Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(41)

Side by Side Diff: source/i18n/uspoof.cpp

Issue 1621843002: ICU 56 update step 1 (Closed) Base URL: https://chromium.googlesource.com/chromium/deps/icu.git@561
Patch Set: Created 4 years, 11 months ago
Use n/p to move between diff chunks; N/P to move between comments. Draft comments are only viewable by you.
Jump to:
View unified diff | Download patch
« no previous file with comments | « source/i18n/usearch.cpp ('k') | source/i18n/uspoof_build.cpp » ('j') | no next file with comments »
Toggle Intra-line Diffs ('i') | Expand Comments ('e') | Collapse Comments ('c') | Show Comments Hide Comments ('s')
OLDNEW
1 /* 1 /*
2 *************************************************************************** 2 ***************************************************************************
3 * Copyright (C) 2008-2014, International Business Machines Corporation 3 * Copyright (C) 2008-2015, International Business Machines Corporation
4 * and others. All Rights Reserved. 4 * and others. All Rights Reserved.
5 *************************************************************************** 5 ***************************************************************************
6 * file name: uspoof.cpp 6 * file name: uspoof.cpp
7 * encoding: US-ASCII 7 * encoding: US-ASCII
8 * tab size: 8 (not used) 8 * tab size: 8 (not used)
9 * indentation:4 9 * indentation:4
10 * 10 *
11 * created on: 2008Feb13 11 * created on: 2008Feb13
12 * created by: Andy Heninger 12 * created by: Andy Heninger
13 * 13 *
(...skipping 19 matching lines...) Expand all
33 33
34 U_NAMESPACE_USE 34 U_NAMESPACE_USE
35 35
36 36
37 // 37 //
38 // Static Objects used by the spoof impl, their thread safe initialization and t heir cleanup. 38 // Static Objects used by the spoof impl, their thread safe initialization and t heir cleanup.
39 // 39 //
40 static UnicodeSet *gInclusionSet = NULL; 40 static UnicodeSet *gInclusionSet = NULL;
41 static UnicodeSet *gRecommendedSet = NULL; 41 static UnicodeSet *gRecommendedSet = NULL;
42 static const Normalizer2 *gNfdNormalizer = NULL; 42 static const Normalizer2 *gNfdNormalizer = NULL;
43 static UInitOnce gSpoofInitOnce = U_INITONCE_INITIALIZER; 43 static SpoofData *gDefaultSpoofData = NULL;
44 static UInitOnce gSpoofInitStaticsOnce = U_INITONCE_INITIALIZER;
45 static UInitOnce gSpoofInitDefaultOnce = U_INITONCE_INITIALIZER;
44 46
45 static UBool U_CALLCONV 47 static UBool U_CALLCONV
46 uspoof_cleanup(void) { 48 uspoof_cleanup(void) {
47 delete gInclusionSet; 49 delete gInclusionSet;
48 gInclusionSet = NULL; 50 gInclusionSet = NULL;
49 delete gRecommendedSet; 51 delete gRecommendedSet;
50 gRecommendedSet = NULL; 52 gRecommendedSet = NULL;
51 gNfdNormalizer = NULL; 53 gNfdNormalizer = NULL;
52 gSpoofInitOnce.reset(); 54 if (gDefaultSpoofData) {
55 gDefaultSpoofData->removeReference(); // Will delete, assuming all use r-level spoof checkers were closed.
56 }
57 gDefaultSpoofData = NULL;
58 gSpoofInitStaticsOnce.reset();
59 gSpoofInitDefaultOnce.reset();
53 return TRUE; 60 return TRUE;
54 } 61 }
55 62
56 static void U_CALLCONV initializeStatics(UErrorCode &status) { 63 static void U_CALLCONV initializeStatics(UErrorCode &status) {
57 static const char *inclusionPat = 64 static const char *inclusionPat =
58 "[\\u0027\\u002d-\\u002e\\u003A\\u00B7\\u0375\\u058A\\u05F3-\\u05F4" 65 "[\\u0027\\u002D-\\u002E\\u003A\\u00B7\\u0375\\u058A\\u05F3-\\u05F4"
59 "\\u06FD-\\u06FE\\u0F0B\\u200C-\\u200D\\u2010\\u2019\\u2027\\u30A0\\u 30FB]"; 66 "\\u06FD-\\u06FE\\u0F0B\\u200C-\\u200D\\u2010\\u2019\\u2027\\u30A0\\u 30FB]";
60 gInclusionSet = new UnicodeSet(UnicodeString(inclusionPat, -1, US_INV), stat us); 67 gInclusionSet = new UnicodeSet(UnicodeString(inclusionPat, -1, US_INV), stat us);
68 gInclusionSet->freeze();
61 69
62 // Note: data from http://unicode.org/Public/security/latest/xidmodification s.txt version 6.3.0 70 // Note: data from http://unicode.org/Public/security/latest/xidmodification s.txt version 8.0.0
71 // There is no tooling to generate this from the .txt file, hand extra cted with editor macros.
