summaryrefslogtreecommitdiffstats
path: root/Modules/_json.c
diff options
context:
space:
mode:
Diffstat (limited to 'Modules/_json.c')
-rw-r--r--Modules/_json.c1072
1 files changed, 227 insertions, 845 deletions
diff --git a/Modules/_json.c b/Modules/_json.c
index 13f5816..01436b6 100644
--- a/Modules/_json.c
+++ b/Modules/_json.c
@@ -20,8 +20,6 @@ typedef int Py_ssize_t;
#define UNUSED
#endif
-#define DEFAULT_ENCODING "utf-8"
-
#define PyScanner_Check(op) PyObject_TypeCheck(op, &PyScannerType)
#define PyScanner_CheckExact(op) (Py_TYPE(op) == &PyScannerType)
#define PyEncoder_Check(op) PyObject_TypeCheck(op, &PyEncoderType)
@@ -32,20 +30,19 @@ static PyTypeObject PyEncoderType;
typedef struct _PyScannerObject {
PyObject_HEAD
- PyObject *encoding;
PyObject *strict;
PyObject *object_hook;
- PyObject *pairs_hook;
+ PyObject *object_pairs_hook;
PyObject *parse_float;
PyObject *parse_int;
PyObject *parse_constant;
+ PyObject *memo;
} PyScannerObject;
static PyMemberDef scanner_members[] = {
- {"encoding", T_OBJECT, offsetof(PyScannerObject, encoding), READONLY, "encoding"},
{"strict", T_OBJECT, offsetof(PyScannerObject, strict), READONLY, "strict"},
{"object_hook", T_OBJECT, offsetof(PyScannerObject, object_hook), READONLY, "object_hook"},
- {"object_pairs_hook", T_OBJECT, offsetof(PyScannerObject, pairs_hook), READONLY, "object_pairs_hook"},
+ {"object_pairs_hook", T_OBJECT, offsetof(PyScannerObject, object_pairs_hook), READONLY},
{"parse_float", T_OBJECT, offsetof(PyScannerObject, parse_float), READONLY, "parse_float"},
{"parse_int", T_OBJECT, offsetof(PyScannerObject, parse_int), READONLY, "parse_int"},
{"parse_constant", T_OBJECT, offsetof(PyScannerObject, parse_constant), READONLY, "parse_constant"},
@@ -78,18 +75,12 @@ static PyMemberDef encoder_members[] = {
{NULL}
};
-static Py_ssize_t
-ascii_escape_char(Py_UNICODE c, char *output, Py_ssize_t chars);
static PyObject *
ascii_escape_unicode(PyObject *pystr);
static PyObject *
-ascii_escape_str(PyObject *pystr);
-static PyObject *
py_encode_basestring_ascii(PyObject* self UNUSED, PyObject *pystr);
void init_json(void);
static PyObject *
-scan_once_str(PyScannerObject *s, PyObject *pystr, Py_ssize_t idx, Py_ssize_t *next_idx_ptr);
-static PyObject *
scan_once_unicode(PyScannerObject *s, PyObject *pystr, Py_ssize_t idx, Py_ssize_t *next_idx_ptr);
static PyObject *
_build_rval_index_tuple(PyObject *rval, Py_ssize_t idx);
@@ -142,7 +133,7 @@ static int
_convertPyInt_AsSsize_t(PyObject *o, Py_ssize_t *size_ptr)
{
/* PyObject to Py_ssize_t converter */
- *size_ptr = PyInt_AsSsize_t(o);
+ *size_ptr = PyLong_AsSsize_t(o);
if (*size_ptr == -1 && PyErr_Occurred())
return 0;
return 1;
@@ -152,19 +143,19 @@ static PyObject *
_convertPyInt_FromSsize_t(Py_ssize_t *size_ptr)
{
/* Py_ssize_t to PyObject converter */
- return PyInt_FromSsize_t(*size_ptr);
+ return PyLong_FromSsize_t(*size_ptr);
}
static Py_ssize_t
-ascii_escape_char(Py_UNICODE c, char *output, Py_ssize_t chars)
+ascii_escape_unichar(Py_UNICODE c, Py_UNICODE *output, Py_ssize_t chars)
{
/* Escape unicode code point c to ASCII escape sequences
in char *output. output must have at least 12 bytes unused to
accommodate an escaped surrogate pair "\uXXXX\uXXXX" */
output[chars++] = '\\';
switch (c) {
- case '\\': output[chars++] = (char)c; break;
- case '"': output[chars++] = (char)c; break;
+ case '\\': output[chars++] = c; break;
+ case '"': output[chars++] = c; break;
case '\b': output[chars++] = 'b'; break;
case '\f': output[chars++] = 'f'; break;
case '\n': output[chars++] = 'n'; break;
@@ -197,14 +188,14 @@ ascii_escape_char(Py_UNICODE c, char *output, Py_ssize_t chars)
