1/*
2**********************************************************************
3* Copyright (C) 2010, International Business Machines Corporation
4* and others.  All Rights Reserved.
5**********************************************************************
6*/
7/**
8 * IntlTestSpoof tests for USpoofDetector
9 */
10
11#include "unicode/utypes.h"
12
13#if !UCONFIG_NO_REGULAR_EXPRESSIONS && !UCONFIG_NO_NORMALIZATION && !UCONFIG_NO_FILE_IO
14
15#include "itspoof.h"
16#include "unicode/uspoof.h"
17#include "unicode/unistr.h"
18#include "unicode/regex.h"
19#include "unicode/normlzr.h"
20#include "cstring.h"
21#include <stdlib.h>
22#include <stdio.h>
23
24#define TEST_ASSERT_SUCCESS(status) {if (U_FAILURE(status)) { \
25    errcheckln(status, "Failure at file %s, line %d, error = %s", __FILE__, __LINE__, u_errorName(status));}}
26
27#define TEST_ASSERT(expr) {if ((expr)==FALSE) { \
28    errln("Test Failure at file %s, line %d: \"%s\" is false.\n", __FILE__, __LINE__, #expr);};}
29
30#define TEST_ASSERT_EQ(a, b) { if ((a) != (b)) { \
31    errln("Test Failure at file %s, line %d: \"%s\" (%d) != \"%s\" (%d) \n", \
32             __FILE__, __LINE__, #a, (a), #b, (b)); }}
33
34#define TEST_ASSERT_NE(a, b) { if ((a) == (b)) { \
35    errln("Test Failure at file %s, line %d: \"%s\" (%d) == \"%s\" (%d) \n", \
36             __FILE__, __LINE__, #a, (a), #b, (b)); }}
37
38/*
39 *   TEST_SETUP and TEST_TEARDOWN
40 *         macros to handle the boilerplate around setting up test case.
41 *         Put arbitrary test code between SETUP and TEARDOWN.
42 *         "sc" is the ready-to-go  SpoofChecker for use in the tests.
43 */
44#define TEST_SETUP {  \
45    UErrorCode status = U_ZERO_ERROR; \
46    USpoofChecker *sc;     \
47    sc = uspoof_open(&status);  \
48    TEST_ASSERT_SUCCESS(status);   \
49    if (U_SUCCESS(status)){
50
51#define TEST_TEARDOWN  \
52    }  \
53    TEST_ASSERT_SUCCESS(status);  \
54    uspoof_close(sc);  \
55}
56
57
58
59
60void IntlTestSpoof::runIndexedTest( int32_t index, UBool exec, const char* &name, char* /*par*/ )
61{
62    if (exec) logln("TestSuite spoof: ");
63    switch (index) {
64        case 0:
65            name = "TestSpoofAPI";
66            if (exec) {
67                testSpoofAPI();
68            }
69            break;
70         case 1:
71            name = "TestSkeleton";
72            if (exec) {
73                testSkeleton();
74            }
75            break;
76         case 2:
77            name = "TestAreConfusable";
78            if (exec) {
79                testAreConfusable();
80            }
81            break;
82          case 3:
83            name = "TestInvisible";
84            if (exec) {
85                testInvisible();
86            }
87            break;
88          case 4:
89            name = "testConfData";
90            if (exec) {
91                testConfData();
92            }
93            break;
94        default: name=""; break;