72 // Ultimately, data will be available as character properties, elimina ting this.
63 // Note: concatenated string constants do not work with UNICODE_STRING_SIMPL E on all platforms. 73 // Note: concatenated string constants do not work with UNICODE_STRING_SIMPL E on all platforms.
64 static const char *recommendedPat = 74 static const char *recommendedPat =
65 "[\\u0030-\\u0039\\u0041-\\u005A\\u005F\\u0061-\\u007A\\u00C0-\\u00D 6\\u00D8-\\u00F6" 75 "[\\u0030-\\u0039\\u0041-\\u005A\\u005F\\u0061-\\u007A\\u00C0-\\u00D 6\\u00D8-\\u00F6"
66 "\\u00F8-\\u0131\\u0134-\\u013E\\u0141-\\u0148\\u014A-\\u017E\\u01A0 -\\u01A1\\u01AF-\\u01B0" 76 "\\u00F8-\\u0131\\u0134-\\u013E\\u0141-\\u0148\\u014A-\\u017E\\u018F \\u01A0-\\u01A1"
67 "\\u01CD-\\u01DC\\u01DE-\\u01E3\\u01E6-\\u01F0\\u01F4-\\u01F5\\u01F8 -\\u021B\\u021E-\\u021F" 77 "\\u01AF-\\u01B0\\u01CD-\\u01DC\\u01DE-\\u01E3\\u01E6-\\u01F0\\u01F4 -\\u01F5\\u01F8-\\u021B"
68 "\\u0226-\\u0233\\u0259\\u02BB-\\u02BC\\u02EC\\u0300-\\u0304\\u0306- \\u030C\\u030F-\\u0311" 78 "\\u021E-\\u021F\\u0226-\\u0233\\u0259\\u02BB-\\u02BC\\u02EC\\u0300- \\u0304\\u0306-\\u030C"
69 "\\u0313-\\u0314\\u031B\\u0323-\\u0328\\u032D-\\u032E\\u0330-\\u0331 \\u0335\\u0338-\\u0339" 79 "\\u030F-\\u0311\\u0313-\\u0314\\u031B\\u0323-\\u0328\\u032D-\\u032E \\u0330-\\u0331"
70 "\\u0342\\u0345\\u037B-\\u037D\\u0386\\u0388-\\u038A\\u038C\\u038E-\ \u03A1\\u03A3-\\u03CE" 80 "\\u0335\\u0338-\\u0339\\u0342\\u0345\\u037B-\\u037D\\u0386\\u0388-\ \u038A\\u038C"
71 "\\u03FC-\\u045F\\u048A-\\u0527\\u0531-\\u0556\\u0559\\u0561-\\u0586 \\u05B4\\u05D0-\\u05EA" 81 "\\u038E-\\u03A1\\u03A3-\\u03CE\\u03FC-\\u045F\\u048A-\\u0529\\u052E -\\u052F\\u0531-\\u0556"
72 "\\u05F0-\\u05F2\\u0620-\\u063F\\u0641-\\u0655\\u0660-\\u0669\\u0670 -\\u0672\\u0674" 82 "\\u0559\\u0561-\\u0586\\u05B4\\u05D0-\\u05EA\\u05F0-\\u05F2\\u0620- \\u063F\\u0641-\\u0655"
73 "\\u0679-\\u068D\\u068F-\\u06D3\\u06D5\\u06E5-\\u06E6\\u06EE-\\u06FC \\u06FF\\u0750-\\u07B1" 83 "\\u0660-\\u0669\\u0670-\\u0672\\u0674\\u0679-\\u068D\\u068F-\\u06D3 \\u06D5\\u06E5-\\u06E6"
74 "\\u08A0\\u08A2-\\u08AC\\u0901-\\u094D\\u094F-\\u0950\\u0956-\\u0957 \\u0960-\\u0963" 84 "\\u06EE-\\u06FC\\u06FF\\u0750-\\u07B1\\u08A0-\\u08AC\\u08B2\\u0901- \\u094D\\u094F-\\u0950"
75 "\\u0966-\\u096F\\u0971-\\u0977\\u0979-\\u097F\\u0981-\\u0983\\u0985 -\\u098C\\u098F-\\u0990" 85 "\\u0956-\\u0957\\u0960-\\u0963\\u0966-\\u096F\\u0971-\\u0977\\u0979 -\\u097F\\u0981-\\u0983"
76 "\\u0993-\\u09A8\\u09AA-\\u09B0\\u09B2\\u09B6-\\u09B9\\u09BC-\\u09C4 \\u09C7-\\u09C8" 86 "\\u0985-\\u098C\\u098F-\\u0990\\u0993-\\u09A8\\u09AA-\\u09B0\\u09B2 \\u09B6-\\u09B9"
