ICU 54.1  54.1
uspoof.h
Go to the documentation of this file.
1 /*
2 ***************************************************************************
3 * Copyright (C) 2008-2014, 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 
144 struct USpoofChecker;
145 typedef struct USpoofChecker USpoofChecker;
154 typedef enum USpoofChecks {
161 
171 
182 
190 
205 
206 #ifndef U_HIDE_DEPRECATED_API
207 
213 #endif /* U_HIDE_DEPRECATED_API */
214 
222 
228 
236 
243 
256  USPOOF_AUX_INFO = 0x40000000
257 
258  } USpoofChecks;
259 
260 
266  typedef enum URestrictionLevel {
272  USPOOF_ASCII = 0x10000000,
273 #ifndef U_HIDE_DRAFT_API
274 
280 #endif /* U_HIDE_DRAFT_API */
281 
306  USPOOF_UNRESTRICTIVE = 0x60000000,
307 #ifndef U_HIDE_DRAFT_API
308 
314 #endif /* U_HIDE_DRAFT_API */
316 
327 U_STABLE USpoofChecker * U_EXPORT2
328 uspoof_open(UErrorCode *status);
329 
330 
352 U_STABLE USpoofChecker * U_EXPORT2
353 uspoof_openFromSerialized(const void *data, int32_t length, int32_t *pActualLength,
354  UErrorCode *pErrorCode);
355 
387 U_STABLE USpoofChecker * U_EXPORT2
388 uspoof_openFromSource(const char *confusables, int32_t confusablesLen,
389  const char *confusablesWholeScript, int32_t confusablesWholeScriptLen,
390  int32_t *errType, UParseError *pe, UErrorCode *status);
391 
392 
398 U_STABLE void U_EXPORT2
400 
401 #if U_SHOW_CPLUSPLUS_API
402 
404 
415 
417 
418 #endif
419 
429 U_STABLE USpoofChecker * U_EXPORT2
430 uspoof_clone(const USpoofChecker *sc, UErrorCode *status);
431 
432 
445 U_STABLE void U_EXPORT2
446 uspoof_setChecks(USpoofChecker *sc, int32_t checks, UErrorCode *status);
447 
459 U_STABLE int32_t U_EXPORT2
460 uspoof_getChecks(const USpoofChecker *sc, UErrorCode *status);
461 
470 U_STABLE void U_EXPORT2
472 
473 
481 U_STABLE URestrictionLevel U_EXPORT2
483 
526 U_STABLE void U_EXPORT2
527 uspoof_setAllowedLocales(USpoofChecker *sc, const char *localesList, UErrorCode *status);
528 
550 U_STABLE const char * U_EXPORT2
552 
553 
572 U_STABLE void U_EXPORT2
573 uspoof_setAllowedChars(USpoofChecker *sc, const USet *chars, UErrorCode *status);
574 
575 
596 U_STABLE const USet * U_EXPORT2
598 
599 
600 #if U_SHOW_CPLUSPLUS_API
601 
619 U_STABLE void U_EXPORT2
621 
622 
643 U_STABLE const icu::UnicodeSet * U_EXPORT2
645 #endif
646 
647 
674 U_STABLE int32_t U_EXPORT2
675 uspoof_check(const USpoofChecker *sc,
676  const UChar *id, int32_t length,
677  int32_t *position,
678  UErrorCode *status);
679 
680 
708 U_STABLE int32_t U_EXPORT2
710  const char *id, int32_t length,
711  int32_t *position,
712  UErrorCode *status);
713 
714 
715 #if U_SHOW_CPLUSPLUS_API
716 
739 U_STABLE int32_t U_EXPORT2
741  const icu::UnicodeString &id,
742  int32_t *position,
743  UErrorCode *status);
744 
745 #endif
746 
747 
787 U_STABLE int32_t U_EXPORT2
789  const UChar *id1, int32_t length1,
790  const UChar *id2, int32_t length2,
791  UErrorCode *status);
792 
793 
794 
820 U_STABLE int32_t U_EXPORT2
822  const char *id1, int32_t length1,
823  const char *id2, int32_t length2,
824  UErrorCode *status);
825 
826 
827 
828 
829 #if U_SHOW_CPLUSPLUS_API
830 
851 U_STABLE int32_t U_EXPORT2
853  const icu::UnicodeString &s1,
854  const icu::UnicodeString &s2,
855  UErrorCode *status);
856 #endif
857 
858 
891 U_STABLE int32_t U_EXPORT2
893  uint32_t type,
894  const UChar *id, int32_t length,
