ICU 55.1  55.1
uspoof.h
Go to the documentation of this file.
1 /*
2 ***************************************************************************
3 * Copyright (C) 2008-2015, International Business Machines Corporation
4 * and others. All Rights Reserved.
5 ***************************************************************************
6 * file name: uspoof.h
7 * encoding: US-ASCII
8 * tab size: 8 (not used)
9 * indentation:4
10 *
11 * created on: 2008Feb13
12 * created by: Andy Heninger
13 *
14 * Unicode Spoof Detection
15 */
16 
17 #ifndef USPOOF_H
18 #define USPOOF_H
19 
20 #include "unicode/utypes.h"
21 #include "unicode/uset.h"
22 #include "unicode/parseerr.h"
23 #include "unicode/localpointer.h"
24 
25 #if !UCONFIG_NO_NORMALIZATION
26 
27 
28 #if U_SHOW_CPLUSPLUS_API
29 #include "unicode/unistr.h"
30 #include "unicode/uniset.h"
31 #endif
32 
33 
152 struct USpoofChecker;
153 typedef struct USpoofChecker USpoofChecker;
162 typedef enum USpoofChecks {
168  USPOOF_SINGLE_SCRIPT_CONFUSABLE = 1,
169 
178  USPOOF_MIXED_SCRIPT_CONFUSABLE = 2,
179 
189  USPOOF_WHOLE_SCRIPT_CONFUSABLE = 4,
190 
197  USPOOF_ANY_CASE = 8,
198 
212  USPOOF_RESTRICTION_LEVEL = 16,
213 
214 #ifndef U_HIDE_DEPRECATED_API
215 
220  USPOOF_SINGLE_SCRIPT = USPOOF_RESTRICTION_LEVEL,
221 #endif /* U_HIDE_DEPRECATED_API */
222 
229  USPOOF_INVISIBLE = 32,
230 
235  USPOOF_CHAR_LIMIT = 64,
236 
243  USPOOF_MIXED_NUMBERS = 128,
244 
250  USPOOF_ALL_CHECKS = 0xFFFF,
251 
264  USPOOF_AUX_INFO = 0x40000000
265 
266  } USpoofChecks;
267 
268 
274  typedef enum URestrictionLevel {
280  USPOOF_ASCII = 0x10000000,
286  USPOOF_SINGLE_SCRIPT_RESTRICTIVE = 0x20000000,
294  USPOOF_HIGHLY_RESTRICTIVE = 0x30000000,
300  USPOOF_MODERATELY_RESTRICTIVE = 0x40000000,
306  USPOOF_MINIMALLY_RESTRICTIVE = 0x50000000,
312  USPOOF_UNRESTRICTIVE = 0x60000000,
318  USPOOF_RESTRICTION_LEVEL_MASK = 0x7F000000
319  } URestrictionLevel;
320 
331 U_STABLE USpoofChecker * U_EXPORT2
332 uspoof_open(UErrorCode *status);
333 
334 
356 U_STABLE USpoofChecker * U_EXPORT2
357 uspoof_openFromSerialized(const void *data, int32_t length, int32_t *pActualLength,
358  UErrorCode *pErrorCode);
359 
391 U_STABLE USpoofChecker * U_EXPORT2
392 uspoof_openFromSource(const char *confusables, int32_t confusablesLen,
393  const char *confusablesWholeScript, int32_t confusablesWholeScriptLen,
394  int32_t *errType, UParseError *pe, UErrorCode *status);
395 
396 
402 U_STABLE void U_EXPORT2
403 uspoof_close(USpoofChecker *sc);
404 
405 #if U_SHOW_CPLUSPLUS_API
406 
407 U_NAMESPACE_BEGIN
408 
418 U_DEFINE_LOCAL_OPEN_POINTER(LocalUSpoofCheckerPointer, USpoofChecker, uspoof_close);
419 
420 U_NAMESPACE_END
421 
422 #endif
423 
433 U_STABLE USpoofChecker * U_EXPORT2
434 uspoof_clone(const USpoofChecker *sc, UErrorCode *status);
435 
436 
449 U_STABLE void U_EXPORT2
450 uspoof_setChecks(USpoofChecker *sc, int32_t checks, UErrorCode *status);
451 
463 U_STABLE int32_t U_EXPORT2
464 uspoof_getChecks(const USpoofChecker *sc, UErrorCode *status);
465 
474 U_STABLE void U_EXPORT2
475 uspoof_setRestrictionLevel(USpoofChecker *sc, URestrictionLevel restrictionLevel);
476 
477 
485 U_STABLE URestrictionLevel U_EXPORT2
486 uspoof_getRestrictionLevel(const USpoofChecker *sc);
487 
530 U_STABLE void U_EXPORT2
531 uspoof_setAllowedLocales(USpoofChecker *sc, const char *localesList, UErrorCode *status);
532 
