1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/icu/source/i18n/regexst.cpp Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,288 @@
1.4 +//
1.5 +// regexst.h
1.6 +//
1.7 +// Copyright (C) 2004-2013, International Business Machines Corporation and others.
1.8 +// All Rights Reserved.
1.9 +//
1.10 +// This file contains class RegexStaticSets
1.11 +//
1.12 +// This class is internal to the regular expression implementation.
1.13 +// For the public Regular Expression API, see the file "unicode/regex.h"
1.14 +//
1.15 +// RegexStaticSets groups together the common UnicodeSets that are needed
1.16 +// for compiling or executing RegularExpressions. This grouping simplifies
1.17 +// the thread safe lazy creation and sharing of these sets across
1.18 +// all instances of regular expressions.
1.19 +//
1.20 +#include "unicode/utypes.h"
1.21 +
1.22 +#if !UCONFIG_NO_REGULAR_EXPRESSIONS
1.23 +
1.24 +#include "unicode/unistr.h"
1.25 +#include "unicode/uniset.h"
1.26 +#include "unicode/uchar.h"
1.27 +#include "unicode/regex.h"
1.28 +#include "uprops.h"
1.29 +#include "cmemory.h"
1.30 +#include "cstring.h"
1.31 +#include "uassert.h"
1.32 +#include "ucln_in.h"
1.33 +#include "umutex.h"
1.34 +
1.35 +#include "regexcst.h" // Contains state table for the regex pattern parser.
1.36 + // generated by a Perl script.
1.37 +#include "regexst.h"
1.38 +
1.39 +
1.40 +
1.41 +U_NAMESPACE_BEGIN
1.42 +
1.43 +
1.44 +//------------------------------------------------------------------------------
1.45 +//
1.46 +// Unicode Set pattern strings for all of the required constant sets.
1.47 +// Initialized with hex values for portability to EBCDIC based machines.
1.48 +// Really ugly, but there's no good way to avoid it.
1.49 +//
1.50 +//------------------------------------------------------------------------------
1.51 +
1.52 +// "Rule Char" Characters are those with no special meaning, and therefore do not
1.53 +// need to be escaped to appear as literals in a regexp. Expressed
1.54 +// as the inverse of those needing escaping -- [^\*\?\+\[\(\)\{\}\^\$\|\\\.]
1.55 +static const UChar gRuleSet_rule_char_pattern[] = {
1.56 + // [ ^ \ * \ ? \ + \ [ \ ( / )
1.57 + 0x5b, 0x5e, 0x5c, 0x2a, 0x5c, 0x3f, 0x5c, 0x2b, 0x5c, 0x5b, 0x5c, 0x28, 0x5c, 0x29,
1.58 + // \ { \ } \ ^ \ $ \ | \ \ \ . ]
1.59 + 0x5c, 0x7b,0x5c, 0x7d, 0x5c, 0x5e, 0x5c, 0x24, 0x5c, 0x7c, 0x5c, 0x5c, 0x5c, 0x2e, 0x5d, 0};
1.60 +
1.61 +
1.62 +static const UChar gRuleSet_digit_char_pattern[] = {
1.63 +// [ 0 - 9 ]
1.64 + 0x5b, 0x30, 0x2d, 0x39, 0x5d, 0};
1.65 +
1.66 +//
1.67 +// Here are the backslash escape characters that ICU's unescape() function
1.68 +// will handle.
