Read UBJSON and BJData containers without recursing per nesting level

get_ubjson_array() and get_ubjson_object() read their elements by calling
back into the value reader, which called them again for a nested container,
so the native call stack grew with the nesting depth of the input. '[' alone
opens a container, so half a million of them crashes the process before the
input runs out (#5104). The optimized forms reach the same path through a
size or type annotation, and in plain UBJSON '[' and '{' are permitted as the
type of an optimized container, so "[$[#i\x01" repeated nests just as deeply
at six bytes a level.

Both readers now only open their container, and parse_ubjson_internal() loops:
it closes the containers that have ended, claims the next element of the
innermost one, reads its key when it is an object, and works out the marker
of the value to read next. That last part is where the formats differ, and
the loop follows what the four element loops used to do:

  - a sized, typed container gives its elements no marker of their own
  - a sized, untyped container reads one for each element
  - a container that ends at a marker has the byte already, from the test
    against ']' or '}'; for an object it is the first byte of the key

The ND-array wrapper and the 'B' binary shortcut stay as they are. Both read
a complete value rather than opening a container, and their elements are
always scalars: BJData does not permit '[' or '{' as an optimized type, which
is also why only plain UBJSON needed the type-marker case above.

A container of no-ops keeps its behaviour of holding no elements while still
announcing its declared size to the SAX parser, by opening it and then
setting its count to zero.

unit-ubjson and unit-bjdata pass unchanged, 1.39 million assertions between
them, and a behaviour comparison against the previous commit over every
container form -- sized, unsized, typed, untyped, empty, no-op, ND-array,
binary, and the forms nested inside one another -- gives identical values,
error codes, messages and byte offsets. 500,000 levels of each vector now
report a parse error instead of crashing, and a well-formed 100,000-level
value is read to completion.

The driver costs about 3 % on parsing 60,000 small objects and one array of a
million integers, for the reason given in the previous commit; reading the
frame once per element rather than per branch halved what it cost before.

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

include/nlohmann/detail/input/binary_reader.hpp

@@ -130,6 +130,7 @@ class binary_reader
                    const cbor_tag_handler_t tag_handler = cbor_tag_handler_t::error)
     {
         sax = sax_;
+        container_stack.clear();
         bool result = false;
 
         switch (format)
@@ -179,6 +180,75 @@ class binary_reader
     }
 
   private:
+    ////////////////////////
+    // nested containers  //
+    ////////////////////////
+
+    /*!
+    @brief a container that has been opened and not closed yet
+
+    The binary readers do not call themselves once per nesting level. Like
+    @ref parser::sax_parse_internal, which does the same for JSON text, they
+    keep the containers they are inside of on a heap-allocated stack, so that
+    the native call stack does not grow with the nesting depth of the input
+    and a deeply nested value is bounded by memory rather than by the stack
+    (see #5104).
+
+    The members are ordered by decreasing alignment, which is the ordering that
+    keeps a struct from growing as members are added to it.
+    */
+    struct container_frame
+    {
+        container_frame(const std::size_t remaining_, const bool is_object_,
+                        const char_int_type type_marker_ = 0) noexcept
+            : remaining(remaining_), type_marker(type_marker_), is_object(is_object_) {}
+
+        /// number of elements that have not been read yet, or npos when the
+        /// container is not sized and ends at a marker instead
+        std::size_t remaining;
+        /// UBJSON/BJData: the type marker of an optimized container, so that
+        /// its elements are read without one of their own; 0 otherwise
+        char_int_type type_marker;
+        /// whether to close this container with end_object() or end_array()
+        bool is_object;
+    };
+
+    /*!
+    @brief open a nested array or object
+
+    Emits the SAX start event and records the container. This is the only
+    place the binary readers start a container, so a check that rejects one
+    can be made here and is then guaranteed to run before the start event.
+
+    @param[in] is_object  whether an object (true) or an array (false) begins
+    @param[in] len        number of elements the container declares
+
+    @return whether the SAX parser accepted the start event
+    */
+    bool enter_container(const bool is_object, const std::size_t len,
+                         const char_int_type type_marker = 0)
+    {
+        if (JSON_HEDLEY_UNLIKELY(is_object ? !sax->start_object(len) : !sax->start_array(len)))
+        {
+            return false;
+        }
+
+        container_stack.emplace_back(len, is_object, type_marker);
+        return true;
+    }
+
+    /// @copydoc enter_container
+    bool enter_array(const std::size_t len, const char_int_type type_marker = 0)
+    {
+        return enter_container(/*is_object*/false, len, type_marker);
+    }
+
+    /// @copydoc enter_container
+    bool enter_object(const std::size_t len, const char_int_type type_marker = 0)
+    {
+        return enter_container(/*is_object*/true, len, type_marker);
+    }
+
     //////////
     // BSON //
     //////////
@@ -511,9 +581,12 @@ class binary_reader
 
     @return whether a valid CBOR value was passed to the SAX parser
     */
-    bool parse_cbor_internal(const bool get_char,
+    bool parse_cbor_value(const bool get_char,
-                             const cbor_tag_handler_t tag_handler)
+                          const cbor_tag_handler_t tag_handler,
+                          bool& tag_pending)
     {
+        tag_pending = false;
+
         switch (get_char ? get() : current)
         {
             // EOF
@@ -705,37 +778,36 @@ class binary_reader
             case 0x95:
             case 0x96:
             case 0x97:
-                return get_cbor_array(
+                return enter_array(conditional_static_cast<std::size_t>(static_cast<unsigned int>(current) & 0x1Fu));
-                           conditional_static_cast<std::size_t>(static_cast<unsigned int>(current) & 0x1Fu), tag_handler);
 
             case 0x98: // array (one-byte uint8_t for n follows)
             {
                 std::uint8_t len{};
-                return get_number(input_format_t::cbor, len) && get_cbor_array(static_cast<std::size_t>(len), tag_handler);
+                return get_number(input_format_t::cbor, len) && enter_array(static_cast<std::size_t>(len));
             }
 
             case 0x99: // array (two-byte uint16_t for n follow)
             {
                 std::uint16_t len{};
-                return get_number(input_format_t::cbor, len) && get_cbor_array(static_cast<std::size_t>(len), tag_handler);
+                return get_number(input_format_t::cbor, len) && enter_array(static_cast<std::size_t>(len));
             }
 
             case 0x9A: // array (four-byte uint32_t for n follow)
             {
                 std::uint32_t len{};
                 std::size_t size{};
-                return get_number(input_format_t::cbor, len) && get_cbor_container_size(len, size, "array") && get_cbor_array(size, tag_handler);
+                return get_number(input_format_t::cbor, len) && get_cbor_container_size(len, size, "array") && enter_array(size);
             }
 
             case 0x9B: // array (eight-byte uint64_t for n follow)
             {
                 std::uint64_t len{};
                 std::size_t size{};
-                return get_number(input_format_t::cbor, len) && get_cbor_container_size(len, size, "array") && get_cbor_array(size, tag_handler);
+                return get_number(input_format_t::cbor, len) && get_cbor_container_size(len, size, "array") && enter_array(size);
             }
 
             case 0x9F: // array (indefinite length)
-                return get_cbor_array(detail::unknown_size(), tag_handler);
+                return enter_array(detail::unknown_size());
 
             // map (0x00..0x17 pairs of data items follow)
             case 0xA0:
@@ -762,36 +834,36 @@ class binary_reader
             case 0xB5:
             case 0xB6:
             case 0xB7:
-                return get_cbor_object(conditional_static_cast<std::size_t>(static_cast<unsigned int>(current) & 0x1Fu), tag_handler);
+                return enter_object(conditional_static_cast<std::size_t>(static_cast<unsigned int>(current) & 0x1Fu));
 
             case 0xB8: // map (one-byte uint8_t for n follows)
             {
                 std::uint8_t len{};
-                return get_number(input_format_t::cbor, len) && get_cbor_object(static_cast<std::size_t>(len), tag_handler);
+                return get_number(input_format_t::cbor, len) && enter_object(static_cast<std::size_t>(len));
             }
 
             case 0xB9: // map (two-byte uint16_t for n follow)
             {
                 std::uint16_t len{};
-                return get_number(input_format_t::cbor, len) && get_cbor_object(static_cast<std::size_t>(len), tag_handler);
+                return get_number(input_format_t::cbor, len) && enter_object(static_cast<std::size_t>(len));
             }
 