554 U_STABLE const char * U_EXPORT2
555 uspoof_getAllowedLocales(USpoofChecker *sc, UErrorCode *status);
556 
557 
576 U_STABLE void U_EXPORT2
577 uspoof_setAllowedChars(USpoofChecker *sc, const USet *chars, UErrorCode *status);
578 
579 
600 U_STABLE const USet * U_EXPORT2
601 uspoof_getAllowedChars(const USpoofChecker *sc, UErrorCode *status);
602 
603 
604 #if U_SHOW_CPLUSPLUS_API
605 
623 U_STABLE void U_EXPORT2
624 uspoof_setAllowedUnicodeSet(USpoofChecker *sc, const icu::UnicodeSet *chars, UErrorCode *status);
625 
626 
647 U_STABLE const icu::UnicodeSet * U_EXPORT2
648 uspoof_getAllowedUnicodeSet(const USpoofChecker *sc, UErrorCode *status);
649 #endif
650 
651 
678 U_STABLE int32_t U_EXPORT2
679 uspoof_check(const USpoofChecker *sc,
680  const UChar *id, int32_t length,
681  int32_t *position,
682  UErrorCode *status);
683 
684 
712 U_STABLE int32_t U_EXPORT2
713 uspoof_checkUTF8(const USpoofChecker *sc,
714  const char *id, int32_t length,
715  int32_t *position,
716  UErrorCode *status);
717 
718 
719 #if U_SHOW_CPLUSPLUS_API
720 
743 U_STABLE int32_t U_EXPORT2
744 uspoof_checkUnicodeString(const USpoofChecker *sc,
745  const icu::UnicodeString &id,
746  int32_t *position,
747  UErrorCode *status);
748 
749 #endif
750 
751 
791 U_STABLE int32_t U_EXPORT2
792 uspoof_areConfusable(const USpoofChecker *sc,
793  const UChar *id1, int32_t length1,
794  const UChar *id2, int32_t length2,
795  UErrorCode *status);
796 
797 
798 
824 U_STABLE int32_t U_EXPORT2
825 uspoof_areConfusableUTF8(const USpoofChecker *sc,
826  const char *id1, int32_t length1,
827  const char *id2, int32_t length2,
828  UErrorCode *status);
829 
830 
831 
832 
833 #if U_SHOW_CPLUSPLUS_API
834 
855 U_STABLE int32_t U_EXPORT2
856 uspoof_areConfusableUnicodeString(const USpoofChecker *sc,
857  const icu::UnicodeString &s1,
858  const icu::UnicodeString &s2,
859  UErrorCode *status);
860 #endif
861 
862 
895 U_STABLE int32_t U_EXPORT2
896 uspoof_getSkeleton(const USpoofChecker *sc,
897  uint32_t type,
898  const UChar *id, int32_t length,
899  UChar *dest, int32_t destCapacity,
900  UErrorCode *status);
901 
937 U_STABLE int32_t U_EXPORT2
938 uspoof_getSkeletonUTF8(const USpoofChecker *sc,
939  uint32_t type,
940  const char *id, int32_t length,
941  char *dest, int32_t destCapacity,
942  UErrorCode *status);
943 
944 #if U_SHOW_CPLUSPLUS_API
945 
970 U_I18N_API icu::UnicodeString & U_EXPORT2
971 uspoof_getSkeletonUnicodeString(const USpoofChecker *sc,
972  uint32_t type,
973  const icu::UnicodeString &id,
974  icu::UnicodeString &dest,
975  UErrorCode *status);
976 #endif /* U_SHOW_CPLUSPLUS_API */
977 
989 U_STABLE const USet * U_EXPORT2
990 uspoof_getInclusionSet(UErrorCode *status);
991 
1003 U_STABLE const USet * U_EXPORT2
1004 uspoof_getRecommendedSet(UErrorCode *status);
1005 
1006 #if U_SHOW_CPLUSPLUS_API
1007 
1019 U_STABLE const icu::UnicodeSet * U_EXPORT2
1020 uspoof_getInclusionUnicodeSet(UErrorCode *status);
1021 
1033 U_STABLE const icu::UnicodeSet * U_EXPORT2
1034 uspoof_getRecommendedUnicodeSet(UErrorCode *status);
1035 
1036 #endif /* U_SHOW_CPLUSPLUS_API */
1037 
1060 U_STABLE int32_t U_EXPORT2
1061 uspoof_serialize(USpoofChecker *sc,
1062  void *data, int32_t capacity,
1063  UErrorCode *status);
1064 
1065 
1066 #endif
1067 
1068 #endif /* USPOOF_H */
USPOOF_ASCII
@ USPOOF_ASCII
Only ASCII characters: U+0000..U+007F.