static PyObject *
ascii_escape_unicode(PyObject *pystr)
{
- /* Take a PyUnicode pystr and return a new ASCII-only escaped PyString */
+ /* Take a PyUnicode pystr and return a new ASCII-only escaped PyUnicode */
Py_ssize_t i;
Py_ssize_t input_chars;
Py_ssize_t output_size;
Py_ssize_t max_output_size;
Py_ssize_t chars;
PyObject *rval;
- char *output;
+ Py_UNICODE *output;
Py_UNICODE *input_unicode;
input_chars = PyUnicode_GET_SIZE(pystr);
@@ -213,20 +204,20 @@ ascii_escape_unicode(PyObject *pystr)
/* One char input can be up to 6 chars output, estimate 4 of these */
output_size = 2 + (MIN_EXPANSION * 4) + input_chars;
max_output_size = 2 + (input_chars * MAX_EXPANSION);
- rval = PyString_FromStringAndSize(NULL, output_size);
+ rval = PyUnicode_FromStringAndSize(NULL, output_size);
if (rval == NULL) {
return NULL;
}
- output = PyString_AS_STRING(rval);
+ output = PyUnicode_AS_UNICODE(rval);
chars = 0;
output[chars++] = '"';
for (i = 0; i < input_chars; i++) {
Py_UNICODE c = input_unicode[i];
if (S_CHAR(c)) {
- output[chars++] = (char)c;
+ output[chars++] = c;
}
else {
- chars = ascii_escape_char(c, output, chars);
+ chars = ascii_escape_unichar(c, output, chars);
}
if (output_size - chars < (1 + MAX_EXPANSION)) {
/* There's more than four, so let's resize by a lot */
@@ -238,101 +229,15 @@ ascii_escape_unicode(PyObject *pystr)
/* Make sure that the output size changed before resizing */
if (new_output_size != output_size) {
output_size = new_output_size;
- if (_PyString_Resize(&rval, output_size) == -1) {
+ if (PyUnicode_Resize(&rval, output_size) == -1) {
return NULL;
}
- output = PyString_AS_STRING(rval);
- }
- }
- }
- output[chars++] = '"';
- if (_PyString_Resize(&rval, chars) == -1) {
- return NULL;
- }
- return rval;
-}
-
-static PyObject *
-ascii_escape_str(PyObject *pystr)
-{
- /* Take a PyString pystr and return a new ASCII-only escaped PyString */
- Py_ssize_t i;
- Py_ssize_t input_chars;
- Py_ssize_t output_size;
- Py_ssize_t chars;
- PyObject *rval;
- char *output;
- char *input_str;
-
- input_chars = PyString_GET_SIZE(pystr);
- input_str = PyString_AS_STRING(pystr);
-
- /* Fast path for a string that's already ASCII */
- for (i = 0; i < input_chars; i++) {
- Py_UNICODE c = (Py_UNICODE)(unsigned char)input_str[i];
- if (!S_CHAR(c)) {
- /* If we have to escape something, scan the string for unicode */
- Py_ssize_t j;
- for (j = i; j < input_chars; j++) {
- c = (Py_UNICODE)(unsigned char)input_str[j];
- if (c > 0x7f) {
- /* We hit a non-ASCII character, bail to unicode mode */
- PyObject *uni;
- uni = PyUnicode_DecodeUTF8(input_str, input_chars, "strict");
- if (uni == NULL) {
- return NULL;
- }
- rval = ascii_escape_unicode(uni);
- Py_DECREF(uni);
- return rval;
- }
- }
- break;
- }
- }
-
- if (i == input_chars) {
- /* Input is already ASCII */
- output_size = 2 + input_chars;
- }
- else {
- /* One char input can be up to 6 chars output, estimate 4 of these */
- output_size = 2 + (MIN_EXPANSION * 4) + input_chars;
- }
- rval = PyString_FromStringAndSize(NULL, output_size);
- if (rval == NULL) {
- return NULL;
- }
- output = PyString_AS_STRING(rval);
- output[0] = '"';
-
- /* We know that everything up to i is ASCII already */
- chars = i + 1;
- memcpy(&output[1], input_str, i);
-
- for (; i < input_chars; i++) {
- Py_UNICODE c = (Py_UNICODE)(unsigned char)input_str[i];
- if (S_CHAR(c)) {
- output[chars++] = (char)c;
- }
- else {
- chars = ascii_escape_char(c, output, chars);
- }
- /* An ASCII char can't possibly expand to a surrogate! */
- if (output_size - chars < (1 + MIN_EXPANSION)) {
- /* There's more than four, so let's resize by a lot */
