12 All declarations are in :file:`jansson.h`, so it's enough to
20 All constants are prefixed with ``JSON_`` (except for those describing
21 the library version, prefixed with ``JANSSON_``). Other identifiers
22 are prefixed with ``json_``. Type names are suffixed with ``_t`` and
23 ``typedef``\ 'd so that the ``struct`` keyword need not be used.
29 The Jansson version is of the form *A.B.C*, where *A* is the major
30 version, *B* is the minor version and *C* is the micro version. If the
31 micro version is zero, it's omitted from the version string, i.e. the
32 version string is just *A.B*.
34 When a new release only fixes bugs and doesn't add new features or
35 functionality, the micro version is incremented. When new features are
36 added in a backwards compatible way, the minor version is incremented
37 and the micro version is set to zero. When there are backwards
38 incompatible changes, the major version is incremented and others are
41 The following preprocessor constants specify the current version of
44 ``JANSSON_VERSION_MAJOR``, ``JANSSON_VERSION_MINOR``, ``JANSSON_VERSION_MICRO``
45 Integers specifying the major, minor and micro versions,
49 A string representation of the current version, e.g. ``"1.2.1"`` or
52 ``JANSSON_VERSION_HEX``
53 A 3-byte hexadecimal representation of the version, e.g.
54 ``0x010201`` for version 1.2.1 and ``0x010300`` for version 1.3.
55 This is useful in numeric comparisions, e.g.::
57 #if JANSSON_VERSION_HEX >= 0x010300
58 /* Code specific to version 1.3 and above */
65 The JSON specification (:rfc:`4627`) defines the following data types:
66 *object*, *array*, *string*, *number*, *boolean*, and *null*. JSON
67 types are used dynamically; arrays and objects can hold any other data
68 type, including themselves. For this reason, Jansson's type system is
69 also dynamic in nature. There's one C type to represent all JSON
70 values, and this structure knows the type of the JSON value it holds.
74 This data structure is used throughout the library to represent all
75 JSON values. It always contains the type of the JSON value it holds
76 and the value's reference count. The rest depends on the type of the
79 Objects of :type:`json_t` are always used through a pointer. There
80 are APIs for querying the type, manipulating the reference count, and
81 for constructing and manipulating values of different types.
83 Unless noted otherwise, all API functions return an error value if an
84 error occurs. Depending on the function's signature, the error value
85 is either *NULL* or -1. Invalid arguments or invalid input are
86 apparent sources for errors. Memory allocation and I/O operations may
93 The type of a JSON value is queried and tested using the following
96 .. type:: enum json_type
98 The type of a JSON value. The following members are defined:
100 +--------------------+
102 +--------------------+
104 +--------------------+
106 +--------------------+
108 +--------------------+
110 +--------------------+
112 +--------------------+
114 +--------------------+
116 +--------------------+
118 These correspond to JSON object, array, string, number, boolean and
119 null. A number is represented by either a value of the type
120 ``JSON_INTEGER`` or of the type ``JSON_REAL``. A true boolean value
121 is represented by a value of the type ``JSON_TRUE`` and false by a
122 value of the type ``JSON_FALSE``.
124 .. function:: int json_typeof(const json_t *json)
126 Return the type of the JSON value (a :type:`json_type` cast to
127 :type:`int`). *json* MUST NOT be *NULL*. This function is actually
128 implemented as a macro for speed.
130 .. function:: json_is_object(const json_t *json)
131 json_is_array(const json_t *json)
132 json_is_string(const json_t *json)
133 json_is_integer(const json_t *json)
134 json_is_real(const json_t *json)
135 json_is_true(const json_t *json)
136 json_is_false(const json_t *json)
137 json_is_null(const json_t *json)
139 These functions (actually macros) return true (non-zero) for values
140 of the given type, and false (zero) for values of other types and
143 .. function:: json_is_number(const json_t *json)
145 Returns true for values of types ``JSON_INTEGER`` and
146 ``JSON_REAL``, and false for other types and for *NULL*.
148 .. function:: json_is_boolean(const json_t *json)
150 Returns true for types ``JSON_TRUE`` and ``JSON_FALSE``, and false
151 for values of other types and for *NULL*.
154 .. _apiref-reference-count:
159 The reference count is used to track whether a value is still in use
160 or not. When a value is created, it's reference count is set to 1. If
161 a reference to a value is kept (e.g. a value is stored somewhere for
162 later use), its reference count is incremented, and when the value is
163 no longer needed, the reference count is decremented. When the
164 reference count drops to zero, there are no references left, and the
165 value can be destroyed.
167 The following functions are used to manipulate the reference count.
169 .. function:: json_t *json_incref(json_t *json)
171 Increment the reference count of *json* if it's not non-*NULL*.
174 .. function:: void json_decref(json_t *json)
176 Decrement the reference count of *json*. As soon as a call to
177 :func:`json_decref()` drops the reference count to zero, the value
178 is destroyed and it can no longer be used.
180 Functions creating new JSON values set the reference count to 1. These
181 functions are said to return a **new reference**. Other functions
182 returning (existing) JSON values do not normally increase the
183 reference count. These functions are said to return a **borrowed
184 reference**. So, if the user will hold a reference to a value returned
185 as a borrowed reference, he must call :func:`json_incref`. As soon as
186 the value is no longer needed, :func:`json_decref` should be called
187 to release the reference.