Definition: uspoof.h:280
USpoofChecks
USpoofChecks
Enum for the kinds of checks that USpoofChecker can perform.
Definition: uspoof.h:162
parseerr.h
C API: Parse Error Information.
uspoof_getRestrictionLevel
URestrictionLevel uspoof_getRestrictionLevel(const USpoofChecker *sc)
Get the Restriction Level that will be tested if the checks include RESTRICTION_LEVEL.
USPOOF_RESTRICTION_LEVEL_MASK
@ USPOOF_RESTRICTION_LEVEL_MASK
Mask for selecting the Restriction Level bits from the return value of uspoof_check().
Definition: uspoof.h:318
utypes.h
Basic definitions for ICU, for both C and C++ APIs.
icu::UnicodeSet
A mutable set of Unicode characters and multicharacter strings.
Definition: uniset.h:276
USPOOF_ANY_CASE
@ USPOOF_ANY_CASE
Any Case Modifier for confusable identifier tests.
Definition: uspoof.h:197
USPOOF_UNRESTRICTIVE
@ USPOOF_UNRESTRICTIVE
Any valid identifiers, including characters outside of the Identifier Profile.
Definition: uspoof.h:312
U_I18N_API
#define U_I18N_API
Set to export library symbols from inside the i18n library, and to import them from outside.
Definition: utypes.h:358
uspoof_getSkeletonUnicodeString
U_I18N_API icu::UnicodeString & uspoof_getSkeletonUnicodeString(const USpoofChecker *sc, uint32_t type, const icu::UnicodeString &id, icu::UnicodeString &dest, UErrorCode *status)
Get the "skeleton" for an identifier.
uspoof_areConfusable
int32_t uspoof_areConfusable(const USpoofChecker *sc, const UChar *id1, int32_t length1, const UChar *id2, int32_t length2, UErrorCode *status)
Check the whether two specified strings are visually confusable.
uspoof_getRecommendedUnicodeSet
const icu::UnicodeSet * uspoof_getRecommendedUnicodeSet(UErrorCode *status)
Get the set of characters from Recommended Scripts for Inclusion in Identifiers, as defined in Unicod...
uspoof_getChecks
int32_t uspoof_getChecks(const USpoofChecker *sc, UErrorCode *status)
Get the set of checks that this Spoof Checker has been configured to perform.
uspoof_getAllowedUnicodeSet
const icu::UnicodeSet * uspoof_getAllowedUnicodeSet(const USpoofChecker *sc, UErrorCode *status)
Get a UnicodeSet for the characters permitted in an identifier.
uspoof_getRecommendedSet
const USet * uspoof_getRecommendedSet(UErrorCode *status)
Get the set of characters from Recommended Scripts for Inclusion in Identifiers, as defined in Unicod...
uspoof_getAllowedChars
const USet * uspoof_getAllowedChars(const USpoofChecker *sc, UErrorCode *status)
Get a USet for the characters permitted in an identifier.
USPOOF_HIGHLY_RESTRICTIVE
@ USPOOF_HIGHLY_RESTRICTIVE
All characters in each identifier must be from a single script, or from the combinations: Latin + Han...
Definition: uspoof.h:294
uspoof_setAllowedUnicodeSet
void uspoof_setAllowedUnicodeSet(USpoofChecker *sc, const icu::UnicodeSet *chars, UErrorCode *status)
Limit the acceptable characters to those specified by a Unicode Set.
uspoof_getInclusionSet
const USet * uspoof_getInclusionSet(UErrorCode *status)
Get the set of Candidate Characters for Inclusion in Identifiers, as defined in Unicode UAX #31,...
uspoof_checkUTF8
int32_t uspoof_checkUTF8(const USpoofChecker *sc, const char *id, int32_t length, int32_t *position, UErrorCode *status)
Check the specified string for possible security issues.