- output_size *= 2;
- if (output_size > 2 + (input_chars * MIN_EXPANSION)) {
- output_size = 2 + (input_chars * MIN_EXPANSION);
- }
- if (_PyString_Resize(&rval, output_size) == -1) {
- return NULL;
+ output = PyUnicode_AS_UNICODE(rval);
}
- output = PyString_AS_STRING(rval);
}
}
output[chars++] = '"';
- if (_PyString_Resize(&rval, chars) == -1) {
+ if (PyUnicode_Resize(&rval, chars) == -1) {
return NULL;
}
return rval;
@@ -365,18 +270,13 @@ static PyObject *
join_list_unicode(PyObject *lst)
{
/* return u''.join(lst) */
- static PyObject *joinfn = NULL;
- if (joinfn == NULL) {
- PyObject *ustr = PyUnicode_FromUnicode(NULL, 0);
- if (ustr == NULL)
- return NULL;
-
- joinfn = PyObject_GetAttrString(ustr, "join");
- Py_DECREF(ustr);
- if (joinfn == NULL)
+ static PyObject *sep = NULL;
+ if (sep == NULL) {
+ sep = PyUnicode_FromStringAndSize("", 0);
+ if (sep == NULL)
return NULL;
}
- return PyObject_CallFunctionObjArgs(joinfn, lst, NULL);
+ return PyUnicode_Join(sep, lst);
}
static PyObject *
@@ -390,7 +290,7 @@ _build_rval_index_tuple(PyObject *rval, Py_ssize_t idx) {
if (rval == NULL) {
return NULL;
}
- pyidx = PyInt_FromSsize_t(idx);
+ pyidx = PyLong_FromSsize_t(idx);
if (pyidx == NULL) {
Py_DECREF(rval);
return NULL;
@@ -406,190 +306,20 @@ _build_rval_index_tuple(PyObject *rval, Py_ssize_t idx) {
return tpl;
}
-static PyObject *
-scanstring_str(PyObject *pystr, Py_ssize_t end, char *encoding, int strict, Py_ssize_t *next_end_ptr)
-{
- /* Read the JSON string from PyString pystr.
- end is the index of the first character after the quote.
- encoding is the encoding of pystr (must be an ASCII superset)
- if strict is zero then literal control characters are allowed
- *next_end_ptr is a return-by-reference index of the character
- after the end quote
-
- Return value is a new PyString (if ASCII-only) or PyUnicode
- */
- PyObject *rval;
- Py_ssize_t len = PyString_GET_SIZE(pystr);
- Py_ssize_t begin = end - 1;
- Py_ssize_t next;
- char *buf = PyString_AS_STRING(pystr);
- PyObject *chunks = PyList_New(0);
- if (chunks == NULL) {
- goto bail;
- }
- if (end < 0 || len <= end) {
- PyErr_SetString(PyExc_ValueError, "end is out of bounds");
- goto bail;
- }
- while (1) {
- /* Find the end of the string or the next escape */
- Py_UNICODE c = 0;
- PyObject *chunk = NULL;
- for (next = end; next < len; next++) {
- c = (unsigned char)buf[next];
- if (c == '"' || c == '\\') {
- break;
- }
- else if (strict && c <= 0x1f) {
- raise_errmsg("Invalid control character at", pystr, next);
- goto bail;
- }
- }
- if (!(c == '"' || c == '\\')) {
- raise_errmsg("Unterminated string starting at", pystr, begin);
- goto bail;
- }
- /* Pick up this chunk if it's not zero length */
- if (next != end) {
- PyObject *strchunk = PyString_FromStringAndSize(&buf[end], next - end);
- if (strchunk == NULL) {
- goto bail;
- }
- chunk = PyUnicode_FromEncodedObject(strchunk, encoding, NULL);
- Py_DECREF(strchunk);
- if (chunk == NULL) {
- goto bail;
- }
- if (PyList_Append(chunks, chunk)) {
- Py_DECREF(chunk);
- goto bail;
- }
- Py_DECREF(chunk);
- }
- next++;
- if (c == '"') {
- end = next;
- break;
- }
- if (next == len) {
- raise_errmsg("Unterminated string starting at", pystr, begin);
- goto bail;
- }
- c = buf[next];
- if (c != 'u') {
- /* Non-unicode backslash escapes */
- end = next + 1;
- switch (c) {
- case '"': break;
- case '\\': break;
- case '/': break;
- case 'b': c = '\b'; break;
- case 'f': c = '\f'; break;
- case 'n': c = '\n'; break;
- case 'r': c = '\r'; break;
- case 't': c = '\t'; break;
- default: c = 0;
- }
- if (c == 0) {
