Merge branch 'develop' into claude/noexception-value-json-pointer-5672

Conflicts:
- tests/src/unit-disabled_exceptions.cpp: kept both new sections (#5672 value(json_pointer) test and develop's ordered_json growth test)

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

docs/mkdocs/docs/api/operator_gtgt.md

@@ -18,6 +18,10 @@ Deserializes an input stream to a JSON value.
 
 the stream `i`
 
+## Exception safety
+
+Strong guarantee: if an exception is thrown, there are no changes in `j`.
+
 ## Exceptions
 
 - Throws [`parse_error.101`](../home/exceptions.md#jsonexceptionparse_error101) in case of an unexpected token, or if
@@ -125,3 +129,5 @@ being read.
   the stream; planned to become the default in version 4.0.0.
 - Fixed a null pointer dereference for an `std::istream` without a stream buffer (now throws `parse_error.101`), and a
   crash (`std::terminate`) when `i` has `eofbit` in its exception mask, in version 3.13.0.
+- Changed to the strong exception safety guarantee in version 3.13.0: `j` is no longer left with a partially parsed
+  value if parsing throws.

docs/mkdocs/docs/features/types/template_parameters.md

@@ -389,6 +389,7 @@ using array_t = ArrayType<basic_json, AllocatorType<basic_json>>;
 | Functionality                                                                                                                     | Additional requirement                                                                                                                                                                       |
 |-----------------------------------------------------------------------------------------------------------------------------------|----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
 | [`diff`](../../api/basic_json/diff.md), [`items`](../../api/basic_json/items.md), [`std::hash`](../../api/basic_json/std_hash.md) | conversion of a `#!cpp std::size_t` to `StringType`: either assignability from the result of `#!cpp std::to_string`, or an ADL overload `#!cpp void int_to_string(StringType&, std::size_t)` |
+| [`operator/(std::size_t)`](../../api/json_pointer/operator_slash.md)                                                              | the same conversion of a `#!cpp std::size_t` to `StringType` as `diff`, `items`, and `std::hash` above                                                                                       |
 | [`std::hash<basic_json>`](../../api/basic_json/std_hash.md)                                                                       | additionally a specialization of `#!cpp std::hash<StringType>`                                                                                                                               |
 | [`to_bson`](../../api/basic_json/to_bson.md)                                                                                      | `find(value_type)` and `npos`                                                                                                                                                                |
 | [`parse`](../../api/basic_json/parse.md) from a `string_t`                                                                        | the input adapters must accept it; otherwise pass a character range                                                                                                                          |

include/nlohmann/detail/json_pointer.hpp

@@ -26,6 +26,7 @@
 #include <nlohmann/detail/macro_scope.hpp>
 #include <nlohmann/detail/string_concat.hpp>
 #include <nlohmann/detail/string_escape.hpp>
+#include <nlohmann/detail/string_utils.hpp>
 #include <nlohmann/detail/value_t.hpp>
 
 NLOHMANN_JSON_NAMESPACE_BEGIN
@@ -116,7 +117,7 @@ class json_pointer
     /// @sa https://json.nlohmann.me/api/json_pointer/operator_slasheq/
     json_pointer& operator/=(std::size_t array_idx)
     {
-        return *this /= std::to_string(array_idx);
+        return *this /= detail::to_string<string_t>(array_idx);
     }
 
     /// @brief create a new JSON pointer by appending the right JSON pointer at the end of the left JSON pointer
@@ -678,19 +679,29 @@ class json_pointer
                         return nullptr;
                     }
 
-                    // may throw parse_error.106/109 for a malformed index; an
-                    // index that is syntactically valid but cannot be
-                    // represented (out_of_range.404/410) is treated like an
-                    // out-of-range index below
-                    typename BasicJsonType::size_type idx{};
-                    JSON_TRY
+                    // tokens that array_index() rejects with parse_error.106/109
+                    // are passed on to it; all other tokens that it would reject
+                    // with out_of_range.404/410 are detected here, so that this
+                    // also works without exceptions
+                    if (JSON_HEDLEY_UNLIKELY(reference_token.size() > 1 && !(reference_token[0] >= '1' && reference_token[0] <= '9')))
                     {
-                        idx = array_index<BasicJsonType>(reference_token);
+                        static_cast<void>(array_index<BasicJsonType>(reference_token)); // throws parse_error.106/109
                     }
-                    JSON_INTERNAL_CATCH (detail::out_of_range&)
+                    if (JSON_HEDLEY_UNLIKELY(reference_token.empty() || !std::all_of(reference_token.begin(), reference_token.end(), [](const char c)
+                {
+                    return c >= '0' && c <= '9';
+                })))
                     {
                         return nullptr;
                     }
+                    errno = 0; // strtoull() does not reset errno on success
+                    char* p_end = nullptr; // NOLINT(misc-const-correctness)
+                    const unsigned long long magnitude = std::strtoull(reference_token.data(), &p_end, 10); // NOLINT(runtime/int)
+                    if (JSON_HEDLEY_UNLIKELY(errno == ERANGE || magnitude >= static_cast<unsigned long long>((std::numeric_limits<typename BasicJsonType::size_type>::max)()))) // NOLINT(runtime/int)
+                    {
+                        return nullptr;
+                    }
+                    const auto idx = static_cast<typename BasicJsonType::size_type>(magnitude);
 
