2 #include "structmember.h"
3 #if PY_VERSION_HEX < 0x02060000 && !defined(Py_TYPE)
4 #define Py_TYPE(ob) (((PyObject*)(ob))->ob_type)
6 #if PY_VERSION_HEX < 0x02050000 && !defined(PY_SSIZE_T_MIN)
7 typedef int Py_ssize_t
;
8 #define PY_SSIZE_T_MAX INT_MAX
9 #define PY_SSIZE_T_MIN INT_MIN
10 #define PyInt_FromSsize_t PyInt_FromLong
11 #define PyInt_AsSsize_t PyInt_AsLong
14 #define Py_IS_FINITE(X) (!Py_IS_INFINITY(X) && !Py_IS_NAN(X))
18 #define UNUSED __attribute__((__unused__))
23 #define DEFAULT_ENCODING "utf-8"
25 #define PyScanner_Check(op) PyObject_TypeCheck(op, &PyScannerType)
26 #define PyScanner_CheckExact(op) (Py_TYPE(op) == &PyScannerType)
27 #define PyEncoder_Check(op) PyObject_TypeCheck(op, &PyEncoderType)
28 #define PyEncoder_CheckExact(op) (Py_TYPE(op) == &PyEncoderType)
30 static PyTypeObject PyScannerType
;
31 static PyTypeObject PyEncoderType
;
33 typedef struct _PyScannerObject
{
37 PyObject
*object_hook
;
39 PyObject
*parse_float
;
41 PyObject
*parse_constant
;
44 static PyMemberDef scanner_members
[] = {
45 {"encoding", T_OBJECT
, offsetof(PyScannerObject
, encoding
), READONLY
, "encoding"},
46 {"strict", T_OBJECT
, offsetof(PyScannerObject
, strict
), READONLY
, "strict"},
47 {"object_hook", T_OBJECT
, offsetof(PyScannerObject
, object_hook
), READONLY
, "object_hook"},
48 {"object_pairs_hook", T_OBJECT
, offsetof(PyScannerObject
, pairs_hook
), READONLY
, "object_pairs_hook"},
49 {"parse_float", T_OBJECT
, offsetof(PyScannerObject
, parse_float
), READONLY
, "parse_float"},
50 {"parse_int", T_OBJECT
, offsetof(PyScannerObject
, parse_int
), READONLY
, "parse_int"},
51 {"parse_constant", T_OBJECT
, offsetof(PyScannerObject
, parse_constant
), READONLY
, "parse_constant"},
55 typedef struct _PyEncoderObject
{
61 PyObject
*key_separator
;
62 PyObject
*item_separator
;
69 static PyMemberDef encoder_members
[] = {
70 {"markers", T_OBJECT
, offsetof(PyEncoderObject
, markers
), READONLY
, "markers"},
71 {"default", T_OBJECT
, offsetof(PyEncoderObject
, defaultfn
), READONLY
, "default"},
72 {"encoder", T_OBJECT
, offsetof(PyEncoderObject
, encoder
), READONLY
, "encoder"},
73 {"indent", T_OBJECT
, offsetof(PyEncoderObject
, indent
), READONLY
, "indent"},
74 {"key_separator", T_OBJECT
, offsetof(PyEncoderObject
, key_separator
), READONLY
, "key_separator"},
75 {"item_separator", T_OBJECT
, offsetof(PyEncoderObject
, item_separator
), READONLY
, "item_separator"},
76 {"sort_keys", T_OBJECT
, offsetof(PyEncoderObject
, sort_keys
), READONLY
, "sort_keys"},
77 {"skipkeys", T_OBJECT
, offsetof(PyEncoderObject
, skipkeys
), READONLY
, "skipkeys"},
82 ascii_escape_char(Py_UNICODE c
, char *output
, Py_ssize_t chars
);
84 ascii_escape_unicode(PyObject
*pystr
);
86 ascii_escape_str(PyObject
*pystr
);
88 py_encode_basestring_ascii(PyObject
* self UNUSED
, PyObject
*pystr
);
91 scan_once_str(PyScannerObject
*s
, PyObject
*pystr
, Py_ssize_t idx
, Py_ssize_t
*next_idx_ptr
);
93 scan_once_unicode(PyScannerObject
*s
, PyObject
*pystr
, Py_ssize_t idx
, Py_ssize_t
*next_idx_ptr
);
95 _build_rval_index_tuple(PyObject
*rval
, Py_ssize_t idx
);
97 scanner_new(PyTypeObject
*type
, PyObject
*args
, PyObject
*kwds
);
99 scanner_init(PyObject
*self
, PyObject
*args
, PyObject
*kwds
);
101 scanner_dealloc(PyObject
*self
);
103 scanner_clear(PyObject
*self
);
105 encoder_new(PyTypeObject
*type
, PyObject
*args
, PyObject
*kwds
);
107 encoder_init(PyObject
*self
, PyObject
*args
, PyObject
*kwds
);
109 encoder_dealloc(PyObject
*self
);
111 encoder_clear(PyObject
*self
);
113 encoder_listencode_list(PyEncoderObject
*s
, PyObject
*rval
, PyObject
*seq
, Py_ssize_t indent_level
);
115 encoder_listencode_obj(PyEncoderObject
*s
, PyObject
*rval
, PyObject
*obj
, Py_ssize_t indent_level
);
117 encoder_listencode_dict(PyEncoderObject
*s
, PyObject
*rval
, PyObject
*dct
, Py_ssize_t indent_level
);
119 _encoded_const(PyObject
*obj
);
121 raise_errmsg(char *msg
, PyObject
*s
, Py_ssize_t end
);
123 encoder_encode_string(PyEncoderObject
*s
, PyObject
*obj
);
125 _convertPyInt_AsSsize_t(PyObject
*o
, Py_ssize_t
*size_ptr
);
127 _convertPyInt_FromSsize_t(Py_ssize_t
*size_ptr
);
129 encoder_encode_float(PyEncoderObject
*s
, PyObject
*obj
);
131 #define S_CHAR(c) (c >= ' ' && c <= '~' && c != '\\' && c != '"')
132 #define IS_WHITESPACE(c) (((c) == ' ') || ((c) == '\t') || ((c) == '\n') || ((c) == '\r'))
134 #define MIN_EXPANSION 6
135 #ifdef Py_UNICODE_WIDE
136 #define MAX_EXPANSION (2 * MIN_EXPANSION)
138 #define MAX_EXPANSION MIN_EXPANSION
142 _convertPyInt_AsSsize_t(PyObject
*o
, Py_ssize_t
*size_ptr
)
144 /* PyObject to Py_ssize_t converter */
145 *size_ptr
= PyInt_AsSsize_t(o
);
146 if (*size_ptr
== -1 && PyErr_Occurred());
152 _convertPyInt_FromSsize_t(Py_ssize_t
*size_ptr
)
154 /* Py_ssize_t to PyObject converter */
155 return PyInt_FromSsize_t(*size_ptr
);
159 ascii_escape_char(Py_UNICODE c
, char *output
, Py_ssize_t chars
)
161 /* Escape unicode code point c to ASCII escape sequences
162 in char *output. output must have at least 12 bytes unused to
163 accommodate an escaped surrogate pair "\uXXXX\uXXXX" */
164 output
[chars
++] = '\\';
166 case '\\': output
[chars
++] = (char)c
; break;
167 case '"': output
[chars
++] = (char)c
; break;
168 case '\b': output
[chars
++] = 'b'; break;
169 case '\f': output
[chars
++] = 'f'; break;
170 case '\n': output
[chars
++] = 'n'; break;
171 case '\r': output
[chars
++] = 'r'; break;
172 case '\t': output
[chars
++] = 't'; break;
174 #ifdef Py_UNICODE_WIDE
176 /* UTF-16 surrogate pair */
177 Py_UNICODE v
= c
- 0x10000;
178 c
= 0xd800 | ((v
>> 10) & 0x3ff);
179 output
[chars
++] = 'u';
180 output
[chars
++] = "0123456789abcdef"[(c
>> 12) & 0xf];
181 output
[chars
++] = "0123456789abcdef"[(c
>> 8) & 0xf];
182 output
[chars
++] = "0123456789abcdef"[(c
>> 4) & 0xf];
183 output
[chars
++] = "0123456789abcdef"[(c
) & 0xf];
184 c
= 0xdc00 | (v
& 0x3ff);
185 output
[chars
++] = '\\';
188 output
[chars
++] = 'u';
189 output
[chars
++] = "0123456789abcdef"[(c
>> 12) & 0xf];
190 output
[chars
++] = "0123456789abcdef"[(c
>> 8) & 0xf];
191 output
[chars
++] = "0123456789abcdef"[(c
>> 4) & 0xf];
192 output
[chars
++] = "0123456789abcdef"[(c
) & 0xf];
198 ascii_escape_unicode(PyObject
*pystr
)
200 /* Take a PyUnicode pystr and return a new ASCII-only escaped PyString */
202 Py_ssize_t input_chars
;
203 Py_ssize_t output_size
;
204 Py_ssize_t max_output_size
;
208 Py_UNICODE
*input_unicode
;
210 input_chars
= PyUnicode_GET_SIZE(pystr
);
211 input_unicode
= PyUnicode_AS_UNICODE(pystr
);
213 /* One char input can be up to 6 chars output, estimate 4 of these */
214 output_size
= 2 + (MIN_EXPANSION
* 4) + input_chars
;
215 max_output_size
= 2 + (input_chars
* MAX_EXPANSION
);
216 rval
= PyString_FromStringAndSize(NULL
, output_size
);
220 output
= PyString_AS_STRING(rval
);
222 output
[chars
++] = '"';
223 for (i
= 0; i
< input_chars
; i
++) {
224 Py_UNICODE c
= input_unicode
[i
];
226 output
[chars
++] = (char)c
;
229 chars
= ascii_escape_char(c
, output
, chars
);
231 if (output_size
- chars
< (1 + MAX_EXPANSION
)) {
232 /* There's more than four, so let's resize by a lot */
233 Py_ssize_t new_output_size
= output_size
* 2;
234 /* This is an upper bound */
235 if (new_output_size
> max_output_size
) {
236 new_output_size
= max_output_size
;