- raise_errmsg("Invalid \\escape", pystr, end - 2);
- goto bail;
- }
- }
- else {
- c = 0;
- next++;
- end = next + 4;
- if (end >= len) {
- raise_errmsg("Invalid \\uXXXX escape", pystr, next - 1);
- goto bail;
- }
- /* Decode 4 hex digits */
- for (; next < end; next++) {
- Py_UNICODE digit = buf[next];
- c <<= 4;
- switch (digit) {
- case '0': case '1': case '2': case '3': case '4':
- case '5': case '6': case '7': case '8': case '9':
- c |= (digit - '0'); break;
- case 'a': case 'b': case 'c': case 'd': case 'e':
- case 'f':
- c |= (digit - 'a' + 10); break;
- case 'A': case 'B': case 'C': case 'D': case 'E':
- case 'F':
- c |= (digit - 'A' + 10); break;
- default:
- raise_errmsg("Invalid \\uXXXX escape", pystr, end - 5);
- goto bail;
- }
- }
-#ifdef Py_UNICODE_WIDE
- /* Surrogate pair */
- if ((c & 0xfc00) == 0xd800) {
- Py_UNICODE c2 = 0;
- if (end + 6 >= len) {
- raise_errmsg("Unpaired high surrogate", pystr, end - 5);
- goto bail;
- }
- if (buf[next++] != '\\' || buf[next++] != 'u') {
- raise_errmsg("Unpaired high surrogate", pystr, end - 5);
- goto bail;
- }
- end += 6;
- /* Decode 4 hex digits */
- for (; next < end; next++) {
- Py_UNICODE digit = buf[next];
- c2 <<= 4;
- switch (digit) {
- case '0': case '1': case '2': case '3': case '4':
- case '5': case '6': case '7': case '8': case '9':
- c2 |= (digit - '0'); break;
- case 'a': case 'b': case 'c': case 'd': case 'e':
- case 'f':
- c2 |= (digit - 'a' + 10); break;
- case 'A': case 'B': case 'C': case 'D': case 'E':
- case 'F':
- c2 |= (digit - 'A' + 10); break;
- default:
- raise_errmsg("Invalid \\uXXXX escape", pystr, end - 5);
- goto bail;
- }
- }
- if ((c2 & 0xfc00) != 0xdc00) {
- raise_errmsg("Unpaired high surrogate", pystr, end - 5);
- goto bail;
- }
- c = 0x10000 + (((c - 0xd800) << 10) | (c2 - 0xdc00));
- }
- else if ((c & 0xfc00) == 0xdc00) {
- raise_errmsg("Unpaired low surrogate", pystr, end - 5);
- goto bail;
- }
-#endif
- }
- chunk = PyUnicode_FromUnicode(&c, 1);
- if (chunk == NULL) {
- goto bail;
- }
- if (PyList_Append(chunks, chunk)) {
- Py_DECREF(chunk);
- goto bail;
- }
- Py_DECREF(chunk);
- }
-
- rval = join_list_unicode(chunks);
- if (rval == NULL) {
- goto bail;
+#define APPEND_OLD_CHUNK \
+ if (chunk != NULL) { \
+ if (chunks == NULL) { \
+ chunks = PyList_New(0); \
+ if (chunks == NULL) { \
+ goto bail; \
+ } \
+ } \
+ if (PyList_Append(chunks, chunk)) { \
+ Py_DECREF(chunk); \
+ goto bail; \
+ } \
+ Py_CLEAR(chunk); \
}
- Py_CLEAR(chunks);
- *next_end_ptr = end;
- return rval;
-bail:
- *next_end_ptr = -1;
- Py_XDECREF(chunks);
- return NULL;
-}
-
static PyObject *
scanstring_unicode(PyObject *pystr, Py_ssize_t end, int strict, Py_ssize_t *next_end_ptr)
@@ -602,15 +332,14 @@ scanstring_unicode(PyObject *pystr, Py_ssize_t end, int strict, Py_ssize_t *next
Return value is a new PyUnicode
*/
- PyObject *rval;
+ PyObject *rval = NULL;
Py_ssize_t len = PyUnicode_GET_SIZE(pystr);
Py_ssize_t begin = end - 1;
- Py_ssize_t next;
+ Py_ssize_t next = begin;
const Py_UNICODE *buf = PyUnicode_AS_UNICODE(pystr);
- PyObject *chunks = PyList_New(0);
- if (chunks == NULL) {
- goto bail;
- }
+ PyObject *chunks = NULL;
+ PyObject *chunk = NULL;
+
if (end < 0 || len <= end) {
PyErr_SetString(PyExc_ValueError, "end is out of bounds");
goto bail;
@@ -618,7 +347,6 @@ scanstring_unicode(PyObject *pystr, Py_ssize_t end, int strict, Py_ssize_t *next
while (1) {
/* Find the end of the string or the next escape */
Py_UNICODE c = 0;
- PyObject *chunk = NULL;
for (next = end; next < len; next++) {
c = buf[next];
if (c == '"' || c == '\\') {
@@ -635,15 +363,11 @@ scanstring_unicode(PyObject *pystr, Py_ssize_t end, int strict, Py_ssize_t *next