UParseError
A UParseError struct is used to returned detailed information about parsing errors.
Definition: parseerr.h:56
USPOOF_CHAR_LIMIT
@ USPOOF_CHAR_LIMIT
Check that an identifier contains only characters from a specified set of acceptable characters.
Definition: uspoof.h:235
USPOOF_ALL_CHECKS
@ USPOOF_ALL_CHECKS
Enable all spoof checks.
Definition: uspoof.h:250
uspoof_serialize
int32_t uspoof_serialize(USpoofChecker *sc, void *data, int32_t capacity, UErrorCode *status)
Serialize the data for a spoof detector into a chunk of memory.
icu::UnicodeString
UnicodeString is a string class that stores Unicode characters directly and provides similar function...
Definition: unistr.h:246
uspoof_setRestrictionLevel
void uspoof_setRestrictionLevel(USpoofChecker *sc, URestrictionLevel restrictionLevel)
Set the loosest restriction level allowed.
uspoof_checkUnicodeString
int32_t uspoof_checkUnicodeString(const USpoofChecker *sc, const icu::UnicodeString &id, int32_t *position, UErrorCode *status)
Check the specified string for possible security issues.
UErrorCode
UErrorCode
Error code to replace exception handling, so that the code is compatible with all C++ compilers,...
Definition: utypes.h:476
USPOOF_SINGLE_SCRIPT_RESTRICTIVE
@ USPOOF_SINGLE_SCRIPT_RESTRICTIVE
All characters in each identifier must be from a single script.
Definition: uspoof.h:286
uspoof_openFromSource
USpoofChecker * uspoof_openFromSource(const char *confusables, int32_t confusablesLen, const char *confusablesWholeScript, int32_t confusablesWholeScriptLen, int32_t *errType, UParseError *pe, UErrorCode *status)
Open a Spoof Checker from the source form of the spoof data.
uspoof_getInclusionUnicodeSet
const icu::UnicodeSet * uspoof_getInclusionUnicodeSet(UErrorCode *status)
Get the set of Candidate Characters for Inclusion in Identifiers, as defined in Unicode UAX #31,...
USPOOF_SINGLE_SCRIPT
@ USPOOF_SINGLE_SCRIPT
Check that an identifier contains only characters from a single script (plus chars from the common an...
Definition: uspoof.h:220
uspoof_open
USpoofChecker * uspoof_open(UErrorCode *status)
Create a Unicode Spoof Checker, configured to perform all checks except for USPOOF_LOCALE_LIMIT and U...
uspoof_getAllowedLocales
const char * uspoof_getAllowedLocales(USpoofChecker *sc, UErrorCode *status)
Get a list of locales for the scripts that are acceptable in strings to be checked.
U_DEFINE_LOCAL_OPEN_POINTER
#define U_DEFINE_LOCAL_OPEN_POINTER(LocalPointerClassName, Type, closeFunction)
"Smart pointer" definition macro, deletes objects via the closeFunction.
Definition: localpointer.h:356
USPOOF_SINGLE_SCRIPT_CONFUSABLE
@ USPOOF_SINGLE_SCRIPT_CONFUSABLE
Single script confusable test.
Definition: uspoof.h:168
USPOOF_MIXED_NUMBERS
@ USPOOF_MIXED_NUMBERS
Check that an identifier does not include decimal digits from more than one numbering system.
Definition: uspoof.h:243
USPOOF_AUX_INFO
@ USPOOF_AUX_INFO
Enable the return of auxillary (non-error) information in the upper bits of the check results value.
