Fix UTF-8 validity checks to not do unaligned reads.

1 files changed, 12 insertions, 2 deletions

diff --git a/src/google/protobuf/stubs/structurally_valid_unittest.cc b/src/google/protobuf/stubs/structurally_valid_unittest.cc
index 22825516..90888885 100644
--- a/src/google/protobuf/stubs/structurally_valid_unittest.cc
+++ b/src/google/protobuf/stubs/structurally_valid_unittest.cc
@@ -13,15 +13,25 @@ TEST(StructurallyValidTest, ValidUTF8String) {
   // On GCC, this string can be written as:
   //   "abcd 1234 - \u2014\u2013\u2212"
   // MSVC seems to interpret \u differently.
-  string valid_str("abcd 1234 - \342\200\224\342\200\223\342\210\222");
+  string valid_str("abcd 1234 - \342\200\224\342\200\223\342\210\222 - xyz789");
   EXPECT_TRUE(IsStructurallyValidUTF8(valid_str.data(),
                                       valid_str.size()));
+  // Additional check for pointer alignment
+  for (int i = 1; i < 8; ++i) {
+    EXPECT_TRUE(IsStructurallyValidUTF8(valid_str.data() + i,
+                                        valid_str.size() - i));
+  }
 }
 
 TEST(StructurallyValidTest, InvalidUTF8String) {
-  string invalid_str("\xA0\xB0");
+  const string invalid_str("abcd\xA0\xB0\xA0\xB0\xA0\xB0 - xyz789");
   EXPECT_FALSE(IsStructurallyValidUTF8(invalid_str.data(),
                                        invalid_str.size()));
+  // Additional check for pointer alignment
+  for (int i = 1; i < 8; ++i) {
+    EXPECT_FALSE(IsStructurallyValidUTF8(invalid_str.data() + i,
+                                         invalid_str.size() - i));
+  }
 }
 
 }  // namespace