189 Normally, all functions accepting a JSON value as an argument will
190 manage the reference, i.e. increase and decrease the reference count
191 as needed. However, some functions **steal** the reference, i.e. they
192 have the same result as if the user called :func:`json_decref()` on
193 the argument right after calling the function. These functions are
194 suffixed with ``_new`` or have ``_new_`` somewhere in their name.
196 For example, the following code creates a new JSON array and appends
199 json_t *array, *integer;
201 array = json_array();
202 integer = json_integer(42);
204 json_array_append(array, integer);
205 json_decref(integer);
207 Note how the caller has to release the reference to the integer value
208 by calling :func:`json_decref()`. By using a reference stealing
209 function :func:`json_array_append_new()` instead of
210 :func:`json_array_append()`, the code becomes much simpler::
212 json_t *array = json_array();
213 json_array_append_new(array, json_integer(42));
215 In this case, the user doesn't have to explicitly release the
216 reference to the integer value, as :func:`json_array_append_new()`
217 steals the reference when appending the value to the array.
219 In the following sections it is clearly documented whether a function
220 will return a new or borrowed reference or steal a reference to its
227 A circular reference is created when an object or an array is,
228 directly or indirectly, inserted inside itself. The direct case is
231 json_t *obj = json_object();
232 json_object_set(obj, "foo", obj);
234 Jansson will refuse to do this, and :func:`json_object_set()` (and
235 all the other such functions for objects and arrays) will return with
236 an error status. The indirect case is the dangerous one::
238 json_t *arr1 = json_array(), *arr2 = json_array();
239 json_array_append(arr1, arr2);
240 json_array_append(arr2, arr1);
242 In this example, the array ``arr2`` is contained in the array
243 ``arr1``, and vice versa. Jansson cannot check for this kind of
244 indirect circular references without a performance hit, so it's up to
245 the user to avoid them.
247 If a circular reference is created, the memory consumed by the values
248 cannot be freed by :func:`json_decref()`. The reference counts never
249 drops to zero because the values are keeping the references to each
250 other. Moreover, trying to encode the values with any of the encoding
251 functions will fail. The encoder detects circular references and
252 returns an error status.
258 These values are implemented as singletons, so each of these functions
259 returns the same value each time.
261 .. function:: json_t *json_true(void)
265 Returns the JSON true value.
267 .. function:: json_t *json_false(void)
271 Returns the JSON false value.
273 .. function:: json_t *json_null(void)
277 Returns the JSON null value.
283 Jansson uses UTF-8 as the character encoding. All JSON strings must be
284 valid UTF-8 (or ASCII, as it's a subset of UTF-8). Normal null
285 terminated C strings are used, so JSON strings may not contain
286 embedded null characters. All other Unicode codepoints U+0001 through
287 U+10FFFF are allowed.
289 .. function:: json_t *json_string(const char *value)
293 Returns a new JSON string, or *NULL* on error. *value* must be a
294 valid UTF-8 encoded Unicode string.
296 .. function:: json_t *json_string_nocheck(const char *value)
300 Like :func:`json_string`, but doesn't check that *value* is valid
301 UTF-8. Use this function only if you are certain that this really
302 is the case (e.g. you have already checked it by other means).
304 .. function:: const char *json_string_value(const json_t *string)
306 Returns the associated value of *string* as a null terminated UTF-8
307 encoded string, or *NULL* if *string* is not a JSON string.
309 The retuned value is read-only and must not be modified or freed by
310 the user. It is valid as long as *string* exists, i.e. as long as
311 its reference count has not dropped to zero.
313 .. function:: int json_string_set(const json_t *string, const char *value)
315 Sets the associated value of *string* to *value*. *value* must be a
316 valid UTF-8 encoded Unicode string. Returns 0 on success and -1 on
319 .. function:: int json_string_set_nocheck(const json_t *string, const char *value)
321 Like :func:`json_string_set`, but doesn't check that *value* is
322 valid UTF-8. Use this function only if you are certain that this
323 really is the case (e.g. you have already checked it by other
330 The JSON specification only contains one numeric type, "number". The C
331 programming language has distinct types for integer and floating-point
332 numbers, so for practical reasons Jansson also has distinct types for
333 the two. They are called "integer" and "real", respectively. For more
334 information, see :ref:`rfc-conformance`.
338 This is the C type that is used to store JSON integer values. It
339 represents the widest integer type available on your system. In
340 practice it's just a typedef of ``long long`` if your compiler
341 supports it, otherwise ``long``.
343 Usually, you can safely use plain ``int`` in place of
344 ``json_int_t``, and the implicit C integer conversion handles the
345 rest. Only when you know that you need the full 64-bit range, you
346 should use ``json_int_t`` explicitly.
