1// © 2016 and later: Unicode, Inc. and others.
2// License & terms of use: http://www.unicode.org/copyright.html
3/*
4***************************************************************************
5* Copyright (C) 2008-2013, International Business Machines Corporation
6* and others. All Rights Reserved.
7***************************************************************************
8*
9* uspoof_impl.h
10*
11* Implemenation header for spoof detection
12*
13*/
14
15#ifndef USPOOFIM_H
16#define USPOOFIM_H
17
18#include "uassert.h"
19#include "unicode/utypes.h"
20#include "unicode/uspoof.h"
21#include "unicode/uscript.h"
22#include "unicode/udata.h"
23#include "udataswp.h"
24#include "utrie2.h"
25
26#if !UCONFIG_NO_NORMALIZATION
27
28#ifdef __cplusplus
29
30#include "capi_helper.h"
31
32U_NAMESPACE_BEGIN
33
34// The maximium length (in UTF-16 UChars) of the skeleton replacement string resulting from
35// a single input code point. This is function of the unicode.org data.
36#define USPOOF_MAX_SKELETON_EXPANSION 20
37
38// The default stack buffer size for copies or conversions or normalizations
39// of input strings being checked. (Used in multiple places.)
40#define USPOOF_STACK_BUFFER_SIZE 100
41
42// Magic number for sanity checking spoof data.
43#define USPOOF_MAGIC 0x3845fdef
44
45// Magic number for sanity checking spoof checkers.
46#define USPOOF_CHECK_MAGIC 0x2734ecde
47
48class ScriptSet;
49class SpoofData;
50struct SpoofDataHeader;
51class ConfusableDataUtils;
52
53/**
54 * Class SpoofImpl corresponds directly to the plain C API opaque type
55 * USpoofChecker. One can be cast to the other.
56 */
57class SpoofImpl : public UObject,
58 public IcuCApiHelper<USpoofChecker, SpoofImpl, USPOOF_MAGIC> {
59public:
60 SpoofImpl(SpoofData *data, UErrorCode& status);
61 SpoofImpl(UErrorCode& status);
62 SpoofImpl();
63 void construct(UErrorCode& status);
64 virtual ~SpoofImpl();
65
66 /** Copy constructor, used by the user level uspoof_clone() function.
67 */
68 SpoofImpl(const SpoofImpl &src, UErrorCode &status);
69
70 USpoofChecker *asUSpoofChecker();
71 static SpoofImpl *validateThis(USpoofChecker *sc, UErrorCode &status);
72 static const SpoofImpl *validateThis(const USpoofChecker *sc, UErrorCode &status);
73
74 /** Set and Get AllowedLocales, implementations of the corresponding API */
75 void setAllowedLocales(const char *localesList, UErrorCode &status);
76 const char * getAllowedLocales(UErrorCode &status);
77
78 // Add (union) to the UnicodeSet all of the characters for the scripts used for
79 // the specified locale. Part of the implementation of setAllowedLocales.
80 void addScriptChars(const char *locale, UnicodeSet *allowedChars, UErrorCode &status);
81
82 // Functions implementing the features of UTS 39 section 5.
83 static void getAugmentedScriptSet(UChar32 codePoint, ScriptSet& result, UErrorCode& status);
84 void getResolvedScriptSet(const UnicodeString& input, ScriptSet& result, UErrorCode& status) const;
85 void getResolvedScriptSetWithout(const UnicodeString& input, UScriptCode script, ScriptSet& result, UErrorCode& status) const;
86 void getNumerics(const UnicodeString& input, UnicodeSet& result, UErrorCode& status) const;
87 URestrictionLevel getRestrictionLevel(const UnicodeString& input, UErrorCode& status) const;
88
89 int32_t findHiddenOverlay(const UnicodeString& input, UErrorCode& status) const;
90 bool isIllegalCombiningDotLeadCharacter(UChar32 cp) const;
91
92 /** parse a hex number. Untility used by the builders. */
93 static UChar32 ScanHex(const UChar *s, int32_t start, int32_t limit, UErrorCode &status);
94
95 static UClassID U_EXPORT2 getStaticClassID(void);
96 virtual UClassID getDynamicClassID(void) const;
97
98 //
99 // Data Members
100 //
101
102 int32_t fChecks; // Bit vector of checks to perform.
103
104 SpoofData *fSpoofData;
105
106 const UnicodeSet *fAllowedCharsSet; // The UnicodeSet of allowed characters.
107 // for this Spoof Checker. Defaults to all chars.
108
109 const char *fAllowedLocales; // The list of allowed locales.