238 /* Make sure that the output size changed before resizing */
239 if (new_output_size
!= output_size
) {
240 output_size
= new_output_size
;
241 if (_PyString_Resize(&rval
, output_size
) == -1) {
244 output
= PyString_AS_STRING(rval
);
248 output
[chars
++] = '"';
249 if (_PyString_Resize(&rval
, chars
) == -1) {
256 ascii_escape_str(PyObject
*pystr
)
258 /* Take a PyString pystr and return a new ASCII-only escaped PyString */
260 Py_ssize_t input_chars
;
261 Py_ssize_t output_size
;
267 input_chars
= PyString_GET_SIZE(pystr
);
268 input_str
= PyString_AS_STRING(pystr
);
270 /* Fast path for a string that's already ASCII */
271 for (i
= 0; i
< input_chars
; i
++) {
272 Py_UNICODE c
= (Py_UNICODE
)(unsigned char)input_str
[i
];
274 /* If we have to escape something, scan the string for unicode */
276 for (j
= i
; j
< input_chars
; j
++) {
277 c
= (Py_UNICODE
)(unsigned char)input_str
[j
];
279 /* We hit a non-ASCII character, bail to unicode mode */
281 uni
= PyUnicode_DecodeUTF8(input_str
, input_chars
, "strict");
285 rval
= ascii_escape_unicode(uni
);
294 if (i
== input_chars
) {
295 /* Input is already ASCII */
296 output_size
= 2 + input_chars
;
299 /* One char input can be up to 6 chars output, estimate 4 of these */
300 output_size
= 2 + (MIN_EXPANSION
* 4) + input_chars
;
302 rval
= PyString_FromStringAndSize(NULL
, output_size
);
306 output
= PyString_AS_STRING(rval
);
309 /* We know that everything up to i is ASCII already */
311 memcpy(&output
[1], input_str
, i
);
313 for (; i
< input_chars
; i
++) {
314 Py_UNICODE c
= (Py_UNICODE
)(unsigned char)input_str
[i
];
316 output
[chars
++] = (char)c
;
319 chars
= ascii_escape_char(c
, output
, chars
);
321 /* An ASCII char can't possibly expand to a surrogate! */
322 if (output_size
- chars
< (1 + MIN_EXPANSION
)) {
323 /* There's more than four, so let's resize by a lot */
325 if (output_size
> 2 + (input_chars
* MIN_EXPANSION
)) {
326 output_size
= 2 + (input_chars
* MIN_EXPANSION
);
328 if (_PyString_Resize(&rval
, output_size
) == -1) {
331 output
= PyString_AS_STRING(rval
);
334 output
[chars
++] = '"';
335 if (_PyString_Resize(&rval
, chars
) == -1) {
342 raise_errmsg(char *msg
, PyObject
*s
, Py_ssize_t end
)
344 /* Use the Python function json.decoder.errmsg to raise a nice
345 looking ValueError exception */
346 static PyObject
*errmsg_fn
= NULL
;
348 if (errmsg_fn
== NULL
) {
349 PyObject
*decoder
= PyImport_ImportModule("json.decoder");
352 errmsg_fn
= PyObject_GetAttrString(decoder
, "errmsg");
354 if (errmsg_fn
== NULL
)
357 pymsg
= PyObject_CallFunction(errmsg_fn
, "(zOO&)", msg
, s
, _convertPyInt_FromSsize_t
, &end
);
359 PyErr_SetObject(PyExc_ValueError
, pymsg
);
365 join_list_unicode(PyObject
*lst
)
367 /* return u''.join(lst) */
368 static PyObject
*joinfn
= NULL
;
369 if (joinfn
== NULL
) {
370 PyObject
*ustr
= PyUnicode_FromUnicode(NULL
, 0);
374 joinfn
= PyObject_GetAttrString(ustr
, "join");
379 return PyObject_CallFunctionObjArgs(joinfn
, lst
, NULL
);
383 join_list_string(PyObject
*lst
)
385 /* return ''.join(lst) */
386 static PyObject
*joinfn
= NULL
;
387 if (joinfn
== NULL
) {
388 PyObject
*ustr
= PyString_FromStringAndSize(NULL
, 0);
392 joinfn
= PyObject_GetAttrString(ustr
, "join");
397 return PyObject_CallFunctionObjArgs(joinfn
, lst
, NULL
);
401 _build_rval_index_tuple(PyObject
*rval
, Py_ssize_t idx
) {
402 /* return (rval, idx) tuple, stealing reference to rval */
406 steal a reference to rval, returns (rval, idx)
411 pyidx
= PyInt_FromSsize_t(idx
);
416 tpl
= PyTuple_New(2);
422 PyTuple_SET_ITEM(tpl
, 0, rval
);
423 PyTuple_SET_ITEM(tpl
, 1, pyidx
);
428 scanstring_str(PyObject
*pystr
, Py_ssize_t end
, char *encoding
, int strict
, Py_ssize_t
*next_end_ptr
)
430 /* Read the JSON string from PyString pystr.
431 end is the index of the first character after the quote.
432 encoding is the encoding of pystr (must be an ASCII superset)
433 if strict is zero then literal control characters are allowed
434 *next_end_ptr is a return-by-reference index of the character
437 Return value is a new PyString (if ASCII-only) or PyUnicode
440 Py_ssize_t len
= PyString_GET_SIZE(pystr
);
441 Py_ssize_t begin
= end
- 1;
442 Py_ssize_t next
= begin
;
444 char *buf
= PyString_AS_STRING(pystr
);
445 PyObject
*chunks
= PyList_New(0);
446 if (chunks
== NULL
) {
449 if (end
< 0 || len
<= end
) {
450 PyErr_SetString(PyExc_ValueError
, "end is out of bounds");
454 /* Find the end of the string or the next escape */
456 PyObject
*chunk
= NULL
;
457 for (next
= end
; next
< len
; next
++) {
458 c
= (unsigned char)buf
[next
];
459 if (c
== '"' || c
== '\\') {
462 else if (strict
&& c
<= 0x1f) {
463 raise_errmsg("Invalid control character at", pystr
, next
);
470 if (!(c
== '"' || c
== '\\')) {
471 raise_errmsg("Unterminated string starting at", pystr
, begin
);
474 /* Pick up this chunk if it's not zero length */
476 PyObject
*strchunk
= PyString_FromStringAndSize(&buf
[end
], next
- end
);
477 if (strchunk
== NULL
) {
481 chunk
= PyUnicode_FromEncodedObject(strchunk
, encoding
, NULL
);
490 if (PyList_Append(chunks
, chunk
)) {
502 raise_errmsg("Unterminated string starting at", pystr
, begin
);
507 /* Non-unicode backslash escapes */
513 case 'b': c
= '\b'; break;
514 case 'f': c
= '\f'; break;
515 case 'n': c
= '\n'; break;
516 case 'r': c
= '\r'; break;
517 case 't': c
= '\t'; break;
521 raise_errmsg("Invalid \\escape", pystr
, end
- 2);
530 raise_errmsg("Invalid \\uXXXX escape", pystr
, next
- 1);
533 /* Decode 4 hex digits */
534 for (; next
< end
; next
++) {
535 Py_UNICODE digit
= buf
[next
];
538 case '0': case '1': case '2': case '3': case '4':
539 case '5': case '6': case '7': case '8': case '9':
540 c
|= (digit
- '0'); break;
541 case 'a': case 'b': case 'c': case 'd': case 'e':
543 c
|= (digit
- 'a' + 10); break;
544 case 'A': case 'B': case 'C': case 'D': case 'E':
546 c
|= (digit
- 'A' + 10); break;
548 raise_errmsg("Invalid \\uXXXX escape", pystr
, end
- 5);
552 #ifdef Py_UNICODE_WIDE
554 if ((c
& 0xfc00) == 0xd800) {
556 if (end
+ 6 >= len
) {
557 raise_errmsg("Unpaired high surrogate", pystr
, end
- 5);
560 if (buf
[next
++] != '\\' || buf
[next
++] != 'u') {
561 raise_errmsg("Unpaired high surrogate", pystr
, end
- 5);
565 /* Decode 4 hex digits */
566 for (; next
< end
; next
++) {
568 Py_UNICODE digit
= buf
[next
];
570 case '0': case '1': case '2': case '3': case '4':
571 case '5': case '6': case '7': case '8': case '9':
572 c2
|= (digit
- '0'); break;
573 case 'a': case 'b': case 'c': case 'd': case 'e':
575 c2
|= (digit
- 'a' + 10); break;
576 case 'A': case 'B': case 'C': case 'D': case 'E':
578 c2
|= (digit
- 'A' + 10); break;
580 raise_errmsg("Invalid \\uXXXX escape", pystr
, end
- 5);
584 if ((c2
& 0xfc00) != 0xdc00) {
585 raise_errmsg("Unpaired high surrogate", pystr
, end
- 5);
588 c
= 0x10000 + (((c
- 0xd800) << 10) | (c2
- 0xdc00));
590 else if ((c
& 0xfc00) == 0xdc00) {
591 raise_errmsg("Unpaired low surrogate", pystr
, end
- 5);
600 chunk
= PyUnicode_FromUnicode(&c
, 1);
606 char c_char
= Py_CHARMASK(c
);
607 chunk
= PyString_FromStringAndSize(&c_char
, 1);
612 if (PyList_Append(chunks
, chunk
)) {
619 rval
= join_list_string(chunks
);
634 scanstring_unicode(PyObject
*pystr
, Py_ssize_t end
, int strict
, Py_ssize_t
*next_end_ptr
)
636 /* Read the JSON string from PyUnicode pystr.
