From a8ad56bd5cf29bd80eee4e22bd301d0a0b76d5e1 Mon Sep 17 00:00:00 2001
From: Stephen Flores <stephen.flores@oracle.com>
Date: Mon, 19 Dec 2011 14:14:14 -0800
Subject: [PATCH] 6990617: Regular expression doesn't match if unicode
 character next to a digit

Updated RemoveQEQuotation() to deal with this case correctly

Reviewed-by: sherman
---
 .../classes/java/util/regex/Pattern.java      | 21 ++++++++++-
 jdk/test/java/util/regex/RegExTest.java       | 37 ++++++++++++++++---
 2 files changed, 51 insertions(+), 7 deletions(-)

diff --git a/jdk/src/share/classes/java/util/regex/Pattern.java b/jdk/src/share/classes/java/util/regex/Pattern.java
index d89eca9845d..626d5b08657 100644
--- a/jdk/src/share/classes/java/util/regex/Pattern.java
+++ b/jdk/src/share/classes/java/util/regex/Pattern.java
@@ -1583,13 +1583,26 @@ loop:   for(int x=0, offset=0; x<nCodePoints; x++, offset+=len) {
             return;
         int j = i;
         i += 2;
-        int[] newtemp = new int[j + 2*(pLen-i) + 2];
+        int[] newtemp = new int[j + 3*(pLen-i) + 2];
         System.arraycopy(temp, 0, newtemp, 0, j);
 
         boolean inQuote = true;
+        boolean beginQuote = true;
         while (i < pLen) {
             int c = temp[i++];
-            if (! ASCII.isAscii(c) || ASCII.isAlnum(c)) {
+            if (!ASCII.isAscii(c) || ASCII.isAlpha(c)) {
+                newtemp[j++] = c;
+            } else if (ASCII.isDigit(c)) {
+                if (beginQuote) {
+                    /*
+                     * A unicode escape \[0xu] could be before this quote,
+                     * and we don't want this numeric char to processed as
+                     * part of the escape.
+                     */
+                    newtemp[j++] = '\\';
+                    newtemp[j++] = 'x';
+                    newtemp[j++] = '3';
+                }
                 newtemp[j++] = c;
             } else if (c != '\\') {
                 if (inQuote) newtemp[j++] = '\\';
@@ -1606,12 +1619,16 @@ loop:   for(int x=0, offset=0; x<nCodePoints; x++, offset+=len) {
                 if (temp[i] == 'Q') {
                     i++;
                     inQuote = true;
+                    beginQuote = true;
+                    continue;
                 } else {
                     newtemp[j++] = c;
                     if (i != pLen)
                         newtemp[j++] = temp[i++];
                 }
             }
+
+            beginQuote = false;
         }
 
         patternLength = j;
diff --git a/jdk/test/java/util/regex/RegExTest.java b/jdk/test/java/util/regex/RegExTest.java
index 10644e491aa..159e9afddd9 100644
--- a/jdk/test/java/util/regex/RegExTest.java
+++ b/jdk/test/java/util/regex/RegExTest.java
@@ -32,7 +32,7 @@
  * 4872664 4803179 4892980 4900747 4945394 4938995 4979006 4994840 4997476
  * 5013885 5003322 4988891 5098443 5110268 6173522 4829857 5027748 6376940
  * 6358731 6178785 6284152 6231989 6497148 6486934 6233084 6504326 6635133
- * 6350801 6676425 6878475 6919132 6931676 6948903 7014645 7039066
+ * 6350801 6676425 6878475 6919132 6931676 6948903 6990617 7014645 7039066
  */
 
 import java.util.regex.*;
@@ -50,6 +50,7 @@ public class RegExTest {
     private static Random generator = new Random();
     private static boolean failure = false;
     private static int failCount = 0;
+    private static String firstFailure = null;
 
     /**
      * Main to interpret arguments and run several tests.
@@ -133,15 +134,19 @@ public class RegExTest {
         hitEndTest();
         toMatchResultTest();
         surrogatesInClassTest();
+        removeQEQuotingTest();
         namedGroupCaptureTest();
         nonBmpClassComplementTest();
         unicodePropertiesTest();
         unicodeHexNotationTest();
         unicodeClassesTest();
-        if (failure)
-            throw new RuntimeException("Failure in the RE handling.");
-        else
+        if (failure) {
+            throw new
+                RuntimeException("RegExTest failed, 1st failure: " +
+                                 firstFailure);
+        } else {
             System.err.println("OKAY: All tests passed.");
+        }
     }
 
     // Utility functions
@@ -215,8 +220,14 @@ public class RegExTest {
         String paddedName = paddedNameBuffer.toString();
         System.err.println(paddedName + ": " +
                            (failCount==0 ? "Passed":"Failed("+failCount+")"));
-        if (failCount > 0)
+        if (failCount > 0) {
             failure = true;
+
+            if (firstFailure == null) {
+                firstFailure = testName;
+            }
+        }
+
         failCount = 0;
     }
 
@@ -295,6 +306,22 @@ public class RegExTest {
         Matcher matcher = pattern.matcher("\ud834\udd22");
         if (!matcher.find())
             failCount++;
+
+        report("Surrogate pair in Unicode escape");
+    }
+
+    // This is for bug6990617
+    // Test if Pattern.RemoveQEQuoting works correctly if the octal unicode
+    // char encoding is only 2 or 3 digits instead of 4 and the first quoted
+    // char is an octal digit.
+    private static void removeQEQuotingTest() throws Exception {
+        Pattern pattern =
+            Pattern.compile("\\011\\Q1sometext\\E\\011\\Q2sometext\\E");
+        Matcher matcher = pattern.matcher("\t1sometext\t2sometext");
+        if (!matcher.find())
+            failCount++;
+
+        report("Remove Q/E Quoting");
     }
 
     // This is for bug 4988891