does not - * point to a valid digit on entry, or if the number to be parsed - * does not fit into a 31-bit unsigned integer. - */ -int32_t ICU_Utility::parseNumber(const UnicodeString& text, - int32_t& pos, int8_t radix) { - // assert(pos[0] >= 0); - // assert(radix >= 2); - // assert(radix <= 36); - int32_t n = 0; - int32_t p = pos; - while (p < text.length()) { - UChar32 ch = text.char32At(p); - int32_t d = u_digit(ch, radix); - if (d < 0) { - break; - } - n = radix*n + d; - // ASSUME that when a 32-bit integer overflows it becomes - // negative. E.g., 214748364 * 10 + 8 => negative value. - if (n < 0) { - return -1; - } - ++p; - } - if (p == pos) { - return -1; - } - pos = p; - return n; -} - /** * Append a character to a rule that is being built up. To flush * the quoteBuf to rule, make one final call with isLiteral == TRUE. @@ -570,7 +369,7 @@ void ICU_Utility::appendToRule(UnicodeString& rule, !((c >= 0x0030/*'0'*/ && c <= 0x0039/*'9'*/) || (c >= 0x0041/*'A'*/ && c <= 0x005A/*'Z'*/) || (c >= 0x0061/*'a'*/ && c <= 0x007A/*'z'*/))) || - uprv_isRuleWhiteSpace(c)) { + PatternProps::isWhiteSpace(c)) { quoteBuf.append(c); // Double ' within a quote if (c == APOSTROPHE) { @@ -609,16 +408,4 @@ void ICU_Utility::appendToRule(UnicodeString& rule, } } -U_CAPI UBool U_EXPORT2 -uprv_isRuleWhiteSpace(UChar32 c) { - /* "white space" in the sense of ICU rule parsers - This is a FIXED LIST that is NOT DEPENDENT ON UNICODE PROPERTIES. - See UTR #31: http://www.unicode.org/reports/tr31/. - U+0009..U+000D, U+0020, U+0085, U+200E..U+200F, and U+2028..U+2029 - */ - return (c >= 0x0009 && c <= 0x2029 && - (c <= 0x000D || c == 0x0020 || c == 0x0085 || - c == 0x200E || c == 0x200F || c >= 0x2028)); -} - -//eof +U_NAMESPACE_END