             case 0xBA: // map (four-byte uint32_t for n follow)
             {
                 std::uint32_t len{};
                 std::size_t size{};
-                return get_number(input_format_t::cbor, len) && get_cbor_container_size(len, size, "map") && get_cbor_object(size, tag_handler);
+                return get_number(input_format_t::cbor, len) && get_cbor_container_size(len, size, "map") && enter_object(size);
             }
 
             case 0xBB: // map (eight-byte uint64_t for n follow)
             {
                 std::uint64_t len{};
                 std::size_t size{};
-                return get_number(input_format_t::cbor, len) && get_cbor_container_size(len, size, "map") && get_cbor_object(size, tag_handler);
+                return get_number(input_format_t::cbor, len) && get_cbor_container_size(len, size, "map") && enter_object(size);
             }
 
             case 0xBF: // map (indefinite length)
-                return get_cbor_object(detail::unknown_size(), tag_handler);
+                return enter_object(detail::unknown_size());
 
             case 0xC0: // tagged item
             case 0xC1:
@@ -875,7 +947,10 @@ class binary_reader
                             default:
                                 break;
                         }
-                        return parse_cbor_internal(true, tag_handler);
+                        // the tagged value follows; it is read by the loop in
+                        // parse_cbor_internal() rather than by recursing here
+                        tag_pending = true;
+                        return true;
                     }
 
                     case cbor_tag_handler_t::store:
@@ -925,7 +1000,11 @@ class binary_reader
                                 break;
                             }
                             default:
-                                return parse_cbor_internal(true, tag_handler);
+                            {
+                                // as above, the tagged value is read by the caller
+                                tag_pending = true;
+                                return true;
+                            }
                         }
                         get();
                         return get_cbor_binary(b) && sax->binary(b);
@@ -1323,96 +1402,110 @@ class binary_reader
     }
 
     /*!
-    @param[in] len  the length of the array or detail::unknown_size() for an
+    @brief read a CBOR value and everything nested inside it
-                    array of indefinite size
+
+    Reads values until the one that was begun here is complete, resuming the
+    enclosing container after each element, so that the nesting depth of the
+    input costs heap rather than native stack (see #5104).
+
+    @param[in] get_char  whether a new character should be retrieved from the
+                         input (true) or whether the last read character
+                         @a current should be considered instead
     @param[in] tag_handler how CBOR tags should be treated
-    @return whether array creation completed
+
+    @return whether reading the value succeeded
     */
-    bool get_cbor_array(const std::size_t len,
+    bool parse_cbor_internal(const bool get_char,
-                        const cbor_tag_handler_t tag_handler)
+                             const cbor_tag_handler_t tag_handler)
     {
-        if (JSON_HEDLEY_UNLIKELY(!sax->start_array(len)))
+        // whether the next value starts at a fresh byte or at the one already
-        {
+        // read into `current`
-            return false;
+        bool fetch = get_char;
-        }
 
-        if (len != detail::unknown_size())
+        // the key currently being read; hoisted out of the loop so that its
+        // capacity is reused across elements and across nesting levels
+        string_t key;
+
+        while (true)
         {
-            for (std::size_t i = 0; i < len; ++i)
+            if (!container_stack.empty())
             {
-                if (JSON_HEDLEY_UNLIKELY(!parse_cbor_internal(true, tag_handler)))
+                // the reference is not held across parse_cbor_value() below,
+                // which can push onto the stack and reallocate it
+                container_frame& top = container_stack.back();
+                bool at_end = false;
+
+                if (top.remaining != npos)
                 {
-                    return false;
+                    // definite length: the container ends once its elements
+                    // have been read
+                    at_end = (top.remaining == 0);
+                    if (!at_end)
+                    {
+                        // claim the element about to be read
+                        --top.remaining;
+                        if (top.is_object)
+                        {
+                            get();
+                        }
+                    }
+                    fetch = true;
                 }
-            }
+                else
-        }
-        else
-        {
-            while (get() != 0xFF)
-            {
-                if (JSON_HEDLEY_UNLIKELY(!parse_cbor_internal(false, tag_handler)))
                 {
-                    return false;
+                    // indefinite length: the container ends at a break marker.
+                    // Testing for it consumes a byte, which is the first byte
+                    // of the next element when it is not one.
+                    at_end = (get() == 0xFF);
+                    fetch = top.is_object;
                 }
-            }
-        }
 
-        return sax->end_array();
+                if (at_end)
-    }
-
-    /*!
-    @param[in] len  the length of the object or detail::unknown_size() for an
-                    object of indefinite size
-    @param[in] tag_handler how CBOR tags should be treated
-    @return whether object creation completed
-    */
-    bool get_cbor_object(const std::size_t len,
-                         const cbor_tag_handler_t tag_handler)
-    {
-        if (JSON_HEDLEY_UNLIKELY(!sax->start_object(len)))
-        {
-            return false;
-        }
-
-        if (len != 0)
-        {
-            string_t key;
-            if (len != detail::unknown_size())
-            {
-                for (std::size_t i = 0; i < len; ++i)
                 {
-                    get();
+                    const bool is_object = top.is_object;
+                    container_stack.pop_back();
+                    if (JSON_HEDLEY_UNLIKELY(is_object ? !sax->end_object() : !sax->end_array()))
+                    {
+                        return false;
+                    }
+                    // the value begun here is complete once its container is
+                    if (container_stack.empty())
+                    {
+                        return true;
+                    }
+                    continue;
+                }
+
+                if (top.is_object)
+                {
+                    key.clear();
                     if (JSON_HEDLEY_UNLIKELY(!get_cbor_string(key) || !sax->key(key)))
                     {
                         return false;
                     }
-
+                    fetch = true;
-                    if (JSON_HEDLEY_UNLIKELY(!parse_cbor_internal(true, tag_handler)))
-                    {
-                        return false;
-                    }
-                    key.clear();
                 }
             }
-            else
-            {
-                while (get() != 0xFF)
-                {
-                    if (JSON_HEDLEY_UNLIKELY(!get_cbor_string(key) || !sax->key(key)))
-                    {
-                        return false;
-                    }
 
-                    if (JSON_HEDLEY_UNLIKELY(!parse_cbor_internal(true, tag_handler)))
+            // a tag is not a value of its own: read on until the tagged value
-                    {
+            bool tag_pending = false;
-                        return false;
+            do
-                    }
+            {
-                    key.clear();
+                if (JSON_HEDLEY_UNLIKELY(!parse_cbor_value(fetch, tag_handler, tag_pending)))
+                {
+                    return false;
                 }
+                fetch = true;
+            }
+            while (tag_pending);
+
+            // a value that opened a container left it on the stack; one that
+            // did not, and that was not inside a container, was the whole value
+            if (container_stack.empty())
+            {
+                return true;
             }
         }
-
-        return sax->end_object();
     }
 
     /////////////
@@ -1422,7 +1515,17 @@ class binary_reader
     /*!
     @return whether a valid MessagePack value was passed to the SAX parser
     */
-    bool parse_msgpack_internal()
+    /*!
+    @brief read one MessagePack value
+
+    Reads a single value and passes it to the SAX parser. A value that begins
+    a container is not read to its end: the container is opened with
+    @ref enter_container and its elements are read by
+    @ref parse_msgpack_internal, so that nesting does not consume native stack.
+
+    @return whether reading the value succeeded
+    */
+    bool parse_msgpack_value()
     {
         switch (get())
         {
@@ -1578,7 +1681,7 @@ class binary_reader
             case 0x8D:
             case 0x8E:
             case 0x8F:
-                return get_msgpack_object(conditional_static_cast<std::size_t>(static_cast<unsigned int>(current) & 0x0Fu));
+                return enter_object(conditional_static_cast<std::size_t>(static_cast<unsigned int>(current) & 0x0Fu));
 
             // fixarray
             case 0x90:
@@ -1597,7 +1700,7 @@ class binary_reader
             case 0x9D:
             case 0x9E:
             case 0x9F:
-                return get_msgpack_array(conditional_static_cast<std::size_t>(static_cast<unsigned int>(current) & 0x0Fu));
+                return enter_array(conditional_static_cast<std::size_t>(static_cast<unsigned int>(current) & 0x0Fu));
 
