Wed, 31 Dec 2014 06:09:35 +0100
Cloned upstream origin tor-browser at tor-browser-31.3.0esr-4.5-1-build1
revision ID fc1c9ff7c1b2defdbc039f12214767608f46423f for hacking purpose.
michael@0 | 1 | /* |
michael@0 | 2 | ******************************************************************************* |
michael@0 | 3 | * |
michael@0 | 4 | * Copyright (C) 2001-2011, International Business Machines |
michael@0 | 5 | * Corporation and others. All Rights Reserved. |
michael@0 | 6 | * |
michael@0 | 7 | ******************************************************************************* |
michael@0 | 8 | * file name: ucol_tok.h |
michael@0 | 9 | * encoding: US-ASCII |
michael@0 | 10 | * tab size: 8 (not used) |
michael@0 | 11 | * indentation:4 |
michael@0 | 12 | * |
michael@0 | 13 | * created 02/22/2001 |
michael@0 | 14 | * created by: Vladimir Weinstein |
michael@0 | 15 | * |
michael@0 | 16 | * This module reads a tailoring rule string and produces a list of |
michael@0 | 17 | * tokens that will be turned into collation elements |
michael@0 | 18 | * |
michael@0 | 19 | */ |
michael@0 | 20 | |
michael@0 | 21 | #ifndef UCOL_TOKENS_H |
michael@0 | 22 | #define UCOL_TOKENS_H |
michael@0 | 23 | |
michael@0 | 24 | #include "unicode/utypes.h" |
michael@0 | 25 | #include "unicode/uset.h" |
michael@0 | 26 | |
michael@0 | 27 | #if !UCONFIG_NO_COLLATION |
michael@0 | 28 | |
michael@0 | 29 | #include "ucol_imp.h" |
michael@0 | 30 | #include "uhash.h" |
michael@0 | 31 | #include "unicode/parseerr.h" |
michael@0 | 32 | |
michael@0 | 33 | #define UCOL_TOK_UNSET 0xFFFFFFFF |
michael@0 | 34 | #define UCOL_TOK_RESET 0xDEADBEEF |
michael@0 | 35 | |
michael@0 | 36 | #define UCOL_TOK_POLARITY_NEGATIVE 0 |
michael@0 | 37 | #define UCOL_TOK_POLARITY_POSITIVE 1 |
michael@0 | 38 | |
michael@0 | 39 | #define UCOL_TOK_TOP 0x04 |
michael@0 | 40 | #define UCOL_TOK_VARIABLE_TOP 0x08 |
michael@0 | 41 | #define UCOL_TOK_BEFORE 0x03 |
michael@0 | 42 | #define UCOL_TOK_SUCCESS 0x10 |
michael@0 | 43 | |
michael@0 | 44 | /* this is space for the extra strings that need to be unquoted */ |
michael@0 | 45 | /* during the parsing of the rules */ |
michael@0 | 46 | #define UCOL_TOK_EXTRA_RULE_SPACE_SIZE 4096 |
michael@0 | 47 | typedef struct UColToken UColToken; |
michael@0 | 48 | |
michael@0 | 49 | typedef struct { |
michael@0 | 50 | UColToken* first; |
michael@0 | 51 | UColToken* last; |
michael@0 | 52 | UColToken* reset; |
michael@0 | 53 | UBool indirect; |
michael@0 | 54 | uint32_t baseCE; |
michael@0 | 55 | uint32_t baseContCE; |
michael@0 | 56 | uint32_t nextCE; |
michael@0 | 57 | uint32_t nextContCE; |
michael@0 | 58 | uint32_t previousCE; |
michael@0 | 59 | uint32_t previousContCE; |
michael@0 | 60 | int32_t pos[UCOL_STRENGTH_LIMIT]; |
michael@0 | 61 | uint32_t gapsLo[3*UCOL_CE_STRENGTH_LIMIT]; |
michael@0 | 62 | uint32_t gapsHi[3*UCOL_CE_STRENGTH_LIMIT]; |
michael@0 | 63 | uint32_t numStr[UCOL_CE_STRENGTH_LIMIT]; |
michael@0 | 64 | UColToken* fStrToken[UCOL_CE_STRENGTH_LIMIT]; |
michael@0 | 65 | UColToken* lStrToken[UCOL_CE_STRENGTH_LIMIT]; |
michael@0 | 66 | } UColTokListHeader; |
michael@0 | 67 | |
michael@0 | 68 | struct UColToken { |
michael@0 | 69 | UChar debugSource; |