77 "\\u09CB-\\u09CE\\u09D7\\u09E0-\\u09E3\\u09E6-\\u09F1\\u0A01-\\u0A03 \\u0A05-\\u0A0A" 87 "\\u09BC-\\u09C4\\u09C7-\\u09C8\\u09CB-\\u09CE\\u09D7\\u09E0-\\u09E3 \\u09E6-\\u09F1"
78 "\\u0A0F-\\u0A10\\u0A13-\\u0A28\\u0A2A-\\u0A30\\u0A32\\u0A35\\u0A38- \\u0A39\\u0A3C" 88 "\\u0A01-\\u0A03\\u0A05-\\u0A0A\\u0A0F-\\u0A10\\u0A13-\\u0A28\\u0A2A -\\u0A30\\u0A32"
79 "\\u0A3E-\\u0A42\\u0A47-\\u0A48\\u0A4B-\\u0A4D\\u0A5C\\u0A66-\\u0A74 \\u0A81-\\u0A83" 89 "\\u0A35\\u0A38-\\u0A39\\u0A3C\\u0A3E-\\u0A42\\u0A47-\\u0A48\\u0A4B- \\u0A4D\\u0A5C"
80 "\\u0A85-\\u0A8D\\u0A8F-\\u0A91\\u0A93-\\u0AA8\\u0AAA-\\u0AB0\\u0AB2 -\\u0AB3\\u0AB5-\\u0AB9" 90 "\\u0A66-\\u0A74\\u0A81-\\u0A83\\u0A85-\\u0A8D\\u0A8F-\\u0A91\\u0A93 -\\u0AA8\\u0AAA-\\u0AB0"
81 "\\u0ABC-\\u0AC5\\u0AC7-\\u0AC9\\u0ACB-\\u0ACD\\u0AD0\\u0AE0-\\u0AE3 \\u0AE6-\\u0AEF" 91 "\\u0AB2-\\u0AB3\\u0AB5-\\u0AB9\\u0ABC-\\u0AC5\\u0AC7-\\u0AC9\\u0ACB -\\u0ACD\\u0AD0"
82 "\\u0B01-\\u0B03\\u0B05-\\u0B0C\\u0B0F-\\u0B10\\u0B13-\\u0B28\\u0B2A -\\u0B30\\u0B32-\\u0B33" 92 "\\u0AE0-\\u0AE3\\u0AE6-\\u0AEF\\u0B01-\\u0B03\\u0B05-\\u0B0C\\u0B0F -\\u0B10\\u0B13-\\u0B28"
83 "\\u0B35-\\u0B39\\u0B3C-\\u0B43\\u0B47-\\u0B48\\u0B4B-\\u0B4D\\u0B56 -\\u0B57\\u0B5F-\\u0B61" 93 "\\u0B2A-\\u0B30\\u0B32-\\u0B33\\u0B35-\\u0B39\\u0B3C-\\u0B43\\u0B47 -\\u0B48\\u0B4B-\\u0B4D"
84 "\\u0B66-\\u0B6F\\u0B71\\u0B82-\\u0B83\\u0B85-\\u0B8A\\u0B8E-\\u0B90 \\u0B92-\\u0B95" 94 "\\u0B56-\\u0B57\\u0B5F-\\u0B61\\u0B66-\\u0B6F\\u0B71\\u0B82-\\u0B83 \\u0B85-\\u0B8A"
85 "\\u0B99-\\u0B9A\\u0B9C\\u0B9E-\\u0B9F\\u0BA3-\\u0BA4\\u0BA8-\\u0BAA \\u0BAE-\\u0BB9" 95 "\\u0B8E-\\u0B90\\u0B92-\\u0B95\\u0B99-\\u0B9A\\u0B9C\\u0B9E-\\u0B9F \\u0BA3-\\u0BA4"
86 "\\u0BBE-\\u0BC2\\u0BC6-\\u0BC8\\u0BCA-\\u0BCD\\u0BD0\\u0BD7\\u0BE6- \\u0BEF\\u0C01-\\u0C03" 96 "\\u0BA8-\\u0BAA\\u0BAE-\\u0BB9\\u0BBE-\\u0BC2\\u0BC6-\\u0BC8\\u0BCA -\\u0BCD\\u0BD0"
87 "\\u0C05-\\u0C0C\\u0C0E-\\u0C10\\u0C12-\\u0C28\\u0C2A-\\u0C33\\u0C35 -\\u0C39\\u0C3D-\\u0C44" 97 "\\u0BD7\\u0BE6-\\u0BEF\\u0C01-\\u0C03\\u0C05-\\u0C0C\\u0C0E-\\u0C10 \\u0C12-\\u0C28"
88 "\\u0C46-\\u0C48\\u0C4A-\\u0C4D\\u0C55-\\u0C56\\u0C60-\\u0C61\\u0C66 -\\u0C6F\\u0C82-\\u0C83" 98 "\\u0C2A-\\u0C33\\u0C35-\\u0C39\\u0C3D-\\u0C44\\u0C46-\\u0C48\\u0C4A -\\u0C4D\\u0C55-\\u0C56"
89 "\\u0C85-\\u0C8C\\u0C8E-\\u0C90\\u0C92-\\u0CA8\\u0CAA-\\u0CB3\\u0CB5 -\\u0CB9\\u0CBC-\\u0CC4" 99 "\\u0C60-\\u0C61\\u0C66-\\u0C6F\\u0C82-\\u0C83\\u0C85-\\u0C8C\\u0C8E -\\u0C90\\u0C92-\\u0CA8"
90 "\\u0CC6-\\u0CC8\\u0CCA-\\u0CCD\\u0CD5-\\u0CD6\\u0CE0-\\u0CE3\\u0CE6 -\\u0CEF\\u0CF1-\\u0CF2" 100 "\\u0CAA-\\u0CB3\\u0CB5-\\u0CB9\\u0CBC-\\u0CC4\\u0CC6-\\u0CC8\\u0CCA -\\u0CCD\\u0CD5-\\u0CD6"