637 end is the index of the first character after the quote.
638 if strict is zero then literal control characters are allowed
639 *next_end_ptr is a return-by-reference index of the character
642 Return value is a new PyUnicode
645 Py_ssize_t len
= PyUnicode_GET_SIZE(pystr
);
646 Py_ssize_t begin
= end
- 1;
647 Py_ssize_t next
= begin
;
648 const Py_UNICODE
*buf
= PyUnicode_AS_UNICODE(pystr
);
649 PyObject
*chunks
= PyList_New(0);
650 if (chunks
== NULL
) {
653 if (end
< 0 || len
<= end
) {
654 PyErr_SetString(PyExc_ValueError
, "end is out of bounds");
658 /* Find the end of the string or the next escape */
660 PyObject
*chunk
= NULL
;
661 for (next
= end
; next
< len
; next
++) {
663 if (c
== '"' || c
== '\\') {
666 else if (strict
&& c
<= 0x1f) {
667 raise_errmsg("Invalid control character at", pystr
, next
);
671 if (!(c
== '"' || c
== '\\')) {
672 raise_errmsg("Unterminated string starting at", pystr
, begin
);
675 /* Pick up this chunk if it's not zero length */
677 chunk
= PyUnicode_FromUnicode(&buf
[end
], next
- end
);
681 if (PyList_Append(chunks
, chunk
)) {
693 raise_errmsg("Unterminated string starting at", pystr
, begin
);
698 /* Non-unicode backslash escapes */
704 case 'b': c
= '\b'; break;
705 case 'f': c
= '\f'; break;
706 case 'n': c
= '\n'; break;
707 case 'r': c
= '\r'; break;
708 case 't': c
= '\t'; break;
712 raise_errmsg("Invalid \\escape", pystr
, end
- 2);
721 raise_errmsg("Invalid \\uXXXX escape", pystr
, next
- 1);
724 /* Decode 4 hex digits */
725 for (; next
< end
; next
++) {
726 Py_UNICODE digit
= buf
[next
];
729 case '0': case '1': case '2': case '3': case '4':
730 case '5': case '6': case '7': case '8': case '9':
731 c
|= (digit
- '0'); break;
732 case 'a': case 'b': case 'c': case 'd': case 'e':
734 c
|= (digit
- 'a' + 10); break;
735 case 'A': case 'B': case 'C': case 'D': case 'E':
737 c
|= (digit
- 'A' + 10); break;
739 raise_errmsg("Invalid \\uXXXX escape", pystr
, end
- 5);
743 #ifdef Py_UNICODE_WIDE
745 if ((c
& 0xfc00) == 0xd800) {
747 if (end
+ 6 >= len
) {
748 raise_errmsg("Unpaired high surrogate", pystr
, end
- 5);
751 if (buf
[next
++] != '\\' || buf
[next
++] != 'u') {
752 raise_errmsg("Unpaired high surrogate", pystr
, end
- 5);
756 /* Decode 4 hex digits */
757 for (; next
< end
; next
++) {
759 Py_UNICODE digit
= buf
[next
];
761 case '0': case '1': case '2': case '3': case '4':
762 case '5': case '6': case '7': case '8': case '9':
763 c2
|= (digit
- '0'); break;
764 case 'a': case 'b': case 'c': case 'd': case 'e':
766 c2
|= (digit
- 'a' + 10); break;
767 case 'A': case 'B': case 'C': case 'D': case 'E':
769 c2
|= (digit
- 'A' + 10); break;
771 raise_errmsg("Invalid \\uXXXX escape", pystr
, end
- 5);
775 if ((c2
& 0xfc00) != 0xdc00) {
776 raise_errmsg("Unpaired high surrogate", pystr
, end
- 5);
779 c
= 0x10000 + (((c
- 0xd800) << 10) | (c2
- 0xdc00));
781 else if ((c
& 0xfc00) == 0xdc00) {
782 raise_errmsg("Unpaired low surrogate", pystr
, end
- 5);
787 chunk
= PyUnicode_FromUnicode(&c
, 1);
791 if (PyList_Append(chunks
, chunk
)) {
798 rval
= join_list_unicode(chunks
);
811 PyDoc_STRVAR(pydoc_scanstring
,
812 "scanstring(basestring, end, encoding, strict=True) -> (str, end)\n"
814 "Scan the string s for a JSON string. End is the index of the\n"
815 "character in s after the quote that started the JSON string.\n"
816 "Unescapes all valid JSON string escape sequences and raises ValueError\n"
817 "on attempt to decode an invalid string. If strict is False then literal\n"
818 "control characters are allowed in the string.\n"
820 "Returns a tuple of the decoded string and the index of the character in s\n"
821 "after the end quote."
825 py_scanstring(PyObject
* self UNUSED
, PyObject
*args
)
830 Py_ssize_t next_end
= -1;
831 char *encoding
= NULL
;
833 if (!PyArg_ParseTuple(args
, "OO&|zi:scanstring", &pystr
, _convertPyInt_AsSsize_t
, &end
, &encoding
, &strict
)) {
836 if (encoding
== NULL
) {
837 encoding
= DEFAULT_ENCODING
;
839 if (PyString_Check(pystr
)) {
840 rval
= scanstring_str(pystr
, end
, encoding
, strict
, &next_end
);
842 else if (PyUnicode_Check(pystr
)) {
843 rval
= scanstring_unicode(pystr
, end
, strict
, &next_end
);
846 PyErr_Format(PyExc_TypeError
,
847 "first argument must be a string, not %.80s",
848 Py_TYPE(pystr
)->tp_name
);
851 return _build_rval_index_tuple(rval
, next_end
);
854 PyDoc_STRVAR(pydoc_encode_basestring_ascii
,
855 "encode_basestring_ascii(basestring) -> str\n"
857 "Return an ASCII-only JSON representation of a Python string"
861 py_encode_basestring_ascii(PyObject
* self UNUSED
, PyObject
*pystr
)
863 /* Return an ASCII-only JSON representation of a Python string */
865 if (PyString_Check(pystr
)) {
866 return ascii_escape_str(pystr
);
868 else if (PyUnicode_Check(pystr
)) {
869 return ascii_escape_unicode(pystr
);
872 PyErr_Format(PyExc_TypeError
,
873 "first argument must be a string, not %.80s",
874 Py_TYPE(pystr
)->tp_name
);
880 scanner_dealloc(PyObject
*self
)
882 /* Deallocate scanner object */
884 Py_TYPE(self
)->tp_free(self
);
888 scanner_traverse(PyObject
*self
, visitproc visit
, void *arg
)
891 assert(PyScanner_Check(self
));
892 s
= (PyScannerObject
*)self
;
893 Py_VISIT(s
->encoding
);
895 Py_VISIT(s
->object_hook
);
896 Py_VISIT(s
->pairs_hook
);
897 Py_VISIT(s
->parse_float
);
898 Py_VISIT(s
->parse_int
);
899 Py_VISIT(s
->parse_constant
);
904 scanner_clear(PyObject
*self
)
907 assert(PyScanner_Check(self
));
908 s
= (PyScannerObject
*)self
;
909 Py_CLEAR(s
->encoding
);
911 Py_CLEAR(s
->object_hook
);
912 Py_CLEAR(s
->pairs_hook
);
913 Py_CLEAR(s
->parse_float
);
914 Py_CLEAR(s
->parse_int
);
915 Py_CLEAR(s
->parse_constant
);
920 _parse_object_str(PyScannerObject
*s
, PyObject
*pystr
, Py_ssize_t idx
, Py_ssize_t
*next_idx_ptr
) {
921 /* Read a JSON object from PyString pystr.
922 idx is the index of the first character after the opening curly brace.
923 *next_idx_ptr is a return-by-reference index to the first character after
924 the closing curly brace.