348 ``JSON_INTEGER_IS_LONG_LONG``
350 This is a preprocessor variable that holds the value 1 if
351 :type:`json_int_t` is ``long long``, and 0 if it's ``long``. It
352 can be used as follows::
354 #if JSON_INTEGER_IS_LONG_LONG
355 /* Code specific for long long */
357 /* Code specific for long */
360 ``JSON_INTEGER_FORMAT``
362 This is a macro that expands to a :func:`printf()` conversion
363 specifier that corresponds to :type:`json_int_t`, without the
364 leading ``%`` sign, i.e. either ``"lld"`` or ``"ld"``. This macro
365 is required because the actual type of :type:`json_int_t` can be
366 either ``long`` or ``long long``, and :func:`printf()` reuiqres
367 different length modifiers for the two.
371 json_int_t x = 123123123;
372 printf("x is %" JSON_INTEGER_FORMAT "\n", x);
375 .. function:: json_t *json_integer(json_int_t value)
379 Returns a new JSON integer, or *NULL* on error.
381 .. function:: json_int_t json_integer_value(const json_t *integer)
383 Returns the associated value of *integer*, or 0 if *json* is not a
386 .. function:: int json_integer_set(const json_t *integer, json_int_t value)
388 Sets the associated value of *integer* to *value*. Returns 0 on
389 success and -1 if *integer* is not a JSON integer.
391 .. function:: json_t *json_real(double value)
395 Returns a new JSON real, or *NULL* on error.
397 .. function:: double json_real_value(const json_t *real)
399 Returns the associated value of *real*, or 0.0 if *real* is not a
402 .. function:: int json_real_set(const json_t *real, double value)
404 Sets the associated value of *real* to *value*. Returns 0 on
405 success and -1 if *real* is not a JSON real.
407 In addition to the functions above, there's a common query function
408 for integers and reals:
410 .. function:: double json_number_value(const json_t *json)
412 Returns the associated value of the JSON integer or JSON real
413 *json*, cast to double regardless of the actual type. If *json* is
414 neither JSON real nor JSON integer, 0.0 is returned.
420 A JSON array is an ordered collection of other JSON values.
422 .. function:: json_t *json_array(void)
426 Returns a new JSON array, or *NULL* on error. Initially, the array
429 .. function:: size_t json_array_size(const json_t *array)
431 Returns the number of elements in *array*, or 0 if *array* is NULL
434 .. function:: json_t *json_array_get(const json_t *array, size_t index)
436 .. refcounting:: borrow
438 Returns the element in *array* at position *index*. The valid range
439 for *index* is from 0 to the return value of
440 :func:`json_array_size()` minus 1. If *array* is not a JSON array,
441 if *array* is *NULL*, or if *index* is out of range, *NULL* is
444 .. function:: int json_array_set(json_t *array, size_t index, json_t *value)
446 Replaces the element in *array* at position *index* with *value*.
447 The valid range for *index* is from 0 to the return value of
448 :func:`json_array_size()` minus 1. Returns 0 on success and -1 on
451 .. function:: int json_array_set_new(json_t *array, size_t index, json_t *value)
453 Like :func:`json_array_set()` but steals the reference to *value*.
454 This is useful when *value* is newly created and not used after
457 .. function:: int json_array_append(json_t *array, json_t *value)
459 Appends *value* to the end of *array*, growing the size of *array*
460 by 1. Returns 0 on success and -1 on error.
462 .. function:: int json_array_append_new(json_t *array, json_t *value)
464 Like :func:`json_array_append()` but steals the reference to
465 *value*. This is useful when *value* is newly created and not used
468 .. function:: int json_array_insert(json_t *array, size_t index, json_t *value)
470 Inserts *value* to *array* at position *index*, shifting the
471 elements at *index* and after it one position towards the end of
472 the array. Returns 0 on success and -1 on error.
474 .. function:: int json_array_insert_new(json_t *array, size_t index, json_t *value)
476 Like :func:`json_array_insert()` but steals the reference to
477 *value*. This is useful when *value* is newly created and not used
480 .. function:: int json_array_remove(json_t *array, size_t index)
482 Removes the element in *array* at position *index*, shifting the
483 elements after *index* one position towards the start of the array.
484 Returns 0 on success and -1 on error. The reference count of the
485 removed value is decremented.
487 .. function:: int json_array_clear(json_t *array)
489 Removes all elements from *array*. Returns 0 on sucess and -1 on
490 error. The reference count of all removed values are decremented.
492 .. function:: int json_array_extend(json_t *array, json_t *other_array)
494 Appends all elements in *other_array* to the end of *array*.
495 Returns 0 on success and -1 on error.
501 A JSON object is a dictionary of key-value pairs, where the key is a
502 Unicode string and the value is any JSON value.
504 .. function:: json_t *json_object(void)
508 Returns a new JSON object, or *NULL* on error. Initially, the
511 .. function:: size_t json_object_size(const json_t *object)
513 Returns the number of elements in *object*, or 0 if *object* is not
516 .. function:: json_t *json_object_get(const json_t *object, const char *key)
518 .. refcounting:: borrow
520 Get a value corresponding to *key* from *object*. Returns *NULL* if
521 *key* is not found and on error.
523 .. function:: int json_object_set(json_t *object, const char *key, json_t *value)
525 Set the value of *key* to *value* in *object*. *key* must be a
526 valid null terminated UTF-8 encoded Unicode string. If there
527 already is a value for *key*, it is replaced by the new value.