91 "\\u0D02-\\u0D03\\u0D05-\\u0D0C\\u0D0E-\\u0D10\\u0D12-\\u0D3A\\u0D3D -\\u0D43\\u0D46-\\u0D48" 101 "\\u0CE0-\\u0CE3\\u0CE6-\\u0CEF\\u0CF1-\\u0CF2\\u0D02-\\u0D03\\u0D05 -\\u0D0C\\u0D0E-\\u0D10"
92 "\\u0D4A-\\u0D4E\\u0D57\\u0D60-\\u0D61\\u0D66-\\u0D6F\\u0D7A-\\u0D7F \\u0D82-\\u0D83" 102 "\\u0D12-\\u0D3A\\u0D3D-\\u0D43\\u0D46-\\u0D48\\u0D4A-\\u0D4E\\u0D57 \\u0D60-\\u0D61"
93 "\\u0D85-\\u0D8E\\u0D91-\\u0D96\\u0D9A-\\u0DA5\\u0DA7-\\u0DB1\\u0DB3 -\\u0DBB\\u0DBD" 103 "\\u0D66-\\u0D6F\\u0D7A-\\u0D7F\\u0D82-\\u0D83\\u0D85-\\u0D8E\\u0D91 -\\u0D96\\u0D9A-\\u0DA5"
94 "\\u0DC0-\\u0DC6\\u0DCA\\u0DCF-\\u0DD4\\u0DD6\\u0DD8-\\u0DDE\\u0DF2\ \u0E01-\\u0E32" 104 "\\u0DA7-\\u0DB1\\u0DB3-\\u0DBB\\u0DBD\\u0DC0-\\u0DC6\\u0DCA\\u0DCF- \\u0DD4\\u0DD6"
95 "\\u0E34-\\u0E3A\\u0E40-\\u0E4E\\u0E50-\\u0E59\\u0E81-\\u0E82\\u0E84 \\u0E87-\\u0E88" 105 "\\u0DD8-\\u0DDE\\u0DF2\\u0E01-\\u0E32\\u0E34-\\u0E3A\\u0E40-\\u0E4E \\u0E50-\\u0E59"
96 "\\u0E8A\\u0E8D\\u0E94-\\u0E97\\u0E99-\\u0E9F\\u0EA1-\\u0EA3\\u0EA5\ \u0EA7\\u0EAA-\\u0EAB" 106 "\\u0E81-\\u0E82\\u0E84\\u0E87-\\u0E88\\u0E8A\\u0E8D\\u0E94-\\u0E97\ \u0E99-\\u0E9F"
97 "\\u0EAD-\\u0EB2\\u0EB4-\\u0EB9\\u0EBB-\\u0EBD\\u0EC0-\\u0EC4\\u0EC6 \\u0EC8-\\u0ECD" 107 "\\u0EA1-\\u0EA3\\u0EA5\\u0EA7\\u0EAA-\\u0EAB\\u0EAD-\\u0EB2\\u0EB4- \\u0EB9\\u0EBB-\\u0EBD"
98 "\\u0ED0-\\u0ED9\\u0EDE-\\u0EDF\\u0F00\\u0F20-\\u0F29\\u0F35\\u0F37\ \u0F3E-\\u0F42" 108 "\\u0EC0-\\u0EC4\\u0EC6\\u0EC8-\\u0ECD\\u0ED0-\\u0ED9\\u0EDE-\\u0EDF \\u0F00\\u0F20-\\u0F29"
99 "\\u0F44-\\u0F47\\u0F49-\\u0F4C\\u0F4E-\\u0F51\\u0F53-\\u0F56\\u0F58 -\\u0F5B\\u0F5D-\\u0F68" 109 "\\u0F35\\u0F37\\u0F3E-\\u0F42\\u0F44-\\u0F47\\u0F49-\\u0F4C\\u0F4E- \\u0F51\\u0F53-\\u0F56"
100 "\\u0F6A-\\u0F6C\\u0F71-\\u0F72\\u0F74\\u0F7A-\\u0F80\\u0F82-\\u0F84 \\u0F86-\\u0F92" 110 "\\u0F58-\\u0F5B\\u0F5D-\\u0F68\\u0F6A-\\u0F6C\\u0F71-\\u0F72\\u0F74 \\u0F7A-\\u0F80"
101 "\\u0F94-\\u0F97\\u0F99-\\u0F9C\\u0F9E-\\u0FA1\\u0FA3-\\u0FA6\\u0FA8 -\\u0FAB\\u0FAD-\\u0FB8" 111 "\\u0F82-\\u0F84\\u0F86-\\u0F92\\u0F94-\\u0F97\\u0F99-\\u0F9C\\u0F9E -\\u0FA1\\u0FA3-\\u0FA6"
102 "\\u0FBA-\\u0FBC\\u0FC6\\u1000-\\u1049\\u1050-\\u109D\\u10C7\\u10CD\ \u10D0-\\u10F0" 112 "\\u0FA8-\\u0FAB\\u0FAD-\\u0FB8\\u0FBA-\\u0FBC\\u0FC6\\u1000-\\u1049 \\u1050-\\u109D"
103 "\\u10F7-\\u10FA\\u10FD-\\u10FF\\u1200-\\u1248\\u124A-\\u124D\\u1250 -\\u1256\\u1258" 113 "\\u10C7\\u10CD\\u10D0-\\u10F0\\u10F7-\\u10FA\\u10FD-\\u10FF\\u1200- \\u1248\\u124A-\\u124D"
104 "\\u125A-\\u125D\\u1260-\\u1288\\u128A-\\u128D\\u1290-\\u12B0\\u12B2 -\\u12B5\\u12B8-\\u12BE" 114 "\\u1250-\\u1256\\u1258\\u125A-\\u125D\\u1260-\\u1288\\u128A-\\u128D \\u1290-\\u12B0"
105 "\\u12C0\\u12C2-\\u12C5\\u12C8-\\u12D6\\u12D8-\\u1310\\u1312-\\u1315 \\u1318-\\u135A" 115 "\\u12B2-\\u12B5\\u12B8-\\u12BE\\u12C0\\u12C2-\\u12C5\\u12C8-\\u12D6 \\u12D8-\\u1310"
