+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
/********************************************************************
* COPYRIGHT:
- * Copyright (c) 2002-2014, International Business Machines Corporation and
+ * Copyright (c) 2002-2016, International Business Machines Corporation and
* others. All Rights Reserved.
********************************************************************/
#include "intltest.h"
#if !UCONFIG_NO_REGULAR_EXPRESSIONS
+#include <stdlib.h>
+#include <stdio.h>
+#include <string.h>
+
+#include "unicode/localpointer.h"
#include "unicode/regex.h"
#include "unicode/uchar.h"
#include "unicode/ucnv.h"
#include "unicode/uniset.h"
#include "unicode/uregex.h"
+#include "unicode/usetiter.h"
#include "unicode/ustring.h"
+#include "unicode/utext.h"
+#include "unicode/utf16.h"
+#include "cstr.h"
#include "regextst.h"
+#include "regexcmp.h"
#include "uvector.h"
#include "util.h"
-#include <stdlib.h>
-#include <string.h>
-#include <stdio.h>
+#include "cmemory.h"
#include "cstring.h"
#include "uinvchar.h"
void RegexTest::runIndexedTest( int32_t index, UBool exec, const char* &name, char* /*par*/ )
{
if (exec) logln("TestSuite RegexTest: ");
- switch (index) {
-
- case 0: name = "Basic";
- if (exec) Basic();
- break;
- case 1: name = "API_Match";
- if (exec) API_Match();
- break;
- case 2: name = "API_Replace";
- if (exec) API_Replace();
- break;
- case 3: name = "API_Pattern";
- if (exec) API_Pattern();
- break;
- case 4:
+ TESTCASE_AUTO_BEGIN;
+ TESTCASE_AUTO(Basic);
+ TESTCASE_AUTO(API_Match);
+ TESTCASE_AUTO(API_Replace);
+ TESTCASE_AUTO(API_Pattern);
#if !UCONFIG_NO_FILE_IO
- name = "Extended";
- if (exec) Extended();
-#else
- name = "skip";
+ TESTCASE_AUTO(Extended);
#endif
- break;
- case 5: name = "Errors";
- if (exec) Errors();
- break;
- case 6: name = "PerlTests";
- if (exec) PerlTests();
- break;
- case 7: name = "Callbacks";
- if (exec) Callbacks();
- break;
- case 8: name = "FindProgressCallbacks";
- if (exec) FindProgressCallbacks();
- break;
- case 9: name = "Bug 6149";
- if (exec) Bug6149();
- break;
- case 10: name = "UTextBasic";
- if (exec) UTextBasic();
- break;
- case 11: name = "API_Match_UTF8";
- if (exec) API_Match_UTF8();
- break;
- case 12: name = "API_Replace_UTF8";
- if (exec) API_Replace_UTF8();
- break;
- case 13: name = "API_Pattern_UTF8";
- if (exec) API_Pattern_UTF8();
- break;
- case 14: name = "PerlTestsUTF8";
- if (exec) PerlTestsUTF8();
- break;
- case 15: name = "PreAllocatedUTextCAPI";
- if (exec) PreAllocatedUTextCAPI();
- break;
- case 16: name = "Bug 7651";
- if (exec) Bug7651();
- break;
- case 17: name = "Bug 7740";
- if (exec) Bug7740();
- break;
- case 18: name = "Bug 8479";
- if (exec) Bug8479();
- break;
- case 19: name = "Bug 7029";
- if (exec) Bug7029();
- break;
- case 20: name = "CheckInvBufSize";
- if (exec) CheckInvBufSize();
- break;
- case 21: name = "Bug 9283";
- if (exec) Bug9283();
- break;
- case 22: name = "Bug10459";
- if (exec) Bug10459();
- break;
-
- default: name = "";
- break; //needed to end loop
- }
+ TESTCASE_AUTO(Errors);
+ TESTCASE_AUTO(PerlTests);
+ TESTCASE_AUTO(Callbacks);
+ TESTCASE_AUTO(FindProgressCallbacks);
+ TESTCASE_AUTO(Bug6149);
+ TESTCASE_AUTO(UTextBasic);
+ TESTCASE_AUTO(API_Match_UTF8);
+ TESTCASE_AUTO(API_Replace_UTF8);
+ TESTCASE_AUTO(API_Pattern_UTF8);
+ TESTCASE_AUTO(PerlTestsUTF8);
+ TESTCASE_AUTO(PreAllocatedUTextCAPI);
+ TESTCASE_AUTO(Bug7651);
+ TESTCASE_AUTO(Bug7740);
+ TESTCASE_AUTO(Bug8479);
+ TESTCASE_AUTO(Bug7029);
+ TESTCASE_AUTO(CheckInvBufSize);
+ TESTCASE_AUTO(Bug9283);
+ TESTCASE_AUTO(Bug10459);
+ TESTCASE_AUTO(TestCaseInsensitiveStarters);
+ TESTCASE_AUTO(TestBug11049);
+ TESTCASE_AUTO(TestBug11371);
+ TESTCASE_AUTO(TestBug11480);
+ TESTCASE_AUTO(NamedCapture);
+ TESTCASE_AUTO(NamedCaptureLimits);
+ TESTCASE_AUTO(TestBug12884);
+ TESTCASE_AUTO(TestBug13631);
+ TESTCASE_AUTO(TestBug13632);
+ TESTCASE_AUTO_END;
}
-
/**
* Calls utext_openUTF8 after, potentially, converting invariant text from the compilation codepage
* into ASCII.
return ASSERT_BUF;
}
-#define LENGTHOF(array) (int32_t)(sizeof(array)/sizeof((array)[0]))
-
-#define REGEX_VERBOSE_TEXT(text) {char buf[200];utextToPrintable(buf,sizeof(buf)/sizeof(buf[0]),text);logln("%s:%d: UText %s=\"%s\"", __FILE__, __LINE__, #text, buf);}
+#define REGEX_VERBOSE_TEXT(text) {char buf[200];utextToPrintable(buf,UPRV_LENGTHOF(buf),text);logln("%s:%d: UText %s=\"%s\"", __FILE__, __LINE__, #text, buf);}
#define REGEX_CHECK_STATUS {if (U_FAILURE(status)) {dataerrln("%s:%d: RegexTest failure. status=%s", \
__FILE__, __LINE__, u_errorName(status)); return;}}
#define REGEX_ASSERT_L(expr, line) {if ((expr)==FALSE) { \
errln("RegexTest failure at line %d, from %d.", __LINE__, (line)); return;}}
-#define REGEX_ASSERT_UNISTR(ustr,inv) {if (!(ustr==inv)) {errln("%s:%d: RegexTest failure: REGEX_ASSERT_UNISTR(%s,%s) failed \n", __FILE__, __LINE__, extractToAssertBuf(ustr),inv);};}
+// expected: const char * , restricted to invariant characters.