Definition: uspoof.h:264
uspoof_areConfusableUnicodeString
int32_t uspoof_areConfusableUnicodeString(const USpoofChecker *sc, const icu::UnicodeString &s1, const icu::UnicodeString &s2, UErrorCode *status)
Check the whether two specified strings are visually confusable.
uspoof_getSkeleton
int32_t uspoof_getSkeleton(const USpoofChecker *sc, uint32_t type, const UChar *id, int32_t length, UChar *dest, int32_t destCapacity, UErrorCode *status)
Get the "skeleton" for an identifier.
uspoof_clone
USpoofChecker * uspoof_clone(const USpoofChecker *sc, UErrorCode *status)
Clone a Spoof Checker.
uset.h
C API: Unicode Set.
localpointer.h
C++ API: "Smart pointers" for use with and in ICU4C C++ code.
uspoof_setAllowedChars
void uspoof_setAllowedChars(USpoofChecker *sc, const USet *chars, UErrorCode *status)
Limit the acceptable characters to those specified by a Unicode Set.
uspoof_setAllowedLocales
void uspoof_setAllowedLocales(USpoofChecker *sc, const char *localesList, UErrorCode *status)
Limit characters that are acceptable in identifiers being checked to those normally used with the lan...
uspoof_openFromSerialized
USpoofChecker * uspoof_openFromSerialized(const void *data, int32_t length, int32_t *pActualLength, UErrorCode *pErrorCode)
Open a Spoof checker from its serialized form, stored in 32-bit-aligned memory.
USPOOF_MINIMALLY_RESTRICTIVE
@ USPOOF_MINIMALLY_RESTRICTIVE
Allow arbitrary mixtures of scripts.
Definition: uspoof.h:306
U_STABLE
#define U_STABLE
This is used to declare a function as a stable public ICU C API.
Definition: umachine.h:109
USpoofChecker
struct USpoofChecker USpoofChecker
typedef for C of USpoofChecker
Definition: uspoof.h:153
UChar
uint16_t UChar
Define UChar to be UCHAR_TYPE, if that is #defined (for example, to char16_t), or wchar_t if that is ...
Definition: umachine.h:312
USPOOF_MODERATELY_RESTRICTIVE
@ USPOOF_MODERATELY_RESTRICTIVE
Allow Latin with other scripts except Cyrillic, Greek, Cherokee Otherwise, the same as Highly Restric...
Definition: uspoof.h:300
USet
struct USet USet
Definition: ucnv.h:67
uspoof_check
int32_t uspoof_check(const USpoofChecker *sc, const UChar *id, int32_t length, int32_t *position, UErrorCode *status)
Check the specified string for possible security issues.
uspoof_setChecks
void uspoof_setChecks(USpoofChecker *sc, int32_t checks, UErrorCode *status)
Specify the set of checks that will be performed by the check functions of this Spoof Checker.
uspoof_areConfusableUTF8
int32_t uspoof_areConfusableUTF8(const USpoofChecker *sc, const char *id1, int32_t length1, const char *id2, int32_t length2, UErrorCode *status)
Check the whether two specified strings are visually confusable.
uniset.h
C++ API: Unicode Set.
U_NAMESPACE_END
#define U_NAMESPACE_END
This is used to end a declaration of a public ICU C++ API.
Definition: uversion.h:130
U_NAMESPACE_BEGIN
#define U_NAMESPACE_BEGIN
This is used to begin a declaration of a public ICU C++ API.
Definition: uversion.h:129
uspoof_close
void uspoof_close(USpoofChecker *sc)
Close a Spoof Checker, freeing any memory that was being held by its implementation.
LocalUSpoofCheckerPointer
"Smart pointer" class, closes a USpoofChecker via uspoof_close().
URestrictionLevel
URestrictionLevel
Constants from UAX #39 for use in setRestrictionLevel(), and for returned identifier restriction leve...
Definition: uspoof.h:274
uspoof_getSkeletonUTF8
int32_t uspoof_getSkeletonUTF8(const USpoofChecker *sc, uint32_t type, const char *id, int32_t length, char *dest, int32_t destCapacity, UErrorCode *status)
Get the "skeleton" for an identifier.
unistr.h
C++ API: Unicode String.
USPOOF_RESTRICTION_LEVEL
@ USPOOF_RESTRICTION_LEVEL
Check that an identifier is no looser than the specified RestrictionLevel.
Definition: uspoof.h:212
USPOOF_WHOLE_SCRIPT_CONFUSABLE
@ USPOOF_WHOLE_SCRIPT_CONFUSABLE
Whole script confusable test.
Definition: uspoof.h:189
USPOOF_INVISIBLE
@ USPOOF_INVISIBLE
Check an identifier for the presence of invisible characters, such as zero-width spaces,...
Definition: uspoof.h:229
USPOOF_MIXED_SCRIPT_CONFUSABLE
@ USPOOF_MIXED_SCRIPT_CONFUSABLE
Mixed script confusable test.
Definition: uspoof.h:178
Generated by doxygen 1.8.20