Report BON8 input that ends after a UTF-8 lead byte as truncated

A lead byte (0xC2..0xF7) inside a string begins either another character
(if a continuation byte follows) or an integer (otherwise). When the input
ended right after the lead byte, the reader took the missing byte as "not
a continuation byte", ended the string before the lead byte, and treated
the lead byte as the start of the next value. With strict=false, a message
cut off there was therefore read as a shorter value: the 11 bytes of
"😀😀é" cut after 9 bytes gave "😀😀", and ["aé"] cut after 3 of its 5
bytes gave ["a"]. With strict=true, the input was rejected with a
misleading message ("expected end of input"), or, for a key, with
parse_error.112 instead of 110.

Either reading of the lead byte leaves the message incomplete: a string at
the end of a message must be terminated by 0xFF, so the lead byte cannot
belong to a following message. Report parse_error.110 (unexpected end of
input) for strings and keys, as the comment on get_bon8_string() already
requires and as the reference decoder (HikoGUI) does.

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

include/nlohmann/detail/input/binary_reader.hpp

@@ -3821,6 +3821,11 @@ class binary_reader
         if (0xC2 <= byte && byte <= 0xF7)
         {
             const auto second = get_bon8();
+            if (second == char_traits<char_type>::eof())
+            {
+                // the input ends inside a character or an integer
+                return unexpect_eof(input_format_t::bon8, "key");
+            }
             unget_bon8(second);
             if (is_bon8_continuation(second))
             {
@@ -3919,6 +3924,12 @@ class binary_reader
             // a lead byte ends the string if no continuation byte follows: it
             // is then the first byte of an integer
             const auto second = get_bon8();
+            if (second == char_traits<char_type>::eof())
+            {
+                // the input ends inside a character or an integer: either
+                // way, the message is incomplete
+                return unexpect_eof(input_format_t::bon8, "string");
+            }
             if (!is_bon8_continuation(second))
             {
                 unget_bon8(second);

single_include/nlohmann/json.hpp

@@ -16588,6 +16588,11 @@ class binary_reader
         if (0xC2 <= byte && byte <= 0xF7)
         {
             const auto second = get_bon8();
+            if (second == char_traits<char_type>::eof())
+            {
+                // the input ends inside a character or an integer
+                return unexpect_eof(input_format_t::bon8, "key");
+            }
             unget_bon8(second);
             if (is_bon8_continuation(second))
             {
@@ -16686,6 +16691,12 @@ class binary_reader
             // a lead byte ends the string if no continuation byte follows: it
             // is then the first byte of an integer
             const auto second = get_bon8();
+            if (second == char_traits<char_type>::eof())
+            {
+                // the input ends inside a character or an integer: either
+                // way, the message is incomplete
+                return unexpect_eof(input_format_t::bon8, "string");
+            }
             if (!is_bon8_continuation(second))
             {
                 unget_bon8(second);

tests/src/unit-bon8.cpp

@@ -531,6 +531,47 @@ TEST_CASE("BON8")
             CHECK_THROWS_WITH_AS(_ = json::from_bon8(bytes{0x87, 'a'}), "[json.exception.parse_error.110] parse error at byte 3: syntax error while parsing BON8 string: unexpected end of input", json::parse_error&);
         }
 