106 "\\u135D-\\u135F\\u1380-\\u138F\\u1780-\\u17A2\\u17A5-\\u17A7\\u17A9 -\\u17B3\\u17B6-\\u17CA" 116 "\\u1312-\\u1315\\u1318-\\u135A\\u135D-\\u135F\\u1380-\\u138F\\u1780 -\\u17A2\\u17A5-\\u17A7"
107 "\\u17D2\\u17D7\\u17DC\\u17E0-\\u17E9\\u1E00-\\u1E99\\u1EBF\\u1F00-\ \u1F15\\u1F18-\\u1F1D" 117 "\\u17A9-\\u17B3\\u17B6-\\u17CA\\u17D2\\u17D7\\u17DC\\u17E0-\\u17E9\ \u1E00-\\u1E99"
108 "\\u1F20-\\u1F45\\u1F48-\\u1F4D\\u1F50-\\u1F57\\u1F59\\u1F5B\\u1F5D\ \u1F5F-\\u1F70" 118 "\\u1E9E\\u1EA0-\\u1EF9\\u1F00-\\u1F15\\u1F18-\\u1F1D\\u1F20-\\u1F45 \\u1F48-\\u1F4D"
109 "\\u1F72\\u1F74\\u1F76\\u1F78\\u1F7A\\u1F7C\\u1F80-\\u1FB4\\u1FB6-\\ u1FBA\\u1FBC\\u1FC2-\\u1FC4" 119 "\\u1F50-\\u1F57\\u1F59\\u1F5B\\u1F5D\\u1F5F-\\u1F70\\u1F72\\u1F74\\ u1F76\\u1F78"
110 "\\u1FC6-\\u1FC8\\u1FCA\\u1FCC\\u1FD0-\\u1FD2\\u1FD6-\\u1FDA\\u1FE0- \\u1FE2\\u1FE4-\\u1FEA" 120 "\\u1F7A\\u1F7C\\u1F80-\\u1FB4\\u1FB6-\\u1FBA\\u1FBC\\u1FC2-\\u1FC4\ \u1FC6-\\u1FC8"
111 "\\u1FEC\\u1FF2-\\u1FF4\\u1FF6-\\u1FF8\\u1FFA\\u1FFC\\u2D27\\u2D2D\\ u2D80-\\u2D96\\u2DA0-\\u2DA6" 121 "\\u1FCA\\u1FCC\\u1FD0-\\u1FD2\\u1FD6-\\u1FDA\\u1FE0-\\u1FE2\\u1FE4- \\u1FEA\\u1FEC"
122 "\\u1FF2-\\u1FF4\\u1FF6-\\u1FF8\\u1FFA\\u1FFC\\u2D27\\u2D2D\\u2D80-\ \u2D96\\u2DA0-\\u2DA6"
112 "\\u2DA8-\\u2DAE\\u2DB0-\\u2DB6\\u2DB8-\\u2DBE\\u2DC0-\\u2DC6\\u2DC8 -\\u2DCE\\u2DD0-\\u2DD6" 123 "\\u2DA8-\\u2DAE\\u2DB0-\\u2DB6\\u2DB8-\\u2DBE\\u2DC0-\\u2DC6\\u2DC8 -\\u2DCE\\u2DD0-\\u2DD6"
113 "\\u2DD8-\\u2DDE\\u3005-\\u3007\\u3041-\\u3096\\u3099-\\u309A\\u309D -\\u309E\\u30A1-\\u30FA" 124 "\\u2DD8-\\u2DDE\\u3005-\\u3007\\u3041-\\u3096\\u3099-\\u309A\\u309D -\\u309E\\u30A1-\\u30FA"
114 "\\u30FC-\\u30FE\\u3105-\\u312D\\u31A0-\\u31BA\\u3400-\\u4DB5\\u4E00 -\\u9FCC\\uA660-\\uA661" 125 "\\u30FC-\\u30FE\\u3105-\\u312D\\u31A0-\\u31BA\\u3400-\\u4DB5\\u4E00 -\\u9FD5\\uA660-\\uA661"
115 "\\uA674-\\uA67B\\uA67F\\uA69F\\uA717-\\uA71F\\uA788\\uA78D-\\uA78E\ \uA790-\\uA793" 126 "\\uA674-\\uA67B\\uA67F\\uA69F\\uA717-\\uA71F\\uA788\\uA78D-\\uA78E\ \uA790-\\uA793"
116 "\\uA7A0-\\uA7AA\\uA7FA\\uA9CF\\uAA60-\\uAA76\\uAA7A-\\uAA7B\\uAB01- \\uAB06\\uAB09-\\uAB0E" 127 "\\uA7A0-\\uA7AA\\uA7FA\\uA9E7-\\uA9FE\\uAA60-\\uAA76\\uAA7A-\\uAA7F \\uAB01-\\uAB06"
117 "\\uAB11-\\uAB16\\uAB20-\\uAB26\\uAB28-\\uAB2E\\uAC00-\\uD7A3\\uFA0E -\\uFA0F\\uFA11" 128 "\\uAB09-\\uAB0E\\uAB11-\\uAB16\\uAB20-\\uAB26\\uAB28-\\uAB2E\\uAC00 -\\uD7A3\\uFA0E-\\uFA0F"
118 "\\uFA13-\\uFA14\\uFA1F\\uFA21\\uFA23-\\uFA24\\uFA27-\\uFA29\\U0001B 000-\\U0001B001\\U00020000-\\U0002A6D6" 129 "\\uFA11\\uFA13-\\uFA14\\uFA1F\\uFA21\\uFA23-\\uFA24\\uFA27-\\uFA29\ \U00020000-\\U0002A6D6"