}
/* Pick up this chunk if it's not zero length */
if (next != end) {
+ APPEND_OLD_CHUNK
chunk = PyUnicode_FromUnicode(&buf[end], next - end);
if (chunk == NULL) {
goto bail;
}
- if (PyList_Append(chunks, chunk)) {
- Py_DECREF(chunk);
- goto bail;
- }
- Py_DECREF(chunk);
}
next++;
if (c == '"') {
@@ -745,32 +469,39 @@ scanstring_unicode(PyObject *pystr, Py_ssize_t end, int strict, Py_ssize_t *next
}
#endif
}
+ APPEND_OLD_CHUNK
chunk = PyUnicode_FromUnicode(&c, 1);
if (chunk == NULL) {
goto bail;
}
- if (PyList_Append(chunks, chunk)) {
- Py_DECREF(chunk);
+ }
+
+ if (chunks == NULL) {
+ if (chunk != NULL)
+ rval = chunk;
+ else
+ rval = PyUnicode_FromStringAndSize("", 0);
+ }
+ else {
+ APPEND_OLD_CHUNK
+ rval = join_list_unicode(chunks);
+ if (rval == NULL) {
goto bail;
}
- Py_DECREF(chunk);
+ Py_CLEAR(chunks);
}
- rval = join_list_unicode(chunks);
- if (rval == NULL) {
- goto bail;
- }
- Py_DECREF(chunks);
*next_end_ptr = end;
return rval;
bail:
*next_end_ptr = -1;
Py_XDECREF(chunks);
+ Py_XDECREF(chunk);
return NULL;
}
PyDoc_STRVAR(pydoc_scanstring,
- "scanstring(basestring, end, encoding, strict=True) -> (str, end)\n"
+ "scanstring(string, end, strict=True) -> (string, end)\n"
"\n"
"Scan the string s for a JSON string. End is the index of the\n"
"character in s after the quote that started the JSON string.\n"
@@ -789,18 +520,11 @@ py_scanstring(PyObject* self UNUSED, PyObject *args)
PyObject *rval;
Py_ssize_t end;
Py_ssize_t next_end = -1;
- char *encoding = NULL;
int strict = 1;
- if (!PyArg_ParseTuple(args, "OO&|zi:scanstring", &pystr, _convertPyInt_AsSsize_t, &end, &encoding, &strict)) {
+ if (!PyArg_ParseTuple(args, "OO&|i:scanstring", &pystr, _convertPyInt_AsSsize_t, &end, &strict)) {
return NULL;
}
- if (encoding == NULL) {
- encoding = DEFAULT_ENCODING;
- }
- if (PyString_Check(pystr)) {
- rval = scanstring_str(pystr, end, encoding, strict, &next_end);
- }
- else if (PyUnicode_Check(pystr)) {
+ if (PyUnicode_Check(pystr)) {
rval = scanstring_unicode(pystr, end, strict, &next_end);
}
else {
@@ -813,7 +537,7 @@ py_scanstring(PyObject* self UNUSED, PyObject *args)
}
PyDoc_STRVAR(pydoc_encode_basestring_ascii,
- "encode_basestring_ascii(basestring) -> str\n"
+ "encode_basestring_ascii(string) -> string\n"
"\n"
"Return an ASCII-only JSON representation of a Python string"
);
@@ -821,13 +545,11 @@ PyDoc_STRVAR(pydoc_encode_basestring_ascii,
static PyObject *
py_encode_basestring_ascii(PyObject* self UNUSED, PyObject *pystr)
{
+ PyObject *rval;
/* Return an ASCII-only JSON representation of a Python string */
/* METH_O */
- if (PyString_Check(pystr)) {
- return ascii_escape_str(pystr);
- }
- else if (PyUnicode_Check(pystr)) {
- return ascii_escape_unicode(pystr);
+ if (PyUnicode_Check(pystr)) {
+ rval = ascii_escape_unicode(pystr);
}
else {
PyErr_Format(PyExc_TypeError,
@@ -835,6 +557,7 @@ py_encode_basestring_ascii(PyObject* self UNUSED, PyObject *pystr)
Py_TYPE(pystr)->tp_name);
return NULL;
}
+ return rval;
}
static void
@@ -851,10 +574,9 @@ scanner_traverse(PyObject *self, visitproc visit, void *arg)
PyScannerObject *s;
assert(PyScanner_Check(self));
s = (PyScannerObject *)self;
- Py_VISIT(s->encoding);
Py_VISIT(s->strict);
Py_VISIT(s->object_hook);
- Py_VISIT(s->pairs_hook);
+ Py_VISIT(s->object_pairs_hook);
Py_VISIT(s->parse_float);
Py_VISIT(s->parse_int);
Py_VISIT(s->parse_constant);
@@ -867,141 +589,17 @@ scanner_clear(PyObject *self)
PyScannerObject *s;
assert(PyScanner_Check(self));
s = (PyScannerObject *)self;
- Py_CLEAR(s->encoding);
Py_CLEAR(s->strict);
Py_CLEAR(s->object_hook);
- Py_CLEAR(s->pairs_hook);