+// actual: const UnicodeString &
+#define REGEX_ASSERT_UNISTR(expected, actual) { \
+ if (UnicodeString(expected, -1, US_INV) != (actual)) { \
+ errln("%s:%d: RegexTest failure: REGEX_ASSERT_UNISTR(%s, %s) failed \n", \
+ __FILE__, __LINE__, expected, extractToAssertBuf(actual));};}
static UBool testUTextEqual(UText *uta, UText *utb) {
if (!testUTextEqual(&expectedText, actual)) {
char buf[201 /*21*/];
char expectedBuf[201];
- utextToPrintable(buf, sizeof(buf)/sizeof(buf[0]), actual);
- utextToPrintable(expectedBuf, sizeof(expectedBuf)/sizeof(expectedBuf[0]), &expectedText);
+ utextToPrintable(buf, UPRV_LENGTHOF(buf), actual);
+ utextToPrintable(expectedBuf, UPRV_LENGTHOF(expectedBuf), &expectedText);
errln("%s:%d: assertUText: Failure: expected \"%s\" (%d chars), got \"%s\" (%d chars)", file, line, expectedBuf, (int)utext_nativeLength(&expectedText), buf, (int)utext_nativeLength(actual));
}
utext_close(&expectedText);
if (!testUTextEqual(&expectedText, actual)) {
char buf[201 /*21*/];
char expectedBuf[201];
- utextToPrintable(buf, sizeof(buf)/sizeof(buf[0]), actual);
- utextToPrintable(expectedBuf, sizeof(expectedBuf)/sizeof(expectedBuf[0]), &expectedText);
+ utextToPrintable(buf, UPRV_LENGTHOF(buf), actual);
+ utextToPrintable(expectedBuf, UPRV_LENGTHOF(expectedBuf), &expectedText);
errln("%s:%d: assertUTextInvariant: Failure: expected \"%s\" (%d uchars), got \"%s\" (%d chars)", file, line, expectedBuf, (int)utext_nativeLength(&expectedText), buf, (int)utext_nativeLength(actual));
}
utext_close(&expectedText);
REGEX_ASSERT(dest == "The value of $1 is bc.defg");
dest = matcher2->replaceFirst("$ by itself, no group number $$$", status);
- REGEX_CHECK_STATUS;
- REGEX_ASSERT(dest == "$ by itself, no group number $$$defg");
+ REGEX_ASSERT(U_FAILURE(status));
+ status = U_ZERO_ERROR;
UnicodeString replacement = UNICODE_STRING_SIMPLE("Supplemental Digit 1 $\\U0001D7CF.");
replacement = replacement.unescape();
utext_openUnicodeString(&destText, &dest, &status);
UText *result;
//const char str_0123456789[] = { 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, 0x38, 0x39, 0x00 }; /* 0123456789 */
- // Test shallow-clone API
+ // Test shallow-clone API
int64_t group_len;
result = matcher->group((UText *)NULL, group_len, status);
REGEX_CHECK_STATUS;
utext_close(&destText);
utext_openUnicodeString(&destText, &dest, &status);
- result = matcher->group(0, NULL, status);
+ int64_t length;
+ result = matcher->group(0, NULL, length, status);
REGEX_CHECK_STATUS;
REGEX_ASSERT_UTEXT_UTF8(str_0123456789, result);
utext_close(result);
- result = matcher->group(0, &destText, status);
+ result = matcher->group(0, &destText, length, status);
REGEX_CHECK_STATUS;
REGEX_ASSERT(result == &destText);
- REGEX_ASSERT_UTEXT_UTF8(str_0123456789, result);
+ REGEX_ASSERT(utext_getNativeIndex(result) == 0);
+ REGEX_ASSERT(length == 10);
+ REGEX_ASSERT_UTEXT_INVARIANT("0123456789", result);
- result = matcher->group(1, NULL, status);
+ // Capture Group 1 == "234567"
+ result = matcher->group(1, NULL, length, status);
REGEX_CHECK_STATUS;
- const char str_234567[] = { 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, 0x00 }; /* 234567 */
- REGEX_ASSERT_UTEXT_UTF8(str_234567, result);
+ REGEX_ASSERT(utext_getNativeIndex(result) == 2);
+ REGEX_ASSERT(length == 6);
+ REGEX_ASSERT_UTEXT_INVARIANT("0123456789", result);
utext_close(result);
- result = matcher->group(1, &destText, status);
+
+ result = matcher->group(1, &destText, length, status);
REGEX_CHECK_STATUS;
REGEX_ASSERT(result == &destText);
- REGEX_ASSERT_UTEXT_UTF8(str_234567, result);
+ REGEX_ASSERT(utext_getNativeIndex(result) == 2);
+ REGEX_ASSERT(length == 6);
+ REGEX_ASSERT_UTEXT_INVARIANT("0123456789", result);
+ utext_close(result);
- result = matcher->group(2, NULL, status);
+ // Capture Group 2 == "45"
+ result = matcher->group(2, NULL, length, status);
REGEX_CHECK_STATUS;
- const char str_45[] = { 0x34, 0x35, 0x00 }; /* 45 */
- REGEX_ASSERT_UTEXT_UTF8(str_45, result);
+ REGEX_ASSERT(utext_getNativeIndex(result) == 4);
+ REGEX_ASSERT(length == 2);
+ REGEX_ASSERT_UTEXT_INVARIANT("0123456789", result);
utext_close(result);
- result = matcher->group(2, &destText, status);
+
+ result = matcher->group(2, &destText, length, status);
REGEX_CHECK_STATUS;
REGEX_ASSERT(result == &destText);
- REGEX_ASSERT_UTEXT_UTF8(str_45, result);
+ REGEX_ASSERT(utext_getNativeIndex(result) == 4);
+ REGEX_ASSERT(length == 2);
+ REGEX_ASSERT_UTEXT_INVARIANT("0123456789", result);
+ utext_close(result);
- result = matcher->group(3, NULL, status);
+ // Capture Group 3 == "89"
+ result = matcher->group(3, NULL, length, status);
REGEX_CHECK_STATUS;
- const char str_89[] = { 0x38, 0x39, 0x00 }; /* 89 */
- REGEX_ASSERT_UTEXT_UTF8(str_89, result);
+ REGEX_ASSERT(utext_getNativeIndex(result) == 8);
+ REGEX_ASSERT(length == 2);
+ REGEX_ASSERT_UTEXT_INVARIANT("0123456789", result);
utext_close(result);
- result = matcher->group(3, &destText, status);
+
+ result = matcher->group(3, &destText, length, status);
REGEX_CHECK_STATUS;
REGEX_ASSERT(result == &destText);
- REGEX_ASSERT_UTEXT_UTF8(str_89, result);
+ REGEX_ASSERT(utext_getNativeIndex(result) == 8);
+ REGEX_ASSERT(length == 2);
+ REGEX_ASSERT_UTEXT_INVARIANT("0123456789", result);
+ utext_close(result);
+ // Capture Group number out of range.