528 Returns 0 on success and -1 on error.
530 .. function:: int json_object_set_nocheck(json_t *object, const char *key, json_t *value)
532 Like :func:`json_object_set`, but doesn't check that *key* is
533 valid UTF-8. Use this function only if you are certain that this
534 really is the case (e.g. you have already checked it by other
537 .. function:: int json_object_set_new(json_t *object, const char *key, json_t *value)
539 Like :func:`json_object_set()` but steals the reference to
540 *value*. This is useful when *value* is newly created and not used
543 .. function:: int json_object_set_new_nocheck(json_t *object, const char *key, json_t *value)
545 Like :func:`json_object_set_new`, but doesn't check that *key* is
546 valid UTF-8. Use this function only if you are certain that this
547 really is the case (e.g. you have already checked it by other
550 .. function:: int json_object_del(json_t *object, const char *key)
552 Delete *key* from *object* if it exists. Returns 0 on success, or
553 -1 if *key* was not found. The reference count of the removed value
556 .. function:: int json_object_clear(json_t *object)
558 Remove all elements from *object*. Returns 0 on success and -1 if
559 *object* is not a JSON object. The reference count of all removed
560 values are decremented.
562 .. function:: int json_object_update(json_t *object, json_t *other)
564 Update *object* with the key-value pairs from *other*, overwriting
565 existing keys. Returns 0 on success or -1 on error.
568 The following functions implement an iteration protocol for objects,
569 allowing to iterate through all key-value pairs in an object. The
570 items are not returned in any particular order, as this would require
571 sorting due to the internal hashtable implementation.
573 .. function:: void *json_object_iter(json_t *object)
575 Returns an opaque iterator which can be used to iterate over all
576 key-value pairs in *object*, or *NULL* if *object* is empty.
578 .. function:: void *json_object_iter_at(json_t *object, const char *key)
580 Like :func:`json_object_iter()`, but returns an iterator to the
581 key-value pair in *object* whose key is equal to *key*, or NULL if
582 *key* is not found in *object*. Iterating forward to the end of
583 *object* only yields all key-value pairs of the object if *key*
584 happens to be the first key in the underlying hash table.
586 .. function:: void *json_object_iter_next(json_t *object, void *iter)
588 Returns an iterator pointing to the next key-value pair in *object*
589 after *iter*, or *NULL* if the whole object has been iterated
592 .. function:: const char *json_object_iter_key(void *iter)
594 Extract the associated key from *iter*.
596 .. function:: json_t *json_object_iter_value(void *iter)
598 .. refcounting:: borrow
600 Extract the associated value from *iter*.
602 .. function:: int json_object_iter_set(json_t *object, void *iter, json_t *value)
604 Set the value of the key-value pair in *object*, that is pointed to
605 by *iter*, to *value*.
607 .. function:: int json_object_iter_set_new(json_t *object, void *iter, json_t *value)
609 Like :func:`json_object_iter_set()`, but steals the reference to
610 *value*. This is useful when *value* is newly created and not used
613 The iteration protocol can be used for example as follows::
615 /* obj is a JSON object */
618 void *iter = json_object_iter(obj);
621 key = json_object_iter_key(iter);
622 value = json_object_iter_value(iter);
623 /* use key and value ... */
624 iter = json_object_iter_next(obj, iter);
631 Jansson uses a single struct type to pass error information to the
632 user. See sections :ref:`apiref-decoding`, :ref:`apiref-pack` and
633 :ref:`apiref-unpack` for functions that pass error information using
636 .. type:: json_error_t
638 .. member:: char text[]
640 The error message (in UTF-8), or an empty string if a message is
643 .. member:: char source[]
645 Source of the error. This can be (a part of) the file name or a
646 special identifier in angle brackers (e.g. ``<string>``).
650 The line number on which the error occurred.
652 .. member:: int column
654 The column on which the error occurred. Note that this is the
655 *character column*, not the byte column, i.e. a multibyte UTF-8
656 character counts as one column.
658 .. member:: size_t position
660 The position in bytes from the start of the input. This is
661 useful for debugging Unicode encoding problems.
663 The normal use of :type:`json_error_t` is to allocate it on the stack,
664 and pass a pointer to a function. Example::
670 json = json_load_file("/path/to/file.json", 0, &error);
672 /* the error variable contains error information */
677 Also note that if the call succeeded (``json != NULL`` in the above
678 example), the contents of ``error`` are generally left unspecified.
679 The decoding functions write to the ``position`` member also on
680 success. See :ref:`apiref-decoding` for more info.
682 All functions also accept *NULL* as the :type:`json_error_t` pointer,
683 in which case no error information is returned to the caller.
689 This sections describes the functions that can be used to encode
690 values to JSON. By default, only objects and arrays can be encoded
691 directly, since they are the only valid *root* values of a JSON text.
692 To encode any JSON value, use the ``JSON_ENCODE_ANY`` flag (see
695 By default, the output has no newlines, and spaces are used between
696 array and object elements for a readable output. This behavior can be
697 altered by using the ``JSON_INDENT`` and ``JSON_COMPACT`` flags
698 described below. A newline is never appended to the end of the encoded
701 Each function takes a *flags* parameter that controls some aspects of
702 how the data is encoded. Its default value is 0. The following macros
703 can be ORed together to obtain *flags*.