119 "\\U0002A700-\\U0002B734\\U0002B740-\\U0002B81D]"; 130 "\\U0002A700-\\U0002B734\\U0002B740-\\U0002B81D\\U0002B820-\\U0002CE A1]";
131
120 gRecommendedSet = new UnicodeSet(UnicodeString(recommendedPat, -1, US_INV), status); 132 gRecommendedSet = new UnicodeSet(UnicodeString(recommendedPat, -1, US_INV), status);
133 gRecommendedSet->freeze();
121 gNfdNormalizer = Normalizer2::getNFDInstance(status); 134 gNfdNormalizer = Normalizer2::getNFDInstance(status);
122 ucln_i18n_registerCleanup(UCLN_I18N_SPOOF, uspoof_cleanup); 135 ucln_i18n_registerCleanup(UCLN_I18N_SPOOF, uspoof_cleanup);
123 } 136 }
124 137
138 static void U_CALLCONV initializeDefaultData(UErrorCode &status) {
139 gDefaultSpoofData = SpoofData::getDefault(status);
140 ucln_i18n_registerCleanup(UCLN_I18N_SPOOF, uspoof_cleanup);
141 }
142
143 U_CFUNC void uspoof_internalInitStatics(UErrorCode *status) {
144 umtx_initOnce(gSpoofInitStaticsOnce, &initializeStatics, *status);
145 }
125 146
126 U_CAPI USpoofChecker * U_EXPORT2 147 U_CAPI USpoofChecker * U_EXPORT2
127 uspoof_open(UErrorCode *status) { 148 uspoof_open(UErrorCode *status) {
149 umtx_initOnce(gSpoofInitStaticsOnce, &initializeStatics, *status);
150 umtx_initOnce(gSpoofInitDefaultOnce, &initializeDefaultData, *status);
128 if (U_FAILURE(*status)) { 151 if (U_FAILURE(*status)) {
129 return NULL; 152 return NULL;
130 } 153 }
131 umtx_initOnce(gSpoofInitOnce, &initializeStatics, *status); 154 SpoofImpl *si = new SpoofImpl(gDefaultSpoofData, *status);
132 SpoofImpl *si = new SpoofImpl(SpoofData::getDefault(*status), *status); 155 if (si) {
156 gDefaultSpoofData->addReference();
157 }
158 if (U_SUCCESS(*status) && si == NULL) {
159 *status = U_MEMORY_ALLOCATION_ERROR;
160 }
133 if (U_FAILURE(*status)) { 161 if (U_FAILURE(*status)) {
134 delete si; 162 delete si;
135 si = NULL; 163 si = NULL;
136 } 164 }
137 return reinterpret_cast<USpoofChecker *>(si); 165 return reinterpret_cast<USpoofChecker *>(si);
138 } 166 }
139 167
140 168
141 U_CAPI USpoofChecker * U_EXPORT2 169 U_CAPI USpoofChecker * U_EXPORT2
142 uspoof_openFromSerialized(const void *data, int32_t length, int32_t *pActualLeng th, 170 uspoof_openFromSerialized(const void *data, int32_t length, int32_t *pActualLeng th,
143 UErrorCode *status) { 171 UErrorCode *status) {
144 if (U_FAILURE(*status)) { 172 if (U_FAILURE(*status)) {
145 return NULL; 173 return NULL;
146 } 174 }
147 umtx_initOnce(gSpoofInitOnce, &initializeStatics, *status); 175 umtx_initOnce(gSpoofInitStaticsOnce, &initializeStatics, *status);
148 SpoofData *sd = new SpoofData(data, length, *status); 176 SpoofData *sd = new SpoofData(data, length, *status);