+ Py_CLEAR(s->object_pairs_hook);
Py_CLEAR(s->parse_float);
Py_CLEAR(s->parse_int);
Py_CLEAR(s->parse_constant);
+ Py_CLEAR(s->memo);
return 0;
}
static PyObject *
-_parse_object_str(PyScannerObject *s, PyObject *pystr, Py_ssize_t idx, Py_ssize_t *next_idx_ptr) {
- /* Read a JSON object from PyString pystr.
- idx is the index of the first character after the opening curly brace.
- *next_idx_ptr is a return-by-reference index to the first character after
- the closing curly brace.
-
- Returns a new PyObject (usually a dict, but object_hook can change that)
- */
- char *str = PyString_AS_STRING(pystr);
- Py_ssize_t end_idx = PyString_GET_SIZE(pystr) - 1;
- PyObject *rval;
- PyObject *pairs;
- PyObject *item;
- PyObject *key = NULL;
- PyObject *val = NULL;
- char *encoding = PyString_AS_STRING(s->encoding);
- int strict = PyObject_IsTrue(s->strict);
- Py_ssize_t next_idx;
-
- pairs = PyList_New(0);
- if (pairs == NULL)
- return NULL;
-
- /* skip whitespace after { */
- while (idx <= end_idx && IS_WHITESPACE(str[idx])) idx++;
-
- /* only loop if the object is non-empty */
- if (idx <= end_idx && str[idx] != '}') {
- while (idx <= end_idx) {
- /* read key */
- if (str[idx] != '"') {
- raise_errmsg("Expecting property name", pystr, idx);
- goto bail;
- }
- key = scanstring_str(pystr, idx + 1, encoding, strict, &next_idx);
- if (key == NULL)
- goto bail;
- idx = next_idx;
-
- /* skip whitespace between key and : delimiter, read :, skip whitespace */
- while (idx <= end_idx && IS_WHITESPACE(str[idx])) idx++;
- if (idx > end_idx || str[idx] != ':') {
- raise_errmsg("Expecting : delimiter", pystr, idx);
- goto bail;
- }
- idx++;
- while (idx <= end_idx && IS_WHITESPACE(str[idx])) idx++;
-
- /* read any JSON data type */
- val = scan_once_str(s, pystr, idx, &next_idx);
- if (val == NULL)
- goto bail;
-
- item = PyTuple_Pack(2, key, val);
- if (item == NULL)
- goto bail;
- Py_CLEAR(key);
- Py_CLEAR(val);
- if (PyList_Append(pairs, item) == -1) {
- Py_DECREF(item);
- goto bail;
- }
- Py_DECREF(item);
- idx = next_idx;
-
- /* skip whitespace before } or , */
- while (idx <= end_idx && IS_WHITESPACE(str[idx])) idx++;
-
- /* bail if the object is closed or we didn't get the , delimiter */
- if (idx > end_idx) break;
- if (str[idx] == '}') {
- break;
- }
- else if (str[idx] != ',') {
- raise_errmsg("Expecting , delimiter", pystr, idx);
- goto bail;
- }
- idx++;
-
- /* skip whitespace after , delimiter */
- while (idx <= end_idx && IS_WHITESPACE(str[idx])) idx++;
- }
- }
- /* verify that idx < end_idx, str[idx] should be '}' */
- if (idx > end_idx || str[idx] != '}') {
- raise_errmsg("Expecting object", pystr, end_idx);
- goto bail;
- }
-
- /* if pairs_hook is not None: rval = object_pairs_hook(pairs) */
- if (s->pairs_hook != Py_None) {
- val = PyObject_CallFunctionObjArgs(s->pairs_hook, pairs, NULL);
- if (val == NULL)
- goto bail;
- Py_DECREF(pairs);
- *next_idx_ptr = idx + 1;
- return val;
- }
-
- rval = PyObject_CallFunctionObjArgs((PyObject *)(&PyDict_Type),
- pairs, NULL);
- if (rval == NULL)
- goto bail;
- Py_CLEAR(pairs);
-
- /* if object_hook is not None: rval = object_hook(rval) */
- if (s->object_hook != Py_None) {
- val = PyObject_CallFunctionObjArgs(s->object_hook, rval, NULL);
- if (val == NULL)
- goto bail;
- Py_DECREF(rval);
- rval = val;
- val = NULL;
- }
- *next_idx_ptr = idx + 1;
- return rval;
-bail:
- Py_XDECREF(key);
- Py_XDECREF(val);
- Py_XDECREF(pairs);
- return NULL;
-}
-
-static PyObject *
_parse_object_unicode(PyScannerObject *s, PyObject *pystr, Py_ssize_t idx, Py_ssize_t *next_idx_ptr) {
/* Read a JSON object from PyUnicode pystr.