706 Pretty-print the result, using newlines between array and object
707 items, and indenting with *n* spaces. The valid range for *n* is
708 between 0 and 31 (inclusive), other values result in an undefined
709 output. If ``JSON_INDENT`` is not used or *n* is 0, no newlines are
710 inserted between array and object items.
713 This flag enables a compact representation, i.e. sets the separator
714 between array and object items to ``","`` and between object keys
715 and values to ``":"``. Without this flag, the corresponding
716 separators are ``", "`` and ``": "`` for more readable output.
718 ``JSON_ENSURE_ASCII``
719 If this flag is used, the output is guaranteed to consist only of
720 ASCII characters. This is achived by escaping all Unicode
721 characters outside the ASCII range.
724 If this flag is used, all the objects in output are sorted by key.
725 This is useful e.g. if two JSON texts are diffed or visually
728 ``JSON_PRESERVE_ORDER``
729 If this flag is used, object keys in the output are sorted into the
730 same order in which they were first inserted to the object. For
731 example, decoding a JSON text and then encoding with this flag
732 preserves the order of object keys.
735 Specifying this flag makes it possible to encode any JSON value on
736 its own. Without it, only objects and arrays can be passed as the
737 *root* value to the encoding functions.
739 **Note:** Encoding any value may be useful in some scenarios, but
740 it's generally discouraged as it violates strict compatiblity with
741 :rfc:`4627`. If you use this flag, don't expect interoperatibility
742 with other JSON systems.
744 .. versionadded:: 2.1
746 The following functions perform the actual JSON encoding. The result
749 .. function:: char *json_dumps(const json_t *root, size_t flags)
751 Returns the JSON representation of *root* as a string, or *NULL* on
752 error. *flags* is described above. The return value must be freed
753 by the caller using :func:`free()`.
755 .. function:: int json_dumpf(const json_t *root, FILE *output, size_t flags)
757 Write the JSON representation of *root* to the stream *output*.
758 *flags* is described above. Returns 0 on success and -1 on error.
759 If an error occurs, something may have already been written to
760 *output*. In this case, the output is undefined and most likely not
763 .. function:: int json_dump_file(const json_t *json, const char *path, size_t flags)
765 Write the JSON representation of *root* to the file *path*. If
766 *path* already exists, it is overwritten. *flags* is described
767 above. Returns 0 on success and -1 on error.
769 .. type:: json_dump_callback_t
771 A typedef for a function that's called by
772 :func:`json_dump_callback()`::
774 typedef int (*json_dump_callback_t)(const char *buffer, size_t size, void *data);
776 *buffer* points to a buffer containing a chunk of output, *size* is
777 the length of the buffer, and *data* is the corresponding
778 :func:`json_dump_callback()` argument passed through.
780 On error, the function should return -1 to stop the encoding
781 process. On success, it should return 0.
783 .. versionadded:: 2.2
785 .. function:: int json_dump_callback(const json_t *json, json_dump_callback_t callback, void *data, size_t flags)
787 Call *callback* repeatedly, passing a chunk of the JSON
788 representation of *root* each time. *flags* is described above.
789 Returns 0 on success and -1 on error.
791 .. versionadded:: 2.2
799 This sections describes the functions that can be used to decode JSON
800 text to the Jansson representation of JSON data. The JSON
801 specification requires that a JSON text is either a serialized array
802 or object, and this requirement is also enforced with the following
803 functions. In other words, the top level value in the JSON text being
804 decoded must be either array or object. To decode any JSON value, use
805 the ``JSON_DECODE_ANY`` flag (see below).
807 See :ref:`rfc-conformance` for a discussion on Jansson's conformance
808 to the JSON specification. It explains many design decisions that
809 affect especially the behavior of the decoder.
811 Each function takes a *flags* parameter that can be used to control
812 the behavior of the decoder. Its default value is 0. The following
813 macros can be ORed together to obtain *flags*.
815 ``JSON_REJECT_DUPLICATES``
816 Issue a decoding error if any JSON object in the input text
817 contains duplicate keys. Without this flag, the value of the last
818 occurence of each key ends up in the result. Key equivalence is
819 checked byte-by-byte, without special Unicode comparison
822 .. versionadded:: 2.1
825 By default, the decoder expects an array or object as the input.
826 With this flag enabled, the decoder accepts any valid JSON value.
828 **Note:** Decoding any value may be useful in some scenarios, but
829 it's generally discouraged as it violates strict compatiblity with
830 :rfc:`4627`. If you use this flag, don't expect interoperatibility
831 with other JSON systems.
833 .. versionadded:: 2.3
835 ``JSON_DISABLE_EOF_CHECK``
836 By default, the decoder expects that its whole input constitutes a
837 valid JSON text, and issues an error if there's extra data after
838 the otherwise valid JSON input. With this flag enabled, the decoder
839 stops after decoding a valid JSON array or object, and thus allows
840 extra data after the JSON text.
842 Normally, reading will stop when the last ``]`` or ``}`` in the
843 JSON input is encountered. If both ``JSON_DISABLE_EOF_CHECK`` and
844 ``JSON_DECODE_ANY`` flags are used, the decoder may read one extra
845 UTF-8 code unit (up to 4 bytes of input). For example, decoding
846 ``4true`` correctly decodes the integer 4, but also reads the
847 ``t``. For this reason, if reading multiple consecutive values that
848 are not arrays or objects, they should be separated by at least one
849 whitespace character.