                     if (JSON_HEDLEY_UNLIKELY(idx >= ptr->m_data.m_value.array->size()))
                     {
@@ -752,7 +763,7 @@ class json_pointer
                         // would throw out_of_range.404 -- contains() must not throw (see #5395)
                         return false;
                     }
-                    if (JSON_HEDLEY_UNLIKELY(reference_token.size() == 1 && !("0" <= reference_token && reference_token <= "9")))
+                    if (JSON_HEDLEY_UNLIKELY(reference_token.size() == 1 && !('0' <= reference_token[0] && reference_token[0] <= '9')))
                     {
                         // invalid char
                         return false;
@@ -780,7 +791,7 @@ class json_pointer
                     // not throw (see #5395), so such a reference token is treated as "not found"
                     errno = 0; // strtoull() does not reset errno on success
                     char* p_end = nullptr; // NOLINT(misc-const-correctness)
-                    const unsigned long long magnitude = std::strtoull(reference_token.c_str(), &p_end, 10); // NOLINT(runtime/int)
+                    const unsigned long long magnitude = std::strtoull(reference_token.data(), &p_end, 10); // NOLINT(runtime/int)
                     if (JSON_HEDLEY_UNLIKELY(errno == ERANGE // the value exceeds ULLONG_MAX
                                              || magnitude >= static_cast<unsigned long long>((std::numeric_limits<typename BasicJsonType::size_type>::max)()))) // NOLINT(runtime/int)
                     {

include/nlohmann/json.hpp

@@ -5092,7 +5092,10 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
     /// @sa https://json.nlohmann.me/api/basic_json/operator_gtgt/
     friend std::istream& operator>>(std::istream& i, basic_json& j)
     {
-        parser(detail::input_adapter(i)).parse(false, j);
+        // parse into a temporary so that j is left unchanged if parsing fails
+        basic_json result;
+        parser(detail::input_adapter(i)).parse(false, result);
+        j = std::move(result);
         return i;
     }
 #endif  // JSON_NO_IO

single_include/nlohmann/json.hpp

@@ -19636,6 +19636,8 @@ NLOHMANN_JSON_NAMESPACE_END
 
 // #include <nlohmann/detail/string_escape.hpp>
 
+// #include <nlohmann/detail/string_utils.hpp>
+
 // #include <nlohmann/detail/value_t.hpp>
 
 
@@ -19727,7 +19729,7 @@ class json_pointer
     /// @sa https://json.nlohmann.me/api/json_pointer/operator_slasheq/
     json_pointer& operator/=(std::size_t array_idx)
     {
-        return *this /= std::to_string(array_idx);
+        return *this /= detail::to_string<string_t>(array_idx);
     }
 
     /// @brief create a new JSON pointer by appending the right JSON pointer at the end of the left JSON pointer
@@ -20289,19 +20291,29 @@ class json_pointer
                         return nullptr;
                     }
 