95    }
96}
97
98void IntlTestSpoof::testSpoofAPI() {
99
100    TEST_SETUP
101        UnicodeString s("xyz");  // Many latin ranges are whole-script confusable with other scripts.
102                                 // If this test starts failing, consult confusablesWholeScript.txt
103        int32_t position = 666;
104        int32_t checkResults = uspoof_checkUnicodeString(sc, s, &position, &status);
105        TEST_ASSERT_SUCCESS(status);
106        TEST_ASSERT_EQ(0, checkResults);
107        TEST_ASSERT_EQ(666, position);
108    TEST_TEARDOWN;
109
110    TEST_SETUP
111        UnicodeString s1("cxs");
112        UnicodeString s2 = UnicodeString("\\u0441\\u0445\\u0455").unescape();  // Cyrillic "cxs"
113        int32_t checkResults = uspoof_areConfusableUnicodeString(sc, s1, s2, &status);
114        TEST_ASSERT_EQ(USPOOF_MIXED_SCRIPT_CONFUSABLE | USPOOF_WHOLE_SCRIPT_CONFUSABLE, checkResults);
115
116    TEST_TEARDOWN;
117
118    TEST_SETUP
119        UnicodeString s("I1l0O");
120        UnicodeString dest;
121        UnicodeString &retStr = uspoof_getSkeletonUnicodeString(sc, USPOOF_ANY_CASE, s, dest, &status);
122        TEST_ASSERT_SUCCESS(status);
123        TEST_ASSERT(UnicodeString("lllOO") == dest);
124        TEST_ASSERT(&dest == &retStr);
125    TEST_TEARDOWN;
126}
127
128
129#define CHECK_SKELETON(type, input, expected) { \
130    checkSkeleton(sc, type, input, expected, __LINE__); \
131    }
132
133
134// testSkeleton.   Spot check a number of confusable skeleton substitutions from the
135//                 Unicode data file confusables.txt
136//                 Test cases chosen for substitutions of various lengths, and
137//                 membership in different mapping tables.
138void IntlTestSpoof::testSkeleton() {
139    const uint32_t ML = 0;
140    const uint32_t SL = USPOOF_SINGLE_SCRIPT_CONFUSABLE;
141    const uint32_t MA = USPOOF_ANY_CASE;
142    const uint32_t SA = USPOOF_SINGLE_SCRIPT_CONFUSABLE | USPOOF_ANY_CASE;
143
144    TEST_SETUP
145        // A long "identifier" that will overflow implementation stack buffers, forcing heap allocations.
146        CHECK_SKELETON(SL, " A 1ong \\u02b9identifier' that will overflow implementation stack buffers, forcing heap allocations."
147                           " A 1ong 'identifier' that will overflow implementation stack buffers, forcing heap allocations."
148                           " A 1ong 'identifier' that will overflow implementation stack buffers, forcing heap allocations."
149                           " A 1ong 'identifier' that will overflow implementation stack buffers, forcing heap allocations.",
150
151               " A long 'identifier' that vvill overflovv irnplernentation stack buffers, forcing heap allocations."
152               " A long 'identifier' that vvill overflovv irnplernentation stack buffers, forcing heap allocations."
153               " A long 'identifier' that vvill overflovv irnplernentation stack buffers, forcing heap allocations."
154               " A long 'identifier' that vvill overflovv irnplernentation stack buffers, forcing heap allocations.")
155
156        // FC5F ;	FE74 0651 ;   ML  #* ARABIC LIGATURE SHADDA WITH KASRATAN ISOLATED FORM to
157        //                                ARABIC KASRATAN ISOLATED FORM, ARABIC SHADDA
158        //    This character NFKD normalizes to \u0020 \u064d \u0651, so its confusable mapping
159        //    is never used in creating a skeleton.
