Fixed uncanonicalization inconsistency.

test/cctest/test-regexp.cc

 // Copyright 2006-2008 the V8 project authors. All rights reserved.
 // Redistribution and use in source and binary forms, with or without
 // modification, are permitted provided that the following conditions are
 // met:
 //
 //     * Redistributions of source code must retain the above copyright
 //       notice, this list of conditions and the following disclaimer.
 //     * Redistributions in binary form must reproduce the above
 //       copyright notice, this list of conditions and the following
 //       disclaimer in the documentation and/or other materials provided
@@ skipping 1068 matching lines @@
   // Check that we arrive at the same results when skipping over
   // canonicalization ranges.
   int next_block = 0;
   while (next_block < CharacterRange::kRangeCanonicalizeMax) {
     uc32 start = CanonRange(next_block);
     CHECK_NE((start & CharacterRange::kStartMarker), 0);
     unsigned dist = start & CharacterRange::kPayloadMask;
     unibrow::uchar first[unibrow::Ecma262UnCanonicalize::kMaxWidth];
     int first_length = un_canonicalize.get(next_block, '\0', first);
     for (unsigned i = 1; i < dist; i++) {
-      CHECK_EQ(i, CanonRange(i));
+      uc32 i_range = CanonRange(next_block + i);
+      if (next_block + i == 880)

[Erik Corry, 2008/11/25 18:24:51]

I don't understand this change.

[Christian Plesner Hansen, 2008/11/26 06:03:45]

That was a mistake.

+        CanonRange(next_block + i);
+      CHECK_EQ(i, i_range);
       unibrow::uchar succ[unibrow::Ecma262UnCanonicalize::kMaxWidth];
       int succ_length = un_canonicalize.get(next_block + i, '\0', succ);
       CHECK_EQ(first_length, succ_length);
       for (int j = 0; j < succ_length; j++) {
         int calc = first[j] + i;
         int found = succ[j];
         CHECK_EQ(calc, found);
       }
     }
     next_block = next_block + dist;
   }
 }
 
 
+TEST(UncanonicalizeEquivalence) {
+  unibrow::Mapping<unibrow::Ecma262UnCanonicalize> un_canonicalize;
+  unibrow::uchar chars[unibrow::Ecma262UnCanonicalize::kMaxWidth];
+  for (int i = 0; i < (1 << 16); i++) {
+    int length = un_canonicalize.get(i, '\0', chars);
+    for (int j = 0; j < length; j++) {
+      unibrow::uchar chars2[unibrow::Ecma262UnCanonicalize::kMaxWidth];
+      int length2 = un_canonicalize.get(chars[j], '\0', chars2);
+      CHECK_EQ(length, length2);
+      for (int k = 0; k < length; k++)
+        CHECK_EQ(static_cast<int>(chars[k]), static_cast<int>(chars2[k]));
+    }
+  }
+}
+
+
 static void TestRangeCaseIndependence(CharacterRange input,
                                       Vector<CharacterRange> expected) {
   ZoneScope zone_scope(DELETE_ON_EXIT);
   int count = expected.length();
   ZoneList<CharacterRange>* list = new ZoneList<CharacterRange>(count);
   input.AddCaseEquivalents(list);
   CHECK_EQ(count, list->length());
   for (int i = 0; i < list->length(); i++) {
     CHECK_EQ(expected[i].from(), list->at(i).from());
     CHECK_EQ(expected[i].to(), list->at(i).to());
@@ skipping 35 matching lines @@
   // whole block at a time.
   TestSimpleRangeCaseIndependence(CharacterRange('A', 'k'),
                                   CharacterRange('a', 'z'));
 }
 
 
 TEST(Graph) {
   V8::Initialize(NULL);
   Execute("(x)?\\1y", "", true);
 }