895  UChar *dest, int32_t destCapacity,
896  UErrorCode *status);
897 
933 U_STABLE int32_t U_EXPORT2
935  uint32_t type,
936  const char *id, int32_t length,
937  char *dest, int32_t destCapacity,
938  UErrorCode *status);
939 
940 #if U_SHOW_CPLUSPLUS_API
941 
966 U_I18N_API icu::UnicodeString & U_EXPORT2
968  uint32_t type,
969  const icu::UnicodeString &id,
970  icu::UnicodeString &dest,
971  UErrorCode *status);
972 #endif /* U_SHOW_CPLUSPLUS_API */
973 
985 U_STABLE const USet * U_EXPORT2
987 
999 U_STABLE const USet * U_EXPORT2
1001 
1002 #if U_SHOW_CPLUSPLUS_API
1003 
1015 U_STABLE const icu::UnicodeSet * U_EXPORT2
1017 
1029 U_STABLE const icu::UnicodeSet * U_EXPORT2
1031 
1032 #endif /* U_SHOW_CPLUSPLUS_API */
1033 
1056 U_STABLE int32_t U_EXPORT2
1058  void *data, int32_t capacity,
1059  UErrorCode *status);
1060 
1061 
1062 #endif
1063 
1064 #endif /* USPOOF_H */
const USet * uspoof_getInclusionSet(UErrorCode *status)
Get the set of Candidate Characters for Inclusion in Identifiers, as defined in Unicode UAX #31...
const icu::UnicodeSet * uspoof_getInclusionUnicodeSet(UErrorCode *status)
Get the set of Candidate Characters for Inclusion in Identifiers, as defined in Unicode UAX #31...
int32_t uspoof_checkUnicodeString(const USpoofChecker *sc, const icu::UnicodeString &id, int32_t *position, UErrorCode *status)
Check the specified string for possible security issues.
void uspoof_close(USpoofChecker *sc)
Close a Spoof Checker, freeing any memory that was being held by its implementation.
URestrictionLevel
Constants from UAX #39 for use in setRestrictionLevel(), and for returned identifier restriction leve...
Definition: uspoof.h:266
Check an identifier for the presence of invisible characters, such as zero-width spaces, or character sequences that are likely not to display, such as multiple occurrences of the same non-spacing mark.
Definition: uspoof.h:221
Check that an identifier contains only characters from a single script (plus chars from the common an...
Definition: uspoof.h:212
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...
USpoofChecker * uspoof_clone(const USpoofChecker *sc, UErrorCode *status)
Clone a Spoof Checker.
Only ASCII characters: U+0000..U+007F.
Definition: uspoof.h:272
Check that an identifier contains only characters from a specified set of acceptable characters...
Definition: uspoof.h:227
Allow arbitrary mixtures of scripts.
Definition: uspoof.h:300
USpoofChecks
Enum for the kinds of checks that USpoofChecker can perform.
Definition: uspoof.h:154
C++ API: Unicode String.
USpoofChecker * uspoof_open(UErrorCode *status)
Create a Unicode Spoof Checker, configured to perform all checks except for USPOOF_LOCALE_LIMIT and U...
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.
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...
int32_t uspoof_getChecks(const USpoofChecker *sc, UErrorCode *status)
Get the set of checks that this Spoof Checker has been configured to perform.
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.
URestrictionLevel uspoof_getRestrictionLevel(const USpoofChecker *sc)
Get the Restriction Level that will be tested if the checks include RESTRICTION_LEVEL.
const icu::UnicodeSet * uspoof_getAllowedUnicodeSet(const USpoofChecker *sc, UErrorCode *status)
Get a UnicodeSet for the characters permitted in an identifier.