1.69 +//
1.70 +static const UChar gUnescapeCharPattern[] = {
1.71 +// [ a c e f n r t u U x ]
1.72 + 0x5b, 0x61, 0x63, 0x65, 0x66, 0x6e, 0x72, 0x74, 0x75, 0x55, 0x78, 0x5d, 0};
1.73 +
1.74 +
1.75 +//
1.76 +// Unicode Set Definitions for Regular Expression \w
1.77 +//
1.78 +static const UChar gIsWordPattern[] = {
1.79 +// [ \ p { A l p h a b e t i c }
1.80 + 0x5b, 0x5c, 0x70, 0x7b, 0x61, 0x6c, 0x70, 0x68, 0x61, 0x62, 0x65, 0x74, 0x69, 0x63, 0x7d,
1.81 +// \ p { M } Mark
1.82 + 0x5c, 0x70, 0x7b, 0x4d, 0x7d,
1.83 +// \ p { N d } Digit_Numeric
1.84 + 0x5c, 0x70, 0x7b, 0x4e, 0x64, 0x7d,
1.85 +// \ p { P c } Connector_Punctuation
1.86 + 0x5c, 0x70, 0x7b, 0x50, 0x63, 0x7d,
1.87 +// \ u 2 0 0 c \ u 2 0 0 d ]
1.88 + 0x5c, 0x75, 0x32, 0x30, 0x30, 0x63, 0x5c, 0x75, 0x32, 0x30, 0x30, 0x64, 0x5d, 0};
1.89 +
1.90 +
1.91 +//
1.92 +// Unicode Set Definitions for Regular Expression \s
1.93 +//
1.94 +static const UChar gIsSpacePattern[] = {
1.95 +// [ \ p { W h i t e S p a c e } ]
1.96 + 0x5b, 0x5c, 0x70, 0x7b, 0x57, 0x68, 0x69, 0x74, 0x65, 0x53, 0x70, 0x61, 0x63, 0x65, 0x7d, 0x5d, 0};
1.97 +
1.98 +
1.99 +//
1.100 +// UnicodeSets used in implementation of Grapheme Cluster detection, \X
1.101 +//
1.102 +static const UChar gGC_ControlPattern[] = {
1.103 +// [ [ : Z l : ] [ : Z p : ]
1.104 + 0x5b, 0x5b, 0x3a, 0x5A, 0x6c, 0x3a, 0x5d, 0x5b, 0x3a, 0x5A, 0x70, 0x3a, 0x5d,
1.105 +// [ : C c : ] [ : C f : ] -
1.106 + 0x5b, 0x3a, 0x43, 0x63, 0x3a, 0x5d, 0x5b, 0x3a, 0x43, 0x66, 0x3a, 0x5d, 0x2d,
1.107 +// [ : G r a p h e m e _
1.108 + 0x5b, 0x3a, 0x47, 0x72, 0x61, 0x70, 0x68, 0x65, 0x6d, 0x65, 0x5f,
1.109 +// E x t e n d : ] ]
1.110 + 0x45, 0x78, 0x74, 0x65, 0x6e, 0x64, 0x3a, 0x5d, 0x5d, 0};
1.111 +
1.112 +static const UChar gGC_ExtendPattern[] = {
1.113 +// [ \ p { G r a p h e m e _
1.114 + 0x5b, 0x5c, 0x70, 0x7b, 0x47, 0x72, 0x61, 0x70, 0x68, 0x65, 0x6d, 0x65, 0x5f,
1.115 +// E x t e n d } ]
1.116 + 0x45, 0x78, 0x74, 0x65, 0x6e, 0x64, 0x7d, 0x5d, 0};
1.117 +
1.118 +static const UChar gGC_LPattern[] = {
1.119 +// [ \ p { H a n g u l _ S y l
1.120 + 0x5b, 0x5c, 0x70, 0x7b, 0x48, 0x61, 0x6e, 0x67, 0x75, 0x6c, 0x5f, 0x53, 0x79, 0x6c,
1.121 +// l a b l e _ T y p e = L } ]
1.122 + 0x6c, 0x61, 0x62, 0x6c, 0x65, 0x5f, 0x54, 0x79, 0x70, 0x65, 0x3d, 0x4c, 0x7d, 0x5d, 0};
1.123 +
1.124 +static const UChar gGC_VPattern[] = {
1.125 +// [ \ p { H a n g u l _ S y l
1.126 + 0x5b, 0x5c, 0x70, 0x7b, 0x48, 0x61, 0x6e, 0x67, 0x75, 0x6c, 0x5f, 0x53, 0x79, 0x6c,
1.127 +// l a b l e _ T y p e = V } ]
1.128 + 0x6c, 0x61, 0x62, 0x6c, 0x65, 0x5f, 0x54, 0x79, 0x70, 0x65, 0x3d, 0x56, 0x7d, 0x5d, 0};
1.129 +
1.130 +static const UChar gGC_TPattern[] = {
1.131 +// [ \ p { H a n g u l _ S y l
1.132 + 0x5b, 0x5c, 0x70, 0x7b, 0x48, 0x61, 0x6e, 0x67, 0x75, 0x6c, 0x5f, 0x53, 0x79, 0x6c,
1.133 +// l a b l e _ T y p e = T } ]
1.134 + 0x6c, 0x61, 0x62, 0x6c, 0x65, 0x5f, 0x54, 0x79, 0x70, 0x65, 0x3d, 0x54, 0x7d, 0x5d, 0};
1.135 +
1.136 +static const UChar gGC_LVPattern[] = {