-                    // may throw parse_error.106/109 for a malformed index; an
-                    // index that is syntactically valid but cannot be
-                    // represented (out_of_range.404/410) is treated like an
-                    // out-of-range index below
-                    typename BasicJsonType::size_type idx{};
-                    JSON_TRY
+                    // tokens that array_index() rejects with parse_error.106/109
+                    // are passed on to it; all other tokens that it would reject
+                    // with out_of_range.404/410 are detected here, so that this
+                    // also works without exceptions
+                    if (JSON_HEDLEY_UNLIKELY(reference_token.size() > 1 && !(reference_token[0] >= '1' && reference_token[0] <= '9')))
                     {
-                        idx = array_index<BasicJsonType>(reference_token);
+                        static_cast<void>(array_index<BasicJsonType>(reference_token)); // throws parse_error.106/109
                     }
-                    JSON_INTERNAL_CATCH (detail::out_of_range&)
+                    if (JSON_HEDLEY_UNLIKELY(reference_token.empty() || !std::all_of(reference_token.begin(), reference_token.end(), [](const char c)
+                {
+                    return c >= '0' && c <= '9';
+                })))
                     {
                         return nullptr;
                     }
+                    errno = 0; // strtoull() does not reset errno on success
+                    char* p_end = nullptr; // NOLINT(misc-const-correctness)
+                    const unsigned long long magnitude = std::strtoull(reference_token.data(), &p_end, 10); // NOLINT(runtime/int)
+                    if (JSON_HEDLEY_UNLIKELY(errno == ERANGE || magnitude >= static_cast<unsigned long long>((std::numeric_limits<typename BasicJsonType::size_type>::max)()))) // NOLINT(runtime/int)
+                    {
+                        return nullptr;
+                    }
+                    const auto idx = static_cast<typename BasicJsonType::size_type>(magnitude);
 
                     if (JSON_HEDLEY_UNLIKELY(idx >= ptr->m_data.m_value.array->size()))
                     {
@@ -20363,7 +20375,7 @@ class json_pointer
                         // would throw out_of_range.404 -- contains() must not throw (see #5395)
                         return false;
                     }
-                    if (JSON_HEDLEY_UNLIKELY(reference_token.size() == 1 && !("0" <= reference_token && reference_token <= "9")))
+                    if (JSON_HEDLEY_UNLIKELY(reference_token.size() == 1 && !('0' <= reference_token[0] && reference_token[0] <= '9')))
                     {
                         // invalid char
                         return false;
@@ -20391,7 +20403,7 @@ class json_pointer
                     // not throw (see #5395), so such a reference token is treated as "not found"
                     errno = 0; // strtoull() does not reset errno on success
                     char* p_end = nullptr; // NOLINT(misc-const-correctness)
-                    const unsigned long long magnitude = std::strtoull(reference_token.c_str(), &p_end, 10); // NOLINT(runtime/int)
+                    const unsigned long long magnitude = std::strtoull(reference_token.data(), &p_end, 10); // NOLINT(runtime/int)
                     if (JSON_HEDLEY_UNLIKELY(errno == ERANGE // the value exceeds ULLONG_MAX
                                              || magnitude >= static_cast<unsigned long long>((std::numeric_limits<typename BasicJsonType::size_type>::max)()))) // NOLINT(runtime/int)
                     {
@@ -32017,7 +32029,10 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
     /// @sa https://json.nlohmann.me/api/basic_json/operator_gtgt/
     friend std::istream& operator>>(std::istream& i, basic_json& j)
     {
-        parser(detail::input_adapter(i)).parse(false, j);
+        // parse into a temporary so that j is left unchanged if parsing fails
+        basic_json result;
+        parser(detail::input_adapter(i)).parse(false, result);
+        j = std::move(result);
         return i;
     }
 #endif  // JSON_NO_IO

tests/src/unit-alt-string.cpp

@@ -352,6 +352,41 @@ TEST_CASE("alternative string type")
         CHECK(j2.flatten().unflatten() == j2);
     }
 
+    SECTION("contains(json_pointer)")
+    {
+        // contains(json_pointer) must compile and work with a string_t that has
+        // no c_str() and no comparison with const char* (see #5666)
+        auto j = alt_json::parse(R"({"foo": ["bar", "baz"]})");
+
+        // present: object key and array indices
+        CHECK(j.contains(alt_json::json_pointer("/foo")));
+        CHECK(j.contains(alt_json::json_pointer("/foo/0")));
+        CHECK(j.contains(alt_json::json_pointer("/foo/1")));
+
+        // missing: absent object key and out-of-range array index
+        CHECK_FALSE(j.contains(alt_json::json_pointer("/bar")));
+        CHECK_FALSE(j.contains(alt_json::json_pointer("/foo/2")));
+
+        // "-" always fails the range check
+        CHECK_FALSE(j.contains(alt_json::json_pointer("/foo/-")));
+
+        // an array index must not have a leading zero
+        CHECK_FALSE(j.contains(alt_json::json_pointer("/foo/01")));
+
+        // a reference token that is not a number
+        CHECK_FALSE(j.contains(alt_json::json_pointer("/foo/bar")));
+    }
+
+    SECTION("operator/(std::size_t)")
+    {
+        // json_pointer::operator/=(std::size_t) must compile without string_t
+        // being constructible from std::string (see #5666)
+        auto j = alt_json::parse(R"({"foo": ["bar", "baz"]})");
+
+        CHECK(j.at(alt_json::json_pointer("/foo") / std::size_t(0)) == j["foo"][0]);
+        CHECK(j.at(alt_json::json_pointer("/foo") / std::size_t(1)) == j["foo"][1]);
+    }
+
     SECTION("patch")
     {
         alt_json const patch1 = alt_json::parse(R"([{ "op": "add", "path": "/a/b", "value": [ "foo", "bar" ] }])");

tests/src/unit-diagnostics.cpp

@@ -19,6 +19,7 @@ using nlohmann::json;
 