160        CHECK_SKELETON(SL, "\\uFC5F", " \\u064d\\u0651");
161
162        CHECK_SKELETON(SL, "nochange", "nochange");
163        CHECK_SKELETON(MA, "love", "love");
164        CHECK_SKELETON(MA, "1ove", "love");   // Digit 1 to letter l
165        CHECK_SKELETON(ML, "OOPS", "OOPS");
166        CHECK_SKELETON(ML, "00PS", "00PS");   // Digit 0 unchanged in lower case mode.
167        CHECK_SKELETON(MA, "OOPS", "OOPS");
168        CHECK_SKELETON(MA, "00PS", "OOPS");   // Digit 0 to letter O in any case mode only
169        CHECK_SKELETON(SL, "\\u059c", "\\u0301");
170        CHECK_SKELETON(SL, "\\u2A74", "\\u003A\\u003A\\u003D");
171        CHECK_SKELETON(SL, "\\u247E", "\\u0028\\u006C\\u006C\\u0029");  // "(ll)"
172        CHECK_SKELETON(SL, "\\uFDFB", "\\u062C\\u0644\\u0020\\u062C\\u0644\\u0627\\u0644\\u0647");
173
174        // This mapping exists in the ML and MA tables, does not exist in SL, SA
175        //0C83 ;	0C03 ;
176        CHECK_SKELETON(SL, "\\u0C83", "\\u0C83");
177        CHECK_SKELETON(SA, "\\u0C83", "\\u0C83");
178        CHECK_SKELETON(ML, "\\u0C83", "\\u0983");
179        CHECK_SKELETON(MA, "\\u0C83", "\\u0983");
180
181        // 0391 ; 0041 ;
182        // This mapping exists only in the MA table.
183        CHECK_SKELETON(MA, "\\u0391", "A");
184        CHECK_SKELETON(SA, "\\u0391", "\\u0391");
185        CHECK_SKELETON(ML, "\\u0391", "\\u0391");
186        CHECK_SKELETON(SL, "\\u0391", "\\u0391");
187
188        // 13CF ;  0062 ;
189        // This mapping exists in the ML and MA tables
190        CHECK_SKELETON(ML, "\\u13CF", "b");
191        CHECK_SKELETON(MA, "\\u13CF", "b");
192        CHECK_SKELETON(SL, "\\u13CF", "\\u13CF");
193        CHECK_SKELETON(SA, "\\u13CF", "\\u13CF");
194
195        // 0022 ;  0027 0027 ;
196        // all tables.
197        CHECK_SKELETON(SL, "\\u0022", "\\u0027\\u0027");
198        CHECK_SKELETON(SA, "\\u0022", "\\u0027\\u0027");
199        CHECK_SKELETON(ML, "\\u0022", "\\u0027\\u0027");
200        CHECK_SKELETON(MA, "\\u0022", "\\u0027\\u0027");
201
202    TEST_TEARDOWN;
203}
204
205
206//
207//  Run a single confusable skeleton transformation test case.
208//
209void IntlTestSpoof::checkSkeleton(const USpoofChecker *sc, uint32_t type,
210                                  const char *input, const char *expected, int32_t lineNum) {
211    UnicodeString uInput = UnicodeString(input).unescape();
212    UnicodeString uExpected = UnicodeString(expected).unescape();
213
214    UErrorCode status = U_ZERO_ERROR;
215    UnicodeString actual;
216    uspoof_getSkeletonUnicodeString(sc, type, uInput, actual, &status);
217    if (U_FAILURE(status)) {
218        errln("File %s, Line %d, Test case from line %d, status is %s", __FILE__, __LINE__, lineNum,
219              u_errorName(status));
220        return;
221    }
222    if (uExpected != actual) {
223        errln("File %s, Line %d, Test case from line %d, Actual and Expected skeletons differ.",
224               __FILE__, __LINE__, lineNum);
225        errln(UnicodeString(" Actual   Skeleton: \"") + actual + UnicodeString("\"\n") +
226              UnicodeString(" Expected Skeleton: \"") + uExpected + UnicodeString("\""));
227    }
228}
229
230void IntlTestSpoof::testAreConfusable() {
231    TEST_SETUP
232        UnicodeString s1("A long string that will overflow stack buffers.  A long string that will overflow stack buffers. "
