ICU-22890 MF2: Add ICU4C test for lone surrogates

Add a test to ICU4C for handling of lone surrogates. Incidentally fix uninitialized-memory bug in MessageFormatter (initialize `errors` to nullptr) Co-authored-by: Frank Tang <ftang@chromium.org>
unicode-org · Sep 19, 2024 · b3efbe2 · b3efbe2
1 parent 3205630
commit b3efbe2
Show file tree

Hide file tree

Showing 3 changed files with 33 additions and 1 deletion.
diff --git a/icu4c/source/i18n/unicode/messageformat2.h b/icu4c/source/i18n/unicode/messageformat2.h
@@ -378,7 +378,7 @@ namespace message2 {
         // Must be a raw pointer to avoid including the internal header file
         // defining StaticErrors
         // Owned by `this`
-        StaticErrors* errors;
+        StaticErrors* errors = nullptr;
 
     }; // class MessageFormatter
 

diff --git a/icu4c/source/test/intltest/messageformat2test.cpp b/icu4c/source/test/intltest/messageformat2test.cpp
@@ -30,6 +30,8 @@ TestMessageFormat2::runIndexedTest(int32_t index, UBool exec,
     TESTCASE_AUTO(testAPI);
     TESTCASE_AUTO(testAPISimple);
     TESTCASE_AUTO(testDataModelAPI);
+    TESTCASE_AUTO(testHighLoneSurrogate);
+    TESTCASE_AUTO(testLowLoneSurrogate);
     TESTCASE_AUTO(dataDrivenTests);
     TESTCASE_AUTO_END;
 }
@@ -263,6 +265,34 @@ void TestMessageFormat2::testAPICustomFunctions() {
     delete person;
 }
 
+// ICU-22890 lone surrogate cause infinity loop
+void TestMessageFormat2::testHighLoneSurrogate() {
+    IcuTestErrorCode errorCode(*this, "testHighLoneSurrogate");
+    UParseError pe = { 0, 0, {0}, {0} };
+    // Lone surrogate with only high surrogate
+    UnicodeString loneSurrogate({0xda02, 0});
+    icu::message2::MessageFormatter msgfmt1 =
+      icu::message2::MessageFormatter::Builder(errorCode)
+      .setPattern(loneSurrogate, pe, errorCode)
+      .build(errorCode);
+    UnicodeString result = msgfmt1.formatToString({}, errorCode);
+    errorCode.expectErrorAndReset(U_MF_SYNTAX_ERROR, "testHighLoneSurrogate");
+}
+
+// ICU-22890 lone surrogate cause infinity loop
+void TestMessageFormat2::testLowLoneSurrogate() {
+    IcuTestErrorCode errorCode(*this, "testLowLoneSurrogate");
+    UParseError pe = { 0, 0, {0}, {0} };
+    // Lone surrogate with only low surrogate
+    UnicodeString loneSurrogate({0xdc02, 0});
+    icu::message2::MessageFormatter msgfmt2 =
+      icu::message2::MessageFormatter::Builder(errorCode)
+      .setPattern(loneSurrogate, pe, errorCode)
+      .build(errorCode);
+    UnicodeString result = msgfmt2.formatToString({}, errorCode);
+    errorCode.expectErrorAndReset(U_MF_SYNTAX_ERROR, "testLowLoneSurrogate");
+}
+
 void TestMessageFormat2::dataDrivenTests() {
     IcuTestErrorCode errorCode(*this, "jsonTests");
 

diff --git a/icu4c/source/test/intltest/messageformat2test.h b/icu4c/source/test/intltest/messageformat2test.h
@@ -87,6 +87,8 @@ class TestMessageFormat2: public IntlTest {
     void testMessageFormatDateTimeSkeleton(message2::TestCase::Builder&, IcuTestErrorCode&);
     void testMf1Behavior(message2::TestCase::Builder&, IcuTestErrorCode&);
 
+    void testHighLoneSurrogate(void);
+    void testLowLoneSurrogate(void);
 }; // class TestMessageFormat2
 
 U_NAMESPACE_BEGIN