851 .. versionadded:: 2.1
853 Each function also takes an optional :type:`json_error_t` parameter
854 that is filled with error information if decoding fails. It's also
855 updated on success; the number of bytes of input read is written to
856 its ``position`` field. This is especially useful when using
857 ``JSON_DISABLE_EOF_CHECK`` to read multiple consecutive JSON texts.
859 .. versionadded:: 2.3
860 Number of bytes of input read is written to the ``position`` field
861 of the :type:`json_error_t` structure.
863 If no error or position information is needed, you can pass *NULL*.
865 The following functions perform the actual JSON decoding.
867 .. function:: json_t *json_loads(const char *input, size_t flags, json_error_t *error)
871 Decodes the JSON string *input* and returns the array or object it
872 contains, or *NULL* on error, in which case *error* is filled with
873 information about the error. *flags* is described above.
875 .. function:: json_t *json_loadb(const char *buffer, size_t buflen, size_t flags, json_error_t *error)
879 Decodes the JSON string *buffer*, whose length is *buflen*, and
880 returns the array or object it contains, or *NULL* on error, in
881 which case *error* is filled with information about the error. This
882 is similar to :func:`json_loads()` except that the string doesn't
883 need to be null-terminated. *flags* is described above.
885 .. versionadded:: 2.1
887 .. function:: json_t *json_loadf(FILE *input, size_t flags, json_error_t *error)
891 Decodes the JSON text in stream *input* and returns the array or
892 object it contains, or *NULL* on error, in which case *error* is
893 filled with information about the error. *flags* is described
896 This function will start reading the input from whatever position
897 the input file was, without attempting to seek first. If an error
898 occurs, the file position will be left indeterminate. On success,
899 the file position will be at EOF, unless ``JSON_DISABLE_EOF_CHECK``
900 flag was used. In this case, the file position will be at the first
901 character after the last ``]`` or ``}`` in the JSON input. This
902 allows calling :func:`json_loadf()` on the same ``FILE`` object
903 multiple times, if the input consists of consecutive JSON texts,
904 possibly separated by whitespace.
906 .. function:: json_t *json_load_file(const char *path, size_t flags, json_error_t *error)
910 Decodes the JSON text in file *path* and returns the array or
911 object it contains, or *NULL* on error, in which case *error* is
912 filled with information about the error. *flags* is described
921 This section describes functions that help to create, or *pack*,
922 complex JSON values, especially nested objects and arrays. Value
923 building is based on a *format string* that is used to tell the
924 functions about the expected arguments.
926 For example, the format string ``"i"`` specifies a single integer
927 value, while the format string ``"[ssb]"`` or the equivalent ``"[s, s,
928 b]"`` specifies an array value with two strings and a boolean as its
931 /* Create the JSON integer 42 */
934 /* Create the JSON array ["foo", "bar", true] */
935 json_pack("[ssb]", "foo", "bar", 1);
937 Here's the full list of format characters. The type in parentheses
938 denotes the resulting JSON type, and the type in brackets (if any)
939 denotes the C type that is expected as the corresponding argument.
941 ``s`` (string) [const char \*]
942 Convert a NULL terminated UTF-8 string to a JSON string.
945 Output a JSON null value. No argument is consumed.
947 ``b`` (boolean) [int]
948 Convert a C :type:`int` to JSON boolean value. Zero is converted
949 to ``false`` and non-zero to ``true``.
951 ``i`` (integer) [int]
952 Convert a C :type:`int` to JSON integer.
954 ``I`` (integer) [json_int_t]
955 Convert a C :type:`json_int_t` to JSON integer.
957 ``f`` (real) [double]
958 Convert a C :type:`double` to JSON real.
960 ``o`` (any value) [json_t \*]
961 Output any given JSON value as-is. If the value is added to an
962 array or object, the reference to the value passed to ``o`` is
963 stolen by the container.
965 ``O`` (any value) [json_t \*]
966 Like ``o``, but the argument's reference count is incremented.
967 This is useful if you pack into an array or object and want to
968 keep the reference for the JSON value consumed by ``O`` to
972 Build an array with contents from the inner format string. ``fmt``
973 may contain objects and arrays, i.e. recursive value building is
977 Build an object with contents from the inner format string
978 ``fmt``. The first, third, etc. format character represent a key,
979 and must be ``s`` (as object keys are always strings). The second,
980 fourth, etc. format character represent a value. Any value may be
981 an object or array, i.e. recursive value building is supported.
983 The following functions compose the value building API:
985 .. function:: json_t *json_pack(const char *fmt, ...)
989 Build a new JSON value according to the format string *fmt*. For
990 each format character (except for ``{}[]n``), one argument is
991 consumed and used to build the corresponding value. Returns *NULL*
994 .. function:: json_t *json_pack_ex(json_error_t *error, size_t flags, const char *fmt, ...)
995 json_t *json_vpack_ex(json_error_t *error, size_t flags, const char *fmt, va_list ap)
999 Like :func:`json_pack()`, but an in the case of an error, an error
1000 message is written to *error*, if it's not *NULL*. The *flags*
1001 parameter is currently unused and should be set to 0.