+ status = U_ZERO_ERROR;
REGEX_ASSERT_FAIL(matcher->group(-1, status), U_INDEX_OUTOFBOUNDS_ERROR);
+ status = U_ZERO_ERROR;
REGEX_ASSERT_FAIL(matcher->group( 4, status), U_INDEX_OUTOFBOUNDS_ERROR);
+ status = U_ZERO_ERROR;
matcher->reset();
REGEX_ASSERT_FAIL(matcher->group( 0, status), U_REGEX_INVALID_STATE);
REGEX_ASSERT(result == &destText);
REGEX_ASSERT_UTEXT_UTF8(str_Thevalueof1isbcdefg, result);
- const char str_byitselfnogroupnumber[] = { 0x24, 0x20, 0x62, 0x79, 0x20, 0x69, 0x74, 0x73, 0x65, 0x6c, 0x66, 0x2c, 0x20, 0x6e, 0x6f, 0x20, 0x67, 0x72, 0x6f, 0x75, 0x70, 0x20, 0x6e, 0x75, 0x6d, 0x62, 0x65, 0x72, 0x20, 0x24, 0x24, 0x24, 0x00 }; /* $ by itself, no group number $$$ */
+ const char str_byitselfnogroupnumber[] = { 0x5c, 0x24, 0x20, 0x62, 0x79, 0x20, 0x69, 0x74, 0x73, 0x65, 0x6c,
+ 0x66, 0x2c, 0x20, 0x6e, 0x6f, 0x20, 0x67, 0x72, 0x6f, 0x75, 0x70, 0x20, 0x6e, 0x75, 0x6d, 0x62,
+ 0x65, 0x72, 0x20, 0x5c, 0x24, 0x5c, 0x24, 0x5c, 0x24, 0x00 }; /* \$ by itself, no group number \$\$\$ */
utext_openUTF8(&replText, str_byitselfnogroupnumber, -1, &status);
result = matcher2->replaceFirst(&replText, NULL, status);
REGEX_CHECK_STATUS;
delete pat1;
+ //
+ // split of a UText based string, with library allocating output UTexts.
+ //
+ {
+ status = U_ZERO_ERROR;
+ RegexMatcher matcher(UnicodeString("(:)"), 0, status);
+ UnicodeString stringToSplit("first:second:third");
+ UText *textToSplit = utext_openUnicodeString(NULL, &stringToSplit, &status);
+ REGEX_CHECK_STATUS;
+
+ UText *splits[10] = {NULL};
+ int32_t numFields = matcher.split(textToSplit, splits, UPRV_LENGTHOF(splits), status);
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(numFields == 5);
+ REGEX_ASSERT_UTEXT_INVARIANT("first", splits[0]);
+ REGEX_ASSERT_UTEXT_INVARIANT(":", splits[1]);
+ REGEX_ASSERT_UTEXT_INVARIANT("second", splits[2]);
+ REGEX_ASSERT_UTEXT_INVARIANT(":", splits[3]);
+ REGEX_ASSERT_UTEXT_INVARIANT("third", splits[4]);
+ REGEX_ASSERT(splits[5] == NULL);
+
+ for (int i=0; i<UPRV_LENGTHOF(splits); i++) {
+ if (splits[i]) {
+ utext_close(splits[i]);
+ splits[i] = NULL;
+ }
+ }
+ utext_close(textToSplit);
+ }
+
+
//
// RegexPattern::pattern() and patternText()
//
regextst_openUTF8FromInvariant(&re1, helloWorldInvariant, -1, &status);
pat1 = RegexPattern::compile(&re1, pe, status);
REGEX_CHECK_STATUS;
- REGEX_ASSERT_UNISTR(pat1->pattern(),"(Hello, world)*");
+ REGEX_ASSERT_UNISTR("(Hello, world)*", pat1->pattern());
REGEX_ASSERT_UTEXT_INVARIANT("(Hello, world)*", pat1->patternText(status));
delete pat1;
if (UTF8Matcher == NULL) {
// UTF-8 does not allow unpaired surrogates, so this could actually happen without being a failure of the engine
- logln("Unable to create UTF-8 matcher, skipping UTF-8 tests for %s:%d", srcPath, line);
+ logln("Unable to create UTF-8 matcher, skipping UTF-8 tests for %s:%d", srcPath, line);
status = U_ZERO_ERROR;
}
}
// Generate native indices for UTF8 versions of region and capture group info
//
if (UTF8Matcher != NULL) {
+ if (flags.indexOf((UChar)0x74) >= 0) { // 't' trace flag
+ UTF8Matcher->setTrace(TRUE);
+ }
if (regionStart>=0) (void) utextOffsetToNative(&inputText, regionStart, regionStartUTF8);
if (regionEnd>=0) (void) utextOffsetToNative(&inputText, regionEnd, regionEndUTF8);
}
}
matcher->setTrace(FALSE);
+ if (UTF8Matcher) {
+ UTF8Matcher->setTrace(FALSE);
+ }
if (U_FAILURE(status)) {
errln("Error at line %d. ICU ErrorCode is %s", u_errorName(status));
}
failed = TRUE;
goto cleanupAndReturn;
}
+ if (isMatch && groupStarts.size() == 0) {
+ errln("Error at line %d: No match expected, but one found at position %d.", line, matcher->start(status));
+ failed = TRUE;
+ }
+ if (UTF8Matcher && isUTF8Match && groupStarts.size() == 0) {
+ errln("Error at line %d: No match expected, but one found at position %d (UTF-8).", line, UTF8Matcher->start(status));
+ failed = TRUE;
+ }
if (flags.indexOf((UChar)0x47 /*G*/) >= 0) {
// Only check for match / no match. Don't check capture groups.
- if (isMatch && groupStarts.size() == 0) {
- errln("Error at line %d: No match expected, but one found.", line);
- failed = TRUE;
- } else if (UTF8Matcher != NULL && isUTF8Match && groupStarts.size() == 0) {
- errln("Error at line %d: No match expected, but one found. (UTF8)", line);
- failed = TRUE;
- }
goto cleanupAndReturn;
}
REGEX_ERR("abc{1,,2}",1,7, U_REGEX_BAD_INTERVAL);
REGEX_ERR("abc{1,2a}",1,8, U_REGEX_BAD_INTERVAL);
REGEX_ERR("abc{222222222222222222222}",1,14, U_REGEX_NUMBER_TOO_BIG);
- REGEX_ERR("abc{5,50000000000}", 1, 17, U_REGEX_NUMBER_TOO_BIG); // Overflows int during scan
+ REGEX_ERR("abc{5,50000000000}", 1, 16, U_REGEX_NUMBER_TOO_BIG); // Overflows int during scan
REGEX_ERR("abc{5,687865858}", 1, 16, U_REGEX_NUMBER_TOO_BIG); // Overflows regex binary format
REGEX_ERR("abc{687865858,687865859}", 1, 24, U_REGEX_NUMBER_TOO_BIG);
REGEX_ASSERT(matcher.matches(status)==FALSE);
REGEX_ASSERT(status == U_REGEX_STOPPED_BY_CALLER);
REGEX_ASSERT(cbInfo.numCalls == 4);
+
+ // A longer running find that the callback function will abort.