michael@0 | 70 | UChar debugExpansion; |
michael@0 | 71 | UChar debugPrefix; |
michael@0 | 72 | uint32_t CEs[128]; |
michael@0 | 73 | uint32_t noOfCEs; |
michael@0 | 74 | uint32_t expCEs[128]; |
michael@0 | 75 | uint32_t noOfExpCEs; |
michael@0 | 76 | uint32_t source; |
michael@0 | 77 | uint32_t expansion; |
michael@0 | 78 | uint32_t prefix; |
michael@0 | 79 | uint32_t strength; |
michael@0 | 80 | uint32_t toInsert; |
michael@0 | 81 | uint32_t polarity; /* 1 for <, <<, <<<, , ; and -1 for >, >>, >>> */ |
michael@0 | 82 | UColTokListHeader *listHeader; |
michael@0 | 83 | UColToken* previous; |
michael@0 | 84 | UColToken* next; |
michael@0 | 85 | UChar **rulesToParseHdl; |
michael@0 | 86 | uint16_t flags; |
michael@0 | 87 | }; |
michael@0 | 88 | |
michael@0 | 89 | /* |
michael@0 | 90 | * This is a token that has been parsed |
michael@0 | 91 | * but not yet processed. Used to reduce |
michael@0 | 92 | * the number of arguments in the parser |
michael@0 | 93 | */ |
michael@0 | 94 | typedef struct { |
michael@0 | 95 | uint32_t strength; |
michael@0 | 96 | uint32_t charsOffset; |
michael@0 | 97 | uint32_t charsLen; |
michael@0 | 98 | uint32_t extensionOffset; |
michael@0 | 99 | uint32_t extensionLen; |
michael@0 | 100 | uint32_t prefixOffset; |
michael@0 | 101 | uint32_t prefixLen; |
michael@0 | 102 | uint16_t flags; |
michael@0 | 103 | uint16_t indirectIndex; |
michael@0 | 104 | } UColParsedToken; |
michael@0 | 105 | |
michael@0 | 106 | |
michael@0 | 107 | typedef struct { |
michael@0 | 108 | UColParsedToken parsedToken; |
michael@0 | 109 | UChar *source; |
michael@0 | 110 | UChar *end; |
michael@0 | 111 | const UChar *current; |
michael@0 | 112 | UChar *sourceCurrent; |
michael@0 | 113 | UChar *extraCurrent; |
michael@0 | 114 | UChar *extraEnd; |
michael@0 | 115 | const InverseUCATableHeader *invUCA; |
michael@0 | 116 | const UCollator *UCA; |
michael@0 | 117 | UHashtable *tailored; |
michael@0 | 118 | UColOptionSet *opts; |
michael@0 | 119 | uint32_t resultLen; |
michael@0 | 120 | uint32_t listCapacity; |
michael@0 | 121 | UColTokListHeader *lh; |
michael@0 | 122 | UColToken *varTop; |
michael@0 | 123 | USet *copySet; |
michael@0 | 124 | USet *removeSet; |
michael@0 | 125 | UBool buildCCTabFlag; /* Tailoring rule requirs building combining class table. */ |
michael@0 | 126 | |
michael@0 | 127 | UChar32 previousCp; /* Previous code point. */ |
michael@0 | 128 | /* For processing starred lists. */ |
michael@0 | 129 | UBool isStarred; /* Are we processing a starred token? */ |
michael@0 | 130 | UBool savedIsStarred; |
michael@0 | 131 | uint32_t currentStarredCharIndex; /* Index of the current charrecter in the starred expression. */ |
michael@0 | 132 | uint32_t lastStarredCharIndex; /* Index to the last character in the starred expression. */ |
michael@0 | 133 | |
michael@0 | 134 | /* For processing ranges. */ |
michael@0 | 135 | UBool inRange; /* Are we in a range? */ |
michael@0 | 136 | UChar32 currentRangeCp; /* Current code point in the range. */ |
michael@0 | 137 | UChar32 lastRangeCp; /* The last code point in the range. */ |
michael@0 | 138 | |
michael@0 | 139 | /* reorder codes for collation reordering */ |
