Case folding mappings

test/cctest/test-regexp.cc

Index: test/cctest/test-regexp.cc
diff --git a/test/cctest/test-regexp.cc b/test/cctest/test-regexp.cc
index 264dab31fd0e0e09579a48883e56eca905ac1bab..a1cc0e326672fb3a41286a5af6c89599fc1f56cb 100644
--- a/test/cctest/test-regexp.cc
+++ b/test/cctest/test-regexp.cc
@@ -707,6 +707,37 @@ TEST(AddInverseToTable) {
 }
 
 
+static uc32 canonicalize(uc32 c) {
+  unibrow::uchar canon[unibrow::kMaxMappingSize];
+  int count = unibrow::Ecma262Canonicalize::Convert(c, '\0', canon, NULL);
+  if (count == 0) {
+    return c;
+  } else {
+    CHECK_EQ(1, count);
+    return canon[0];
+  }
+}
+
+
+TEST(LatinCanonicalize) {
+  unibrow::Mapping<unibrow::Ecma262UnCanonicalize> un_canonicalize;
+  for (char lower = 'a'; lower <= 'z'; lower++) {
+    char upper = lower + ('A' - 'a');

[Lasse Reichstein, 2008/11/17 10:30:42]

Would it be reasonable to check every code point in the range 0..0xffff, to see
that it agrees with toUpperCase?

[Christian Plesner Hansen, 2008/11/17 10:47:33]

Certainly would; I've added it.

+    CHECK_EQ(canonicalize(lower), canonicalize(upper));
+    unibrow::uchar uncanon[unibrow::kMaxMappingSize];
+    int length = un_canonicalize.get(lower, '\0', uncanon);
+    CHECK_EQ(2, length);
+    CHECK_EQ(upper, uncanon[0]);
+    CHECK_EQ(lower, uncanon[1]);
+  }
+  for (uc32 c = 128; c < (1 << 21); c++) {
+    // These exceptions are caused by a known bug in the implementation.
+    if (c != 0x026B && c != 0x027D)
+      CHECK(canonicalize(c) >= 128);
+  }
+}
+
+
 TEST(Graph) {
   Execute("fo[ob]ar|[ba]z|x[yz]*", "", true);
 }