+ status = U_ZERO_ERROR;
+ cbInfo.reset(4);
+ s = "aaaaaaaaaaaaaaaaaaaaaaab";
+ matcher.reset(s);
+ REGEX_ASSERT(matcher.find(status)==FALSE);
+ REGEX_ASSERT(status == U_REGEX_STOPPED_BY_CALLER);
+ REGEX_ASSERT(cbInfo.numCalls == 4);
}
void reset(int32_t max) {maxCalls=max; numCalls=0;lastIndex=0;};
};
+// call-back function for find().
+// Return TRUE to continue the find().
+// Return FALSE to stop the find().
U_CDECL_BEGIN
static UBool U_CALLCONV
testProgressCallBackFn(const void *context, int64_t matchIndex) {
const void *returnedContext;
URegexFindProgressCallback *returnedFn;
UErrorCode status = U_ZERO_ERROR;
- RegexMatcher matcher(UNICODE_STRING_SIMPLE("((.)+\\2)+x"), 0, status); // A pattern that can run long.
+ RegexMatcher matcher(UNICODE_STRING_SIMPLE("((.)\\2)x"), 0, status);
REGEX_CHECK_STATUS;
matcher.setFindProgressCallback(testProgressCallBackFn, &cbInfo, status);
REGEX_CHECK_STATUS;
REGEX_ASSERT(returnedFn == testProgressCallBackFn);
REGEX_ASSERT(returnedContext == &cbInfo);
- // A short-running match should NOT invoke the callback.
+ // A find that matches on the initial position does NOT invoke the callback.
status = U_ZERO_ERROR;
cbInfo.reset(100);
- UnicodeString s = "abxxx";
+ UnicodeString s = "aaxxx";
matcher.reset(s);
#if 0
matcher.setTrace(TRUE);
REGEX_CHECK_STATUS;
REGEX_ASSERT(cbInfo.numCalls == 0);
- // A medium running match that causes matcher.find() to invoke our callback for each index.
+ // A medium running find() that causes matcher.find() to invoke our callback for each index,
+ // but not so many times that we interrupt the operation.
status = U_ZERO_ERROR;
s = "aaaaaaaaaaaaaaaaaaab";
cbInfo.reset(s.length()); // Some upper limit for number of calls that is greater than size of our input string
cbInfo.reset(s1.length() - 5); // Bail early somewhere near the end of input string
matcher.reset(s1);
REGEX_ASSERT(matcher.find(0, status)==FALSE);
- REGEX_CHECK_STATUS;
+ REGEX_ASSERT(status == U_REGEX_STOPPED_BY_CALLER);
REGEX_ASSERT(cbInfo.numCalls == s1.length() - 5);
-#if 0
// Now a match that will succeed, but after an interruption
status = U_ZERO_ERROR;
UnicodeString s2 = "aaaaaaaaaaaaaa aaaaaaaaab xxx";
cbInfo.reset(s2.length() - 10); // Bail early somewhere near the end of input string
matcher.reset(s2);
REGEX_ASSERT(matcher.find(0, status)==FALSE);
- REGEX_CHECK_STATUS;
+ REGEX_ASSERT(status == U_REGEX_STOPPED_BY_CALLER);
// Now retry the match from where left off
cbInfo.maxCalls = 100; // No callback limit
+ status = U_ZERO_ERROR;
REGEX_ASSERT(matcher.find(cbInfo.lastIndex, status));
REGEX_CHECK_STATUS;
-#endif
}
UChar text1[80];
UText *actual;
UBool result;
- u_uastrncpy(text1, "noise abc interior def, and this is off the end", sizeof(text1)/2);
+ int64_t length = 0;
+
+ u_uastrncpy(text1, "noise abc interior def, and this is off the end", UPRV_LENGTHOF(text1));
+ // 012345678901234567890123456789012345678901234567
+ // 0 1 2 3 4
status = U_ZERO_ERROR;
re = uregex_openC("abc(.*?)def", 0, NULL, &status);
result = uregex_find(re, 0, &status);
REGEX_ASSERT(result==TRUE);
- /* Capture Group 0, the full match. Should succeed. */
+ /* Capture Group 0, the full match. Should succeed. "abc interior def" */
status = U_ZERO_ERROR;
- actual = uregex_groupUTextDeep(re, 0, &bufferText, &status);
+ actual = uregex_groupUText(re, 0, &bufferText, &length, &status);
REGEX_CHECK_STATUS;
REGEX_ASSERT(actual == &bufferText);
- REGEX_ASSERT_UTEXT_INVARIANT("abc interior def", actual);
+ REGEX_ASSERT(utext_getNativeIndex(actual) == 6);
+ REGEX_ASSERT(length == 16);
+ REGEX_ASSERT(utext_nativeLength(actual) == 47);
- /* Capture group #1. Should succeed. */
+ /* Capture group #1. Should succeed, matching " interior ". */
status = U_ZERO_ERROR;
- actual = uregex_groupUTextDeep(re, 1, &bufferText, &status);
+ actual = uregex_groupUText(re, 1, &bufferText, &length, &status);
REGEX_CHECK_STATUS;
REGEX_ASSERT(actual == &bufferText);
- REGEX_ASSERT_UTEXT_INVARIANT(" interior ", actual);
+ REGEX_ASSERT(utext_getNativeIndex(actual) == 9); // position of " interior "
+ REGEX_ASSERT(length == 10);
+ REGEX_ASSERT(utext_nativeLength(actual) == 47);
/* Capture group out of range. Error. */
status = U_ZERO_ERROR;
- actual = uregex_groupUTextDeep(re, 2, &bufferText, &status);
+ actual = uregex_groupUText(re, 2, &bufferText, &length, &status);
REGEX_ASSERT(status == U_INDEX_OUTOFBOUNDS_ERROR);
REGEX_ASSERT(actual == &bufferText);
-
uregex_close(re);
}
UChar text2[80];
UText replText = UTEXT_INITIALIZER;
UText *result;
+ status = U_ZERO_ERROR;
+ utext_openUnicodeString(&bufferText, &buffer, &status);
status = U_ZERO_ERROR;
- u_uastrncpy(text1, "Replace xaax x1x x...x.", sizeof(text1)/2);
- u_uastrncpy(text2, "No match here.", sizeof(text2)/2);
+ u_uastrncpy(text1, "Replace xaax x1x x...x.", UPRV_LENGTHOF(text1));
+ u_uastrncpy(text2, "No match here.", UPRV_LENGTHOF(text2)/2);
regextst_openUTF8FromInvariant(&replText, "<$1>", -1, &status);
re = uregex_openC("x(.*?)x", 0, NULL, &status);
/* Normal case, with match */
uregex_setText(re, text1, -1, &status);
+ REGEX_CHECK_STATUS;
utext_replace(&bufferText, 0, utext_nativeLength(&bufferText), NULL, 0, &status);
+ REGEX_CHECK_STATUS;
result = uregex_replaceFirstUText(re, &replText, &bufferText, &status);
REGEX_CHECK_STATUS;
REGEX_ASSERT(result == &bufferText);
/* Unicode escapes */
uregex_setText(re, text1, -1, &status);
- regextst_openUTF8FromInvariant(&replText, "\\\\\\u0041$1\\U00000042$\\a", -1, &status);