1.137 +// [ \ p { H a n g u l _ S y l
1.138 + 0x5b, 0x5c, 0x70, 0x7b, 0x48, 0x61, 0x6e, 0x67, 0x75, 0x6c, 0x5f, 0x53, 0x79, 0x6c,
1.139 +// l a b l e _ T y p e = L V } ]
1.140 + 0x6c, 0x61, 0x62, 0x6c, 0x65, 0x5f, 0x54, 0x79, 0x70, 0x65, 0x3d, 0x4c, 0x56, 0x7d, 0x5d, 0};
1.141 +
1.142 +static const UChar gGC_LVTPattern[] = {
1.143 +// [ \ p { H a n g u l _ S y l
1.144 + 0x5b, 0x5c, 0x70, 0x7b, 0x48, 0x61, 0x6e, 0x67, 0x75, 0x6c, 0x5f, 0x53, 0x79, 0x6c,
1.145 +// l a b l e _ T y p e = L V T } ]
1.146 + 0x6c, 0x61, 0x62, 0x6c, 0x65, 0x5f, 0x54, 0x79, 0x70, 0x65, 0x3d, 0x4c, 0x56, 0x54, 0x7d, 0x5d, 0};
1.147 +
1.148 +
1.149 +RegexStaticSets *RegexStaticSets::gStaticSets = NULL;
1.150 +UInitOnce gStaticSetsInitOnce = U_INITONCE_INITIALIZER;
1.151 +
1.152 +RegexStaticSets::RegexStaticSets(UErrorCode *status)
1.153 +:
1.154 +fUnescapeCharSet(UnicodeString(TRUE, gUnescapeCharPattern, -1), *status),
1.155 +fRuleDigitsAlias(NULL),
1.156 +fEmptyText(NULL)
1.157 +{
1.158 + // First zero out everything
1.159 + int i;
1.160 + for (i=0; i<URX_LAST_SET; i++) {
1.161 + fPropSets[i] = NULL;
1.162 + }
1.163 + // Then init the sets to their correct values.
1.164 + fPropSets[URX_ISWORD_SET] = new UnicodeSet(UnicodeString(TRUE, gIsWordPattern, -1), *status);
1.165 + fPropSets[URX_ISSPACE_SET] = new UnicodeSet(UnicodeString(TRUE, gIsSpacePattern, -1), *status);
1.166 + fPropSets[URX_GC_EXTEND] = new UnicodeSet(UnicodeString(TRUE, gGC_ExtendPattern, -1), *status);
1.167 + fPropSets[URX_GC_CONTROL] = new UnicodeSet(UnicodeString(TRUE, gGC_ControlPattern, -1), *status);
1.168 + fPropSets[URX_GC_L] = new UnicodeSet(UnicodeString(TRUE, gGC_LPattern, -1), *status);
1.169 + fPropSets[URX_GC_V] = new UnicodeSet(UnicodeString(TRUE, gGC_VPattern, -1), *status);
1.170 + fPropSets[URX_GC_T] = new UnicodeSet(UnicodeString(TRUE, gGC_TPattern, -1), *status);
1.171 + fPropSets[URX_GC_LV] = new UnicodeSet(UnicodeString(TRUE, gGC_LVPattern, -1), *status);
1.172 + fPropSets[URX_GC_LVT] = new UnicodeSet(UnicodeString(TRUE, gGC_LVTPattern, -1), *status);
1.173 +
1.174 + // Check for null pointers
1.175 + if (fPropSets[URX_ISWORD_SET] == NULL || fPropSets[URX_ISSPACE_SET] == NULL || fPropSets[URX_GC_EXTEND] == NULL ||
1.176 + fPropSets[URX_GC_CONTROL] == NULL || fPropSets[URX_GC_L] == NULL || fPropSets[URX_GC_V] == NULL ||
1.177 + fPropSets[URX_GC_T] == NULL || fPropSets[URX_GC_LV] == NULL || fPropSets[URX_GC_LVT] == NULL) {
1.178 + goto ExitConstrDeleteAll;
1.179 + }
1.180 + if (U_FAILURE(*status)) {
1.181 + // Bail out if we were unable to create the above sets.
1.182 + // The rest of the initialization needs them, so we cannot proceed.
1.183 + return;
1.184 + }
1.185 +
1.186 +
1.187 + //
1.188 + // The following sets are dynamically constructed, because their
1.189 + // initialization strings would be unreasonable.
1.190 + //
1.191 +
1.192 +
1.193 + //
1.194 + // "Normal" is the set of characters that don't need special handling
1.195 + // when finding grapheme cluster boundaries.