110 URestrictionLevel fRestrictionLevel; // The maximum restriction level for an acceptable identifier.
111};
112
113/**
114 * Class CheckResult corresponds directly to the plain C API opaque type
115 * USpoofCheckResult. One can be cast to the other.
116 */
117class CheckResult : public UObject,
118 public IcuCApiHelper<USpoofCheckResult, CheckResult, USPOOF_CHECK_MAGIC> {
119public:
120 CheckResult();
121 virtual ~CheckResult();
122
123 USpoofCheckResult *asUSpoofCheckResult();
124 static CheckResult *validateThis(USpoofCheckResult *ptr, UErrorCode &status);
125 static const CheckResult *validateThis(const USpoofCheckResult *ptr, UErrorCode &status);
126
127 void clear();
128
129 // Used to convert this CheckResult to the older int32_t return value API
130 int32_t toCombinedBitmask(int32_t expectedChecks);
131
132 // Data Members
133 int32_t fChecks; // Bit vector of checks that were failed.
134 UnicodeSet fNumerics; // Set of numerics found in the string.
135 URestrictionLevel fRestrictionLevel; // The restriction level of the string.
136};
137
138
139//
140// Confusable Mappings Data Structures, version 2.0
141//
142// For the confusable data, we are essentially implementing a map,
143// key: a code point
144// value: a string. Most commonly one char in length, but can be more.
145//
146// The keys are stored as a sorted array of 32 bit ints.
147// bits 0-23 a code point value
148// bits 24-31 length of value string, in UChars (between 1 and 256 UChars).
149// The key table is sorted in ascending code point order. (not on the
150// 32 bit int value, the flag bits do not participate in the sorting.)
151//
152// Lookup is done by means of a binary search in the key table.
153//
154// The corresponding values are kept in a parallel array of 16 bit ints.
155// If the value string is of length 1, it is literally in the value array.
156// For longer strings, the value array contains an index into the strings table.
157//
158// String Table:
159// The strings table contains all of the value strings (those of length two or greater)
160// concatentated together into one long UChar (UTF-16) array.
161//
162// There is no nul character or other mark between adjacent strings.
163//
164//----------------------------------------------------------------------------
165//
166// Changes from format version 1 to format version 2:
167// 1) Removal of the whole-script confusable data tables.
168// 2) Removal of the SL/SA/ML/MA and multi-table flags in the key bitmask.
169// 3) Expansion of string length value in the key bitmask from 2 bits to 8 bits.
170// 4) Removal of the string lengths table since 8 bits is sufficient for the
171// lengths of all entries in confusables.txt.
172
173
174
175// Internal functions for manipulating confusable data table keys
176#define USPOOF_CONFUSABLE_DATA_FORMAT_VERSION 2 // version for ICU 58
177class ConfusableDataUtils {
178public:
179 inline static UChar32 keyToCodePoint(int32_t key) {
180 return key & 0x00ffffff;
181 }
182 inline static int32_t keyToLength(int32_t key) {
183 return ((key & 0xff000000) >> 24) + 1;
184 }
185 inline static int32_t codePointAndLengthToKey(UChar32 codePoint, int32_t length) {
186 U_ASSERT((codePoint & 0x00ffffff) == codePoint);
187 U_ASSERT(length <= 256);
188 return codePoint | ((length - 1) << 24);
189 }
190};
191
192
193//-------------------------------------------------------------------------------------
194//
195// SpoofData
196//
197// A small class that wraps the raw (usually memory mapped) spoof data.
198// Serves two primary functions:
199// 1. Convenience. Contains real pointers to the data, to avoid dealing with
200// the offsets in the raw data.
201// 2. Reference counting. When a spoof checker is cloned, the raw data is shared
202// and must be retained until all checkers using the data are closed.
203// Nothing in this struct includes state that is specific to any particular
204// USpoofDetector object.
205//
206//---------------------------------------------------------------------------------------
207class SpoofData: public UMemory {
208 public:
209 static SpoofData* getDefault(UErrorCode &status); // Get standard ICU spoof data.
210 static void releaseDefault(); // Cleanup reference to default spoof data.
211
212 SpoofData(UErrorCode &status); // Create new spoof data wrapper.
213 // Only used when building new data from rules.
214
215 // Constructor for use when creating from prebuilt default data.
216 // A UDataMemory is what the ICU internal data loading functions provide.
217 // The udm is adopted by the SpoofData.
218 SpoofData(UDataMemory *udm, UErrorCode &status);
219
220 // Constructor for use when creating from serialized data.
221 //
222 SpoofData(const void *serializedData, int32_t length, UErrorCode &status);
223
224 // Check raw Spoof Data Version compatibility.