             // fixstr
             case 0xA0:
@@ -1728,25 +1831,25 @@ class binary_reader
             case 0xDC: // array 16
             {
                 std::uint16_t len{};
-                return get_number(input_format_t::msgpack, len) && get_msgpack_array(static_cast<std::size_t>(len));
+                return get_number(input_format_t::msgpack, len) && enter_array(static_cast<std::size_t>(len));
             }
 
             case 0xDD: // array 32
             {
                 std::uint32_t len{};
-                return get_number(input_format_t::msgpack, len) && get_msgpack_array(conditional_static_cast<std::size_t>(len));
+                return get_number(input_format_t::msgpack, len) && enter_array(conditional_static_cast<std::size_t>(len));
             }
 
             case 0xDE: // map 16
             {
                 std::uint16_t len{};
-                return get_number(input_format_t::msgpack, len) && get_msgpack_object(static_cast<std::size_t>(len));
+                return get_number(input_format_t::msgpack, len) && enter_object(static_cast<std::size_t>(len));
             }
 
             case 0xDF: // map 32
             {
                 std::uint32_t len{};
-                return get_number(input_format_t::msgpack, len) && get_msgpack_object(conditional_static_cast<std::size_t>(len));
+                return get_number(input_format_t::msgpack, len) && enter_object(conditional_static_cast<std::size_t>(len));
             }
 
             // negative fixint
@@ -1994,55 +2097,69 @@ class binary_reader
     }
 
     /*!
-    @param[in] len  the length of the array
+    @brief read a MessagePack value and everything nested inside it
-    @return whether array creation completed
+
+    Reads values until the one that was begun here is complete, resuming the
+    enclosing container each time an element ends, so that the nesting depth
+    of the input costs heap rather than native stack (see #5104).
+
+    @return whether reading the value succeeded
     */
-    bool get_msgpack_array(const std::size_t len)
+    bool parse_msgpack_internal()
     {
-        if (JSON_HEDLEY_UNLIKELY(!sax->start_array(len)))
+        // the key currently being read; hoisted out of the loop so that its
-        {
+        // capacity is reused across elements and across nesting levels
-            return false;
-        }
-
-        for (std::size_t i = 0; i < len; ++i)
-        {
-            if (JSON_HEDLEY_UNLIKELY(!parse_msgpack_internal()))
-            {
-                return false;
-            }
-        }
-
-        return sax->end_array();
-    }
-
-    /*!
-    @param[in] len  the length of the object
-    @return whether object creation completed
-    */
-    bool get_msgpack_object(const std::size_t len)
-    {
-        if (JSON_HEDLEY_UNLIKELY(!sax->start_object(len)))
-        {
-            return false;
-        }
-
         string_t key;
-        for (std::size_t i = 0; i < len; ++i)
+
+        while (true)
         {
-            get();
+            if (!container_stack.empty())
-            if (JSON_HEDLEY_UNLIKELY(!get_msgpack_string(key) || !sax->key(key)))
+            {
+                // copied out before anything can push onto the stack and
+                // invalidate a reference into it
+                const bool is_object = container_stack.back().is_object;
+
+                if (container_stack.back().remaining == 0)
+                {
+                    container_stack.pop_back();
+                    if (JSON_HEDLEY_UNLIKELY(is_object ? !sax->end_object() : !sax->end_array()))
+                    {
+                        return false;
+                    }
+                    // the value begun here is complete once its container is
+                    if (container_stack.empty())
+                    {
+                        return true;
+                    }
+                    continue;
+                }
+
+                // claim the element about to be read
+                --container_stack.back().remaining;
+
+                if (is_object)
+                {
+                    get();
+                    key.clear();
+                    if (JSON_HEDLEY_UNLIKELY(!get_msgpack_string(key) || !sax->key(key)))
+                    {
+                        return false;
+                    }
+                }
+            }
+
+            if (JSON_HEDLEY_UNLIKELY(!parse_msgpack_value()))
             {
                 return false;
             }
 
-            if (JSON_HEDLEY_UNLIKELY(!parse_msgpack_internal()))
+            // a value that opened a container left it on the stack; one that
+            // did not, and that was not inside a container, was the whole value
+            if (container_stack.empty())
             {
-                return false;
+                return true;
             }
-            key.clear();
         }
-
-        return sax->end_object();
     }
 
     ////////////
@@ -2058,7 +2175,103 @@ class binary_reader
     */
     bool parse_ubjson_internal(const bool get_char = true)
     {
-        return get_ubjson_value(get_char ? get_ignore_noop() : current);
+        // the key currently being read; hoisted out of the loop so that its
+        // capacity is reused across elements and across nesting levels
+        string_t key;
+
+        // the type marker of the value to read next
+        char_int_type prefix = get_char ? get_ignore_noop() : current;
+
+        while (true)
+        {
+            const std::size_t depth = container_stack.size();
+
+            if (JSON_HEDLEY_UNLIKELY(!get_ubjson_value(prefix)))
+            {
+                return false;
+            }
+
+            // the value begun here is complete once it is not inside anything
+            if (container_stack.empty())
+            {
+                return true;
+            }
+
+            // a value was completed rather than a container opened; a
+            // container that ends at a marker needs the next byte to test
+            if (container_stack.size() == depth && container_stack.back().remaining == npos)
+            {
+                get_ignore_noop();
+            }
+
+            // advance to the next element, closing the containers that ended.
+            // The reference is not held across get_ubjson_value() above, which
+            // can push onto the stack and reallocate it.
+            for (;;)
+            {
+                container_frame& top = container_stack.back();
+
+                if (top.remaining != npos)
+                {
+                    if (top.remaining != 0)
+                    {
+                        --top.remaining;
+                        if (top.is_object)
+                        {
+                            key.clear();
+                            if (JSON_HEDLEY_UNLIKELY(!get_ubjson_string(key) || !sax->key(key)))
+                            {
+                                return false;
+                            }
+                        }
+                        // an optimized container gives its elements no marker
+                        prefix = (top.type_marker != 0) ? top.type_marker : get_ignore_noop();
+                        break;
+                    }
+                }
+                // the end marker is compared against a literal rather than
+                // against a conditional expression, because char_int_type is
+                // unsigned for some input adapters and MSVC then reports the
+                // comparison as a signed/unsigned mismatch
+                else if (top.is_object ? (current != '}') : (current != ']'))
+                {
+                    // a container that ends at a marker is never optimized, so
+                    // every element carries its own marker; for an object the
+                    // byte tested above is the first byte of the key
+                    if (top.is_object)
+                    {
+                        key.clear();
+                        if (JSON_HEDLEY_UNLIKELY(!get_ubjson_string(key, false) || !sax->key(key)))
+                        {
+                            return false;
+                        }
+                        prefix = get_ignore_noop();
+                    }
+                    else
+                    {
+                        prefix = current;
+                    }
+                    break;
+                }
+
+                const bool is_object = top.is_object;
+                container_stack.pop_back();
+                if (JSON_HEDLEY_UNLIKELY(is_object ? !sax->end_object() : !sax->end_array()))
+                {
+                    return false;
+                }
+                if (container_stack.empty())
+                {
+                    return true;
+                }
+                // the container that just ended was an element of the one
+                // below it, which may need the next byte for its own test
+                if (container_stack.back().remaining == npos)
+                {
+                    get_ignore_noop();
+                }
+            }
+        }
     }
 
     /*!
@@ -2830,53 +3043,22 @@ class binary_reader
                                         exception_message(input_format, "excessive array size", "size"), nullptr));
             }
 
-            if (JSON_HEDLEY_UNLIKELY(!sax->start_array(size_and_type.first)))
+            if (JSON_HEDLEY_UNLIKELY(!enter_array(size_and_type.first, size_and_type.second)))
             {
                 return false;
             }
 
-            if (size_and_type.second != 0)
+            if (size_and_type.second == 'N')
             {
-                if (size_and_type.second != 'N')
+                // a no-op is not a value, so a container of them holds none;
-                {
+                // the declared size has already been passed to the SAX parser
-                    for (std::size_t i = 0; i < size_and_type.first; ++i)
+                container_stack.back().remaining = 0;
-                    {
-                        if (JSON_HEDLEY_UNLIKELY(!get_ubjson_value(size_and_type.second)))
-                        {
-                            return false;
-                        }
-                    }
-                }
-            }
-            else
-            {
-                for (std::size_t i = 0; i < size_and_type.first; ++i)
-                {
-                    if (JSON_HEDLEY_UNLIKELY(!parse_ubjson_internal()))
-                    {
-                        return false;
-                    }
-                }
-            }
-        }
-        else
-        {
-            if (JSON_HEDLEY_UNLIKELY(!sax->start_array(detail::unknown_size())))
-            {
-                return false;
             }
 
-            while (current != ']')
+            return true;
-            {
-                if (JSON_HEDLEY_UNLIKELY(!parse_ubjson_internal(false)))
-                {
-                    return false;
-                }
-                get_ignore_noop();
-            }
         }
 
-        return sax->end_array();
+        return enter_array(detail::unknown_size());
     }
 