+        SECTION("input that ends after a UTF-8 lead byte")
+        {
+            // the lead byte begins either a character or an integer; both are
+            // incomplete, so the lead byte must not end the string before it
+            for (const bool strict :
+                    {
+                        true, false
+                    })
+            {
+                CAPTURE(strict)
+                CHECK_THROWS_WITH_AS(_ = json::from_bon8(bytes{'a', 0xC3}, strict), "[json.exception.parse_error.110] parse error at byte 3: syntax error while parsing BON8 string: unexpected end of input", json::parse_error&);
+                CHECK_THROWS_WITH_AS(_ = json::from_bon8(bytes{0x81, 'a', 0xC3}, strict), "[json.exception.parse_error.110] parse error at byte 4: syntax error while parsing BON8 string: unexpected end of input", json::parse_error&);
+                CHECK_THROWS_WITH_AS(_ = json::from_bon8(bytes{0x81, 'a', 0xE2}, strict), "[json.exception.parse_error.110] parse error at byte 4: syntax error while parsing BON8 string: unexpected end of input", json::parse_error&);
+                CHECK_THROWS_WITH_AS(_ = json::from_bon8(bytes{0x81, 'a', 0xF0}, strict), "[json.exception.parse_error.110] parse error at byte 4: syntax error while parsing BON8 string: unexpected end of input", json::parse_error&);
+                CHECK_THROWS_WITH_AS(_ = json::from_bon8(bytes{0x81, 0xC3, 0xA9, 0xC3}, strict), "[json.exception.parse_error.110] parse error at byte 5: syntax error while parsing BON8 string: unexpected end of input", json::parse_error&);
+                CHECK_THROWS_WITH_AS(_ = json::from_bon8(bytes{0x87, 'a', 0xC3}, strict), "[json.exception.parse_error.110] parse error at byte 4: syntax error while parsing BON8 string: unexpected end of input", json::parse_error&);
+                CHECK_THROWS_WITH_AS(_ = json::from_bon8(bytes{0x87, 0xC3}, strict), "[json.exception.parse_error.110] parse error at byte 3: syntax error while parsing BON8 key: unexpected end of input", json::parse_error&);
+                CHECK_THROWS_WITH_AS(_ = json::from_bon8(bytes{0x88, 'a', 0x91, 0xE2}, strict), "[json.exception.parse_error.110] parse error at byte 5: syntax error while parsing BON8 key: unexpected end of input", json::parse_error&);
+            }
+        }
+
+        SECTION("a message that is cut off is not read as a shorter value")
+        {
+            const json values = {"\xC3\xA9", "a\xE2\x82\xAC", "\xF0\x9F\x98\x80\xC3\xA9", {"a\xC3\xA9"}, {{"\xC3\xA9", "\xE2\x82\xAC"}}, {{"a", {"b\xC3\xA9", 1}}}};
+            for (const auto& j : values)
+            {
+                const bytes message = json::to_bon8(j);
+                for (std::size_t length = 0; length < message.size(); ++length)
+                {
+                    CAPTURE(j)
+                    CAPTURE(length)
+                    bytes prefix = message;
+                    prefix.resize(length);
+                    CHECK(json::from_bon8(prefix, false, false).is_discarded());
+                    // a stream is read byte by byte rather than in bulk
+                    std::istringstream stream(str(prefix));
+                    CHECK(json::from_bon8(stream, false, false).is_discarded());
+                }
+            }
+        }
+
         SECTION("invalid UTF-8")
         {
             // overlong

tests/src/unit-regression2.cpp

@@ -766,25 +766,6 @@ TEST_CASE("regression tests 2")
         CHECK(j == k);
     }
 
-#ifdef JSON_HAS_CPP_17
-    SECTION("issue #5066 - MSVC converts json to std::variant<json> via the conversion operator")
-    {
-        // std::variant<json> must not be retrievable via get<>(), because otherwise the
-        // implicit conversion operator becomes a candidate that MSVC picks over the variant's
-        // converting constructor, routing a number through the string from_json overload
-        static_assert(!nlohmann::detail::is_detected<nlohmann::detail::get_template_function, const json&, std::variant<json>>::value,
-                      "std::variant<json> must not be retrievable via get<>()");
-
-        // clang before 7 cannot instantiate libstdc++'s std::variant<json>
-#if !(defined(__clang__) && __clang_major__ < 7)
-        // push_back, not emplace_back: #5066 needs the implicit conversion
-        // from json to the vector's value type
-        std::vector<std::variant<json>> v;
-        v.push_back(json(1)); // NOLINT(hicpp-use-emplace,modernize-use-emplace)
-        CHECK(std::get<0>(v[0]) == 1);
-#endif
-    }
-#endif
 }
 
 TEST_CASE("regression test - parser callback must not lose a duplicate key's prior value")