 #include <map>
 #include <unordered_map>
+#include <sstream>
 
 TEST_CASE("Better diagnostics")
 {
@@ -492,6 +493,21 @@ TEST_CASE("Regression tests for extended diagnostics")
             CHECK(copy == j);
         }
     }
+
+    SECTION("Regression test for issue #5652 - operator>> leaves a partial value in its target on a parse error")
+    {
+        json j = "old value";
+        std::istringstream is("[1, x");
+        CHECK_THROWS_WITH_AS(is >> j, "[json.exception.parse_error.101] parse error at line 1, column 5: syntax error while parsing value - invalid literal; last read: '1, x'", json::parse_error);
+
+        // j must be left unchanged, as json::parse() guarantees for its result
+        CHECK(j == "old value");
+
+        // copying j must not trigger assert_invariant(): a failed parse must
+        // not leave array/object elements without a parent pointer
+        json const copy = j; // NOLINT(performance-unnecessary-copy-initialization)
+        CHECK(copy == j);
+    }
 }
 
 TEST_CASE("Better diagnostics past the descent bound of update() and merge_patch()")

tests/src/unit-disabled_exceptions.cpp

@@ -47,6 +47,21 @@ TEST_CASE("Tests with disabled exceptions")
         delete sax_no_exception::error_string; // NOLINT(cppcoreguidelines-owning-memory)
     }
 
+    SECTION("issue #5672 - value(json_pointer, default) must not abort for array tokens that are not a valid index")
+    {
+        const json j = {1, 2, 3};
+
+        // a syntactically valid index that is out of range for this array
+        CHECK(j.value("/7"_json_pointer, 42) == 42);
+        // a reference token that is not a number at all
+        CHECK(j.value("/1a"_json_pointer, 42) == 42);
+        // the empty reference token (JSON pointer "/")
+        CHECK(j.value("/"_json_pointer, 42) == 42);
+        // an index whose magnitude does not fit into size_type
+        CHECK(j.value("/99999999999999999999999"_json_pointer, 42) == 42);
+        CHECK(j.value("/18446744073709551615"_json_pointer, 42) == 42);
+    }
+
     SECTION("growing an ordered_json object")
     {
         auto j = nlohmann::ordered_json::object();

tests/src/unit-element_access2.cpp

@@ -516,6 +516,21 @@ TEST_CASE_TEMPLATE("element access 2", Json, nlohmann::json, nlohmann::ordered_j
                     CHECK(j_array.value("/-"_json_pointer, 42) == 42);
                     CHECK(j_array_const.value("/-"_json_pointer, 42) == 42);
 
+                    // Test an index with a non-digit after a valid leading digit; this is
+                    // out_of_range (not parse_error) and must not throw (see #5672)
+                    CHECK(j_array.value("/1a"_json_pointer, 42) == 42);
+                    CHECK(j_array_const.value("/1a"_json_pointer, 42) == 42);
+
+                    // Test the empty reference token (JSON pointer "/"); see #5672
+                    CHECK(j_array.value("/"_json_pointer, 42) == 42);
+                    CHECK(j_array_const.value("/"_json_pointer, 42) == 42);
+
+                    // Test an index whose magnitude does not fit into size_type (see #5672)
+                    CHECK(j_array.value("/99999999999999999999999"_json_pointer, 42) == 42);
+                    CHECK(j_array_const.value("/99999999999999999999999"_json_pointer, 42) == 42);
+                    CHECK(j_array.value("/18446744073709551615"_json_pointer, 42) == 42);
+                    CHECK(j_array_const.value("/18446744073709551615"_json_pointer, 42) == 42);
+
 #if !defined(JSON_NOEXCEPTION)
                     // Test malformed index (non-numeric) throws parse_error
                     CHECK_THROWS_WITH_AS(j_array.value("/foo"_json_pointer, 1), "[json.exception.parse_error.109] parse error: array index 'foo' is not a number", typename Json::parse_error&);