926 Returns a new PyObject (usually a dict, but object_hook can change that)
928 char *str
= PyString_AS_STRING(pystr
);
929 Py_ssize_t end_idx
= PyString_GET_SIZE(pystr
) - 1;
933 PyObject
*key
= NULL
;
934 PyObject
*val
= NULL
;
935 char *encoding
= PyString_AS_STRING(s
->encoding
);
936 int strict
= PyObject_IsTrue(s
->strict
);
939 pairs
= PyList_New(0);
943 /* skip whitespace after { */
944 while (idx
<= end_idx
&& IS_WHITESPACE(str
[idx
])) idx
++;
946 /* only loop if the object is non-empty */
947 if (idx
<= end_idx
&& str
[idx
] != '}') {
948 while (idx
<= end_idx
) {
950 if (str
[idx
] != '"') {
951 raise_errmsg("Expecting property name", pystr
, idx
);
954 key
= scanstring_str(pystr
, idx
+ 1, encoding
, strict
, &next_idx
);
959 /* skip whitespace between key and : delimiter, read :, skip whitespace */
960 while (idx
<= end_idx
&& IS_WHITESPACE(str
[idx
])) idx
++;
961 if (idx
> end_idx
|| str
[idx
] != ':') {
962 raise_errmsg("Expecting : delimiter", pystr
, idx
);
966 while (idx
<= end_idx
&& IS_WHITESPACE(str
[idx
])) idx
++;
968 /* read any JSON data type */
969 val
= scan_once_str(s
, pystr
, idx
, &next_idx
);
973 item
= PyTuple_Pack(2, key
, val
);
978 if (PyList_Append(pairs
, item
) == -1) {
985 /* skip whitespace before } or , */
986 while (idx
<= end_idx
&& IS_WHITESPACE(str
[idx
])) idx
++;
988 /* bail if the object is closed or we didn't get the , delimiter */
989 if (idx
> end_idx
) break;
990 if (str
[idx
] == '}') {
993 else if (str
[idx
] != ',') {
994 raise_errmsg("Expecting , delimiter", pystr
, idx
);
999 /* skip whitespace after , delimiter */
1000 while (idx
<= end_idx
&& IS_WHITESPACE(str
[idx
])) idx
++;
1003 /* verify that idx < end_idx, str[idx] should be '}' */
1004 if (idx
> end_idx
|| str
[idx
] != '}') {
1005 raise_errmsg("Expecting object", pystr
, end_idx
);
1009 /* if pairs_hook is not None: rval = object_pairs_hook(pairs) */
1010 if (s
->pairs_hook
!= Py_None
) {
1011 val
= PyObject_CallFunctionObjArgs(s
->pairs_hook
, pairs
, NULL
);
1015 *next_idx_ptr
= idx
+ 1;
1019 rval
= PyObject_CallFunctionObjArgs((PyObject
*)(&PyDict_Type
),
1025 /* if object_hook is not None: rval = object_hook(rval) */
1026 if (s
->object_hook
!= Py_None
) {
1027 val
= PyObject_CallFunctionObjArgs(s
->object_hook
, rval
, NULL
);
1034 *next_idx_ptr
= idx
+ 1;
1044 _parse_object_unicode(PyScannerObject
*s
, PyObject
*pystr
, Py_ssize_t idx
, Py_ssize_t
*next_idx_ptr
) {
1045 /* Read a JSON object from PyUnicode pystr.
1046 idx is the index of the first character after the opening curly brace.
1047 *next_idx_ptr is a return-by-reference index to the first character after
1048 the closing curly brace.
1050 Returns a new PyObject (usually a dict, but object_hook can change that)
1052 Py_UNICODE
*str
= PyUnicode_AS_UNICODE(pystr
);
1053 Py_ssize_t end_idx
= PyUnicode_GET_SIZE(pystr
) - 1;
1057 PyObject
*key
= NULL
;
1058 PyObject
*val
= NULL
;
1059 int strict
= PyObject_IsTrue(s
->strict
);
1060 Py_ssize_t next_idx
;
1062 pairs
= PyList_New(0);
1066 /* skip whitespace after { */
1067 while (idx
<= end_idx
&& IS_WHITESPACE(str
[idx
])) idx
++;
1069 /* only loop if the object is non-empty */
1070 if (idx
<= end_idx
&& str
[idx
] != '}') {
1071 while (idx
<= end_idx
) {
1073 if (str
[idx
] != '"') {
1074 raise_errmsg("Expecting property name", pystr
, idx
);
1077 key
= scanstring_unicode(pystr
, idx
+ 1, strict
, &next_idx
);
1082 /* skip whitespace between key and : delimiter, read :, skip whitespace */
1083 while (idx
<= end_idx
&& IS_WHITESPACE(str
[idx
])) idx
++;
1084 if (idx
> end_idx
|| str
[idx
] != ':') {
1085 raise_errmsg("Expecting : delimiter", pystr
, idx
);
1089 while (idx
<= end_idx
&& IS_WHITESPACE(str
[idx
])) idx
++;
1091 /* read any JSON term */
1092 val
= scan_once_unicode(s
, pystr
, idx
, &next_idx
);
1096 item
= PyTuple_Pack(2, key
, val
);
1101 if (PyList_Append(pairs
, item
) == -1) {
1108 /* skip whitespace before } or , */
1109 while (idx
<= end_idx
&& IS_WHITESPACE(str
[idx
])) idx
++;
1111 /* bail if the object is closed or we didn't get the , delimiter */
1112 if (idx
> end_idx
) break;
1113 if (str
[idx
] == '}') {
1116 else if (str
[idx
] != ',') {
1117 raise_errmsg("Expecting , delimiter", pystr
, idx
);
1122 /* skip whitespace after , delimiter */
1123 while (idx
<= end_idx
&& IS_WHITESPACE(str
[idx
])) idx
++;
1127 /* verify that idx < end_idx, str[idx] should be '}' */
1128 if (idx
> end_idx
|| str
[idx
] != '}') {
1129 raise_errmsg("Expecting object", pystr
, end_idx
);
1133 /* if pairs_hook is not None: rval = object_pairs_hook(pairs) */
1134 if (s
->pairs_hook
!= Py_None
) {
1135 val
= PyObject_CallFunctionObjArgs(s
->pairs_hook
, pairs
, NULL
);
1139 *next_idx_ptr
= idx
+ 1;
1143 rval
= PyObject_CallFunctionObjArgs((PyObject
*)(&PyDict_Type
),
1149 /* if object_hook is not None: rval = object_hook(rval) */
1150 if (s
->object_hook
!= Py_None
) {
1151 val
= PyObject_CallFunctionObjArgs(s
->object_hook
, rval
, NULL
);
1158 *next_idx_ptr
= idx
+ 1;
1168 _parse_array_str(PyScannerObject
*s
, PyObject
*pystr
, Py_ssize_t idx
, Py_ssize_t
*next_idx_ptr
) {
1169 /* Read a JSON array from PyString pystr.
1170 idx is the index of the first character after the opening brace.
1171 *next_idx_ptr is a return-by-reference index to the first character after
1174 Returns a new PyList
1176 char *str
= PyString_AS_STRING(pystr
);
1177 Py_ssize_t end_idx
= PyString_GET_SIZE(pystr
) - 1;
1178 PyObject
*val
= NULL
;
1179 PyObject
*rval
= PyList_New(0);
1180 Py_ssize_t next_idx
;
1184 /* skip whitespace after [ */
1185 while (idx
<= end_idx
&& IS_WHITESPACE(str
[idx
])) idx
++;
1187 /* only loop if the array is non-empty */
1188 if (idx
<= end_idx
&& str
[idx
] != ']') {
1189 while (idx
<= end_idx
) {
1191 /* read any JSON term and de-tuplefy the (rval, idx) */
1192 val
= scan_once_str(s
, pystr
, idx
, &next_idx
);
1196 if (PyList_Append(rval
, val
) == -1)
1202 /* skip whitespace between term and , */
1203 while (idx
<= end_idx
&& IS_WHITESPACE(str
[idx
])) idx
++;
1205 /* bail if the array is closed or we didn't get the , delimiter */
1206 if (idx
> end_idx
) break;
1207 if (str
[idx
] == ']') {
1210 else if (str
[idx
] != ',') {
1211 raise_errmsg("Expecting , delimiter", pystr
, idx
);
1216 /* skip whitespace after , */
1217 while (idx
<= end_idx
&& IS_WHITESPACE(str
[idx
])) idx
++;
1221 /* verify that idx < end_idx, str[idx] should be ']' */
1222 if (idx
> end_idx
|| str
[idx
] != ']') {
1223 raise_errmsg("Expecting object", pystr
, end_idx
);
1226 *next_idx_ptr
= idx
+ 1;
1235 _parse_array_unicode(PyScannerObject
*s
, PyObject
*pystr
, Py_ssize_t idx
, Py_ssize_t
*next_idx_ptr
) {
1236 /* Read a JSON array from PyString pystr.
1237 idx is the index of the first character after the opening brace.