+ regextst_openUTF8FromInvariant(&replText, "\\\\\\u0041$1\\U00000042\\$\\a", -1, &status);
utext_replace(&bufferText, 0, utext_nativeLength(&bufferText), NULL, 0, &status);
result = uregex_replaceFirstUText(re, &replText, &bufferText, &status);
REGEX_CHECK_STATUS;
utext_close(&patternText);
}
+
+//--------------------------------------------------------------
+//
+// NamedCapture Check basic named capture group functionality
+//
+//--------------------------------------------------------------
+void RegexTest::NamedCapture() {
+ UErrorCode status = U_ZERO_ERROR;
+ RegexPattern *pat = RegexPattern::compile(UnicodeString(
+ "abc()()(?<three>xyz)(de)(?<five>hmm)(?<six>oh)f\\k<five>"), 0, status);
+ REGEX_CHECK_STATUS;
+ int32_t group = pat->groupNumberFromName("five", -1, status);
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(5 == group);
+ group = pat->groupNumberFromName("three", -1, status);
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(3 == group);
+
+ status = U_ZERO_ERROR;
+ group = pat->groupNumberFromName(UnicodeString("six"), status);
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(6 == group);
+
+ status = U_ZERO_ERROR;
+ group = pat->groupNumberFromName(UnicodeString("nosuch"), status);
+ U_ASSERT(status == U_REGEX_INVALID_CAPTURE_GROUP_NAME);
+
+ status = U_ZERO_ERROR;
+
+ // After copying a pattern, named capture should still work in the copy.
+ RegexPattern *copiedPat = new RegexPattern(*pat);
+ REGEX_ASSERT(*copiedPat == *pat);
+ delete pat; pat = NULL; // Delete original, copy should have no references back to it.
+
+ group = copiedPat->groupNumberFromName("five", -1, status);
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(5 == group);
+ group = copiedPat->groupNumberFromName("three", -1, status);
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(3 == group);
+ delete copiedPat;
+
+ // ReplaceAll with named capture group.
+ status = U_ZERO_ERROR;
+ UnicodeString text("Substitution of <<quotes>> for <<double brackets>>");
+ RegexMatcher *m = new RegexMatcher(UnicodeString("<<(?<mid>.+?)>>"), text, 0, status);
+ REGEX_CHECK_STATUS;
+ // m.pattern().dumpPattern();
+ UnicodeString replacedText = m->replaceAll("'${mid}'", status);
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(UnicodeString("Substitution of 'quotes' for 'double brackets'") == replacedText);
+ delete m;
+
+ // ReplaceAll, allowed capture group numbers.
+ text = UnicodeString("abcmxyz");
+ m = new RegexMatcher(UnicodeString("..(?<one>m)(.)(.)"), text, 0, status);
+ REGEX_CHECK_STATUS;
+
+ status = U_ZERO_ERROR;
+ replacedText = m->replaceAll(UnicodeString("<$0>"), status); // group 0, full match, is allowed.
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(UnicodeString("a<bcmxy>z") == replacedText);
+
+ status = U_ZERO_ERROR;
+ replacedText = m->replaceAll(UnicodeString("<$1>"), status); // group 1 by number.
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(UnicodeString("a<m>z") == replacedText);
+
+ status = U_ZERO_ERROR;
+ replacedText = m->replaceAll(UnicodeString("<${one}>"), status); // group 1 by name.
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(UnicodeString("a<m>z") == replacedText);
+
+ status = U_ZERO_ERROR;
+ replacedText = m->replaceAll(UnicodeString("<$2>"), status); // group 2.
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(UnicodeString("a<x>z") == replacedText);
+
+ status = U_ZERO_ERROR;
+ replacedText = m->replaceAll(UnicodeString("<$3>"), status);
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(UnicodeString("a<y>z") == replacedText);
+
+ status = U_ZERO_ERROR;
+ replacedText = m->replaceAll(UnicodeString("<$4>"), status);
+ REGEX_ASSERT(status == U_INDEX_OUTOFBOUNDS_ERROR);
+
+ status = U_ZERO_ERROR;
+ replacedText = m->replaceAll(UnicodeString("<$04>"), status); // group 0, leading 0,
+ REGEX_CHECK_STATUS; // trailing out-of-range 4 passes through.
+ REGEX_ASSERT(UnicodeString("a<bcmxy4>z") == replacedText);
+
+ status = U_ZERO_ERROR;
+ replacedText = m->replaceAll(UnicodeString("<$000016>"), status); // Consume leading zeroes. Don't consume digits
+ REGEX_CHECK_STATUS; // that push group num out of range.
+ REGEX_ASSERT(UnicodeString("a<m6>z") == replacedText); // This is group 1.
+
+ status = U_ZERO_ERROR;
+ replacedText = m->replaceAll(UnicodeString("<$3$2$1${one}>"), status);
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(UnicodeString("a<yxmm>z") == replacedText);
+
+ status = U_ZERO_ERROR;
+ replacedText = m->replaceAll(UnicodeString("$3$2$1${one}"), status);
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(UnicodeString("ayxmmz") == replacedText);
+
+ status = U_ZERO_ERROR;
+ replacedText = m->replaceAll(UnicodeString("<${noSuchName}>"), status);
+ REGEX_ASSERT(status == U_REGEX_INVALID_CAPTURE_GROUP_NAME);
+
+ status = U_ZERO_ERROR;
+ replacedText = m->replaceAll(UnicodeString("<${invalid-name}>"), status);
+ REGEX_ASSERT(status == U_REGEX_INVALID_CAPTURE_GROUP_NAME);
+
+ status = U_ZERO_ERROR;
+ replacedText = m->replaceAll(UnicodeString("<${one"), status);
+ REGEX_ASSERT(status == U_REGEX_INVALID_CAPTURE_GROUP_NAME);
+
+ status = U_ZERO_ERROR;
+ replacedText = m->replaceAll(UnicodeString("$not a capture group"), status);
+ REGEX_ASSERT(status == U_REGEX_INVALID_CAPTURE_GROUP_NAME);
+
+ delete m;
+
+ // Repeat the above replaceAll() tests using the plain C API, which
+ // has a separate implementation internally.
