src/java.base/share/classes/java/util/regex/Pattern.java

rev 58483 : [mq]: 8237599-Greedy-matching-against-supplementary-chars-does-not-respect-the-region

*** 4338,4355 ****
          CharPropertyGreedy(CharProperty cp, int cmin) {
              this.predicate = cp.predicate;
              this.cmin = cmin;
          }
          boolean match(Matcher matcher, int i, CharSequence seq) {
              int n = 0;
              int to = matcher.to;
              // greedy, all the way down
              while (i < to) {
                  int ch = Character.codePointAt(seq, i);
                  if (!predicate.is(ch))
                     break;
!                 i += Character.charCount(ch);
                  n++;
              }
              if (i >= to) {
                  matcher.hitEnd = true;
              }
--- 4338,4363 ----
          CharPropertyGreedy(CharProperty cp, int cmin) {
              this.predicate = cp.predicate;
              this.cmin = cmin;
          }
          boolean match(Matcher matcher, int i, CharSequence seq) {
+             int starti = i;
              int n = 0;
              int to = matcher.to;
              // greedy, all the way down
              while (i < to) {
                  int ch = Character.codePointAt(seq, i);
+                 int len = Character.charCount(ch);
+                 if (i + len > to) {
+                     // the region cut off the high half of a surrogate pair
+                     matcher.hitEnd = true;
+                     ch = seq.charAt(i);
+                     len = 1;
+                 }
                  if (!predicate.is(ch))
                      break;
!                 i += len;
                  n++;
              }
              if (i >= to) {
                  matcher.hitEnd = true;
              }
*** 4358,4368 ****
                      return true;
                  if (n == cmin)
                      return false;
                   // backing off if match fails
                  int ch = Character.codePointBefore(seq, i);
!                 i -= Character.charCount(ch);
                  n--;
              }
              return false;
          }
  
--- 4366,4377 ----
                      return true;
                  if (n == cmin)
                      return false;
                  // backing off if match fails
                  int ch = Character.codePointBefore(seq, i);
!                 // check if the region cut off the low half of a surrogate pair
!                 i = Math.max(starti, i - Character.charCount(ch));
                  n--;
              }
              return false;
          }