     /*!
@@ -2898,68 +3080,12 @@ class binary_reader
                                     exception_message(input_format, "BJData object does not support ND-array size in optimized format", "object"), nullptr));
         }
 
-        string_t key;
         if (size_and_type.first != npos)
         {
-            if (JSON_HEDLEY_UNLIKELY(!sax->start_object(size_and_type.first)))
+            return enter_object(size_and_type.first, size_and_type.second);
-            {
-                return false;
-            }
-
-            if (size_and_type.second != 0)
-            {
-                for (std::size_t i = 0; i < size_and_type.first; ++i)
-                {
-                    if (JSON_HEDLEY_UNLIKELY(!get_ubjson_string(key) || !sax->key(key)))
-                    {
-                        return false;
-                    }
-                    if (JSON_HEDLEY_UNLIKELY(!get_ubjson_value(size_and_type.second)))
-                    {
-                        return false;
-                    }
-                    key.clear();
-                }
-            }
-            else
-            {
-                for (std::size_t i = 0; i < size_and_type.first; ++i)
-                {
-                    if (JSON_HEDLEY_UNLIKELY(!get_ubjson_string(key) || !sax->key(key)))
-                    {
-                        return false;
-                    }
-                    if (JSON_HEDLEY_UNLIKELY(!parse_ubjson_internal()))
-                    {
-                        return false;
-                    }
-                    key.clear();
-                }
-            }
-        }
-        else
-        {
-            if (JSON_HEDLEY_UNLIKELY(!sax->start_object(detail::unknown_size())))
-            {
-                return false;
-            }
-
-            while (current != '}')
-            {
-                if (JSON_HEDLEY_UNLIKELY(!get_ubjson_string(key, false) || !sax->key(key)))
-                {
-                    return false;
-                }
-                if (JSON_HEDLEY_UNLIKELY(!parse_ubjson_internal()))
-                {
-                    return false;
-                }
-                get_ignore_noop();
-                key.clear();
-            }
         }
 
-        return sax->end_object();
+        return enter_object(detail::unknown_size());
     }
 
     // Note, no reader for UBJSON binary types is implemented because they do
@@ -3347,6 +3473,9 @@ class binary_reader
     /// the SAX parser
     json_sax_t* sax = nullptr;
 
+    /// the containers that have been opened and not closed yet; see @ref container_frame
+    std::vector<container_frame> container_stack{};
+
     // excluded markers in bjdata optimized type
 #define JSON_BINARY_READER_MAKE_BJD_OPTIMIZED_TYPE_MARKERS_ \
     make_array<char_int_type>('F', 'H', 'N', 'S', 'T', 'Z', '[', '{')

single_include/nlohmann/json.hpp

@@ -10817,6 +10817,7 @@ class binary_reader
                    const cbor_tag_handler_t tag_handler = cbor_tag_handler_t::error)
     {
         sax = sax_;
+        container_stack.clear();
         bool result = false;
 
         switch (format)
@@ -10866,6 +10867,75 @@ class binary_reader
     }
 
   private:
+    ////////////////////////
+    // nested containers  //
+    ////////////////////////
+
+    /*!
+    @brief a container that has been opened and not closed yet
+
+    The binary readers do not call themselves once per nesting level. Like
+    @ref parser::sax_parse_internal, which does the same for JSON text, they
+    keep the containers they are inside of on a heap-allocated stack, so that
+    the native call stack does not grow with the nesting depth of the input
+    and a deeply nested value is bounded by memory rather than by the stack
+    (see #5104).
+
+    The members are ordered by decreasing alignment, which is the ordering that
+    keeps a struct from growing as members are added to it.
+    */
+    struct container_frame
+    {
+        container_frame(const std::size_t remaining_, const bool is_object_,
+                        const char_int_type type_marker_ = 0) noexcept
+            : remaining(remaining_), type_marker(type_marker_), is_object(is_object_) {}
+
+        /// number of elements that have not been read yet, or npos when the
+        /// container is not sized and ends at a marker instead
+        std::size_t remaining;
+        /// UBJSON/BJData: the type marker of an optimized container, so that
+        /// its elements are read without one of their own; 0 otherwise
+        char_int_type type_marker;
+        /// whether to close this container with end_object() or end_array()
+        bool is_object;
+    };
+
+    /*!
+    @brief open a nested array or object
+
+    Emits the SAX start event and records the container. This is the only
+    place the binary readers start a container, so a check that rejects one
+    can be made here and is then guaranteed to run before the start event.
+
+    @param[in] is_object  whether an object (true) or an array (false) begins
+    @param[in] len        number of elements the container declares
+
+    @return whether the SAX parser accepted the start event
+    */
+    bool enter_container(const bool is_object, const std::size_t len,
+                         const char_int_type type_marker = 0)
+    {
+        if (JSON_HEDLEY_UNLIKELY(is_object ? !sax->start_object(len) : !sax->start_array(len)))
+        {
+            return false;
+        }
+
+        container_stack.emplace_back(len, is_object, type_marker);
+        return true;
+    }
+
+    /// @copydoc enter_container
+    bool enter_array(const std::size_t len, const char_int_type type_marker = 0)
+    {
+        return enter_container(/*is_object*/false, len, type_marker);
+    }
+
+    /// @copydoc enter_container
+    bool enter_object(const std::size_t len, const char_int_type type_marker = 0)
+    {
+        return enter_container(/*is_object*/true, len, type_marker);
+    }
+
     //////////
     // BSON //
     //////////
@@ -11198,9 +11268,12 @@ class binary_reader
 
     @return whether a valid CBOR value was passed to the SAX parser
     */
-    bool parse_cbor_internal(const bool get_char,
+    bool parse_cbor_value(const bool get_char,
-                             const cbor_tag_handler_t tag_handler)
+                          const cbor_tag_handler_t tag_handler,
+                          bool& tag_pending)
     {
+        tag_pending = false;
+
         switch (get_char ? get() : current)
         {
             // EOF
@@ -11392,37 +11465,36 @@ class binary_reader
             case 0x95:
             case 0x96:
             case 0x97:
-                return get_cbor_array(
+                return enter_array(conditional_static_cast<std::size_t>(static_cast<unsigned int>(current) & 0x1Fu));
-                           conditional_static_cast<std::size_t>(static_cast<unsigned int>(current) & 0x1Fu), tag_handler);
 
             case 0x98: // array (one-byte uint8_t for n follows)
             {
                 std::uint8_t len{};
-                return get_number(input_format_t::cbor, len) && get_cbor_array(static_cast<std::size_t>(len), tag_handler);
+                return get_number(input_format_t::cbor, len) && enter_array(static_cast<std::size_t>(len));
             }
 
             case 0x99: // array (two-byte uint16_t for n follow)
             {
                 std::uint16_t len{};
-                return get_number(input_format_t::cbor, len) && get_cbor_array(static_cast<std::size_t>(len), tag_handler);
+                return get_number(input_format_t::cbor, len) && enter_array(static_cast<std::size_t>(len));
             }
 
             case 0x9A: // array (four-byte uint32_t for n follow)
             {
                 std::uint32_t len{};
                 std::size_t size{};
-                return get_number(input_format_t::cbor, len) && get_cbor_container_size(len, size, "array") && get_cbor_array(size, tag_handler);
+                return get_number(input_format_t::cbor, len) && get_cbor_container_size(len, size, "array") && enter_array(size);
             }
 
             case 0x9B: // array (eight-byte uint64_t for n follow)
             {
                 std::uint64_t len{};
                 std::size_t size{};
-                return get_number(input_format_t::cbor, len) && get_cbor_container_size(len, size, "array") && get_cbor_array(size, tag_handler);
+                return get_number(input_format_t::cbor, len) && get_cbor_container_size(len, size, "array") && enter_array(size);
             }
 
             case 0x9F: // array (indefinite length)
-                return get_cbor_array(detail::unknown_size(), tag_handler);
+                return enter_array(detail::unknown_size());
 