1238 *next_idx_ptr is a return-by-reference index to the first character after
1241 Returns a new PyList
1243 Py_UNICODE
*str
= PyUnicode_AS_UNICODE(pystr
);
1244 Py_ssize_t end_idx
= PyUnicode_GET_SIZE(pystr
) - 1;
1245 PyObject
*val
= NULL
;
1246 PyObject
*rval
= PyList_New(0);
1247 Py_ssize_t next_idx
;
1251 /* skip whitespace after [ */
1252 while (idx
<= end_idx
&& IS_WHITESPACE(str
[idx
])) idx
++;
1254 /* only loop if the array is non-empty */
1255 if (idx
<= end_idx
&& str
[idx
] != ']') {
1256 while (idx
<= end_idx
) {
1258 /* read any JSON term */
1259 val
= scan_once_unicode(s
, pystr
, idx
, &next_idx
);
1263 if (PyList_Append(rval
, val
) == -1)
1269 /* skip whitespace between term and , */
1270 while (idx
<= end_idx
&& IS_WHITESPACE(str
[idx
])) idx
++;
1272 /* bail if the array is closed or we didn't get the , delimiter */
1273 if (idx
> end_idx
) break;
1274 if (str
[idx
] == ']') {
1277 else if (str
[idx
] != ',') {
1278 raise_errmsg("Expecting , delimiter", pystr
, idx
);
1283 /* skip whitespace after , */
1284 while (idx
<= end_idx
&& IS_WHITESPACE(str
[idx
])) idx
++;
1288 /* verify that idx < end_idx, str[idx] should be ']' */
1289 if (idx
> end_idx
|| str
[idx
] != ']') {
1290 raise_errmsg("Expecting object", pystr
, end_idx
);
1293 *next_idx_ptr
= idx
+ 1;
1302 _parse_constant(PyScannerObject
*s
, char *constant
, Py_ssize_t idx
, Py_ssize_t
*next_idx_ptr
) {
1303 /* Read a JSON constant from PyString pystr.
1304 constant is the constant string that was found
1305 ("NaN", "Infinity", "-Infinity").
1306 idx is the index of the first character of the constant
1307 *next_idx_ptr is a return-by-reference index to the first character after
1310 Returns the result of parse_constant
1314 /* constant is "NaN", "Infinity", or "-Infinity" */
1315 cstr
= PyString_InternFromString(constant
);
1319 /* rval = parse_constant(constant) */
1320 rval
= PyObject_CallFunctionObjArgs(s
->parse_constant
, cstr
, NULL
);
1321 idx
+= PyString_GET_SIZE(cstr
);
1323 *next_idx_ptr
= idx
;
1328 _match_number_str(PyScannerObject
*s
, PyObject
*pystr
, Py_ssize_t start
, Py_ssize_t
*next_idx_ptr
) {
1329 /* Read a JSON number from PyString pystr.
1330 idx is the index of the first character of the number
1331 *next_idx_ptr is a return-by-reference index to the first character after
1334 Returns a new PyObject representation of that number:
1335 PyInt, PyLong, or PyFloat.
1336 May return other types if parse_int or parse_float are set
1338 char *str
= PyString_AS_STRING(pystr
);
1339 Py_ssize_t end_idx
= PyString_GET_SIZE(pystr
) - 1;
1340 Py_ssize_t idx
= start
;
1345 /* read a sign if it's there, make sure it's not the end of the string */
1346 if (str
[idx
] == '-') {
1348 if (idx
> end_idx
) {
1349 PyErr_SetNone(PyExc_StopIteration
);
1354 /* read as many integer digits as we find as long as it doesn't start with 0 */
1355 if (str
[idx
] >= '1' && str
[idx
] <= '9') {
1357 while (idx
<= end_idx
&& str
[idx
] >= '0' && str
[idx
] <= '9') idx
++;
1359 /* if it starts with 0 we only expect one integer digit */
1360 else if (str
[idx
] == '0') {
1363 /* no integer digits, error */
1365 PyErr_SetNone(PyExc_StopIteration
);
1369 /* if the next char is '.' followed by a digit then read all float digits */
1370 if (idx
< end_idx
&& str
[idx
] == '.' && str
[idx
+ 1] >= '0' && str
[idx
+ 1] <= '9') {
1373 while (idx
<= end_idx
&& str
[idx
] >= '0' && str
[idx
] <= '9') idx
++;
1376 /* if the next char is 'e' or 'E' then maybe read the exponent (or backtrack) */
1377 if (idx
< end_idx
&& (str
[idx
] == 'e' || str
[idx
] == 'E')) {
1379 /* save the index of the 'e' or 'E' just in case we need to backtrack */
1380 Py_ssize_t e_start
= idx
;
1383 /* read an exponent sign if present */
1384 if (idx
< end_idx
&& (str
[idx
] == '-' || str
[idx
] == '+')) idx
++;
1386 /* read all digits */
1387 while (idx
<= end_idx
&& str
[idx
] >= '0' && str
[idx
] <= '9') idx
++;
1389 /* if we got a digit, then parse as float. if not, backtrack */
1390 if (str
[idx
- 1] >= '0' && str
[idx
- 1] <= '9') {
1398 /* copy the section we determined to be a number */
1399 numstr
= PyString_FromStringAndSize(&str
[start
], idx
- start
);
1403 /* parse as a float using a fast path if available, otherwise call user defined method */
1404 if (s
->parse_float
!= (PyObject
*)&PyFloat_Type
) {
1405 rval
= PyObject_CallFunctionObjArgs(s
->parse_float
, numstr
, NULL
);
1408 rval
= PyFloat_FromDouble(PyOS_ascii_atof(PyString_AS_STRING(numstr
)));
1412 /* parse as an int using a fast path if available, otherwise call user defined method */
1413 if (s
->parse_int
!= (PyObject
*)&PyInt_Type
) {
1414 rval
= PyObject_CallFunctionObjArgs(s
->parse_int
, numstr
, NULL
);
1417 rval
= PyInt_FromString(PyString_AS_STRING(numstr
), NULL
, 10);
1421 *next_idx_ptr
= idx
;
1426 _match_number_unicode(PyScannerObject
*s
, PyObject
*pystr
, Py_ssize_t start
, Py_ssize_t
*next_idx_ptr
) {
1427 /* Read a JSON number from PyUnicode pystr.
1428 idx is the index of the first character of the number
1429 *next_idx_ptr is a return-by-reference index to the first character after
1432 Returns a new PyObject representation of that number:
1433 PyInt, PyLong, or PyFloat.
1434 May return other types if parse_int or parse_float are set
1436 Py_UNICODE
*str
= PyUnicode_AS_UNICODE(pystr
);
1437 Py_ssize_t end_idx
= PyUnicode_GET_SIZE(pystr
) - 1;
1438 Py_ssize_t idx
= start
;
1443 /* read a sign if it's there, make sure it's not the end of the string */
1444 if (str
[idx
] == '-') {
1446 if (idx
> end_idx
) {
1447 PyErr_SetNone(PyExc_StopIteration
);
1452 /* read as many integer digits as we find as long as it doesn't start with 0 */
1453 if (str
[idx
] >= '1' && str
[idx
] <= '9') {
1455 while (idx
<= end_idx
&& str
[idx
] >= '0' && str
[idx
] <= '9') idx
++;
1457 /* if it starts with 0 we only expect one integer digit */
1458 else if (str
[idx
] == '0') {
1461 /* no integer digits, error */
1463 PyErr_SetNone(PyExc_StopIteration
);
1467 /* if the next char is '.' followed by a digit then read all float digits */
1468 if (idx
< end_idx
&& str
[idx
] == '.' && str
[idx
+ 1] >= '0' && str
[idx
+ 1] <= '9') {
1471 while (idx
<= end_idx
&& str
[idx
] >= '0' && str
[idx
] <= '9') idx
++;
1474 /* if the next char is 'e' or 'E' then maybe read the exponent (or backtrack) */
1475 if (idx
< end_idx
&& (str
[idx
] == 'e' || str
[idx
] == 'E')) {
1476 Py_ssize_t e_start
= idx
;
1479 /* read an exponent sign if present */
1480 if (idx
< end_idx
&& (str
[idx
] == '-' || str
[idx
] == '+')) idx
++;
1482 /* read all digits */
1483 while (idx
<= end_idx
&& str
[idx
] >= '0' && str
[idx
] <= '9') idx
++;
1485 /* if we got a digit, then parse as float. if not, backtrack */
1486 if (str
[idx
- 1] >= '0' && str
[idx
- 1] <= '9') {
1494 /* copy the section we determined to be a number */
1495 numstr
= PyUnicode_FromUnicode(&str
[start
], idx
- start
);
1499 /* parse as a float using a fast path if available, otherwise call user defined method */
1500 if (s
->parse_float
!= (PyObject
*)&PyFloat_Type
) {
1501 rval
= PyObject_CallFunctionObjArgs(s
->parse_float
, numstr
, NULL
);
1504 rval
= PyFloat_FromString(numstr
, NULL
);
1508 /* no fast path for unicode -> int, just call */
1509 rval
= PyObject_CallFunctionObjArgs(s
->parse_int
, numstr
, NULL
);
1512 *next_idx_ptr
= idx
;
1517 scan_once_str(PyScannerObject
*s
, PyObject
*pystr
, Py_ssize_t idx
, Py_ssize_t
*next_idx_ptr
)
1519 /* Read one JSON term (of any kind) from PyString pystr.