idx is the index of the first character after the opening curly brace.
@@ -1012,16 +610,18 @@ _parse_object_unicode(PyScannerObject *s, PyObject *pystr, Py_ssize_t idx, Py_ss
*/
Py_UNICODE *str = PyUnicode_AS_UNICODE(pystr);
Py_ssize_t end_idx = PyUnicode_GET_SIZE(pystr) - 1;
- PyObject *rval;
- PyObject *pairs;
- PyObject *item;
- PyObject *key = NULL;
PyObject *val = NULL;
+ PyObject *rval = NULL;
+ PyObject *key = NULL;
int strict = PyObject_IsTrue(s->strict);
+ int has_pairs_hook = (s->object_pairs_hook != Py_None);
Py_ssize_t next_idx;
- pairs = PyList_New(0);
- if (pairs == NULL)
+ if (has_pairs_hook)
+ rval = PyList_New(0);
+ else
+ rval = PyDict_New();
+ if (rval == NULL)
return NULL;
/* skip whitespace after { */
@@ -1030,6 +630,8 @@ _parse_object_unicode(PyScannerObject *s, PyObject *pystr, Py_ssize_t idx, Py_ss
/* only loop if the object is non-empty */
if (idx <= end_idx && str[idx] != '}') {
while (idx <= end_idx) {
+ PyObject *memokey;
+
/* read key */
if (str[idx] != '"') {
raise_errmsg("Expecting property name enclosed in double quotes", pystr, idx);
@@ -1038,6 +640,16 @@ _parse_object_unicode(PyScannerObject *s, PyObject *pystr, Py_ssize_t idx, Py_ss
key = scanstring_unicode(pystr, idx + 1, strict, &next_idx);
if (key == NULL)
goto bail;
+ memokey = PyDict_GetItem(s->memo, key);
+ if (memokey != NULL) {
+ Py_INCREF(memokey);
+ Py_DECREF(key);
+ key = memokey;
+ }
+ else {
+ if (PyDict_SetItem(s->memo, key, key) < 0)
+ goto bail;
+ }
idx = next_idx;
/* skip whitespace between key and : delimiter, read :, skip whitespace */
@@ -1054,16 +666,24 @@ _parse_object_unicode(PyScannerObject *s, PyObject *pystr, Py_ssize_t idx, Py_ss
if (val == NULL)
goto bail;
- item = PyTuple_Pack(2, key, val);
- if (item == NULL)
- goto bail;
- Py_CLEAR(key);
- Py_CLEAR(val);
- if (PyList_Append(pairs, item) == -1) {
+ if (has_pairs_hook) {
+ PyObject *item = PyTuple_Pack(2, key, val);
+ if (item == NULL)
+ goto bail;
+ Py_CLEAR(key);
+ Py_CLEAR(val);
+ if (PyList_Append(rval, item) == -1) {
+ Py_DECREF(item);
+ goto bail;
+ }
Py_DECREF(item);
- goto bail;
}
- Py_DECREF(item);
+ else {
+ if (PyDict_SetItem(rval, key, val) < 0)
+ goto bail;
+ Py_CLEAR(key);
+ Py_CLEAR(val);
+ }
idx = next_idx;
/* skip whitespace before } or , */
@@ -1091,104 +711,25 @@ _parse_object_unicode(PyScannerObject *s, PyObject *pystr, Py_ssize_t idx, Py_ss
goto bail;
}
- /* if pairs_hook is not None: rval = object_pairs_hook(pairs) */
- if (s->pairs_hook != Py_None) {
- val = PyObject_CallFunctionObjArgs(s->pairs_hook, pairs, NULL);
- if (val == NULL)
- goto bail;
- Py_DECREF(pairs);
- *next_idx_ptr = idx + 1;
+ *next_idx_ptr = idx + 1;
+
+ if (has_pairs_hook) {
+ val = PyObject_CallFunctionObjArgs(s->object_pairs_hook, rval, NULL);
+ Py_DECREF(rval);
return val;
}
- rval = PyObject_CallFunctionObjArgs((PyObject *)(&PyDict_Type),
- pairs, NULL);
- if (rval == NULL)
- goto bail;
- Py_CLEAR(pairs);
-