             // map (0x00..0x17 pairs of data items follow)
             case 0xA0:
@@ -11449,36 +11521,36 @@ class binary_reader
             case 0xB5:
             case 0xB6:
             case 0xB7:
-                return get_cbor_object(conditional_static_cast<std::size_t>(static_cast<unsigned int>(current) & 0x1Fu), tag_handler);
+                return enter_object(conditional_static_cast<std::size_t>(static_cast<unsigned int>(current) & 0x1Fu));
 
             case 0xB8: // map (one-byte uint8_t for n follows)
             {
                 std::uint8_t len{};
-                return get_number(input_format_t::cbor, len) && get_cbor_object(static_cast<std::size_t>(len), tag_handler);
+                return get_number(input_format_t::cbor, len) && enter_object(static_cast<std::size_t>(len));
             }
 
             case 0xB9: // map (two-byte uint16_t for n follow)
             {
                 std::uint16_t len{};
-                return get_number(input_format_t::cbor, len) && get_cbor_object(static_cast<std::size_t>(len), tag_handler);
+                return get_number(input_format_t::cbor, len) && enter_object(static_cast<std::size_t>(len));
             }
 
             case 0xBA: // map (four-byte uint32_t for n follow)
             {
                 std::uint32_t len{};
                 std::size_t size{};
-                return get_number(input_format_t::cbor, len) && get_cbor_container_size(len, size, "map") && get_cbor_object(size, tag_handler);
+                return get_number(input_format_t::cbor, len) && get_cbor_container_size(len, size, "map") && enter_object(size);
             }
 
             case 0xBB: // map (eight-byte uint64_t for n follow)
             {
                 std::uint64_t len{};
                 std::size_t size{};
-                return get_number(input_format_t::cbor, len) && get_cbor_container_size(len, size, "map") && get_cbor_object(size, tag_handler);
+                return get_number(input_format_t::cbor, len) && get_cbor_container_size(len, size, "map") && enter_object(size);
             }
 
             case 0xBF: // map (indefinite length)
-                return get_cbor_object(detail::unknown_size(), tag_handler);
+                return enter_object(detail::unknown_size());
 
             case 0xC0: // tagged item
             case 0xC1:
@@ -11562,7 +11634,10 @@ class binary_reader
                             default:
                                 break;
                         }
-                        return parse_cbor_internal(true, tag_handler);
+                        // the tagged value follows; it is read by the loop in
+                        // parse_cbor_internal() rather than by recursing here
+                        tag_pending = true;
+                        return true;
                     }
 
                     case cbor_tag_handler_t::store:
@@ -11612,7 +11687,11 @@ class binary_reader
                                 break;
                             }
                             default:
-                                return parse_cbor_internal(true, tag_handler);
+                            {
+                                // as above, the tagged value is read by the caller
+                                tag_pending = true;
+                                return true;
+                            }
                         }
                         get();
                         return get_cbor_binary(b) && sax->binary(b);
@@ -12010,96 +12089,110 @@ class binary_reader
     }
 
     /*!
-    @param[in] len  the length of the array or detail::unknown_size() for an
+    @brief read a CBOR value and everything nested inside it
-                    array of indefinite size
+
+    Reads values until the one that was begun here is complete, resuming the
+    enclosing container after each element, so that the nesting depth of the
+    input costs heap rather than native stack (see #5104).
+
+    @param[in] get_char  whether a new character should be retrieved from the
+                         input (true) or whether the last read character
+                         @a current should be considered instead
     @param[in] tag_handler how CBOR tags should be treated
-    @return whether array creation completed
+
+    @return whether reading the value succeeded
     */
-    bool get_cbor_array(const std::size_t len,
+    bool parse_cbor_internal(const bool get_char,
-                        const cbor_tag_handler_t tag_handler)
+                             const cbor_tag_handler_t tag_handler)
     {
-        if (JSON_HEDLEY_UNLIKELY(!sax->start_array(len)))
+        // whether the next value starts at a fresh byte or at the one already
-        {
+        // read into `current`
-            return false;
+        bool fetch = get_char;
-        }
 
-        if (len != detail::unknown_size())
+        // the key currently being read; hoisted out of the loop so that its
+        // capacity is reused across elements and across nesting levels
+        string_t key;
+
+        while (true)
         {
-            for (std::size_t i = 0; i < len; ++i)
+            if (!container_stack.empty())
             {
-                if (JSON_HEDLEY_UNLIKELY(!parse_cbor_internal(true, tag_handler)))
+                // the reference is not held across parse_cbor_value() below,
+                // which can push onto the stack and reallocate it
+                container_frame& top = container_stack.back();
+                bool at_end = false;
+
+                if (top.remaining != npos)
                 {
-                    return false;
+                    // definite length: the container ends once its elements
+                    // have been read
+                    at_end = (top.remaining == 0);
+                    if (!at_end)
+                    {
+                        // claim the element about to be read
+                        --top.remaining;
+                        if (top.is_object)
+                        {
+                            get();
+                        }
+                    }
+                    fetch = true;
                 }
-            }
+                else
-        }
-        else
-        {
-            while (get() != 0xFF)
-            {
-                if (JSON_HEDLEY_UNLIKELY(!parse_cbor_internal(false, tag_handler)))
                 {
-                    return false;
+                    // indefinite length: the container ends at a break marker.
+                    // Testing for it consumes a byte, which is the first byte
+                    // of the next element when it is not one.
+                    at_end = (get() == 0xFF);
+                    fetch = top.is_object;
                 }
-            }
-        }
 
-        return sax->end_array();
+                if (at_end)
-    }
-
-    /*!
-    @param[in] len  the length of the object or detail::unknown_size() for an
-                    object of indefinite size
-    @param[in] tag_handler how CBOR tags should be treated
-    @return whether object creation completed
-    */
-    bool get_cbor_object(const std::size_t len,
-                         const cbor_tag_handler_t tag_handler)
-    {
-        if (JSON_HEDLEY_UNLIKELY(!sax->start_object(len)))
-        {
-            return false;
-        }
-
-        if (len != 0)
-        {
-            string_t key;
-            if (len != detail::unknown_size())
-            {
-                for (std::size_t i = 0; i < len; ++i)
                 {
-                    get();
+                    const bool is_object = top.is_object;
+                    container_stack.pop_back();
+                    if (JSON_HEDLEY_UNLIKELY(is_object ? !sax->end_object() : !sax->end_array()))
+                    {
+                        return false;
+                    }
+                    // the value begun here is complete once its container is
+                    if (container_stack.empty())
+                    {
+                        return true;
+                    }
+                    continue;
+                }
+
+                if (top.is_object)
+                {
+                    key.clear();
                     if (JSON_HEDLEY_UNLIKELY(!get_cbor_string(key) || !sax->key(key)))
                     {
                         return false;
                     }
-
+                    fetch = true;
-                    if (JSON_HEDLEY_UNLIKELY(!parse_cbor_internal(true, tag_handler)))
-                    {
-                        return false;
-                    }
-                    key.clear();
                 }
             }
-            else
-            {
-                while (get() != 0xFF)
-                {
-                    if (JSON_HEDLEY_UNLIKELY(!get_cbor_string(key) || !sax->key(key)))
-                    {
-                        return false;
-                    }
 
-                    if (JSON_HEDLEY_UNLIKELY(!parse_cbor_internal(true, tag_handler)))
+            // a tag is not a value of its own: read on until the tagged value
-                    {
+            bool tag_pending = false;
-                        return false;
+            do
-                    }
+            {
-                    key.clear();
+                if (JSON_HEDLEY_UNLIKELY(!parse_cbor_value(fetch, tag_handler, tag_pending)))
+                {
+                    return false;
                 }
+                fetch = true;
+            }
+            while (tag_pending);
+
+            // a value that opened a container left it on the stack; one that
+            // did not, and that was not inside a container, was the whole value
+            if (container_stack.empty())
+            {
+                return true;
             }
         }
-
-        return sax->end_object();
     }
 
     /////////////
@@ -12109,7 +12202,17 @@ class binary_reader
     /*!
     @return whether a valid MessagePack value was passed to the SAX parser
     */
-    bool parse_msgpack_internal()
+    /*!
+    @brief read one MessagePack value
+
+    Reads a single value and passes it to the SAX parser. A value that begins
+    a container is not read to its end: the container is opened with
+    @ref enter_container and its elements are read by
+    @ref parse_msgpack_internal, so that nesting does not consume native stack.
+
+    @return whether reading the value succeeded
+    */
+    bool parse_msgpack_value()
     {
         switch (get())
         {
@@ -12265,7 +12368,7 @@ class binary_reader
             case 0x8D:
             case 0x8E:
             case 0x8F:
-                return get_msgpack_object(conditional_static_cast<std::size_t>(static_cast<unsigned int>(current) & 0x0Fu));
+                return enter_object(conditional_static_cast<std::size_t>(static_cast<unsigned int>(current) & 0x0Fu));
 