149 SpoofImpl *si = new SpoofImpl(sd, *status); 177 SpoofImpl *si = new SpoofImpl(sd, *status);
150 if (U_FAILURE(*status)) { 178 if (U_FAILURE(*status)) {
151 delete sd; 179 delete sd;
152 delete si; 180 delete si;
153 return NULL; 181 return NULL;
154 } 182 }
155 if (sd == NULL || si == NULL) { 183 if (sd == NULL || si == NULL) {
156 *status = U_MEMORY_ALLOCATION_ERROR; 184 *status = U_MEMORY_ALLOCATION_ERROR;
157 delete sd; 185 delete sd;
(...skipping 240 matching lines...) Expand 10 before | Expand all | Expand 10 after
398 } 426 }
399 int32_t flagsForSkeleton = This->fChecks & USPOOF_ANY_CASE; 427 int32_t flagsForSkeleton = This->fChecks & USPOOF_ANY_CASE;
400 428
401 int32_t result = 0; 429 int32_t result = 0;
402 IdentifierInfo *identifierInfo = This->getIdentifierInfo(*status); 430 IdentifierInfo *identifierInfo = This->getIdentifierInfo(*status);
403 if (U_FAILURE(*status)) { 431 if (U_FAILURE(*status)) {
404 return 0; 432 return 0;
405 } 433 }
406 identifierInfo->setIdentifier(id1, *status); 434 identifierInfo->setIdentifier(id1, *status);
407 int32_t id1ScriptCount = identifierInfo->getScriptCount(); 435 int32_t id1ScriptCount = identifierInfo->getScriptCount();
436 int32_t id1FirstScript = identifierInfo->getScripts()->nextSetBit(0);
408 identifierInfo->setIdentifier(id2, *status); 437 identifierInfo->setIdentifier(id2, *status);
409 int32_t id2ScriptCount = identifierInfo->getScriptCount(); 438 int32_t id2ScriptCount = identifierInfo->getScriptCount();
439 int32_t id2FirstScript = identifierInfo->getScripts()->nextSetBit(0);
410 This->releaseIdentifierInfo(identifierInfo); 440 This->releaseIdentifierInfo(identifierInfo);
411 identifierInfo = NULL; 441 identifierInfo = NULL;
412 442
413 if (This->fChecks & USPOOF_SINGLE_SCRIPT_CONFUSABLE) { 443 if (This->fChecks & USPOOF_SINGLE_SCRIPT_CONFUSABLE) {
414 UnicodeString id1Skeleton; 444 UnicodeString id1Skeleton;
415 UnicodeString id2Skeleton; 445 UnicodeString id2Skeleton;
416 if (id1ScriptCount <= 1 && id2ScriptCount <= 1) { 446 if (id1ScriptCount <= 1 && id2ScriptCount <= 1 && id1FirstScript == id2F irstScript) {
417 flagsForSkeleton |= USPOOF_SINGLE_SCRIPT_CONFUSABLE; 447 flagsForSkeleton |= USPOOF_SINGLE_SCRIPT_CONFUSABLE;
418 uspoof_getSkeletonUnicodeString(sc, flagsForSkeleton, id1, id1Skelet on, status); 448 uspoof_getSkeletonUnicodeString(sc, flagsForSkeleton, id1, id1Skelet on, status);
419 uspoof_getSkeletonUnicodeString(sc, flagsForSkeleton, id2, id2Skelet on, status); 449 uspoof_getSkeletonUnicodeString(sc, flagsForSkeleton, id2, id2Skelet on, status);
420 if (id1Skeleton == id2Skeleton) { 450 if (id1Skeleton == id2Skeleton) {