233                         "A long string that will overflow stack buffers.  A long string that will overflow stack buffers. ");
234        UnicodeString s2("A long string that wi11 overflow stack buffers.  A long string that will overflow stack buffers. "
235                         "A long string that wi11 overflow stack buffers.  A long string that will overflow stack buffers. ");
236        TEST_ASSERT_EQ(USPOOF_SINGLE_SCRIPT_CONFUSABLE, uspoof_areConfusableUnicodeString(sc, s1, s2, &status));
237        TEST_ASSERT_SUCCESS(status);
238
239    TEST_TEARDOWN;
240}
241
242void IntlTestSpoof::testInvisible() {
243    TEST_SETUP
244        UnicodeString  s = UnicodeString("abcd\\u0301ef").unescape();
245        int32_t position = -42;
246        TEST_ASSERT_EQ(0, uspoof_checkUnicodeString(sc, s, &position, &status));
247        TEST_ASSERT_SUCCESS(status);
248        TEST_ASSERT(position == -42);
249
250        UnicodeString  s2 = UnicodeString("abcd\\u0301\\u0302\\u0301ef").unescape();
251        TEST_ASSERT_EQ(USPOOF_INVISIBLE, uspoof_checkUnicodeString(sc, s2, &position, &status));
252        TEST_ASSERT_SUCCESS(status);
253        TEST_ASSERT_EQ(7, position);
254
255        // Tow acute accents, one from the composed a with acute accent, \u00e1,
256        // and one separate.
257        position = -42;
258        UnicodeString  s3 = UnicodeString("abcd\\u00e1\\u0301xyz").unescape();
259        TEST_ASSERT_EQ(USPOOF_INVISIBLE, uspoof_checkUnicodeString(sc, s3, &position, &status));
260        TEST_ASSERT_SUCCESS(status);
261        TEST_ASSERT_EQ(7, position);
262    TEST_TEARDOWN;
263}
264
265
266static UnicodeString parseHex(const UnicodeString &in) {
267    // Convert a series of hex numbers in a Unicode String to a string with the
268    // corresponding characters.
269    // The conversion is _really_ annoying.  There must be some function to just do it.
270    UnicodeString result;
271    UChar32 cc = 0;
272    for (int32_t i=0; i<in.length(); i++) {
273        UChar c = in.charAt(i);
274        if (c == 0x20) {   // Space
275            if (cc > 0) {
276               result.append(cc);
277               cc = 0;
278            }
279        } else if (c>=0x30 && c<=0x39) {
280            cc = (cc<<4) + (c - 0x30);
281        } else if ((c>=0x41 && c<=0x46) || (c>=0x61 && c<=0x66)) {
282            cc = (cc<<4) + (c & 0x0f)+9;
283        }
284        // else do something with bad input.
285    }
286    if (cc > 0) {
287        result.append(cc);
288    }
289    return result;
290}
291
292
293//
294// Append the hex form of a UChar32 to a UnicodeString.
295// Used in formatting error messages.
296// Match the formatting of numbers in confusables.txt
297// Minimum of 4 digits, no leading zeroes for positions 5 and up.
298//
299static void appendHexUChar(UnicodeString &dest, UChar32 c) {
300    UBool   doZeroes = FALSE;
301    for (int bitNum=28; bitNum>=0; bitNum-=4) {
302        if (bitNum <= 12) {
303            doZeroes = TRUE;
304        }
305        int hexDigit = (c>>bitNum) & 0x0f;
306        if (hexDigit != 0 || doZeroes) {
307            doZeroes = TRUE;
308            dest.append((UChar)(hexDigit<=9? hexDigit + 0x30: hexDigit -10 + 0x41));
309        }
310    }
311    dest.append((UChar)0x20);
312}
313
314U_DEFINE_LOCAL_OPEN_POINTER(LocalStdioFilePointer, FILE, fclose);
315
316//  testConfData - Check each data item from the Unicode confusables.txt file,
317//                 verify that it transforms correctly in a skeleton.