1520 idx is the index of the first character of the term
1521 *next_idx_ptr is a return-by-reference index to the first character after
1524 Returns a new PyObject representation of the term.
1526 char *str
= PyString_AS_STRING(pystr
);
1527 Py_ssize_t length
= PyString_GET_SIZE(pystr
);
1528 if (idx
>= length
) {
1529 PyErr_SetNone(PyExc_StopIteration
);
1535 return scanstring_str(pystr
, idx
+ 1,
1536 PyString_AS_STRING(s
->encoding
),
1537 PyObject_IsTrue(s
->strict
),
1541 return _parse_object_str(s
, pystr
, idx
+ 1, next_idx_ptr
);
1544 return _parse_array_str(s
, pystr
, idx
+ 1, next_idx_ptr
);
1547 if ((idx
+ 3 < length
) && str
[idx
+ 1] == 'u' && str
[idx
+ 2] == 'l' && str
[idx
+ 3] == 'l') {
1549 *next_idx_ptr
= idx
+ 4;
1555 if ((idx
+ 3 < length
) && str
[idx
+ 1] == 'r' && str
[idx
+ 2] == 'u' && str
[idx
+ 3] == 'e') {
1557 *next_idx_ptr
= idx
+ 4;
1563 if ((idx
+ 4 < length
) && str
[idx
+ 1] == 'a' && str
[idx
+ 2] == 'l' && str
[idx
+ 3] == 's' && str
[idx
+ 4] == 'e') {
1564 Py_INCREF(Py_False
);
1565 *next_idx_ptr
= idx
+ 5;
1571 if ((idx
+ 2 < length
) && str
[idx
+ 1] == 'a' && str
[idx
+ 2] == 'N') {
1572 return _parse_constant(s
, "NaN", idx
, next_idx_ptr
);
1577 if ((idx
+ 7 < length
) && str
[idx
+ 1] == 'n' && str
[idx
+ 2] == 'f' && str
[idx
+ 3] == 'i' && str
[idx
+ 4] == 'n' && str
[idx
+ 5] == 'i' && str
[idx
+ 6] == 't' && str
[idx
+ 7] == 'y') {
1578 return _parse_constant(s
, "Infinity", idx
, next_idx_ptr
);
1583 if ((idx
+ 8 < length
) && str
[idx
+ 1] == 'I' && str
[idx
+ 2] == 'n' && str
[idx
+ 3] == 'f' && str
[idx
+ 4] == 'i' && str
[idx
+ 5] == 'n' && str
[idx
+ 6] == 'i' && str
[idx
+ 7] == 't' && str
[idx
+ 8] == 'y') {
1584 return _parse_constant(s
, "-Infinity", idx
, next_idx_ptr
);
1588 /* Didn't find a string, object, array, or named constant. Look for a number. */
1589 return _match_number_str(s
, pystr
, idx
, next_idx_ptr
);
1593 scan_once_unicode(PyScannerObject
*s
, PyObject
*pystr
, Py_ssize_t idx
, Py_ssize_t
*next_idx_ptr
)
1595 /* Read one JSON term (of any kind) from PyUnicode pystr.
1596 idx is the index of the first character of the term
1597 *next_idx_ptr is a return-by-reference index to the first character after
1600 Returns a new PyObject representation of the term.
1602 Py_UNICODE
*str
= PyUnicode_AS_UNICODE(pystr
);
1603 Py_ssize_t length
= PyUnicode_GET_SIZE(pystr
);
1604 if (idx
>= length
) {
1605 PyErr_SetNone(PyExc_StopIteration
);
1611 return scanstring_unicode(pystr
, idx
+ 1,
1612 PyObject_IsTrue(s
->strict
),
1616 return _parse_object_unicode(s
, pystr
, idx
+ 1, next_idx_ptr
);
1619 return _parse_array_unicode(s
, pystr
, idx
+ 1, next_idx_ptr
);
1622 if ((idx
+ 3 < length
) && str
[idx
+ 1] == 'u' && str
[idx
+ 2] == 'l' && str
[idx
+ 3] == 'l') {
1624 *next_idx_ptr
= idx
+ 4;
1630 if ((idx
+ 3 < length
) && str
[idx
+ 1] == 'r' && str
[idx
+ 2] == 'u' && str
[idx
+ 3] == 'e') {
1632 *next_idx_ptr
= idx
+ 4;
1638 if ((idx
+ 4 < length
) && str
[idx
+ 1] == 'a' && str
[idx
+ 2] == 'l' && str
[idx
+ 3] == 's' && str
[idx
+ 4] == 'e') {
1639 Py_INCREF(Py_False
);
1640 *next_idx_ptr
= idx
+ 5;
1646 if ((idx
+ 2 < length
) && str
[idx
+ 1] == 'a' && str
[idx
+ 2] == 'N') {
1647 return _parse_constant(s
, "NaN", idx
, next_idx_ptr
);
1652 if ((idx
+ 7 < length
) && str
[idx
+ 1] == 'n' && str
[idx
+ 2] == 'f' && str
[idx
+ 3] == 'i' && str
[idx
+ 4] == 'n' && str
[idx
+ 5] == 'i' && str
[idx
+ 6] == 't' && str
[idx
+ 7] == 'y') {
1653 return _parse_constant(s
, "Infinity", idx
, next_idx_ptr
);
1658 if ((idx
+ 8 < length
) && str
[idx
+ 1] == 'I' && str
[idx
+ 2] == 'n' && str
[idx
+ 3] == 'f' && str
[idx
+ 4] == 'i' && str
[idx
+ 5] == 'n' && str
[idx
+ 6] == 'i' && str
[idx
+ 7] == 't' && str
[idx
+ 8] == 'y') {
1659 return _parse_constant(s
, "-Infinity", idx
, next_idx_ptr
);
1663 /* Didn't find a string, object, array, or named constant. Look for a number. */
1664 return _match_number_unicode(s
, pystr
, idx
, next_idx_ptr
);
1668 scanner_call(PyObject
*self
, PyObject
*args
, PyObject
*kwds
)
1670 /* Python callable interface to scan_once_{str,unicode} */
1674 Py_ssize_t next_idx
= -1;
1675 static char *kwlist
[] = {"string", "idx", NULL
};
1677 assert(PyScanner_Check(self
));
1678 s
= (PyScannerObject
*)self
;
1679 if (!PyArg_ParseTupleAndKeywords(args
, kwds
, "OO&:scan_once", kwlist
, &pystr
, _convertPyInt_AsSsize_t
, &idx
))
1682 if (PyString_Check(pystr
)) {
1683 rval
= scan_once_str(s
, pystr
, idx
, &next_idx
);
1685 else if (PyUnicode_Check(pystr
)) {
1686 rval
= scan_once_unicode(s
, pystr
, idx
, &next_idx
);
1689 PyErr_Format(PyExc_TypeError
,
1690 "first argument must be a string, not %.80s",
1691 Py_TYPE(pystr
)->tp_name
);
1694 return _build_rval_index_tuple(rval
, next_idx
);
1698 scanner_new(PyTypeObject
*type
, PyObject
*args
, PyObject
*kwds
)
1701 s
= (PyScannerObject
*)type
->tp_alloc(type
, 0);
1705 s
->object_hook
= NULL
;
1706 s
->pairs_hook
= NULL
;
1707 s
->parse_float
= NULL
;
1708 s
->parse_int
= NULL
;
1709 s
->parse_constant
= NULL
;
1711 return (PyObject
*)s
;
1715 scanner_init(PyObject
*self
, PyObject
*args
, PyObject
*kwds
)
1717 /* Initialize Scanner object */
1719 static char *kwlist
[] = {"context", NULL
};
1722 assert(PyScanner_Check(self
));
1723 s
= (PyScannerObject
*)self
;
1725 if (!PyArg_ParseTupleAndKeywords(args
, kwds
, "O:make_scanner", kwlist
, &ctx
))
1728 /* PyString_AS_STRING is used on encoding */
1729 s
->encoding
= PyObject_GetAttrString(ctx
, "encoding");
1730 if (s
->encoding
== Py_None
) {
1732 s
->encoding
= PyString_InternFromString(DEFAULT_ENCODING
);
1734 else if (PyUnicode_Check(s
->encoding
)) {
1735 PyObject
*tmp
= PyUnicode_AsEncodedString(s
->encoding
, NULL
, NULL
);
1736 Py_DECREF(s
->encoding
);
1739 if (s
->encoding
== NULL
|| !PyString_Check(s
->encoding
))
1742 /* All of these will fail "gracefully" so we don't need to verify them */
1743 s
->strict
= PyObject_GetAttrString(ctx
, "strict");
1744 if (s
->strict
== NULL
)
1746 s
->object_hook
= PyObject_GetAttrString(ctx
, "object_hook");
1747 if (s
->object_hook
== NULL
)
1749 s
->pairs_hook
= PyObject_GetAttrString(ctx
, "object_pairs_hook");
1750 if (s
->object_hook
== NULL
)
1752 s
->parse_float
= PyObject_GetAttrString(ctx
, "parse_float");
1753 if (s
->parse_float
== NULL
)
1755 s
->parse_int
= PyObject_GetAttrString(ctx
, "parse_int");
1756 if (s
->parse_int
== NULL
)
1758 s
->parse_constant
= PyObject_GetAttrString(ctx
, "parse_constant");
1759 if (s
->parse_constant
== NULL
)
1765 Py_CLEAR(s
->encoding
);
1766 Py_CLEAR(s
->strict
);
1767 Py_CLEAR(s
->object_hook
);
1768 Py_CLEAR(s
->pairs_hook
