AuroraMiddleware/scuffed-code
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

ICU-12635 fix ranges overrun in collation weight allocator

Browse Source 
X-SVN-Rev: 40113
This commit is contained in:
Markus Scherer 2017-05-05 19:36:48 +00:00
parent 61b0f07f39
commit 89c546d69f
5 changed files with 70 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
icu4c/source/i18n/collationweights.cpp
View File

@ -527,7 +527,7 @@ CollationWeights::allocWeights(uint32_t lowerLimit, uint32_t upperLimit, int32_t
#ifdef UCOL_DEBUG
printf("lengthen the short ranges from %ld bytes to %ld and iterate\n", minLength, minLength+1);
#endif
for(int32_t i=0; ranges[i].length==minLength; ++i) {
for(int32_t i=0; i<rangeCount && ranges[i].length==minLength; ++i) {
lengthenRange(ranges[i]);
}
}

36
icu4c/source/test/intltest/apicoll.cpp
View File

@ -2470,6 +2470,41 @@ void CollationAPITest::TestBadKeywords() {
}
}
void CollationAPITest::TestGapTooSmall() {
IcuTestErrorCode errorCode(*this, "TestGapTooSmall");
// Try to tailor >20k characters into a too-small primary gap between symbols
// that have 3-byte primary weights.
// In FractionalUCA.txt:
// 263A; [0C BA D0, 05, 05] # Zyyy So [084A.0020.0002] * WHITE SMILING FACE
// 263B; [0C BA D7, 05, 05] # Zyyy So [084B.0020.0002] * BLACK SMILING FACE
{
RuleBasedCollator(u"&☺<*\u4E00-\u9FFF", errorCode);
if(errorCode.isSuccess()) {
errln("no exception for primary-gap overflow");
} else if(errorCode.get() == U_BUFFER_OVERFLOW_ERROR) {
// This is the expected error.
// assertTrue("exception message mentions 'gap'", e.getMessage().contains("gap"));
} else {
errln("unexpected error for primary-gap overflow: %s", errorCode.errorName());
}
errorCode.reset();
}
// CLDR 32/ICU 60 FractionalUCA.txt makes room at the end of the symbols range
// for several 2-byte primaries, or a large number of 3-byters.
// The reset point is primary-before what should be
// the special currency-first-primary contraction,
// which is hopefully fairly stable, but not guaranteed stable.
// In FractionalUCA.txt:
// FDD1 20AC; [0D 70 02, 05, 05] # CURRENCY first primary
{
RuleBasedCollator coll(u"&[before 1]\uFDD1€<*\u4E00-\u9FFF", errorCode);
assertTrue("tailored Han before currency", coll.compare(u"\u4E00", u"$", errorCode) < 0);
errorCode.logIfFailureAndReset(
"unexpected exception for tailoring many characters at the end of symbols");
}
}
void CollationAPITest::dump(UnicodeString msg, RuleBasedCollator* c, UErrorCode& status) {
const char* bigone = "One";
const char* littleone = "one";
@ -2512,6 +2547,7 @@ void CollationAPITest::runIndexedTest( int32_t index, UBool exec, const char* &n
TESTCASE_AUTO(TestCloneBinary);
TESTCASE_AUTO(TestIterNumeric);
TESTCASE_AUTO(TestBadKeywords);
TESTCASE_AUTO(TestGapTooSmall);
TESTCASE_AUTO_END;
}

1
icu4c/source/test/intltest/apicoll.h
View File

@ -173,6 +173,7 @@ public:
void TestCloneBinary();
void TestIterNumeric();
void TestBadKeywords();
void TestGapTooSmall();
private:
// If this is too small for the test data, just increase it.

2
icu4j/main/classes/collate/src/com/ibm/icu/impl/coll/CollationWeights.java
View File

@ -110,7 +110,7 @@ public final class CollationWeights {
/* no good match, lengthen all minLength ranges and iterate */
// printf("lengthen the short ranges from %ld bytes to %ld and iterate\n", minLength, minLength+1);
for(int i=0; ranges[i].length==minLength; ++i) {
for (int i = 0; i < rangeCount && ranges[i].length == minLength; ++i) {
lengthenRange(ranges[i]);
}
}

31
icu4j/main/tests/collate/src/com/ibm/icu/dev/test/collator/CollationAPITest.java
View File

@ -1655,4 +1655,35 @@ public class CollationAPITest extends TestFmwk {
errln("Collator.getInstance(" + localeID + ") did not fail as expected - " + other);
}
}
@Test
public void TestGapTooSmall() {
// Try to tailor >20k characters into a too-small primary gap between symbols
// that have 3-byte primary weights.
// In FractionalUCA.txt:
// 263A; [0C BA D0, 05, 05] # Zyyy So [084A.0020.0002] * WHITE SMILING FACE
// 263B; [0C BA D7, 05, 05] # Zyyy So [084B.0020.0002] * BLACK SMILING FACE
try {
new RuleBasedCollator("&☺<*\u4E00-\u9FFF");
errln("no exception for primary-gap overflow");
} catch (UnsupportedOperationException e) {
assertTrue("exception message mentions 'gap'", e.getMessage().contains("gap"));
} catch (Exception e) {
errln("unexpected exception for primary-gap overflow: " + e);
}
// CLDR 32/ICU 60 FractionalUCA.txt makes room at the end of the symbols range
// for several 2-byte primaries, or a large number of 3-byters.
// The reset point is primary-before what should be
// the special currency-first-primary contraction,
// which is hopefully fairly stable, but not guaranteed stable.
// In FractionalUCA.txt:
// FDD1 20AC; [0D 70 02, 05, 05] # CURRENCY first primary
try {
Collator coll = new RuleBasedCollator("&[before 1]\uFDD1€<*\u4E00-\u9FFF");
assertTrue("tailored Han before currency", coll.compare("\u4E00", "$") < 0);
} catch (Exception e) {
errln("unexpected exception for tailoring many characters at the end of symbols: " + e);
}
}
}