Make JSON UTF-8 boundary check inclusive of the largest possible UTF-…

…8 character. UTF-8 acceptable codepoints are 0x0000-0x10FFFF, previously we treated this as an inclusive-exclusive bound, but UTF-8 does include it's topmost character. PiperOrigin-RevId: 572111435
protocolbuffers · Oct 10, 2023 · e80b8ec · e80b8ec
1 parent d0d52fe
commit e80b8ec
Show file tree

Hide file tree

Showing 2 changed files with 10 additions and 1 deletion.
diff --git a/src/google/protobuf/json/internal/lexer.cc b/src/google/protobuf/json/internal/lexer.cc
@@ -335,7 +335,7 @@ absl::StatusOr<size_t> JsonLexer::ParseUnicodeEscape(char out_utf8[4]) {
     out_utf8[1] = ((rune >> 6) & 0x3f) | 0x80;
     out_utf8[2] = ((rune >> 0) & 0x3f) | 0x80;
     return 3;
-  } else if (rune < 0x10ffff) {
+  } else if (rune <= 0x10ffff) {
     out_utf8[0] = ((rune >> 18) & 0x07) | 0xF0;
     out_utf8[1] = ((rune >> 12) & 0x3f) | 0x80;
     out_utf8[2] = ((rune >> 6) & 0x3f) | 0x80;

diff --git a/src/google/protobuf/json/internal/lexer_test.cc b/src/google/protobuf/json/internal/lexer_test.cc
@@ -294,6 +294,15 @@ TEST(LexerTest, SimpleString) {
   });
 }
 
+TEST(LexerTest, UTFBoundaries) {
+  Do(R"json("\u0001\u07FF\uFFFF\uDBFF\uDFFF")json",
+     [](io::ZeroCopyInputStream* stream) {
+       EXPECT_THAT(Value::Parse(stream),
+                   IsOkAndHolds(ValueIs<std::string>(
+                       "\x01\xdf\xbf\xef\xbf\xbf\xf4\x8f\xbf\xbf")));
+     });
+}
+
 TEST(NonStandard, SingleQuoteString) {
   DoLegacy(R"json('My String')json", [=](const Value& value) {
     EXPECT_THAT(value, ValueIs<std::string>("My String"));