             // fixarray
             case 0x90:
@@ -12284,7 +12387,7 @@ class binary_reader
             case 0x9D:
             case 0x9E:
             case 0x9F:
-                return get_msgpack_array(conditional_static_cast<std::size_t>(static_cast<unsigned int>(current) & 0x0Fu));
+                return enter_array(conditional_static_cast<std::size_t>(static_cast<unsigned int>(current) & 0x0Fu));
 
             // fixstr
             case 0xA0:
@@ -12415,25 +12518,25 @@ class binary_reader
             case 0xDC: // array 16
             {
                 std::uint16_t len{};
-                return get_number(input_format_t::msgpack, len) && get_msgpack_array(static_cast<std::size_t>(len));
+                return get_number(input_format_t::msgpack, len) && enter_array(static_cast<std::size_t>(len));
             }
 
             case 0xDD: // array 32
             {
                 std::uint32_t len{};
-                return get_number(input_format_t::msgpack, len) && get_msgpack_array(conditional_static_cast<std::size_t>(len));
+                return get_number(input_format_t::msgpack, len) && enter_array(conditional_static_cast<std::size_t>(len));
             }
 
             case 0xDE: // map 16
             {
                 std::uint16_t len{};
-                return get_number(input_format_t::msgpack, len) && get_msgpack_object(static_cast<std::size_t>(len));
+                return get_number(input_format_t::msgpack, len) && enter_object(static_cast<std::size_t>(len));
             }
 
             case 0xDF: // map 32
             {
                 std::uint32_t len{};
-                return get_number(input_format_t::msgpack, len) && get_msgpack_object(conditional_static_cast<std::size_t>(len));
+                return get_number(input_format_t::msgpack, len) && enter_object(conditional_static_cast<std::size_t>(len));
             }
 
             // negative fixint
@@ -12681,55 +12784,69 @@ class binary_reader
     }
 
     /*!
-    @param[in] len  the length of the array
+    @brief read a MessagePack value and everything nested inside it
-    @return whether array creation completed
+
+    Reads values until the one that was begun here is complete, resuming the
+    enclosing container each time an element ends, so that the nesting depth
+    of the input costs heap rather than native stack (see #5104).
+
+    @return whether reading the value succeeded
     */
-    bool get_msgpack_array(const std::size_t len)
+    bool parse_msgpack_internal()
     {
-        if (JSON_HEDLEY_UNLIKELY(!sax->start_array(len)))
+        // the key currently being read; hoisted out of the loop so that its
-        {
+        // capacity is reused across elements and across nesting levels
-            return false;
-        }
-
-        for (std::size_t i = 0; i < len; ++i)
-        {
-            if (JSON_HEDLEY_UNLIKELY(!parse_msgpack_internal()))
-            {
-                return false;
-            }
-        }
-
-        return sax->end_array();
-    }
-
-    /*!
-    @param[in] len  the length of the object
-    @return whether object creation completed
-    */
-    bool get_msgpack_object(const std::size_t len)
-    {
-        if (JSON_HEDLEY_UNLIKELY(!sax->start_object(len)))
-        {
-            return false;
-        }
-
         string_t key;
-        for (std::size_t i = 0; i < len; ++i)
+
+        while (true)
         {
-            get();
+            if (!container_stack.empty())
-            if (JSON_HEDLEY_UNLIKELY(!get_msgpack_string(key) || !sax->key(key)))
+            {
+                // copied out before anything can push onto the stack and
+                // invalidate a reference into it
+                const bool is_object = container_stack.back().is_object;
+
+                if (container_stack.back().remaining == 0)
+                {
+                    container_stack.pop_back();
+                    if (JSON_HEDLEY_UNLIKELY(is_object ? !sax->end_object() : !sax->end_array()))
+                    {
+                        return false;
+                    }
+                    // the value begun here is complete once its container is
+                    if (container_stack.empty())
+                    {
+                        return true;
+                    }
+                    continue;
+                }
+
+                // claim the element about to be read
+                --container_stack.back().remaining;
+
+                if (is_object)
+                {
+                    get();
+                    key.clear();
+                    if (JSON_HEDLEY_UNLIKELY(!get_msgpack_string(key) || !sax->key(key)))
+                    {
+                        return false;
+                    }
+                }
+            }
+
+            if (JSON_HEDLEY_UNLIKELY(!parse_msgpack_value()))
             {
                 return false;
             }
 
-            if (JSON_HEDLEY_UNLIKELY(!parse_msgpack_internal()))
+            // a value that opened a container left it on the stack; one that
+            // did not, and that was not inside a container, was the whole value
+            if (container_stack.empty())
             {
-                return false;
+                return true;
             }
-            key.clear();
         }
-
-        return sax->end_object();
     }
 
     ////////////
@@ -12745,7 +12862,103 @@ class binary_reader
     */
     bool parse_ubjson_internal(const bool get_char = true)
     {
-        return get_ubjson_value(get_char ? get_ignore_noop() : current);
+        // the key currently being read; hoisted out of the loop so that its
+        // capacity is reused across elements and across nesting levels
+        string_t key;
+
+        // the type marker of the value to read next
+        char_int_type prefix = get_char ? get_ignore_noop() : current;
+
+        while (true)
+        {
+            const std::size_t depth = container_stack.size();
+
+            if (JSON_HEDLEY_UNLIKELY(!get_ubjson_value(prefix)))
+            {
+                return false;
+            }
+
+            // the value begun here is complete once it is not inside anything
+            if (container_stack.empty())
+            {
+                return true;
+            }
+
+            // a value was completed rather than a container opened; a
+            // container that ends at a marker needs the next byte to test
+            if (container_stack.size() == depth && container_stack.back().remaining == npos)
+            {
+                get_ignore_noop();
+            }
+
+            // advance to the next element, closing the containers that ended.
+            // The reference is not held across get_ubjson_value() above, which
+            // can push onto the stack and reallocate it.
+            for (;;)
+            {
+                container_frame& top = container_stack.back();
+
+                if (top.remaining != npos)
+                {
+                    if (top.remaining != 0)
+                    {
+                        --top.remaining;
+                        if (top.is_object)
+                        {
+                            key.clear();
+                            if (JSON_HEDLEY_UNLIKELY(!get_ubjson_string(key) || !sax->key(key)))
+                            {
+                                return false;
+                            }
+                        }
+                        // an optimized container gives its elements no marker
+                        prefix = (top.type_marker != 0) ? top.type_marker : get_ignore_noop();
+                        break;
+                    }
+                }
+                // the end marker is compared against a literal rather than
+                // against a conditional expression, because char_int_type is
+                // unsigned for some input adapters and MSVC then reports the
+                // comparison as a signed/unsigned mismatch
+                else if (top.is_object ? (current != '}') : (current != ']'))
+                {
+                    // a container that ends at a marker is never optimized, so
+                    // every element carries its own marker; for an object the
+                    // byte tested above is the first byte of the key
+                    if (top.is_object)
+                    {
+                        key.clear();
+                        if (JSON_HEDLEY_UNLIKELY(!get_ubjson_string(key, false) || !sax->key(key)))
+                        {
+                            return false;
+                        }
+                        prefix = get_ignore_noop();
+                    }
+                    else
+                    {
+                        prefix = current;
+                    }
+                    break;
+                }
+
+                const bool is_object = top.is_object;
+                container_stack.pop_back();
+                if (JSON_HEDLEY_UNLIKELY(is_object ? !sax->end_object() : !sax->end_array()))
+                {
+                    return false;
+                }
+                if (container_stack.empty())
+                {
+                    return true;
+                }
+                // the container that just ended was an element of the one
+                // below it, which may need the next byte for its own test
+                if (container_stack.back().remaining == npos)
+                {
+                    get_ignore_noop();
+                }
+            }
+        }
     }
 
     /*!
@@ -13517,53 +13730,22 @@ class binary_reader
                                         exception_message(input_format, "excessive array size", "size"), nullptr));
             }
 
-            if (JSON_HEDLEY_UNLIKELY(!sax->start_array(size_and_type.first)))
+            if (JSON_HEDLEY_UNLIKELY(!enter_array(size_and_type.first, size_and_type.second)))
             {
                 return false;
             }
 
-            if (size_and_type.second != 0)
+            if (size_and_type.second == 'N')
             {
-                if (size_and_type.second != 'N')
+                // a no-op is not a value, so a container of them holds none;
-                {
+                // the declared size has already been passed to the SAX parser
-                    for (std::size_t i = 0; i < size_and_type.first; ++i)
+                container_stack.back().remaining = 0;
-                    {
-                        if (JSON_HEDLEY_UNLIKELY(!get_ubjson_value(size_and_type.second)))
-                        {
-                            return false;
-                        }
-                    }
-                }
-            }
-            else
-            {
-                for (std::size_t i = 0; i < size_and_type.first; ++i)
-                {
-                    if (JSON_HEDLEY_UNLIKELY(!parse_ubjson_internal()))
-                    {
-                        return false;
-                    }
-                }
-            }
-        }
-        else
-        {
-            if (JSON_HEDLEY_UNLIKELY(!sax->start_array(detail::unknown_size())))
-            {
-                return false;
             }
 