);
1769 Py_CLEAR(s
->parse_float
);
1770 Py_CLEAR(s
->parse_int
);
1771 Py_CLEAR(s
->parse_constant
);
1775 PyDoc_STRVAR(scanner_doc
, "JSON scanner object");
1778 PyTypeObject PyScannerType
= {
1779 PyObject_HEAD_INIT(NULL
)
1780 0, /* tp_internal */
1781 "_json.Scanner", /* tp_name */
1782 sizeof(PyScannerObject
), /* tp_basicsize */
1783 0, /* tp_itemsize */
1784 scanner_dealloc
, /* tp_dealloc */
1790 0, /* tp_as_number */
1791 0, /* tp_as_sequence */
1792 0, /* tp_as_mapping */
1794 scanner_call
, /* tp_call */
1796 0,/* PyObject_GenericGetAttr, */ /* tp_getattro */
1797 0,/* PyObject_GenericSetAttr, */ /* tp_setattro */
1798 0, /* tp_as_buffer */
1799 Py_TPFLAGS_DEFAULT
| Py_TPFLAGS_HAVE_GC
, /* tp_flags */
1800 scanner_doc
, /* tp_doc */
1801 scanner_traverse
, /* tp_traverse */
1802 scanner_clear
, /* tp_clear */
1803 0, /* tp_richcompare */
1804 0, /* tp_weaklistoffset */
1806 0, /* tp_iternext */
1808 scanner_members
, /* tp_members */
1812 0, /* tp_descr_get */
1813 0, /* tp_descr_set */
1814 0, /* tp_dictoffset */
1815 scanner_init
, /* tp_init */
1816 0,/* PyType_GenericAlloc, */ /* tp_alloc */
1817 scanner_new
, /* tp_new */
1818 0,/* PyObject_GC_Del, */ /* tp_free */
1822 encoder_new(PyTypeObject
*type
, PyObject
*args
, PyObject
*kwds
)
1825 s
= (PyEncoderObject
*)type
->tp_alloc(type
, 0);
1828 s
->defaultfn
= NULL
;
1831 s
->key_separator
= NULL
;
1832 s
->item_separator
= NULL
;
1833 s
->sort_keys
= NULL
;
1836 return (PyObject
*)s
;
1840 encoder_init(PyObject
*self
, PyObject
*args
, PyObject
*kwds
)
1842 /* initialize Encoder object */
1843 static char *kwlist
[] = {"markers", "default", "encoder", "indent", "key_separator", "item_separator", "sort_keys", "skipkeys", "allow_nan", NULL
};
1846 PyObject
*allow_nan
;
1848 assert(PyEncoder_Check(self
));
1849 s
= (PyEncoderObject
*)self
;
1851 if (!PyArg_ParseTupleAndKeywords(args
, kwds
, "OOOOOOOOO:make_encoder", kwlist
,
1852 &s
->markers
, &s
->defaultfn
, &s
->encoder
, &s
->indent
, &s
->key_separator
, &s
->item_separator
, &s
->sort_keys
, &s
->skipkeys
, &allow_nan
))
1855 Py_INCREF(s
->markers
);
1856 Py_INCREF(s
->defaultfn
);
1857 Py_INCREF(s
->encoder
);
1858 Py_INCREF(s
->indent
);
1859 Py_INCREF(s
->key_separator
);
1860 Py_INCREF(s
->item_separator
);
1861 Py_INCREF(s
->sort_keys
);
1862 Py_INCREF(s
->skipkeys
);
1863 s
->fast_encode
= (PyCFunction_Check(s
->encoder
) && PyCFunction_GetFunction(s
->encoder
) == (PyCFunction
)py_encode_basestring_ascii
);
1864 s
->allow_nan
= PyObject_IsTrue(allow_nan
);
1869 encoder_call(PyObject
*self
, PyObject
*args
, PyObject
*kwds
)
1871 /* Python callable interface to encode_listencode_obj */
1872 static char *kwlist
[] = {"obj", "_current_indent_level", NULL
};
1875 Py_ssize_t indent_level
;
1877 assert(PyEncoder_Check(self
));
1878 s
= (PyEncoderObject
*)self
;
1879 if (!PyArg_ParseTupleAndKeywords(args
, kwds
, "OO&:_iterencode", kwlist
,
1880 &obj
, _convertPyInt_AsSsize_t
, &indent_level
))
1882 rval
= PyList_New(0);
1885 if (encoder_listencode_obj(s
, rval
, obj
, indent_level
)) {
1893 _encoded_const(PyObject
*obj
)
1895 /* Return the JSON string representation of None, True, False */
1896 if (obj
== Py_None
) {
1897 static PyObject
*s_null
= NULL
;
1898 if (s_null
== NULL
) {
1899 s_null
= PyString_InternFromString("null");
1904 else if (obj
== Py_True
) {
1905 static PyObject
*s_true
= NULL
;
1906 if (s_true
== NULL
) {
1907 s_true
= PyString_InternFromString("true");
1912 else if (obj
== Py_False
) {
1913 static PyObject
*s_false
= NULL
;
1914 if (s_false
== NULL
) {
1915 s_false
= PyString_InternFromString("false");
1921 PyErr_SetString(PyExc_ValueError
, "not a const");
1927 encoder_encode_float(PyEncoderObject
*s
, PyObject
*obj
)
1929 /* Return the JSON representation of a PyFloat */
1930 double i
= PyFloat_AS_DOUBLE(obj
);
1931 if (!Py_IS_FINITE(i
)) {
1932 if (!s
->allow_nan
) {
1933 PyErr_SetString(PyExc_ValueError
, "Out of range float values are not JSON compliant");
1937 return PyString_FromString("Infinity");
1940 return PyString_FromString("-Infinity");
1943 return PyString_FromString("NaN");
1946 /* Use a better float format here? */
1947 return PyObject_Repr(obj
);
1951 encoder_encode_string(PyEncoderObject
*s
, PyObject
*obj
)
1953 /* Return the JSON representation of a string */
1955 return py_encode_basestring_ascii(NULL
, obj
);
1957 return PyObject_CallFunctionObjArgs(s
->encoder
, obj
, NULL
);
1961 _steal_list_append(PyObject
*lst
, PyObject
*stolen
)
1963 /* Append stolen and then decrement its reference count */
1964 int rval
= PyList_Append(lst
, stolen
);
1970 encoder_listencode_obj(PyEncoderObject
*s
, PyObject
*rval
, PyObject
*obj
, Py_ssize_t indent_level
)
1972 /* Encode Python object obj to a JSON term, rval is a PyList */
1976 if (obj
== Py_None
|| obj
== Py_True
|| obj
== Py_False
) {
1977 PyObject
*cstr
= _encoded_const(obj
);
1980 return _steal_list_append(rval
, cstr
);
1982 else if (PyString_Check(obj
) || PyUnicode_Check(obj
))
1984 PyObject
*encoded
= encoder_encode_string(s
, obj
);
1985 if (encoded
== NULL
)
1987 return _steal_list_append(rval
, encoded
);
1989 else if (PyInt_Check(obj
) || PyLong_Check(obj
)) {
1990 PyObject
*encoded
= PyObject_Str(obj
);
1991 if (encoded
== NULL
)
1993 return _steal_list_append(rval
, encoded
);
1995 else if (PyFloat_Check(obj
)) {
1996 PyObject
*encoded
= encoder_encode_float(s
, obj
);
1997 if (encoded
== NULL
)
1999 return _steal_list_append(rval
, encoded
);
2001 else if (PyList_Check(obj
) || PyTuple_Check(obj
)) {
2002 return encoder_listencode_list(s
, rval
, obj
, indent_level
);
2004 else if (PyDict_Check(obj
)) {
2005 return encoder_listencode_dict(s
, rval
, obj
, indent_level
);
2008 PyObject
*ident
= NULL
;
2009 if (s
->markers
!= Py_None
) {
2011 ident
= PyLong_FromVoidPtr(obj
);
2014 has_key
= PyDict_Contains(s
->markers
, ident
);
2017 PyErr_SetString(PyExc_ValueError
, "Circular reference detected");
2021 if (PyDict_SetItem(s
->markers
, ident
, obj
)) {
2026 newobj
= PyObject_CallFunctionObjArgs(s
->defaultfn
, obj
, NULL
);
2027 if (newobj
== NULL
) {
2031 rv
= encoder_listencode_obj(s
, rval
, newobj
, indent_level
);
2037 if (ident
!= NULL
) {
2038 if (PyDict_DelItem(s
->markers
, ident
)) {
2049 encoder_listencode_dict(PyEncoderObject
*s
, PyObject
*rval
, PyObject
*dct
, Py_ssize_t indent_level
)
2051 /* Encode Python dict dct a JSON term, rval is a PyList */
2052 static PyObject
*open_dict
= NULL
;
2053 static PyObject
*close_dict
= NULL
;
2054 static PyObject
*empty_dict
= NULL
;
2055 PyObject
*kstr
= NULL
;
2056 PyObject
*ident
= NULL
;
2057 PyObject
*key
