Use and extend the detail helpers to remove duplicated code (#5783)

* Use and extend the detail helpers to remove duplicated code

Library:
- binary_reader: format bytes with hex_byte() instead of snprintf
- add throw_type_must_be() for the 20 copies of type_error.302
- binary_reader: add last_byte_error()/unexpected_byte() for the 32
  "parse error at the last read byte" sites (replaces bon8_error)
- json_sax: add check_container_size() for out_of_range.408 and
  diagnostic_positions::set_container_start/_end()
- json_pointer: add throw_no_parent() (405) and throw_unresolved() (404)

Tests:
- unit-class_parser uses the shared utils::SaxCountdown
- move SaxEventLogger (and its ExitAfter* variants) from unit-class_parser
  and unit-deserialization into the new tests/src/test_sax.hpp

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

* Factor out more repeated error paths and test boilerplate

Library:
- add throw_cannot_use_with() for the 34 copies of type_error.304-312
  "cannot use X with Y"
- iter_impl: add throw_cannot_get_value() (invalid_iterator.214)
- parser: add syntax_error() for the 12 parse_error.101 sites
- ordered_map: share the four at() bodies via at_impl()
- json_sax_dom_callback_parser: add pop_container() for end_object()
  and end_array()
- binary_reader: build the two UBJSON/BJData length-type messages with
  concat() and last_byte_error()

Tests:
- move same_value(), the NDEBUG guard, and step 0 (parse without
  exceptions) of the seven fuzzer drivers into tests/src/fuzzer_common.hpp

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

* Name the exception ids and address review comments

Add detail::exception_id, a scoped enum with one named enumerator per
documented exception id, and use it for every id in the library. The
create() functions get an overload for it; the int overloads stay for
user code.

Following the review of #5783: add binary_reader::invalid_byte() and
length_type_error(), basic_json::throw_subscript_wrong_type(), move the
fuzzer includes and the using-declaration into fuzzer_common.hpp, and
rename test_sax.hpp to sax_event_loggers.hpp.

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

* Silence -Wweak-vtables for the SAX event loggers

The loggers moved from anonymous namespaces in the test files into
sax_event_loggers.hpp, so clang now warns that their vtables are
emitted in every translation unit.

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

* Silence MSVC 2015 C4100 in parser::syntax_error for static SAX::parse_error

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

* Inline invalid_byte() into its call sites

The wrapper only fixed the message string of unexpected_byte(), which is
the same kind of per-argument helper that was declined for
throw_type_must_be(). Call unexpected_byte("invalid byte", ...) directly.

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

---------

Signed-off-by: Niels Lohmann <mail@nlohmann.me>
This commit is contained in:
Niels Lohmann authored and GitHub committed 2026-10-11 08:21:13 +02:00
1 parent f8b47ff6f6
commit 1649d9eda4
26 files changed
+1329 -1391