1003 As only the errors in format string (and out-of-memory errors) can
1004 be caught by the packer, these two functions are most likely only
1005 useful for debugging format strings.
1009 /* Build an empty JSON object */
1012 /* Build the JSON object {"foo": 42, "bar": 7} */
1013 json_pack("{sisi}", "foo", 42, "bar", 7);
1015 /* Like above, ':', ',' and whitespace are ignored */
1016 json_pack("{s:i, s:i}", "foo", 42, "bar", 7);
1018 /* Build the JSON array [[1, 2], {"cool": true}] */
1019 json_pack("[[i,i],{s:b]]", 1, 2, "cool", 1);
1024 Parsing and Validating Values
1025 =============================
1027 This sectinon describes functions that help to validate complex values
1028 and extract, or *unpack*, data from them. Like :ref:`building values
1029 <apiref-pack>`, this is also based on format strings.
1031 While a JSON value is unpacked, the type specified in the format
1032 string is checked to match that of the JSON value. This is the
1033 validation part of the process. In addition to this, the unpacking
1034 functions can also check that all items of arrays and objects are
1035 unpacked. This check be enabled with the format character ``!`` or by
1036 using the flag ``JSON_STRICT``. See below for details.
1038 Here's the full list of format characters. The type in parentheses
1039 denotes the JSON type, and the type in brackets (if any) denotes the C
1040 type whose address should be passed.
1042 ``s`` (string) [const char \*]
1043 Convert a JSON string to a pointer to a NULL terminated UTF-8
1044 string. The resulting string is extracted by using
1045 :func:`json_string_value()` internally, so it exists as long as
1046 there are still references to the corresponding JSON string.
1049 Expect a JSON null value. Nothing is extracted.
1051 ``b`` (boolean) [int]
1052 Convert a JSON boolean value to a C :type:`int`, so that ``true``
1053 is converted to 1 and ``false`` to 0.
1055 ``i`` (integer) [int]
1056 Convert a JSON integer to C :type:`int`.
1058 ``I`` (integer) [json_int_t]
1059 Convert a JSON integer to C :type:`json_int_t`.
1061 ``f`` (real) [double]
1062 Convert a JSON real to C :type:`double`.
1064 ``F`` (integer or real) [double]
1065 Convert a JSON number (integer or real) to C :type:`double`.
1067 ``o`` (any value) [json_t \*]
1068 Store a JSON value with no conversion to a :type:`json_t` pointer.
1070 ``O`` (any value) [json_t \*]
1071 Like ``O``, but the JSON value's reference count is incremented.
1074 Convert each item in the JSON array according to the inner format
1075 string. ``fmt`` may contain objects and arrays, i.e. recursive
1076 value extraction is supporetd.
1079 Convert each item in the JSON object according to the inner format
1080 string ``fmt``. The first, third, etc. format character represent
1081 a key, and must be ``s``. The corresponding argument to unpack
1082 functions is read as the object key. The second fourth, etc.
1083 format character represent a value and is written to the address
1084 given as the corresponding argument. **Note** that every other
1085 argument is read from and every other is written to.
1087 ``fmt`` may contain objects and arrays as values, i.e. recursive
1088 value extraction is supporetd.
1091 This special format character is used to enable the check that
1092 all object and array items are accessed, on a per-value basis. It
1093 must appear inside an array or object as the last format character
1094 before the closing bracket or brace. To enable the check globally,
1095 use the ``JSON_STRICT`` unpacking flag.
1098 This special format character is the opposite of ``!``. If the
1099 ``JSON_STRICT`` flag is used, ``*`` can be used to disable the
1100 strict check on a per-value basis. It must appear inside an array
1101 or object as the last format character before the closing bracket
1104 The following functions compose the parsing and validation API:
1106 .. function:: int json_unpack(json_t *root, const char *fmt, ...)
1108 Validate and unpack the JSON value *root* according to the format
1109 string *fmt*. Returns 0 on success and -1 on failure.
1111 .. function:: int json_unpack_ex(json_t *root, json_error_t *error, size_t flags, const char *fmt, ...)
1112 int json_vunpack_ex(json_t *root, json_error_t *error, size_t flags, const char *fmt, va_list ap)
1114 Validate and unpack the JSON value *root* according to the format
1115 string *fmt*. If an error occurs and *error* is not *NULL*, write
1116 error information to *error*. *flags* can be used to control the
1117 behaviour of the unpacker, see below for the flags. Returns 0 on
1118 success and -1 on failure.
1122 The first argument of all unpack functions is ``json_t *root``
1123 instead of ``const json_t *root``, because the use of ``O`` format
1124 character causes the reference count of ``root``, or some value
1125 reachable from ``root``, to be increased. Furthermore, the ``o``
1126 format character may be used to extract a value as-is, which allows
1127 modifying the structure or contents of a value reachable from
1130 If the ``O`` and ``o`` format character are not used, it's
1131 perfectly safe to cast a ``const json_t *`` variable to plain
1132 ``json_t *`` when used with these functions.