1.196 + //
1.197 + fPropSets[URX_GC_NORMAL] = new UnicodeSet(0, UnicodeSet::MAX_VALUE);
1.198 + // Null pointer check
1.199 + if (fPropSets[URX_GC_NORMAL] == NULL) {
1.200 + goto ExitConstrDeleteAll;
1.201 + }
1.202 + fPropSets[URX_GC_NORMAL]->remove(0xac00, 0xd7a4);
1.203 + fPropSets[URX_GC_NORMAL]->removeAll(*fPropSets[URX_GC_CONTROL]);
1.204 + fPropSets[URX_GC_NORMAL]->removeAll(*fPropSets[URX_GC_L]);
1.205 + fPropSets[URX_GC_NORMAL]->removeAll(*fPropSets[URX_GC_V]);
1.206 + fPropSets[URX_GC_NORMAL]->removeAll(*fPropSets[URX_GC_T]);
1.207 +
1.208 + // Initialize the 8-bit fast bit sets from the parallel full
1.209 + // UnicodeSets.
1.210 + for (i=0; i<URX_LAST_SET; i++) {
1.211 + if (fPropSets[i]) {
1.212 + fPropSets[i]->compact();
1.213 + fPropSets8[i].init(fPropSets[i]);
1.214 + }
1.215 + }
1.216 +
1.217 + // Sets used while parsing rules, but not referenced from the parse state table
1.218 + fRuleSets[kRuleSet_rule_char-128] = UnicodeSet(UnicodeString(TRUE, gRuleSet_rule_char_pattern, -1), *status);
1.219 + fRuleSets[kRuleSet_digit_char-128] = UnicodeSet(UnicodeString(TRUE, gRuleSet_digit_char_pattern, -1), *status);
1.220 + fRuleDigitsAlias = &fRuleSets[kRuleSet_digit_char-128];
1.221 + for (i=0; i<(int32_t)(sizeof(fRuleSets)/sizeof(fRuleSets[0])); i++) {
1.222 + fRuleSets[i].compact();
1.223 + }
1.224 +
1.225 + // Finally, initialize an empty string for utility purposes
1.226 + fEmptyText = utext_openUChars(NULL, NULL, 0, status);
1.227 +
1.228 + return; // If we reached this point, everything is fine so just exit
1.229 +
1.230 +ExitConstrDeleteAll: // Remove fPropSets and fRuleSets and return error
1.231 + for (i=0; i<URX_LAST_SET; i++) {
1.232 + delete fPropSets[i];
1.233 + fPropSets[i] = NULL;
1.234 + }
1.235 + *status = U_MEMORY_ALLOCATION_ERROR;
1.236 +}
1.237 +
1.238 +
1.239 +RegexStaticSets::~RegexStaticSets() {
1.240 + int32_t i;
1.241 +
1.242 + for (i=0; i<URX_LAST_SET; i++) {
1.243 + delete fPropSets[i];
1.244 + fPropSets[i] = NULL;
1.245 + }
1.246 + fRuleDigitsAlias = NULL;
1.247 +
1.248 + utext_close(fEmptyText);
1.249 +}
1.250 +
1.251 +
1.252 +//------------------------------------------------------------------------------
1.253 +//
1.254 +// regex_cleanup Memory cleanup function, free/delete all
1.255 +// cached memory. Called by ICU's u_cleanup() function.
1.256 +//
1.257 +//------------------------------------------------------------------------------
1.258 +UBool
1.259 +RegexStaticSets::cleanup(void) {
1.260 + delete RegexStaticSets::gStaticSets;
1.261 + RegexStaticSets::gStaticSets = NULL;
1.262 + gStaticSetsInitOnce.reset();
1.263 + return TRUE;
1.264 +}
1.265 +
1.266 +U_CDECL_BEGIN
1.267 +static UBool U_CALLCONV
1.268 +regex_cleanup(void) {
1.269 + return RegexStaticSets::cleanup();
1.270 +}
1.271 +
1.272 +static void U_CALLCONV initStaticSets(UErrorCode &status) {
1.273 + U_ASSERT(RegexStaticSets::gStaticSets == NULL);
1.274 + ucln_i18n_registerCleanup(UCLN_I18N_REGEX, regex_cleanup);
1.275 + RegexStaticSets::gStaticSets = new RegexStaticSets(&status);
1.276 + if (U_FAILURE(status)) {
1.277 + delete RegexStaticSets::gStaticSets;
1.278 + RegexStaticSets::gStaticSets = NULL;
1.279 + }
1.280 + if (RegexStaticSets::gStaticSets == NULL && U_SUCCESS(status)) {
1.281 + status = U_MEMORY_ALLOCATION_ERROR;
1.282 + }
1.283 +}
1.284 +U_CDECL_END
1.285 +
1.286 +void RegexStaticSets::initGlobals(UErrorCode *status) {
1.287 + umtx_initOnce(gStaticSetsInitOnce, &initStaticSets, *status);
1.288 +}
1.289 +
1.290 +U_NAMESPACE_END
1.291 +#endif // !UCONFIG_NO_REGULAR_EXPRESSIONS