-            while (current != ']')
+            return true;
-            {
-                if (JSON_HEDLEY_UNLIKELY(!parse_ubjson_internal(false)))
-                {
-                    return false;
-                }
-                get_ignore_noop();
-            }
         }
 
-        return sax->end_array();
+        return enter_array(detail::unknown_size());
     }
 
     /*!
@@ -13585,68 +13767,12 @@ class binary_reader
                                     exception_message(input_format, "BJData object does not support ND-array size in optimized format", "object"), nullptr));
         }
 
-        string_t key;
         if (size_and_type.first != npos)
         {
-            if (JSON_HEDLEY_UNLIKELY(!sax->start_object(size_and_type.first)))
+            return enter_object(size_and_type.first, size_and_type.second);
-            {
-                return false;
-            }
-
-            if (size_and_type.second != 0)
-            {
-                for (std::size_t i = 0; i < size_and_type.first; ++i)
-                {
-                    if (JSON_HEDLEY_UNLIKELY(!get_ubjson_string(key) || !sax->key(key)))
-                    {
-                        return false;
-                    }
-                    if (JSON_HEDLEY_UNLIKELY(!get_ubjson_value(size_and_type.second)))
-                    {
-                        return false;
-                    }
-                    key.clear();
-                }
-            }
-            else
-            {
-                for (std::size_t i = 0; i < size_and_type.first; ++i)
-                {
-                    if (JSON_HEDLEY_UNLIKELY(!get_ubjson_string(key) || !sax->key(key)))
-                    {
-                        return false;
-                    }
-                    if (JSON_HEDLEY_UNLIKELY(!parse_ubjson_internal()))
-                    {
-                        return false;
-                    }
-                    key.clear();
-                }
-            }
-        }
-        else
-        {
-            if (JSON_HEDLEY_UNLIKELY(!sax->start_object(detail::unknown_size())))
-            {
-                return false;
-            }
-
-            while (current != '}')
-            {
-                if (JSON_HEDLEY_UNLIKELY(!get_ubjson_string(key, false) || !sax->key(key)))
-                {
-                    return false;
-                }
-                if (JSON_HEDLEY_UNLIKELY(!parse_ubjson_internal()))
-                {
-                    return false;
-                }
-                get_ignore_noop();
-                key.clear();
-            }
         }
 
-        return sax->end_object();
+        return enter_object(detail::unknown_size());
     }
 
     // Note, no reader for UBJSON binary types is implemented because they do
@@ -14034,6 +14160,9 @@ class binary_reader
     /// the SAX parser
     json_sax_t* sax = nullptr;
 
+    /// the containers that have been opened and not closed yet; see @ref container_frame
+    std::vector<container_frame> container_stack{};
+
     // excluded markers in bjdata optimized type
 #define JSON_BINARY_READER_MAKE_BJD_OPTIMIZED_TYPE_MARKERS_ \
     make_array<char_int_type>('F', 'H', 'N', 'S', 'T', 'Z', '[', '{')

tests/src/unit-cbor.cpp

@@ -2035,6 +2035,93 @@ TEST_CASE("CBOR definite length equal to the indefinite-length sentinel")
     }
 }
 
+TEST_CASE("CBOR nesting does not consume the call stack")
+{
+    // Containers used to be read by calling back into the value reader once
+    // per element, and a tag by calling it for the tagged value, so the native
+    // call stack grew with the nesting depth of the input. Each of the three
+    // costs a single byte to encode -- 0x9F, 0x81 and 0xC2 -- so a payload of
+    // repeated bytes crashed the process (#5104). The containers are kept on a
+    // heap stack now, and a tag is read in a loop.
+    //
+    // Deeply nested values must not be compared, copied or dumped here: those
+    // operations are still recursive and would reintroduce the crash.
+    json _;
+
+    SECTION("indefinite-length containers")
+    {
+        const std::vector<uint8_t> input(500000, 0x9F);
+        CHECK_THROWS_WITH_AS(_ = json::from_cbor(input), "[json.exception.parse_error.110] parse error at byte 500001: syntax error while parsing CBOR value: unexpected end of input", json::parse_error&);
+        CHECK(json::from_cbor(input, true, false).is_discarded());
+    }
+
+    SECTION("definite-length containers")
+    {
+        const std::vector<uint8_t> input(500000, 0x81);
+        CHECK_THROWS_WITH_AS(_ = json::from_cbor(input), "[json.exception.parse_error.110] parse error at byte 500001: syntax error while parsing CBOR value: unexpected end of input", json::parse_error&);
+        CHECK(json::from_cbor(input, true, false).is_discarded());
+    }
+
+    SECTION("tags")
+    {
+        // a tag is not a value of its own, so a chain of them used to recurse
+        const std::vector<uint8_t> input(500000, 0xC2);
+        CHECK_THROWS_WITH_AS(_ = json::from_cbor(input, true, true, json::cbor_tag_handler_t::ignore), "[json.exception.parse_error.110] parse error at byte 500001: syntax error while parsing CBOR value: unexpected end of input", json::parse_error&);
+        CHECK(json::from_cbor(input, true, false, json::cbor_tag_handler_t::ignore).is_discarded());
+    }
+
+    SECTION("a well-formed deep value is read through the SAX interface")
+    {
+        std::vector<uint8_t> input(200000, 0x9F);
+        input.insert(input.end(), 200000, 0xFF);
+
+        SaxCountdown accept_all(1000000);
+        CHECK(json::sax_parse(input, &accept_all, json::input_format_t::cbor));
+    }
+
+    SECTION("a well-formed deep value is read into a value")
+    {
+        const std::size_t depth = 10000;
+        std::vector<uint8_t> input(depth, 0x81);
+        input.push_back(0x00);
+
+        json j = json::from_cbor(input);
+
+        std::size_t measured = 0;
+        const json* p = &j;
+        while (p->is_array() && !p->empty())
+        {
+            p = &p->front();
+            ++measured;
+        }
+        CHECK(measured == depth);
+        CHECK(p->is_number());
+    }
+
+    SECTION("containers are still read the same way")
+    {
+        CHECK(json::from_cbor(std::vector<uint8_t>({0x80})) == json::array());
+        CHECK(json::from_cbor(std::vector<uint8_t>({0xA0})) == json::object());
+        CHECK(json::from_cbor(std::vector<uint8_t>({0x9F, 0xFF})) == json::array());
+        CHECK(json::from_cbor(std::vector<uint8_t>({0xBF, 0xFF})) == json::object());
+        CHECK(json::from_cbor(std::vector<uint8_t>({0x9F, 0x01, 0x02, 0xFF})) == json({1, 2}));
+        CHECK(json::from_cbor(std::vector<uint8_t>({0xBF, 0x61, 'a', 0x01, 0xFF})) == json({{"a", 1}}));
+        // definite and indefinite forms nested inside each other
+        CHECK(json::from_cbor(std::vector<uint8_t>({0x9F, 0x82, 0x01, 0x02, 0xA1, 0x61, 'k', 0xBF, 0xFF, 0xFF})) == json({{1, 2}, {{"k", json::object()}}}));
+    }
+
+    SECTION("tagged values are still read the same way")
+    {
+        const auto ignore = json::cbor_tag_handler_t::ignore;
+        CHECK(json::from_cbor(std::vector<uint8_t>({0xC2, 0x01}), true, true, ignore) == json(1));
+        // a chain of tags resolves to the value that follows it
+        CHECK(json::from_cbor(std::vector<uint8_t>({0xC2, 0xC2, 0xC2, 0x01}), true, true, ignore) == json(1));
+        // a tag inside a container, and one in front of a container
+        CHECK(json::from_cbor(std::vector<uint8_t>({0x82, 0xC2, 0x01, 0x02}), true, true, ignore) == json({1, 2}));
+        CHECK(json::from_cbor(std::vector<uint8_t>({0xC2, 0x82, 0x01, 0x02}), true, true, ignore) == json({1, 2}));
+    }
+}
+
 TEST_CASE("CBOR indefinite-length strings do not recurse per chunk")
 {
     // Reading an indefinite-length string or byte array used to call itself