1134 The following unpacking flags are available:
1137 Enable the extra validation step checking that all object and
1138 array items are unpacked. This is equivalent to appending the
1139 format character ``!`` to the end of every array and object in the
1142 ``JSON_VALIDATE_ONLY``
1143 Don't extract any data, just validate the JSON value against the
1144 given format string. Note that object keys must still be specified
1145 after the format string.
1149 /* root is the JSON integer 42 */
1151 json_unpack(root, "i", &myint);
1152 assert(myint == 42);
1154 /* root is the JSON object {"foo": "bar", "quux": true} */
1157 json_unpack(root, "{s:s, s:b}", "foo", &str, "quux", &boolean);
1158 assert(strcmp(str, "bar") == 0 && boolean == 1);
1160 /* root is the JSON array [[1, 2], {"baz": null} */
1162 json_unpack_ex(root, &error, JSON_VALIDATE_ONLY, "[[i,i], {s:n}]", "baz");
1163 /* returns 0 for validation success, nothing is extracted */
1165 /* root is the JSON array [1, 2, 3, 4, 5] */
1167 json_unpack(root, "[ii!]", &myint1, &myint2);
1168 /* returns -1 for failed validation */
1174 Testing for equality of two JSON values cannot, in general, be
1175 achieved using the ``==`` operator. Equality in the terms of the
1176 ``==`` operator states that the two :type:`json_t` pointers point to
1177 exactly the same JSON value. However, two JSON values can be equal not
1178 only if they are exactly the same value, but also if they have equal
1181 * Two integer or real values are equal if their contained numeric
1182 values are equal. An integer value is never equal to a real value,
1185 * Two strings are equal if their contained UTF-8 strings are equal,
1186 byte by byte. Unicode comparison algorithms are not implemented.
1188 * Two arrays are equal if they have the same number of elements and
1189 each element in the first array is equal to the corresponding
1190 element in the second array.
1192 * Two objects are equal if they have exactly the same keys and the
1193 value for each key in the first object is equal to the value of the
1194 corresponding key in the second object.
1196 * Two true, false or null values have no "contents", so they are equal
1197 if their types are equal. (Because these values are singletons,
1198 their equality can actually be tested with ``==``.)
1200 The following function can be used to test whether two JSON values are
1203 .. function:: int json_equal(json_t *value1, json_t *value2)
1205 Returns 1 if *value1* and *value2* are equal, as defined above.
1206 Returns 0 if they are inequal or one or both of the pointers are
1213 Because of reference counting, passing JSON values around doesn't
1214 require copying them. But sometimes a fresh copy of a JSON value is
1215 needed. For example, if you need to modify an array, but still want to
1216 use the original afterwards, you should take a copy of it first.
1218 Jansson supports two kinds of copying: shallow and deep. There is a
1219 difference between these methods only for arrays and objects. Shallow
1220 copying only copies the first level value (array or object) and uses
1221 the same child values in the copied value. Deep copying makes a fresh
1222 copy of the child values, too. Moreover, all the child values are deep
1223 copied in a recursive fashion.
1225 .. function:: json_t *json_copy(json_t *value)
1227 .. refcounting:: new
1229 Returns a shallow copy of *value*, or *NULL* on error.
1231 .. function:: json_t *json_deep_copy(json_t *value)
1233 .. refcounting:: new
1235 Returns a deep copy of *value*, or *NULL* on error.
1238 .. _apiref-custom-memory-allocation:
1240 Custom Memory Allocation
1241 ========================
1243 By default, Jansson uses :func:`malloc()` and :func:`free()` for
1244 memory allocation. These functions can be overridden if custom
1247 .. type:: json_malloc_t
1249 A typedef for a function pointer with :func:`malloc()`'s
1252 typedef void *(*json_malloc_t)(size_t);
1254 .. type:: json_free_t
1256 A typedef for a function pointer with :func:`free()`'s
1259 typedef void (*json_free_t)(void *);
1261 .. function:: void json_set_alloc_funcs(json_malloc_t malloc_fn, json_free_t free_fn)
1263 Use *malloc_fn* instead of :func:`malloc()` and *free_fn* instead
1264 of :func:`free()`. This function has to be called before any other
1265 Jansson's API functions to ensure that all memory operations use
1270 Use the `Boehm's conservative garbage collector`_ for memory
1273 json_set_alloc_funcs(GC_malloc, GC_free);
1275 .. _Boehm's conservative garbage collector: http://www.hpl.hp.com/personal/Hans_Boehm/gc/
1277 Allow storing sensitive data (e.g. passwords or encryption keys) in
1278 JSON structures by zeroing all memory when freed::
1280 static void *secure_malloc(size_t size)
1282 /* Store the memory area size in the beginning of the block */
1283 void *ptr = malloc(size + 8);
1284 *((size_t *)ptr) = size;
1288 static void secure_free(void *ptr)
1293 size = *((size_t *)ptr);
1295 guaranteed_memset(ptr, 0, size);
1301 json_set_alloc_funcs(secure_malloc, secure_free);
1305 For more information about the issues of storing sensitive data in
1307 http://www.dwheeler.com/secure-programs/Secure-Programs-HOWTO/protect-secrets.html.
1308 The page also explains the :func:`guaranteed_memset()` function used
1309 in the example and gives a sample implementation for it.