+ // TODO: factor out the test data.
+
+ status = U_ZERO_ERROR;
+ URegularExpression *re = uregex_openC("..(?<one>m)(.)(.)", 0, NULL, &status);
+ REGEX_CHECK_STATUS;
+ text = UnicodeString("abcmxyz");
+ uregex_setText(re, text.getBuffer(), text.length(), &status);
+ REGEX_CHECK_STATUS;
+
+ UChar resultBuf[100];
+ int32_t resultLength;
+ UnicodeString repl;
+
+ status = U_ZERO_ERROR;
+ repl = UnicodeString("<$0>");
+ resultLength = uregex_replaceAll(re, repl.getBuffer(), repl.length(), resultBuf, UPRV_LENGTHOF(resultBuf), &status);
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(UnicodeString("a<bcmxy>z") == UnicodeString(resultBuf, resultLength));
+
+ status = U_ZERO_ERROR;
+ repl = UnicodeString("<$1>");
+ resultLength = uregex_replaceAll(re, repl.getBuffer(), repl.length(), resultBuf, UPRV_LENGTHOF(resultBuf), &status);
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(UnicodeString("a<m>z") == UnicodeString(resultBuf, resultLength));
+
+ status = U_ZERO_ERROR;
+ repl = UnicodeString("<${one}>");
+ resultLength = uregex_replaceAll(re, repl.getBuffer(), repl.length(), resultBuf, UPRV_LENGTHOF(resultBuf), &status);
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(UnicodeString("a<m>z") == UnicodeString(resultBuf, resultLength));
+
+ status = U_ZERO_ERROR;
+ repl = UnicodeString("<$2>");
+ resultLength = uregex_replaceAll(re, repl.getBuffer(), repl.length(), resultBuf, UPRV_LENGTHOF(resultBuf), &status);
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(UnicodeString("a<x>z") == UnicodeString(resultBuf, resultLength));
+
+ status = U_ZERO_ERROR;
+ repl = UnicodeString("<$3>");
+ resultLength = uregex_replaceAll(re, repl.getBuffer(), repl.length(), resultBuf, UPRV_LENGTHOF(resultBuf), &status);
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(UnicodeString("a<y>z") == UnicodeString(resultBuf, resultLength));
+
+ status = U_ZERO_ERROR;
+ repl = UnicodeString("<$4>");
+ resultLength = uregex_replaceAll(re, repl.getBuffer(), repl.length(), resultBuf, UPRV_LENGTHOF(resultBuf), &status);
+ REGEX_ASSERT(status == U_INDEX_OUTOFBOUNDS_ERROR);
+
+ status = U_ZERO_ERROR;
+ repl = UnicodeString("<$04>");
+ resultLength = uregex_replaceAll(re, repl.getBuffer(), repl.length(), resultBuf, UPRV_LENGTHOF(resultBuf), &status);
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(UnicodeString("a<bcmxy4>z") == UnicodeString(resultBuf, resultLength));
+
+ status = U_ZERO_ERROR;
+ repl = UnicodeString("<$000016>");
+ resultLength = uregex_replaceAll(re, repl.getBuffer(), repl.length(), resultBuf, UPRV_LENGTHOF(resultBuf), &status);
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(UnicodeString("a<m6>z") == UnicodeString(resultBuf, resultLength));
+
+ status = U_ZERO_ERROR;
+ repl = UnicodeString("<$3$2$1${one}>");
+ resultLength = uregex_replaceAll(re, repl.getBuffer(), repl.length(), resultBuf, UPRV_LENGTHOF(resultBuf), &status);
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(UnicodeString("a<yxmm>z") == UnicodeString(resultBuf, resultLength));
+
+ status = U_ZERO_ERROR;
+ repl = UnicodeString("$3$2$1${one}");
+ resultLength = uregex_replaceAll(re, repl.getBuffer(), repl.length(), resultBuf, UPRV_LENGTHOF(resultBuf), &status);
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(UnicodeString("ayxmmz") == UnicodeString(resultBuf, resultLength));
+
+ status = U_ZERO_ERROR;
+ repl = UnicodeString("<${noSuchName}>");
+ resultLength = uregex_replaceAll(re, repl.getBuffer(), repl.length(), resultBuf, UPRV_LENGTHOF(resultBuf), &status);
+ REGEX_ASSERT(status == U_REGEX_INVALID_CAPTURE_GROUP_NAME);
+
+ status = U_ZERO_ERROR;
+ repl = UnicodeString("<${invalid-name}>");
+ resultLength = uregex_replaceAll(re, repl.getBuffer(), repl.length(), resultBuf, UPRV_LENGTHOF(resultBuf), &status);
+ REGEX_ASSERT(status == U_REGEX_INVALID_CAPTURE_GROUP_NAME);
+
+ status = U_ZERO_ERROR;
+ repl = UnicodeString("<${one");
+ resultLength = uregex_replaceAll(re, repl.getBuffer(), repl.length(), resultBuf, UPRV_LENGTHOF(resultBuf), &status);
+ REGEX_ASSERT(status == U_REGEX_INVALID_CAPTURE_GROUP_NAME);
+
+ status = U_ZERO_ERROR;
+ repl = UnicodeString("$not a capture group");
+ resultLength = uregex_replaceAll(re, repl.getBuffer(), repl.length(), resultBuf, UPRV_LENGTHOF(resultBuf), &status);
+ REGEX_ASSERT(status == U_REGEX_INVALID_CAPTURE_GROUP_NAME);
+
+ uregex_close(re);
+}
+
+//--------------------------------------------------------------
+//
+// NamedCaptureLimits Patterns with huge numbers of named capture groups.
+// The point is not so much what the exact limit is,
+// but that a largish number doesn't hit bad non-linear performance,
+// and that exceeding the limit fails cleanly.
+//
+//--------------------------------------------------------------
+void RegexTest::NamedCaptureLimits() {
+ if (quick) {
+ logln("Skipping test. Runs in exhuastive mode only.");
+ return;
+ }
+ const int32_t goodLimit = 1000000; // Pattern w this many groups builds successfully.
+ const int32_t failLimit = 10000000; // Pattern exceeds internal limits, fails to compile.