tests/src/unit-msgpack.cpp

@@ -1598,6 +1598,67 @@ TEST_CASE("MessagePack")
 }
 
 // use this testcase outside [hide] to run it with Valgrind
+TEST_CASE("MessagePack nesting does not consume the call stack")
+{
+    // Reading a container used to call back into the value reader once per
+    // element, so the native call stack grew with the nesting depth of the
+    // input: one frame per byte for repeated 0x91 (a one-element array), which
+    // crashes the process long before the input is exhausted (#5104). The
+    // containers are kept on a heap stack now.
+    //
+    // Note that deeply nested values must not be compared, copied or dumped
+    // here: those operations are still recursive, and would reintroduce the
+    // very crash this checks for. Depth is measured by descending instead.
+
+    SECTION("an unterminated chain is reported, not crashed on")
+    {
+        json _;
+        const std::vector<uint8_t> input(300000, 0x91);
+        CHECK_THROWS_WITH_AS(_ = json::from_msgpack(input), "[json.exception.parse_error.110] parse error at byte 300001: syntax error while parsing MessagePack value: unexpected end of input", json::parse_error&);
+        CHECK(json::from_msgpack(input, true, false).is_discarded());
+    }
+
+    SECTION("a well-formed deep value is read through the SAX interface")
+    {
+        std::vector<uint8_t> input(300000, 0x91);
+        input.push_back(0x01); // innermost value
+
+        SaxCountdown accept_all(600001);
+        CHECK(json::sax_parse(input, &accept_all, json::input_format_t::msgpack));
+    }
+
+    SECTION("a well-formed deep value is read into a value")
+    {
+        const std::size_t depth = 10000;
+        std::vector<uint8_t> input(depth, 0x91);
+        input.push_back(0x01);
+
+        json j = json::from_msgpack(input);
+
+        std::size_t measured = 0;
+        const json* p = &j;
+        while (p->is_array() && !p->empty())
+        {
+            p = &p->front();
+            ++measured;
+        }
+        CHECK(measured == depth);
+        CHECK(p->is_number());
+    }
+
+    SECTION("containers are still read the same way")
+    {
+        CHECK(json::from_msgpack(std::vector<uint8_t>({0x90})) == json::array());
+        CHECK(json::from_msgpack(std::vector<uint8_t>({0x80})) == json::object());
+        CHECK(json::from_msgpack(std::vector<uint8_t>({0x92, 0x90, 0x80})) == json({json::array(), json::object()}));
+        CHECK(json::from_msgpack(std::vector<uint8_t>({0x91, 0x91, 0x91, 0x90})) == json({{{json::array()}}}));
+        CHECK(json::from_msgpack(std::vector<uint8_t>({0x81, 0xA1, 'a', 0x81, 0xA1, 'b', 0x92, 0x01, 0x02})) == json({{"a", {{"b", {1, 2}}}}}));
+        // array 16 and map 32, i.e. the counted forms
+        CHECK(json::from_msgpack(std::vector<uint8_t>({0xDC, 0x00, 0x02, 0x01, 0x02})) == json({1, 2}));
+        CHECK(json::from_msgpack(std::vector<uint8_t>({0xDF, 0x00, 0x00, 0x00, 0x01, 0xA1, 'k', 0xC3})) == json({{"k", true}}));
+    }
+}
+
 TEST_CASE("single MessagePack roundtrip")
 {
     SECTION("sample.json")

tests/src/unit-ubjson.cpp

@@ -2149,6 +2149,111 @@ TEST_CASE("UBJSON")
     }
 }
 
+TEST_CASE("UBJSON nesting does not consume the call stack")
+{
+    // Containers used to be read by calling back into the value reader once
+    // per element, so the native call stack grew with the nesting depth of the
+    // input. '[' alone opens a container, so a payload of repeated '[' crashed
+    // the process (#5104), as did the optimized forms, which reach the same
+    // path through a type or size annotation. The containers are kept on a
+    // heap stack now.
+    //
+    // Deeply nested values must not be compared, copied or dumped here: those
+    // operations are still recursive and would reintroduce the crash.
+    json _;
+
+    SECTION("containers that end at a marker")
+    {
+        const std::vector<uint8_t> input(500000, '[');
+        CHECK_THROWS_WITH_AS(_ = json::from_ubjson(input), "[json.exception.parse_error.110] parse error at byte 500001: syntax error while parsing UBJSON value: unexpected end of input", json::parse_error&);
+        CHECK(json::from_ubjson(input, true, false).is_discarded());
+    }
+
+    SECTION("containers with a size")
+    {
+        std::vector<uint8_t> input;
+        for (std::size_t i = 0; i < 100000; ++i)
+        {
+            input.push_back('[');
+            input.push_back('#');
+            input.push_back('i');
+            input.push_back(1);
+        }
+        CHECK_THROWS_AS(_ = json::from_ubjson(input), json::parse_error&);
+        CHECK(json::from_ubjson(input, true, false).is_discarded());
+    }
+
+    SECTION("containers with a type and a size")
+    {
+        // '[' is a permitted optimized type in UBJSON, so each element of such
+        // a container is itself a container, read without a marker of its own
+        std::vector<uint8_t> input;
+        for (std::size_t i = 0; i < 100000; ++i)
+        {
+            const std::vector<uint8_t> level = {'[', '$', '[', '#', 'i', 1};
+            input.insert(input.end(), level.begin(), level.end());
+        }
+        CHECK_THROWS_AS(_ = json::from_ubjson(input), json::parse_error&);
+        CHECK(json::from_ubjson(input, true, false).is_discarded());
+    }
+
+    SECTION("a well-formed deep value is read through the SAX interface")
+    {
+        std::vector<uint8_t> input(100000, '[');
+        input.insert(input.end(), 100000, ']');
+
+        SaxCountdown accept_all(1000000);
+        CHECK(json::sax_parse(input, &accept_all, json::input_format_t::ubjson));
+    }
+
+    SECTION("a well-formed deep value is read into a value")
+    {
+        const std::size_t depth = 10000;
+        std::vector<uint8_t> input(depth, '[');
+        input.insert(input.end(), depth, ']');
+
+        json j = json::from_ubjson(input);
+
+        std::size_t measured = 0;
+        const json* p = &j;
+        while (p->is_array() && !p->empty())
+        {
+            p = &p->front();
+            ++measured;
+        }
+        // the innermost array is empty, so the descent stops one level short
+        CHECK(measured == depth - 1);
+    }
+
+    SECTION("containers are still read the same way")
+    {
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'[', ']'})) == json::array());
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'{', '}'})) == json::object());
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'[', '#', 'i', 0})) == json::array());
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'{', '#', 'i', 0})) == json::object());
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'[', '$', 'i', '#', 'i', 2, 1, 2})) == json({1, 2}));
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'[', '#', 'i', 2, 'i', 1, 'i', 2})) == json({1, 2}));
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'{', '$', 'i', '#', 'i', 1, 'i', 1, 'a', 1})) == json({{"a", 1}}));
+        // a no-op is not a value, so a container of them holds none
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'[', '$', 'N', '#', 'i', 2})) == json::array());
+        // sized and unsized forms nested inside one another
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'[', '[', '#', 'i', 2, 'i', 1, 'i', 2, ']'})) == json({{1, 2}}));
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'[', '#', 'i', 1, '[', 'i', 1, ']'})) == json({{1}}));
+        // an optimized container of containers
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'[', '$', '[', '#', 'i', 2, 'i', 1, ']', 'i', 2, ']'})) == json({{1}, {2}}));
+    }
+
+    SECTION("BJData containers are still read the same way")
+    {
+        // the ND-array wrapper and the binary shortcut are complete values,
+        // not containers the reader descends into
+        CHECK(json::from_bjdata(std::vector<uint8_t>({'[', '$', 'U', '#', '[', '$', 'i', '#', 'i', 2, 2, 3, 1, 2, 3, 4, 5, 6})) ==
+        json({{"_ArrayType_", "uint8"}, {"_ArraySize_", {2, 3}}, {"_ArrayData_", {1, 2, 3, 4, 5, 6}}}));
+        CHECK(json::from_bjdata(std::vector<uint8_t>({'[', '$', 'i', '#', 'i', 2, 1, 2})) == json({1, 2}));
+        CHECK(json::from_bjdata(std::vector<uint8_t>({'[', '[', 'i', 1, ']', ']'})) == json({{1}}));
+    }
+}
+
 TEST_CASE("UBJSON optimized arrays of a valueless type are bounded")
 {
     // An element of type 'Z', 'T' or 'F' is encoded by its marker alone, so an