, *value
;
2062 if (open_dict
== NULL
|| close_dict
== NULL
|| empty_dict
== NULL
) {
2063 open_dict
= PyString_InternFromString("{");
2064 close_dict
= PyString_InternFromString("}");
2065 empty_dict
= PyString_InternFromString("{}");
2066 if (open_dict
== NULL
|| close_dict
== NULL
|| empty_dict
== NULL
)
2069 if (PyDict_Size(dct
) == 0)
2070 return PyList_Append(rval
, empty_dict
);
2072 if (s
->markers
!= Py_None
) {
2074 ident
= PyLong_FromVoidPtr(dct
);
2077 has_key
= PyDict_Contains(s
->markers
, ident
);
2080 PyErr_SetString(PyExc_ValueError
, "Circular reference detected");
2083 if (PyDict_SetItem(s
->markers
, ident
, dct
)) {
2088 if (PyList_Append(rval
, open_dict
))
2091 if (s
->indent
!= Py_None
) {
2092 /* TODO: DOES NOT RUN */
2095 newline_indent = '\n' + (' ' * (_indent * _current_indent_level))
2096 separator = _item_separator + newline_indent
2097 buf += newline_indent
2101 /* TODO: C speedup not implemented for sort_keys */
2104 skipkeys
= PyObject_IsTrue(s
->skipkeys
);
2106 while (PyDict_Next(dct
, &pos
, &key
, &value
)) {
2109 if (PyString_Check(key
) || PyUnicode_Check(key
)) {
2113 else if (PyFloat_Check(key
)) {
2114 kstr
= encoder_encode_float(s
, key
);
2118 else if (PyInt_Check(key
) || PyLong_Check(key
)) {
2119 kstr
= PyObject_Str(key
);
2123 else if (key
== Py_True
|| key
== Py_False
|| key
== Py_None
) {
2124 kstr
= _encoded_const(key
);
2128 else if (skipkeys
) {
2132 /* TODO: include repr of key */
2133 PyErr_SetString(PyExc_ValueError
, "keys must be a string");
2138 if (PyList_Append(rval
, s
->item_separator
))
2142 encoded
= encoder_encode_string(s
, kstr
);
2144 if (encoded
== NULL
)
2146 if (PyList_Append(rval
, encoded
)) {
2151 if (PyList_Append(rval
, s
->key_separator
))
2153 if (encoder_listencode_obj(s
, rval
, value
, indent_level
))
2157 if (ident
!= NULL
) {
2158 if (PyDict_DelItem(s
->markers
, ident
))
2162 if (s
->indent
!= Py_None
) {
2163 /* TODO: DOES NOT RUN */
2166 yield '\n' + (' ' * (_indent * _current_indent_level))
2169 if (PyList_Append(rval
, close_dict
))
2181 encoder_listencode_list(PyEncoderObject
*s
, PyObject
*rval
, PyObject
*seq
, Py_ssize_t indent_level
)
2183 /* Encode Python list seq to a JSON term, rval is a PyList */
2184 static PyObject
*open_array
= NULL
;
2185 static PyObject
*close_array
= NULL
;
2186 static PyObject
*empty_array
= NULL
;
2187 PyObject
*ident
= NULL
;
2188 PyObject
*s_fast
= NULL
;
2189 Py_ssize_t num_items
;
2190 PyObject
**seq_items
;
2193 if (open_array
== NULL
|| close_array
== NULL
|| empty_array
== NULL
) {
2194 open_array
= PyString_InternFromString("[");
2195 close_array
= PyString_InternFromString("]");
2196 empty_array
= PyString_InternFromString("[]");
2197 if (open_array
== NULL
|| close_array
== NULL
|| empty_array
== NULL
)
2201 s_fast
= PySequence_Fast(seq
, "_iterencode_list needs a sequence");
2204 num_items
= PySequence_Fast_GET_SIZE(s_fast
);
2205 if (num_items
== 0) {
2207 return PyList_Append(rval
, empty_array
);
2210 if (s
->markers
!= Py_None
) {
2212 ident
= PyLong_FromVoidPtr(seq
);
2215 has_key
= PyDict_Contains(s
->markers
, ident
);
2218 PyErr_SetString(PyExc_ValueError
, "Circular reference detected");
2221 if (PyDict_SetItem(s
->markers
, ident
, seq
)) {
2226 seq_items
= PySequence_Fast_ITEMS(s_fast
);
2227 if (PyList_Append(rval
, open_array
))
2229 if (s
->indent
!= Py_None
) {
2230 /* TODO: DOES NOT RUN */
2233 newline_indent = '\n' + (' ' * (_indent * _current_indent_level))
2234 separator = _item_separator + newline_indent
2235 buf += newline_indent
2238 for (i
= 0; i
< num_items
; i
++) {
2239 PyObject
*obj
= seq_items
[i
];
2241 if (PyList_Append(rval
, s
->item_separator
))
2244 if (encoder_listencode_obj(s
, rval
, obj
, indent_level
))
2247 if (ident
!= NULL
) {
2248 if (PyDict_DelItem(s
->markers
, ident
))
2252 if (s
->indent
!= Py_None
) {
2253 /* TODO: DOES NOT RUN */
2256 yield '\n' + (' ' * (_indent * _current_indent_level))
2259 if (PyList_Append(rval
, close_array
))
2271 encoder_dealloc(PyObject
*self
)
2273 /* Deallocate Encoder */
2274 encoder_clear(self
);
2275 Py_TYPE(self
)->tp_free(self
);
2279 encoder_traverse(PyObject
*self
, visitproc visit
, void *arg
)
2282 assert(PyEncoder_Check(self
));
2283 s
= (PyEncoderObject
*)self
;
2284 Py_VISIT(s
->markers
);
2285 Py_VISIT(s
->defaultfn
);
2286 Py_VISIT(s
->encoder
);
2287 Py_VISIT(s
->indent
);
2288 Py_VISIT(s
->key_separator
);
2289 Py_VISIT(s
->item_separator
);
2290 Py_VISIT(s
->sort_keys
);
2291 Py_VISIT(s
->skipkeys
);
2296 encoder_clear(PyObject
*self
)
2298 /* Deallocate Encoder */
2300 assert(PyEncoder_Check(self
));
2301 s
= (PyEncoderObject
*)self
;
2302 Py_CLEAR(s
->markers
);
2303 Py_CLEAR(s
->defaultfn
);
2304 Py_CLEAR(s
->encoder
);
2305 Py_CLEAR(s
->indent
);
2306 Py_CLEAR(s
->key_separator
);
2307 Py_CLEAR(s
->item_separator
);
2308 Py_CLEAR(s
->sort_keys
);
2309 Py_CLEAR(s
->skipkeys
);
2313 PyDoc_STRVAR(encoder_doc
, "_iterencode(obj, _current_indent_level) -> iterable");
2316 PyTypeObject PyEncoderType
= {
2317 PyObject_HEAD_INIT(NULL
)
2318 0, /* tp_internal */
2319 "_json.Encoder", /* tp_name */
2320 sizeof(PyEncoderObject
), /* tp_basicsize */
2321 0, /* tp_itemsize */
2322 encoder_dealloc
, /* tp_dealloc */
2328 0, /* tp_as_number */
2329 0, /* tp_as_sequence */
2330 0, /* tp_as_mapping */
2332 encoder_call
, /* tp_call */
2334 0, /* tp_getattro */
2335 0, /* tp_setattro */
2336 0, /* tp_as_buffer */
2337 Py_TPFLAGS_DEFAULT
| Py_TPFLAGS_HAVE_GC
, /* tp_flags */
2338 encoder_doc
, /* tp_doc */
2339 encoder_traverse
, /* tp_traverse */
2340 encoder_clear
, /* tp_clear */
2341 0, /* tp_richcompare */
2342 0, /* tp_weaklistoffset */
2344 0, /* tp_iternext */
2346 encoder_members
, /* tp_members */
2350 0, /* tp_descr_get */
2351 0, /* tp_descr_set */
2352 0, /* tp_dictoffset */
2353 encoder_init
, /* tp_init */
2355 encoder_new
, /* tp_new */
2359 static PyMethodDef speedups_methods
[] = {
2360 {"encode_basestring_ascii",
2361 (PyCFunction
)py_encode_basestring_ascii
,
2363 pydoc_encode_basestring_ascii
},
2365 (PyCFunction
)py_scanstring
,
2368 {NULL
, NULL
, 0, NULL
}
2371 PyDoc_STRVAR(module_doc
,
2378 PyScannerType
.tp_new
= PyType_GenericNew
;
2379 if (PyType_Ready(&PyScannerType
) < 0)
2381 PyEncoderType
.tp_new
= PyType_GenericNew
;
2382 if (PyType_Ready(&PyEncoderType
) < 0)
2384 m
= Py_InitModule3("_json", speedups_methods
, module_doc
);
2385 Py_INCREF((PyObject
*)&PyScannerType
);
2386 PyModule_AddObject(m
, "make_scanner", (PyObject
*)&PyScannerType
);
2387 Py_INCREF((PyObject
*)&PyEncoderType
);
2388 PyModule_AddObject(m
, "make_encoder", (PyObject
*)&PyEncoderType
);