7189363: Regex Pattern compilation buggy for special sequences
Summary: fixed the incorrect implementation in expr(...)
Reviewed-by: psandoz, alanb
--- a/jdk/src/share/classes/java/util/regex/Pattern.java Thu Aug 09 14:52:12 2012 +0400
+++ b/jdk/src/share/classes/java/util/regex/Pattern.java Thu Aug 09 10:15:26 2012 -0700
@@ -1983,6 +1983,7 @@
private Node expr(Node end) {
Node prev = null;
Node firstTail = null;
+ Branch branch = null;
Node branchConn = null;
for (;;) {
@@ -2006,8 +2007,8 @@
// the "tail.next" of each atom goes to branchConn
nodeTail.next = branchConn;
}
- if (prev instanceof Branch) {
- ((Branch)prev).add(node);
+ if (prev == branch) {
+ branch.add(node);
} else {
if (prev == end) {
prev = null;
@@ -2016,7 +2017,7 @@
// when put the "prev" into the branch as the first atom.
firstTail.next = branchConn;
}
- prev = new Branch(prev, node, branchConn);
+ prev = branch = new Branch(prev, node, branchConn);
}
}
if (peek() != '|') {
--- a/jdk/test/java/util/regex/RegExTest.java Thu Aug 09 14:52:12 2012 +0400
+++ b/jdk/test/java/util/regex/RegExTest.java Thu Aug 09 10:15:26 2012 -0700
@@ -33,7 +33,7 @@
* 5013885 5003322 4988891 5098443 5110268 6173522 4829857 5027748 6376940
* 6358731 6178785 6284152 6231989 6497148 6486934 6233084 6504326 6635133
* 6350801 6676425 6878475 6919132 6931676 6948903 6990617 7014645 7039066
- * 7067045 7014640
+ * 7067045 7014640 7189363
*/
import java.util.regex.*;
@@ -143,6 +143,7 @@
unicodeClassesTest();
horizontalAndVerticalWSTest();
linebreakTest();
+ branchTest();
if (failure) {
throw new
RuntimeException("RegExTest failed, 1st failure: " +
@@ -3918,4 +3919,32 @@
report("linebreakTest");
}
+ // #7189363
+ private static void branchTest() throws Exception {
+ if (!Pattern.compile("(a)?bc|d").matcher("d").find() || // greedy
+ !Pattern.compile("(a)+bc|d").matcher("d").find() ||
+ !Pattern.compile("(a)*bc|d").matcher("d").find() ||
+ !Pattern.compile("(a)??bc|d").matcher("d").find() || // reluctant
+ !Pattern.compile("(a)+?bc|d").matcher("d").find() ||
+ !Pattern.compile("(a)*?bc|d").matcher("d").find() ||
+ !Pattern.compile("(a)?+bc|d").matcher("d").find() || // possessive
+ !Pattern.compile("(a)++bc|d").matcher("d").find() ||
+ !Pattern.compile("(a)*+bc|d").matcher("d").find() ||
+ !Pattern.compile("(a)?bc|d").matcher("d").matches() || // greedy
+ !Pattern.compile("(a)+bc|d").matcher("d").matches() ||
+ !Pattern.compile("(a)*bc|d").matcher("d").matches() ||
+ !Pattern.compile("(a)??bc|d").matcher("d").matches() || // reluctant
+ !Pattern.compile("(a)+?bc|d").matcher("d").matches() ||
+ !Pattern.compile("(a)*?bc|d").matcher("d").matches() ||
+ !Pattern.compile("(a)?+bc|d").matcher("d").matches() || // possessive
+ !Pattern.compile("(a)++bc|d").matcher("d").matches() ||
+ !Pattern.compile("(a)*+bc|d").matcher("d").matches() ||
+ !Pattern.compile("(a)?bc|de").matcher("de").find() || // others
+ !Pattern.compile("(a)??bc|de").matcher("de").find() ||
+ !Pattern.compile("(a)?bc|de").matcher("de").matches() ||
+ !Pattern.compile("(a)??bc|de").matcher("de").matches())
+ failCount++;
+ report("branchTest");
+ }
+
}