From 4da7ffaa3648386704e1e4c6cbd1447bbac1678f Mon Sep 17 00:00:00 2001
From: Frank Tang <ftang@chromium.org>
Date: Fri, 8 Dec 2023 12:41:52 -0800
Subject: [PATCH] ICU-22580 Address infinity loop in RBBI

ICU-22580 Fix tests
---
 icu4c/source/common/rbbiscan.cpp              |  3 ++
 icu4c/source/test/intltest/rbbiapts.cpp       | 30 ++++++++++++++++++
 icu4c/source/test/intltest/rbbiapts.h         |  1 +
 .../com/ibm/icu/text/RBBIRuleScanner.java     |  5 ++-
 .../ibm/icu/dev/test/rbbi/RBBIAPITest.java    | 31 +++++++++++++++++++
 5 files changed, 69 insertions(+), 1 deletion(-)

diff --git a/icu4c/source/common/rbbiscan.cpp b/icu4c/source/common/rbbiscan.cpp
index 336cd943702..df75212a191 100644
--- a/icu4c/source/common/rbbiscan.cpp
+++ b/icu4c/source/common/rbbiscan.cpp
@@ -934,6 +934,9 @@ void RBBIRuleScanner::nextChar(RBBIRuleChar &c) {
         }
     }
 
+    if (c.fChar == (UChar32)-1) {
+        return;
+    }
     if (fQuoteMode) {
         c.fEscaped = true;
     }
diff --git a/icu4c/source/test/intltest/rbbiapts.cpp b/icu4c/source/test/intltest/rbbiapts.cpp
index d6d1a890d49..9013f18bb7d 100644
--- a/icu4c/source/test/intltest/rbbiapts.cpp
+++ b/icu4c/source/test/intltest/rbbiapts.cpp
@@ -881,6 +881,35 @@ void RBBIAPITest::TestBug2190() {
      delete bi;
 }
 
+void RBBIAPITest::TestBug22580() {
+     UParseError    parseError;
+     // Test single ' will not cause infinity loop
+     {
+         UnicodeString rulesString = u"'";
+         UErrorCode status=U_ZERO_ERROR;
+         RuleBasedBreakIterator(rulesString, parseError, status);
+     }
+     if (quick) {
+         return;
+     }
+     // Test any 1 or 2 ASCII chars as rule will not cause infinity loop.
+     // only in exhaust mode
+     for (char16_t u1 = u' '; u1 <= u'~'; u1++) {
+         {
+             UnicodeString rule = u1;
+             UErrorCode status=U_ZERO_ERROR;
+             RuleBasedBreakIterator bi (rule, parseError, status);
+         }
+         for (char16_t u2 = u' '; u2 <= u'~'; u2++) {
+             {
+                 UnicodeString rule;
+                 rule.append(u1).append(u2);
+                 UErrorCode status=U_ZERO_ERROR;
+                 RuleBasedBreakIterator bi (rule, parseError, status);
+             }
+         }
+     }
+}
 
 void RBBIAPITest::TestRegistration() {
 #if !UCONFIG_NO_SERVICE
@@ -1428,6 +1457,7 @@ void RBBIAPITest::runIndexedTest( int32_t index, UBool exec, const char* &name,
     TESTCASE_AUTO(TestQuoteGrouping);
     TESTCASE_AUTO(TestRuleStatusVec);
     TESTCASE_AUTO(TestBug2190);
+    TESTCASE_AUTO(TestBug22580);
 #if !UCONFIG_NO_FILE_IO
     TESTCASE_AUTO(TestRegistration);
     TESTCASE_AUTO(TestBoilerPlate);
diff --git a/icu4c/source/test/intltest/rbbiapts.h b/icu4c/source/test/intltest/rbbiapts.h
index d65a2bc5219..dd0613817ac 100644
--- a/icu4c/source/test/intltest/rbbiapts.h
+++ b/icu4c/source/test/intltest/rbbiapts.h
@@ -83,6 +83,7 @@ public:
     void TestRuleStatusVec();
 
     void TestBug2190();
+    void TestBug22580();
 
     void TestBoilerPlate();
 
diff --git a/icu4j/main/core/src/main/java/com/ibm/icu/text/RBBIRuleScanner.java b/icu4j/main/core/src/main/java/com/ibm/icu/text/RBBIRuleScanner.java
index bc114b2d2d1..0b90431a3ab 100644
--- a/icu4j/main/core/src/main/java/com/ibm/icu/text/RBBIRuleScanner.java
+++ b/icu4j/main/core/src/main/java/com/ibm/icu/text/RBBIRuleScanner.java
@@ -770,7 +770,7 @@ class RBBIRuleScanner {
         //  These are recognized in all contexts, whether in quoted text or not.
         //
         if (c.fChar == '\'') {
-            if (UTF16.charAt(fRB.fRules, fNextIndex) == '\'') {
+            if (fNextIndex < fRB.fRules.length() && UTF16.charAt(fRB.fRules, fNextIndex) == '\'') {
                 c.fChar = nextCharLL(); // get nextChar officially so character counts
                 c.fEscaped = true; //   stay correct.
             } else {
@@ -787,6 +787,9 @@ class RBBIRuleScanner {
                 return;
             }
         }
+        if (c.fChar == -1) {
+            return;
+        }
 
         if (fQuoteMode) {
             c.fEscaped = true;
diff --git a/icu4j/main/core/src/test/java/com/ibm/icu/dev/test/rbbi/RBBIAPITest.java b/icu4j/main/core/src/test/java/com/ibm/icu/dev/test/rbbi/RBBIAPITest.java
index 8064bf841fa..3fdd1449d7a 100644
--- a/icu4j/main/core/src/test/java/com/ibm/icu/dev/test/rbbi/RBBIAPITest.java
+++ b/icu4j/main/core/src/test/java/com/ibm/icu/dev/test/rbbi/RBBIAPITest.java
@@ -485,4 +485,35 @@ public class RBBIAPITest extends CoreTestFmwk {
         bi.setText("Here is some Text");
         TestFmwk.assertEquals("Title instance break iterator not correctly instantiated", bi.first(), 0);
     }
+
+    @Test
+    public void TestBug22580() {
+        try {
+            RuleBasedBreakIterator bi = new RuleBasedBreakIterator("'");
+        } catch(IllegalArgumentException e) {
+            // nothing.
+        }
+        boolean quick = TestFmwk.getExhaustiveness() <= 5;
+        if (quick) {
+            return;
+        }
+        // Test any 1 or 2 ASCII chars as rule will not cause infinity loop.
+        // only in exhaust mode
+        for (char u1 = ' '; u1 < '~'; u1++) {
+            try {
+                char array[] = {u1};
+                RuleBasedBreakIterator bi = new RuleBasedBreakIterator(new String(array));
+            } catch(IllegalArgumentException e) {
+                // nothing.
+            }
+            for (char u2 = ' '; u2 < '~'; u2++) {
+                try {
+                    char array[] = {u1, u2};
+                    RuleBasedBreakIterator bi = new RuleBasedBreakIterator(new String(array));
+                } catch(IllegalArgumentException e) {
+                // nothing.
+                }
+            }
+        }
+    }
 }