[GR-54824] TRegex: fix OracleDB base case fold table.

PullRequest: graal/18073
oracle · Jun 19, 2024 · 895eb28 · 895eb28
2 parents 65d8848 + 32634a9
commit 895eb28
Show file tree

Hide file tree

Showing 3 changed files with 11 additions and 1 deletion.
diff --git a/...com.oracle.truffle.regex.test/src/com/oracle/truffle/regex/tregex/test/OracleDBTests.java b/...com.oracle.truffle.regex.test/src/com/oracle/truffle/regex/tregex/test/OracleDBTests.java
@@ -1370,6 +1370,9 @@ public void generatedTests() {
  test("[\ua762-\ua78d]", "i", "\ua7ae", 0, false);
  test("[\u1f1d-\ua7ab]", "i", "\u2c63", 0, false);
  test("[\u10b1-\u1cac]", "i", "\u1f04", 0, false);
+ test("($)*\\s*", "m", "\n ", 0, true, 0, 0, 0, 0);
+ test("$*\\s*", "m", "\n ", 0, true, 0, 0);
+ test("(^|(|a))b\\z", "", "b", 0, true, 0, 1, 0, 0, -1, -1);
  test("(a{1100,1100})\\1", "i", "a".repeat(2400), 0, true, 0, 2200, 0, 1100);
 
  /* GENERATED CODE END - KEEP THIS MARKER FOR AUTOMATIC UPDATES */

diff --git a/regex/src/com.oracle.truffle.regex/src/com/oracle/truffle/regex/tregex/dfa/DFAGenerator.java b/regex/src/com.oracle.truffle.regex/src/com/oracle/truffle/regex/tregex/dfa/DFAGenerator.java
@@ -98,6 +98,7 @@
 import com.oracle.truffle.regex.tregex.parser.Counter;
 import com.oracle.truffle.regex.tregex.parser.RegexProperties;
 import com.oracle.truffle.regex.tregex.parser.ast.CharacterClass;
+import com.oracle.truffle.regex.tregex.parser.ast.Group;
 import com.oracle.truffle.regex.tregex.parser.ast.GroupBoundaries;
 import com.oracle.truffle.regex.tregex.parser.ast.RegexAST;
 import com.oracle.truffle.regex.tregex.parser.ast.RegexASTNode;
@@ -591,8 +592,13 @@ private DFAStateNodeBuilder createState(TransitionSet<NFA, NFAState, NFAStateTra
 
  private void optimizeDFA() {
  RegexProperties props = nfa.getAst().getProperties();
+ Group root = nfa.getAst().getRoot();
 
- doSimpleCG = (isForward() || !nfa.getAst().getProperties().hasQuantifiers() && !nfa.getAst().getProperties().hasEmptyCaptureGroups()) &&
+ boolean doSimpleCGBackward = !props.hasQuantifiers() && !props.hasEmptyCaptureGroups() &&
+ (!root.hasCaret() || root.startsWithCaret()) &&
+ (!root.hasDollar() || root.endsWithDollar());
+
+ doSimpleCG = (isForward() || doSimpleCGBackward) &&
  !isBooleanMatch() &&
  executorProps.isAllowSimpleCG() &&
  !hasAmbiguousStates &&

diff --git a/...src/com.oracle.truffle.regex/src/com/oracle/truffle/regex/tregex/parser/CaseFoldData.java b/...src/com.oracle.truffle.regex/src/com/oracle/truffle/regex/tregex/parser/CaseFoldData.java
@@ -1932,6 +1932,7 @@ public void appendRangesTo(RangesBuffer buffer, int startIndex, int endIndex) {
  0x00ff10, 0x00ff19, INTEGER_OFFSET, -65248,
  0x00ff21, 0x00ff3a, INTEGER_OFFSET, -65216,
  0x00ff41, 0x00ff5a, INTEGER_OFFSET, -65248,
+ 0x010400, 0x010425, INTEGER_OFFSET, 40,
  });
  public static final CodePointSet FOLDABLE_CHARACTERS = rangeSet(0x000041, 0x00005a, 0x0000b5, 0x0000b5, 0x0000c0, 0x0000d6, 0x0000d8, 0x0000de, 0x000100, 0x000100, 0x000102, 0x000102, 0x000104,
  0x000104, 0x000106, 0x000106, 0x000108, 0x000108, 0x00010a, 0x00010a, 0x00010c, 0x00010c, 0x00010e, 0x00010e, 0x000110, 0x000110, 0x000112, 0x000112, 0x000114, 0x000114, 0x000116,