michael@0 | 140 | int32_t* reorderCodes; |
michael@0 | 141 | int32_t reorderCodesLength; |
michael@0 | 142 | |
michael@0 | 143 | } UColTokenParser; |
michael@0 | 144 | |
michael@0 | 145 | typedef struct { |
michael@0 | 146 | const UChar *subName; |
michael@0 | 147 | int32_t subLen; |
michael@0 | 148 | UColAttributeValue attrVal; |
michael@0 | 149 | } ucolTokSuboption; |
michael@0 | 150 | |
michael@0 | 151 | typedef struct { |
michael@0 | 152 | const UChar *optionName; |
michael@0 | 153 | int32_t optionLen; |
michael@0 | 154 | const ucolTokSuboption *subopts; |
michael@0 | 155 | int32_t subSize; |
michael@0 | 156 | UColAttribute attr; |
michael@0 | 157 | } ucolTokOption; |
michael@0 | 158 | |
michael@0 | 159 | #define ucol_tok_isSpecialChar(ch) \ |
michael@0 | 160 | (((((ch) <= 0x002F) && ((ch) >= 0x0020)) || \ |
michael@0 | 161 | (((ch) <= 0x003F) && ((ch) >= 0x003A)) || \ |
michael@0 | 162 | (((ch) <= 0x0060) && ((ch) >= 0x005B)) || \ |
michael@0 | 163 | (((ch) <= 0x007E) && ((ch) >= 0x007D)) || \ |
michael@0 | 164 | (ch) == 0x007B)) |
michael@0 | 165 | |
michael@0 | 166 | |
michael@0 | 167 | U_CFUNC |
michael@0 | 168 | uint32_t ucol_tok_assembleTokenList(UColTokenParser *src, |
michael@0 | 169 | UParseError *parseError, |
michael@0 | 170 | UErrorCode *status); |
michael@0 | 171 | |
michael@0 | 172 | U_CFUNC |
michael@0 | 173 | void ucol_tok_initTokenList(UColTokenParser *src, |
michael@0 | 174 | const UChar *rules, |
michael@0 | 175 | const uint32_t rulesLength, |
michael@0 | 176 | const UCollator *UCA, |
michael@0 | 177 | GetCollationRulesFunction importFunc, |
michael@0 | 178 | void* context, |
michael@0 | 179 | UErrorCode *status); |
michael@0 | 180 | |
michael@0 | 181 | U_CFUNC void ucol_tok_closeTokenList(UColTokenParser *src); |
michael@0 | 182 | |
michael@0 | 183 | U_CAPI const UChar* U_EXPORT2 ucol_tok_parseNextToken(UColTokenParser *src, |
michael@0 | 184 | UBool startOfRules, |
michael@0 | 185 | UParseError *parseError, |
michael@0 | 186 | UErrorCode *status); |
michael@0 | 187 | |
michael@0 | 188 | |
michael@0 | 189 | U_CAPI const UChar * U_EXPORT2 |
michael@0 | 190 | ucol_tok_getNextArgument(const UChar *start, const UChar *end, |
michael@0 | 191 | UColAttribute *attrib, UColAttributeValue *value, |
michael@0 | 192 | UErrorCode *status); |
michael@0 | 193 | U_CAPI int32_t U_EXPORT2 ucol_inv_getNextCE(const UColTokenParser *src, |
michael@0 | 194 | uint32_t CE, uint32_t contCE, |
michael@0 | 195 | uint32_t *nextCE, uint32_t *nextContCE, |
michael@0 | 196 | uint32_t strength); |
michael@0 | 197 | U_CFUNC int32_t U_EXPORT2 ucol_inv_getPrevCE(const UColTokenParser *src, |
michael@0 | 198 | uint32_t CE, uint32_t contCE, |
michael@0 | 199 | uint32_t *prevCE, uint32_t *prevContCE, |
michael@0 | 200 | uint32_t strength); |
michael@0 | 201 | |
michael@0 | 202 | const UChar* U_CALLCONV ucol_tok_getRulesFromBundle( |
michael@0 | 203 | void* context, |
michael@0 | 204 | const char* locale, |
michael@0 | 205 | const char* type, |
michael@0 | 206 | int32_t* pLength, |
michael@0 | 207 | UErrorCode* status); |
michael@0 | 208 | |
michael@0 | 209 | #endif /* #if !UCONFIG_NO_COLLATION */ |
michael@0 | 210 | |
michael@0 | 211 | #endif |