C API: Unicode Set.
Enable all spoof checks.
Definition: uspoof.h:242
#define U_I18N_API
Set to export library symbols from inside the i18n library, and to import them from outside...
Definition: utypes.h:358
const icu::UnicodeSet * uspoof_getRecommendedUnicodeSet(UErrorCode *status)
Get the set of characters from Recommended Scripts for Inclusion in Identifiers, as defined in Unicod...
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.
#define U_NAMESPACE_BEGIN
This is used to begin a declaration of a public ICU C++ API.
Definition: uversion.h:129
Any Case Modifier for confusable identifier tests.
Definition: uspoof.h:189
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.
Whole script confusable test.
Definition: uspoof.h:181
#define U_DEFINE_LOCAL_OPEN_POINTER(LocalPointerClassName, Type, closeFunction)
"Smart pointer" definition macro, deletes objects via the closeFunction.
Definition: localpointer.h:290
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.
C++ API: "Smart pointers" for use with and in ICU4C C++ code.
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.
Allow Latin with other scripts except Cyrillic, Greek, Cherokee Otherwise, the same as Highly Restric...
Definition: uspoof.h:294
Check that an identifier does not include decimal digits from more than one numbering system...
Definition: uspoof.h:235
void uspoof_setAllowedChars(USpoofChecker *sc, const USet *chars, UErrorCode *status)
Limit the acceptable characters to those specified by a Unicode Set.
All characters in each identifier must be from a single script.
Definition: uspoof.h:279
All characters in each identifier must be from a single script, or from the combinations: Latin + Han...
Definition: uspoof.h:288
const USet * uspoof_getAllowedChars(const USpoofChecker *sc, UErrorCode *status)
Get a USet for the characters permitted in an identifier.
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.
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.
A mutable set of Unicode characters and multicharacter strings.
Definition: uniset.h:276
struct USpoofChecker USpoofChecker
typedef for C of USpoofChecker
Definition: uspoof.h:145
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.
"Smart pointer" class, closes a USpoofChecker via uspoof_close().
void uspoof_setRestrictionLevel(USpoofChecker *sc, URestrictionLevel restrictionLevel)
Set the loosest restriction level allowed.
Single script confusable test.
Definition: uspoof.h:160
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
#define U_NAMESPACE_END
This is used to end a declaration of a public ICU C++ API.
Definition: uversion.h:130
struct USet USet
Definition: ucnv.h:67
C API: Parse Error Information.
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.
const char * uspoof_getAllowedLocales(USpoofChecker *sc, UErrorCode *status)
Get a list of locales for the scripts that are acceptable in strings to be checked.
UErrorCode
Error code to replace exception handling, so that the code is compatible with all C++ compilers...
Definition: utypes.h:476
Mixed script confusable test.
Definition: uspoof.h:170
Enable the return of auxillary (non-error) information in the upper bits of the check results value...
Definition: uspoof.h:256
Check that an identifier is no looser than the specified RestrictionLevel.
Definition: uspoof.h:204
void uspoof_setAllowedUnicodeSet(USpoofChecker *sc, const icu::UnicodeSet *chars, UErrorCode *status)
Limit the acceptable characters to those specified by a Unicode Set.
A UParseError struct is used to returned detailed information about parsing errors.
Definition: parseerr.h:56
Basic definitions for ICU, for both C and C++ APIs.
Mask for selecting the Restriction Level bits from the return value of uspoof_check().
Definition: uspoof.h:313
const USet * uspoof_getRecommendedSet(UErrorCode *status)
Get the set of characters from Recommended Scripts for Inclusion in Identifiers, as defined in Unicod...
Any valid identifiers, including characters outside of the Identifier Profile.
Definition: uspoof.h:306
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.
UnicodeString is a string class that stores Unicode characters directly and provides similar function...
Definition: unistr.h:245
#define U_STABLE
This is used to declare a function as a stable public ICU C API.
Definition: umachine.h:109
C++ API: Unicode Set.