318//
319void IntlTestSpoof::testConfData() {
320    UErrorCode status = U_ZERO_ERROR;
321
322    const char *testDataDir = IntlTest::getSourceTestData(status);
323    TEST_ASSERT_SUCCESS(status);
324    char buffer[2000];
325    uprv_strcpy(buffer, testDataDir);
326    uprv_strcat(buffer, "confusables.txt");
327
328    LocalStdioFilePointer f(fopen(buffer, "rb"));
329    if (f.isNull()) {
330        errln("Skipping test spoof/testConfData.  File confusables.txt not accessible.");
331        return;
332    }
333    fseek(f.getAlias(), 0, SEEK_END);
334    int32_t  fileSize = ftell(f.getAlias());
335    LocalArray<char> fileBuf(new char[fileSize]);
336    fseek(f.getAlias(), 0, SEEK_SET);
337    int32_t amt_read = fread(fileBuf.getAlias(), 1, fileSize, f.getAlias());
338    TEST_ASSERT_EQ(amt_read, fileSize);
339    TEST_ASSERT(fileSize>0);
340    if (amt_read != fileSize || fileSize <=0) {
341        return;
342    }
343    UnicodeString confusablesTxt = UnicodeString::fromUTF8(StringPiece(fileBuf.getAlias(), fileSize));
344
345    LocalUSpoofCheckerPointer sc(uspoof_open(&status));
346    TEST_ASSERT_SUCCESS(status);
347
348    // Parse lines from the confusables.txt file.  Example Line:
349    // FF44 ;	0064 ;	SL	# ( d -> d ) FULLWIDTH ....
350    // Three fields.  The hex fields can contain more than one character,
351    //                and each character may be more than 4 digits (for supplemntals)
352    // This regular expression matches lines and splits the fields into capture groups.
353    RegexMatcher parseLine("(?m)^([0-9A-F]{4}[^#;]*?);([^#;]*?);([^#]*)", confusablesTxt, 0, status);
354    TEST_ASSERT_SUCCESS(status);
355    while (parseLine.find()) {
356        UnicodeString from = parseHex(parseLine.group(1, status));
357        if (!Normalizer::isNormalized(from, UNORM_NFKD, status)) {
358            // The source character was not NFKD.
359            // Skip this case; the first step in obtaining a skeleton is to NFKD the input,
360            //  so the mapping in this line of confusables.txt will never be applied.
361            continue;
362        }
363
364        UnicodeString rawExpected = parseHex(parseLine.group(2, status));
365        UnicodeString expected;
366        Normalizer::decompose(rawExpected, TRUE, 0, expected, status);
367        TEST_ASSERT_SUCCESS(status);
368
369        int32_t skeletonType = 0;
370        UnicodeString tableType = parseLine.group(3, status);
371        TEST_ASSERT_SUCCESS(status);
372        if (tableType.indexOf("SL") >= 0) {
373            skeletonType = USPOOF_SINGLE_SCRIPT_CONFUSABLE;
374        } else if (tableType.indexOf("SA") >= 0) {
375            skeletonType = USPOOF_SINGLE_SCRIPT_CONFUSABLE | USPOOF_ANY_CASE;
376        } else if (tableType.indexOf("ML") >= 0) {
377            skeletonType = 0;
378        } else if (tableType.indexOf("MA") >= 0) {
379            skeletonType = USPOOF_ANY_CASE;
380        }
381
382        UnicodeString actual;
383        uspoof_getSkeletonUnicodeString(sc.getAlias(), skeletonType, from, actual, &status);
384        TEST_ASSERT_SUCCESS(status);
385        TEST_ASSERT(actual == expected);
386        if (actual != expected) {
387            errln(parseLine.group(0, status));
388            UnicodeString line = "Actual: ";
389            int i = 0;
390            while (i < actual.length()) {
391                appendHexUChar(line, actual.char32At(i));
392                i = actual.moveIndex32(i, 1);
393            }
394            errln(line);
395        }
396        if (U_FAILURE(status)) {
397            break;
398        }
399    }
400}
401#endif // UCONFIG_NO_REGULAR_EXPRESSIONS
402
403