421 result |= USPOOF_SINGLE_SCRIPT_CONFUSABLE; 451 result |= USPOOF_SINGLE_SCRIPT_CONFUSABLE;
422 } 452 }
423 } 453 }
424 } 454 }
425 455
426 if (result & USPOOF_SINGLE_SCRIPT_CONFUSABLE) { 456 if (result & USPOOF_SINGLE_SCRIPT_CONFUSABLE) {
(...skipping 308 matching lines...) Expand 10 before | Expand all | Expand 10 after
735 if (capacity < dataSize) { 765 if (capacity < dataSize) {
736 *status = U_BUFFER_OVERFLOW_ERROR; 766 *status = U_BUFFER_OVERFLOW_ERROR;
737 return dataSize; 767 return dataSize;
738 } 768 }
739 uprv_memcpy(buf, This->fSpoofData->fRawData, dataSize); 769 uprv_memcpy(buf, This->fSpoofData->fRawData, dataSize);
740 return dataSize; 770 return dataSize;
741 } 771 }
742 772
743 U_CAPI const USet * U_EXPORT2 773 U_CAPI const USet * U_EXPORT2
744 uspoof_getInclusionSet(UErrorCode *status) { 774 uspoof_getInclusionSet(UErrorCode *status) {
745 umtx_initOnce(gSpoofInitOnce, &initializeStatics, *status); 775 umtx_initOnce(gSpoofInitStaticsOnce, &initializeStatics, *status);
746 return gInclusionSet->toUSet(); 776 return gInclusionSet->toUSet();
747 } 777 }
748 778
749 U_CAPI const USet * U_EXPORT2 779 U_CAPI const USet * U_EXPORT2
750 uspoof_getRecommendedSet(UErrorCode *status) { 780 uspoof_getRecommendedSet(UErrorCode *status) {
751 umtx_initOnce(gSpoofInitOnce, &initializeStatics, *status); 781 umtx_initOnce(gSpoofInitStaticsOnce, &initializeStatics, *status);
752 return gRecommendedSet->toUSet(); 782 return gRecommendedSet->toUSet();
753 } 783 }
754 784
755 U_I18N_API const UnicodeSet * U_EXPORT2 785 U_I18N_API const UnicodeSet * U_EXPORT2
756 uspoof_getInclusionUnicodeSet(UErrorCode *status) { 786 uspoof_getInclusionUnicodeSet(UErrorCode *status) {
757 umtx_initOnce(gSpoofInitOnce, &initializeStatics, *status); 787 umtx_initOnce(gSpoofInitStaticsOnce, &initializeStatics, *status);
758 return gInclusionSet; 788 return gInclusionSet;
759 } 789 }
760 790
761 U_I18N_API const UnicodeSet * U_EXPORT2 791 U_I18N_API const UnicodeSet * U_EXPORT2
762 uspoof_getRecommendedUnicodeSet(UErrorCode *status) { 792 uspoof_getRecommendedUnicodeSet(UErrorCode *status) {
763 umtx_initOnce(gSpoofInitOnce, &initializeStatics, *status); 793 umtx_initOnce(gSpoofInitStaticsOnce, &initializeStatics, *status);
764 return gRecommendedSet; 794 return gRecommendedSet;
765 } 795 }
766 796
767 797
768 798
769 #endif // !UCONFIG_NO_NORMALIZATION 799 #endif // !UCONFIG_NO_NORMALIZATION
OLDNEW
« no previous file with comments | « source/i18n/usearch.cpp ('k') | source/i18n/uspoof_build.cpp » ('j') | no next file with comments »

Powered by Google App Engine
This is Rietveld 408576698