+ char nnbuf[100];
+ UnicodeString pattern;
+ int32_t nn;
+
+ for (nn=1; nn<goodLimit; nn++) {
+ sprintf(nnbuf, "(?<nn%d>)", nn);
+ pattern.append(UnicodeString(nnbuf, -1, US_INV));
+ }
+ UErrorCode status = U_ZERO_ERROR;
+ RegexPattern *pat = RegexPattern::compile(pattern, 0, status);
+ REGEX_CHECK_STATUS;
+ for (nn=1; nn<goodLimit; nn++) {
+ sprintf(nnbuf, "nn%d", nn);
+ int32_t groupNum = pat->groupNumberFromName(nnbuf, -1, status);
+ REGEX_ASSERT(nn == groupNum);
+ if (nn != groupNum) {
+ break;
+ }
+ }
+ delete pat;
+
+ pattern.remove();
+ for (nn=1; nn<failLimit; nn++) {
+ sprintf(nnbuf, "(?<nn%d>)", nn);
+ pattern.append(UnicodeString(nnbuf, -1, US_INV));
+ }
+ status = U_ZERO_ERROR;
+ pat = RegexPattern::compile(pattern, 0, status);
+ REGEX_ASSERT(status == U_REGEX_PATTERN_TOO_BIG);
+ delete pat;
+}
+
+
//--------------------------------------------------------------
//
// Bug7651 Regex pattern that exceeds default operator stack depth in matcher.
// It should set an U_REGEX_INVALID_STATE.
UChar buf[100];
- int32_t len = uregex_group(icu_re, 0, buf, LENGTHOF(buf), &status);
+ int32_t len = uregex_group(icu_re, 0, buf, UPRV_LENGTHOF(buf), &status);
REGEX_ASSERT(status == U_REGEX_INVALID_STATE);
REGEX_ASSERT(len == 0);
utext_close(utext_txt);
}
-#endif /* !UCONFIG_NO_REGULAR_EXPRESSIONS */
+void RegexTest::TestCaseInsensitiveStarters() {
+ // Test that the data used by RegexCompile::findCaseInsensitiveStarters() hasn't
+ // become stale because of new Unicode characters.
+ // If it is stale, rerun the generation tool
+ // svn+ssh://source.icu-project.org/repos/icu/tools/trunk/unicode/c/genregexcasing
+ // and replace the embedded data in i18n/regexcmp.cpp
+ for (UChar32 cp=0; cp<=0x10ffff; cp++) {
+ if (!u_hasBinaryProperty(cp, UCHAR_CASE_SENSITIVE)) {
+ continue;
+ }
+ UnicodeSet s(cp, cp);
+ s.closeOver(USET_CASE_INSENSITIVE);
+ UnicodeSetIterator setIter(s);
+ while (setIter.next()) {
+ if (!setIter.isString()) {
+ continue;
+ }
+ const UnicodeString &str = setIter.getString();
+ UChar32 firstChar = str.char32At(0);
+ UnicodeSet starters;
+ RegexCompile::findCaseInsensitiveStarters(firstChar, &starters);
+ if (!starters.contains(cp)) {
+ errln("CaseInsensitiveStarters for \\u%x is missing character \\u%x.", cp, firstChar);
+ return;
+ }
+ }
+ }
+}
+
+
+void RegexTest::TestBug11049() {
+ // Original bug report: pattern with match start consisting of one of several individual characters,
+ // and the text being matched ending with a supplementary character. find() would read past the
+ // end of the input text when searching for potential match starting points.
+
+ // To see the problem, the text must exactly fill an allocated buffer, so that valgrind will
+ // detect the bad read.
+
+ TestCase11049("A|B|C", "a string \\ud800\\udc00", FALSE, __LINE__);
+ TestCase11049("A|B|C", "string matches at end C", TRUE, __LINE__);
+
+ // Test again with a pattern starting with a single character,
+ // which takes a different code path than starting with an OR expression,
+ // but with similar logic.
+ TestCase11049("C", "a string \\ud800\\udc00", FALSE, __LINE__);
+ TestCase11049("C", "string matches at end C", TRUE, __LINE__);
+}
+
+// Run a single test case from TestBug11049(). Internal function.
+void RegexTest::TestCase11049(const char *pattern, const char *data, UBool expectMatch, int32_t lineNumber) {
+ UErrorCode status = U_ZERO_ERROR;
+ UnicodeString patternString = UnicodeString(pattern).unescape();
+ LocalPointer<RegexPattern> compiledPat(RegexPattern::compile(patternString, 0, status));
+
+ UnicodeString dataString = UnicodeString(data).unescape();
+ UChar *exactBuffer = new UChar[dataString.length()];
+ dataString.extract(exactBuffer, dataString.length(), status);
+ UText *ut = utext_openUChars(NULL, exactBuffer, dataString.length(), &status);
+
+ LocalPointer<RegexMatcher> matcher(compiledPat->matcher(status));
+ REGEX_CHECK_STATUS;
+ matcher->reset(ut);
+ UBool result = matcher->find();
+ if (result != expectMatch) {
+ errln("File %s, line %d: expected %d, got %d. Pattern = \"%s\", text = \"%s\"",
+ __FILE__, lineNumber, expectMatch, result, pattern, data);
+ }
+
+ // Rerun test with UTF-8 input text. Won't see buffer overreads, but could see
+ // off-by-one on find() with match at the last code point.
+ // Size of the original char * data (invariant charset) will be <= than the equivalent UTF-8
+ // because string.unescape() will only shrink it.
+ char * utf8Buffer = new char[uprv_strlen(data)+1];
+ u_strToUTF8(utf8Buffer, uprv_strlen(data)+1, NULL, dataString.getBuffer(), dataString.length(), &status);
+ REGEX_CHECK_STATUS;
+ ut = utext_openUTF8(ut, utf8Buffer, -1, &status);
+ REGEX_CHECK_STATUS;
+ matcher->reset(ut);
+ result = matcher->find();
+ if (result != expectMatch) {
+ errln("File %s, line %d (UTF-8 check): expected %d, got %d. Pattern = \"%s\", text = \"%s\"",
+ __FILE__, lineNumber, expectMatch, result, pattern, data);
+ }
+ delete [] utf8Buffer;
+
+ utext_close(ut);
+ delete [] exactBuffer;
+}
+
+
+void RegexTest::TestBug11371() {
+ if (quick) {
+ logln("Skipping test. Runs in exhuastive mode only.");
+ return;
+ }
+ UErrorCode status = U_ZERO_ERROR;
+ UnicodeString patternString;
+
+ for (int i=0; i<8000000; i++) {
+ patternString.append(UnicodeString("()"));
+ }
+ LocalPointer<RegexPattern> compiledPat(RegexPattern::compile(patternString, 0, status));
+ if (status != U_REGEX_PATTERN_TOO_BIG) {
+ errln("File %s, line %d expected status=U_REGEX_PATTERN_TOO_BIG; got %s.",
+ __FILE__, __LINE__, u_errorName(status));
+ }
+
+ status = U_ZERO_ERROR;
+ patternString = "(";
+ for (int i=0; i<20000000; i++) {
+ patternString.append(UnicodeString("A++"));
+ }
+ patternString.append(UnicodeString("){0}B++"));
+ LocalPointer<RegexPattern> compiledPat2(RegexPattern::compile(patternString, 0, status));
+ if (status != U_REGEX_PATTERN_TOO_BIG) {
+ errln("File %s, line %d expected status=U_REGEX_PATTERN_TOO_BIG; got %s.",
+ __FILE__, __LINE__, u_errorName(status));
+ }
+
+ // Pattern with too much string data, such that string indexes overflow operand data field size
+ // in compiled instruction.