No files matched your search

+83 -76
View File
@@ -176,6 +176,27 @@ template<typename ArrayType>
inline void reserve_array(ArrayType& /*arr*/, std::size_t /*len*/, priority_tag<0> /*unused*/)
{}
/*!
@brief reports an object or array whose announced size exceeds max_size()
Shared by json_sax_dom_parser and json_sax_dom_callback_parser.
@param[in] sax the SAX parser to report the error to
@param[in] len the number of elements announced by the input, or unknown_size()
@param[in] kind "object" or "array"
@param[in] ref the object or array that was just created
@return whether parsing should continue (false after reporting out_of_range.408)
*/
template<typename SAX, typename BasicJsonType>
bool check_container_size(SAX& sax, std::size_t len, const char* kind, BasicJsonType* ref)
{
if (JSON_HEDLEY_UNLIKELY(len != detail::unknown_size() && len > ref->max_size()))
{
return sax.parse_error(0, "", out_of_range::create(exception_id::container_too_large, concat("excessive ", kind, " size: ", std::to_string(len)), ref));
}
return true;
}
#if JSON_DIAGNOSTIC_POSITIONS
/*!
@brief set the diagnostic positions of a value the DOM SAX parsers just stored
@@ -185,6 +206,35 @@ befriends this struct, as the position members are private.
*/
struct diagnostic_positions
{
/*!
@param[in,out] v the object or array whose opening brace or bracket was just read
@param[in] lexer the lexer that read it, or nullptr to leave @a v alone
*/
template<typename BasicJsonType, typename LexerType>
static void set_container_start(BasicJsonType& v, LexerType* lexer)
{
if (lexer)
{
// Lexer has read the first character of the container, so
// subtract 1 from the position to get the correct start position.
v.start_position = lexer->get_position() - 1;
}
}
/*!
@param[in,out] v the object or array whose closing brace or bracket was just read
@param[in] lexer the lexer that read it, or nullptr to leave @a v alone
*/
template<typename BasicJsonType, typename LexerType>
static void set_container_end(BasicJsonType& v, LexerType* lexer)
{
if (lexer)
{
// Lexer's position is past the closing brace or bracket, so set that as the end position.
v.end_position = lexer->get_position();
}
}
/*!
@param[in,out] v the value that was just parsed
@param[in] lexer the lexer that read it, or nullptr to leave @a v alone
@@ -349,22 +399,11 @@ class json_sax_dom_parser
ref_stack.push_back(handle_value(BasicJsonType::value_t::object));
#if JSON_DIAGNOSTIC_POSITIONS
// Manually set the start position of the object here.
// Ensure this is after the call to handle_value to ensure correct start position.
if (m_lexer_ref)
{
// Lexer has read the first character of the object, so
// subtract 1 from the position to get the correct start position.
ref_stack.back()->start_position = m_lexer_ref->get_position() - 1;
}
diagnostic_positions::set_container_start(*ref_stack.back(), m_lexer_ref);
#endif
if (JSON_HEDLEY_UNLIKELY(len != detail::unknown_size() && len > ref_stack.back()->max_size()))
{
return parse_error(0, "", out_of_range::create(408, concat("excessive object size: ", std::to_string(len)), ref_stack.back()));
}
return true;
return check_container_size(*this, len, "object", ref_stack.back());
}
bool key(string_t& val)
@@ -383,11 +422,7 @@ class json_sax_dom_parser
JSON_ASSERT(ref_stack.back()->is_object());
#if JSON_DIAGNOSTIC_POSITIONS
if (m_lexer_ref)
{
// Lexer's position is past the closing brace, so set that as the end position.
ref_stack.back()->end_position = m_lexer_ref->get_position();
}
diagnostic_positions::set_container_end(*ref_stack.back(), m_lexer_ref);
#endif
ref_stack.back()->set_parents();
@@ -400,17 +435,13 @@ class json_sax_dom_parser
ref_stack.push_back(handle_value(BasicJsonType::value_t::array));
#if JSON_DIAGNOSTIC_POSITIONS
// Manually set the start position of the array here.
// Ensure this is after the call to handle_value to ensure correct start position.
if (m_lexer_ref)
{
ref_stack.back()->start_position = m_lexer_ref->get_position() - 1;
}
diagnostic_positions::set_container_start(*ref_stack.back(), m_lexer_ref);
#endif
if (JSON_HEDLEY_UNLIKELY(len != detail::unknown_size() && len > ref_stack.back()->max_size()))
if (JSON_HEDLEY_UNLIKELY(!check_container_size(*this, len, "array", ref_stack.back())))
{
return parse_error(0, "", out_of_range::create(408, concat("excessive array size: ", std::to_string(len)), ref_stack.back()));
return false;
}
if (len != detail::unknown_size())