225 // Return TRUE it looks good.
226 UBool validateDataVersion(UErrorCode &status) const;
227
228 ~SpoofData(); // Destructor not normally used.
229 // Use removeReference() instead.
230 // Reference Counting functions.
231 // Clone of a user-level spoof detector increments the ref count on the data.
232 // Close of a user-level spoof detector decrements the ref count.
233 // If the data is owned by us, it will be deleted when count goes to zero.
234 SpoofData *addReference();
235 void removeReference();
236
237 // Reset all fields to an initial state.
238 // Called from the top of all constructors.
239 void reset();
240
241 // Copy this instance's raw data buffer to the specified address.
242 int32_t serialize(void *buf, int32_t capacity, UErrorCode &status) const;
243
244 // Get the total number of bytes of data backed by this SpoofData.
245 // Not to be confused with length, which returns the number of confusable entries.
246 int32_t size() const;
247
248 // Get the confusable skeleton transform for a single code point.
249 // The result is a string with a length between 1 and 18 as of Unicode 9.
250 // This is the main public endpoint for this class.
251 // @return The length in UTF-16 code units of the substition string.
252 int32_t confusableLookup(UChar32 inChar, UnicodeString &dest) const;
253
254 // Get the number of confusable entries in this SpoofData.
255 int32_t length() const;
256
257 // Get the code point (key) at the specified index.
258 UChar32 codePointAt(int32_t index) const;
259
260 // Get the confusable skeleton (value) at the specified index.
261 // Append it to the specified UnicodeString&.
262 // @return The length in UTF-16 code units of the skeleton string.
263 int32_t appendValueTo(int32_t index, UnicodeString& dest) const;
264
265 private:
266 // Reserve space in the raw data. For use by builder when putting together a
267 // new set of data. Init the new storage to zero, to prevent inconsistent
268 // results if it is not all otherwise set by the requester.
269 // Return:
270 // pointer to the new space that was added by this function.
271 void *reserveSpace(int32_t numBytes, UErrorCode &status);
272
273 // initialize the pointers from this object to the raw data.
274 void initPtrs(UErrorCode &status);
275
276 SpoofDataHeader *fRawData; // Ptr to the raw memory-mapped data
277 UBool fDataOwned; // True if the raw data is owned, and needs
278 // to be deleted when refcount goes to zero.
279 UDataMemory *fUDM; // If not NULL, our data came from a
280 // UDataMemory, which we must close when
281 // we are done.
282
283 uint32_t fMemLimit; // Limit of available raw data space
284 u_atomic_int32_t fRefCount;
285
286 // Confusable data
287 int32_t *fCFUKeys;
288 uint16_t *fCFUValues;
289 UChar *fCFUStrings;
290
291 friend class ConfusabledataBuilder;
292};
293
294//---------------------------------------------------------------------------------------
295//
296// Raw Binary Data Formats, as loaded from the ICU data file,
297// or as built by the builder.
298//
299//---------------------------------------------------------------------------------------
300struct SpoofDataHeader {
301 int32_t fMagic; // (0x3845fdef)
302 uint8_t fFormatVersion[4]; // Data Format. Same as the value in struct UDataInfo
303 // if there is one associated with this data.
304 int32_t fLength; // Total lenght in bytes of this spoof data,
305 // including all sections, not just the header.
306
307 // The following four sections refer to data representing the confusable data
308 // from the Unicode.org data from "confusables.txt"
309
310 int32_t fCFUKeys; // byte offset to Keys table (from SpoofDataHeader *)
311 int32_t fCFUKeysSize; // number of entries in keys table (32 bits each)
312
313 // TODO: change name to fCFUValues, for consistency.
314 int32_t fCFUStringIndex; // byte offset to String Indexes table
315 int32_t fCFUStringIndexSize; // number of entries in String Indexes table (16 bits each)
316 // (number of entries must be same as in Keys table
317
318 int32_t fCFUStringTable; // byte offset of String table
319 int32_t fCFUStringTableLen; // length of string table (in 16 bit UChars)
320
321 // The following sections are for data from xidmodifications.txt
322
323 int32_t unused[15]; // Padding, Room for Expansion
324};
325
326
327
328U_NAMESPACE_END
329#endif /* __cplusplus */
330
331/**
332 * Endianness swap function for binary spoof data.
333 * @internal
334 */
335U_CAPI int32_t U_EXPORT2
336uspoof_swap(const UDataSwapper *ds, const void *inData, int32_t length, void *outData,
337 UErrorCode *status);
338
339
340#endif
341
342#endif /* USPOOFIM_H */
343
344