+ status = U_ZERO_ERROR;
+ patternString = "";
+ while (patternString.length() < 0x00ffffff) {
+ patternString.append(UnicodeString("stuff and things dont you know, these are a few of my favorite strings\n"));
+ }
+ patternString.append(UnicodeString("X? trailing string"));
+ LocalPointer<RegexPattern> compiledPat3(RegexPattern::compile(patternString, 0, status));
+ if (status != U_REGEX_PATTERN_TOO_BIG) {
+ errln("File %s, line %d expected status=U_REGEX_PATTERN_TOO_BIG; got %s.",
+ __FILE__, __LINE__, u_errorName(status));
+ }
+}
+
+void RegexTest::TestBug11480() {
+ // C API, get capture group of a group that does not participate in the match.
+ // (Returns a zero length string, with nul termination,
+ // indistinguishable from a group with a zero length match.)
+
+ UErrorCode status = U_ZERO_ERROR;
+ URegularExpression *re = uregex_openC("(A)|(B)", 0, NULL, &status);
+ REGEX_CHECK_STATUS;
+ UnicodeString text = UNICODE_STRING_SIMPLE("A");
+ uregex_setText(re, text.getBuffer(), text.length(), &status);
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(uregex_lookingAt(re, 0, &status));
+ UChar buf[10] = {(UChar)13, (UChar)13, (UChar)13, (UChar)13};
+ int32_t length = uregex_group(re, 2, buf+1, UPRV_LENGTHOF(buf)-1, &status);
+ REGEX_ASSERT(length == 0);
+ REGEX_ASSERT(buf[0] == 13);
+ REGEX_ASSERT(buf[1] == 0);
+ REGEX_ASSERT(buf[2] == 13);
+ uregex_close(re);
+
+ // UText C++ API, length of match is 0 for non-participating matches.
+ UText ut = UTEXT_INITIALIZER;
+ utext_openUnicodeString(&ut, &text, &status);
+ RegexMatcher matcher(UnicodeString("(A)|(B)"), 0, status);
+ REGEX_CHECK_STATUS;
+ matcher.reset(&ut);
+ REGEX_ASSERT(matcher.lookingAt(0, status));
+
+ // UText C++ API, Capture group 1 matches "A", position 0, length 1.
+ int64_t groupLen = -666;
+ UText group = UTEXT_INITIALIZER;
+ matcher.group(1, &group, groupLen, status);
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(groupLen == 1);
+ REGEX_ASSERT(utext_getNativeIndex(&group) == 0);
+
+ // Capture group 2, the (B), does not participate in the match.
+ matcher.group(2, &group, groupLen, status);
+ REGEX_CHECK_STATUS;
+ REGEX_ASSERT(groupLen == 0);
+ REGEX_ASSERT(matcher.start(2, status) == -1);
+ REGEX_CHECK_STATUS;
+}
+
+void RegexTest::TestBug12884() {
+ // setTimeLimit() was not effective for empty sub-patterns with large {minimum counts}
+ UnicodeString pattern(u"(((((((){120}){11}){11}){11}){80}){11}){4}");
+ UnicodeString text(u"hello");
+ UErrorCode status = U_ZERO_ERROR;
+ RegexMatcher m(pattern, text, 0, status);
+ REGEX_CHECK_STATUS;
+ m.setTimeLimit(5, status);
+ m.find(status);
+ REGEX_ASSERT(status == U_REGEX_TIME_OUT);
+
+ // Non-greedy loops. They take a different code path during matching.
+ UnicodeString ngPattern(u"(((((((){120}?){11}?){11}?){11}?){80}?){11}?){4}?");
+ status = U_ZERO_ERROR;
+ RegexMatcher ngM(ngPattern, text, 0, status);
+ REGEX_CHECK_STATUS;
+ ngM.setTimeLimit(5, status);
+ ngM.find(status);
+ REGEX_ASSERT(status == U_REGEX_TIME_OUT);
+
+ // UText, wrapping non-UTF-16 text, also takes a different execution path.
+ const char *text8 = u8"¿Qué es Unicode? Unicode proporciona un número único para cada"
+ "carácter, sin importar la plataforma, sin importar el programa,"
+ "sin importar el idioma.";
+ status = U_ZERO_ERROR;
+ LocalUTextPointer ut(utext_openUTF8(NULL, text8, -1, &status));
+ REGEX_CHECK_STATUS;
+ m.reset(ut.getAlias());
+ m.find(status);
+ REGEX_ASSERT(status == U_REGEX_TIME_OUT);
+
+ status = U_ZERO_ERROR;
+ ngM.reset(ut.getAlias());
+ ngM.find(status);
+ REGEX_ASSERT(status == U_REGEX_TIME_OUT);
+}
+
+// Bug 13631. A find() of a pattern with a zero length look-behind assertions
+// can cause a read past the end of the input text.
+// The failure is seen when running this test with Clang's Addresss Sanitizer.
+
+void RegexTest::TestBug13631() {
+ const UChar *pats[] = { u"(?<!^)",
+ u"(?<=^)",
+ nullptr
+ };
+ for (const UChar **pat=pats; *pat; ++pat) {
+ UErrorCode status = U_ZERO_ERROR;
+ UnicodeString upat(*pat);
+ RegexMatcher matcher(upat, 0, status);
+ const UChar s =u'a';
+ UText *ut = utext_openUChars(nullptr, &s, 1, &status);
+ REGEX_CHECK_STATUS;
+ matcher.reset(ut);
+ while (matcher.find()) {
+ }
+ utext_close(ut);
+ }
+}
+
+// Bug 13632 Out of bounds memory reference if a replacement string ends with a '$',
+// where a following group specification would be expected.
+// Failure shows when running the test under Clang's Address Sanitizer.
+
+void RegexTest::TestBug13632() {
+ UErrorCode status = U_ZERO_ERROR;
+ URegularExpression *re = uregex_openC(" ", 0, nullptr, &status);
+ const char16_t *sourceString = u"Hello, world.";
+ uregex_setText(re, sourceString, u_strlen(sourceString), &status);
+
+ const int32_t destCap = 20;
+ char16_t dest[destCap] = {};
+ const char16_t replacement[] = {u'x', u'$'}; // Not nul terminated string.
+ uregex_replaceAll(re, replacement, 2, dest, destCap, &status);
+
+ assertEquals("", U_REGEX_INVALID_CAPTURE_GROUP_NAME, status);
+ uregex_close(re);
+}
+
+#endif /* !UCONFIG_NO_REGULAR_EXPRESSIONS */
projects / apple / icu.git / blobdiff