@@ -427,11 +458,7 @@ class json_sax_dom_parser
JSON_ASSERT(ref_stack.back()->is_array());
#if JSON_DIAGNOSTIC_POSITIONS
if (m_lexer_ref)
{
// Lexer's position is past the closing bracket, so set that as the end position.
ref_stack.back()->end_position = m_lexer_ref->get_position();
}
diagnostic_positions::set_container_end(*ref_stack.back(), m_lexer_ref);
#endif
ref_stack.back()->set_parents();
@@ -611,21 +638,11 @@ class json_sax_dom_callback_parser
{
#if JSON_DIAGNOSTIC_POSITIONS
// Manually set the start position of the object here.
// Ensure this is after the call to handle_value to ensure correct start position.
if (m_lexer_ref)
{
// Lexer has read the first character of the object, so
// subtract 1 from the position to get the correct start position.
ref_stack.back()->start_position = m_lexer_ref->get_position() - 1;
}
diagnostic_positions::set_container_start(*ref_stack.back(), m_lexer_ref);
#endif
// check object limit
if (JSON_HEDLEY_UNLIKELY(len != detail::unknown_size() && len > ref_stack.back()->max_size()))
{
return parse_error(0, "", out_of_range::create(408, concat("excessive object size: ", std::to_string(len)), ref_stack.back()));
}
return check_container_size(*this, len, "object", ref_stack.back());
}
return true;
}
@@ -695,11 +712,7 @@ class json_sax_dom_callback_parser
{
#if JSON_DIAGNOSTIC_POSITIONS
if (m_lexer_ref)
{
// Lexer's position is past the closing brace, so set that as the end position.
ref_stack.back()->end_position = m_lexer_ref->get_position();
}
diagnostic_positions::set_container_end(*ref_stack.back(), m_lexer_ref);
#endif
ref_stack.back()->set_parents();
@@ -710,13 +723,7 @@ class json_sax_dom_callback_parser
}
}
JSON_ASSERT(!ref_stack.empty());
JSON_ASSERT(!keep_stack.empty());
JSON_ASSERT(!container_key_stack.empty());
ref_stack.pop_back();
keep_stack.pop_back();
const string_t object_key = std::move(container_key_stack.back());
container_key_stack.pop_back();
const string_t object_key = pop_container();
if (!ref_stack.empty() && ref_stack.back() && ref_stack.back()->is_structured())
{
@@ -742,20 +749,13 @@ class json_sax_dom_callback_parser
{
#if JSON_DIAGNOSTIC_POSITIONS
// Manually set the start position of the array here.
// Ensure this is after the call to handle_value to ensure correct start position.
if (m_lexer_ref)
{
// Lexer has read the first character of the array, so
// subtract 1 from the position to get the correct start position.
ref_stack.back()->start_position = m_lexer_ref->get_position() - 1;
}
diagnostic_positions::set_container_start(*ref_stack.back(), m_lexer_ref);
#endif
// check array limit
if (JSON_HEDLEY_UNLIKELY(len != detail::unknown_size() && len > ref_stack.back()->max_size()))
if (JSON_HEDLEY_UNLIKELY(!check_container_size(*this, len, "array", ref_stack.back())))
{
return parse_error(0, "", out_of_range::create(408, concat("excessive array size: ", std::to_string(len)), ref_stack.back()));
return false;
}
if (len != detail::unknown_size())
@@ -779,11 +779,7 @@ class json_sax_dom_callback_parser
{
#if JSON_DIAGNOSTIC_POSITIONS
if (m_lexer_ref)
{
// Lexer's position is past the closing bracket, so set that as the end position.
ref_stack.back()->end_position = m_lexer_ref->get_position();
}
diagnostic_positions::set_container_end(*ref_stack.back(), m_lexer_ref);
#endif
ref_stack.back()->set_parents();
@@ -809,13 +805,7 @@ class json_sax_dom_callback_parser
}
}
JSON_ASSERT(!ref_stack.empty());
JSON_ASSERT(!keep_stack.empty());
JSON_ASSERT(!container_key_stack.empty());
ref_stack.pop_back();
keep_stack.pop_back();
const string_t object_key = std::move(container_key_stack.back());
container_key_stack.pop_back();
const string_t object_key = pop_container();
// remove discarded value
if (!ref_stack.empty() && ref_stack.back())
@@ -902,6 +892,23 @@ class json_sax_dom_callback_parser
return string_t{};
}
/*!
@brief leave the object or array that end_object()/end_array() just closed
@return the key the container is stored under in its parent (see
container_key_stack)
*/
string_t pop_container()
{
JSON_ASSERT(!ref_stack.empty());
JSON_ASSERT(!keep_stack.empty());
JSON_ASSERT(!container_key_stack.empty());
ref_stack.pop_back();
keep_stack.pop_back();
string_t object_key = std::move(container_key_stack.back());
container_key_stack.pop_back();
return object_key;
}
/*!
@brief remove the discarded value the callback rejected from its parent,
unless it is a duplicate key's slot with a stashed previous value, in