/* if object_hook is not None: rval = object_hook(rval) */
if (s->object_hook != Py_None) {
val = PyObject_CallFunctionObjArgs(s->object_hook, rval, NULL);
- if (val == NULL)
- goto bail;
Py_DECREF(rval);
- rval = val;
- val = NULL;
+ return val;
}
- *next_idx_ptr = idx + 1;
return rval;
bail:
Py_XDECREF(key);
Py_XDECREF(val);
- Py_XDECREF(pairs);
- return NULL;
-}
-
-static PyObject *
-_parse_array_str(PyScannerObject *s, PyObject *pystr, Py_ssize_t idx, Py_ssize_t *next_idx_ptr) {
- /* Read a JSON array from PyString pystr.
- idx is the index of the first character after the opening brace.
- *next_idx_ptr is a return-by-reference index to the first character after
- the closing brace.
-
- Returns a new PyList
- */
- char *str = PyString_AS_STRING(pystr);
- Py_ssize_t end_idx = PyString_GET_SIZE(pystr) - 1;
- PyObject *val = NULL;
- PyObject *rval = PyList_New(0);
- Py_ssize_t next_idx;
- if (rval == NULL)
- return NULL;
-
- /* skip whitespace after [ */
- while (idx <= end_idx && IS_WHITESPACE(str[idx])) idx++;
-
- /* only loop if the array is non-empty */
- if (idx <= end_idx && str[idx] != ']') {
- while (idx <= end_idx) {
-
- /* read any JSON term and de-tuplefy the (rval, idx) */
- val = scan_once_str(s, pystr, idx, &next_idx);
- if (val == NULL)
- goto bail;
-
- if (PyList_Append(rval, val) == -1)
- goto bail;
-
- Py_CLEAR(val);
- idx = next_idx;
-
- /* skip whitespace between term and , */
- while (idx <= end_idx && IS_WHITESPACE(str[idx])) idx++;
-
- /* bail if the array is closed or we didn't get the , delimiter */
- if (idx > end_idx) break;
- if (str[idx] == ']') {
- break;
- }
- else if (str[idx] != ',') {
- raise_errmsg("Expecting , delimiter", pystr, idx);
- goto bail;
- }
- idx++;
-
- /* skip whitespace after , */
- while (idx <= end_idx && IS_WHITESPACE(str[idx])) idx++;
- }
- }
-
- /* verify that idx < end_idx, str[idx] should be ']' */
- if (idx > end_idx || str[idx] != ']') {
- raise_errmsg("Expecting object", pystr, end_idx);
- goto bail;
- }
- *next_idx_ptr = idx + 1;
- return rval;
-bail:
- Py_XDECREF(val);
- Py_DECREF(rval);
+ Py_XDECREF(rval);
return NULL;
}
@@ -1273,121 +814,19 @@ _parse_constant(PyScannerObject *s, char *constant, Py_ssize_t idx, Py_ssize_t *
PyObject *cstr;
PyObject *rval;
/* constant is "NaN", "Infinity", or "-Infinity" */
- cstr = PyString_InternFromString(constant);
+ cstr = PyUnicode_InternFromString(constant);
if (cstr == NULL)
return NULL;
/* rval = parse_constant(constant) */
rval = PyObject_CallFunctionObjArgs(s->parse_constant, cstr, NULL);
- idx += PyString_GET_SIZE(cstr);
+ idx += PyUnicode_GET_SIZE(cstr);
Py_DECREF(cstr);
*next_idx_ptr = idx;
return rval;
}
static PyObject *
-_match_number_str(PyScannerObject *s, PyObject *pystr, Py_ssize_t start, Py_ssize_t *next_idx_ptr) {
- /* Read a JSON number from PyString pystr.
- idx is the index of the first character of the number
- *next_idx_ptr is a return-by-reference index to the first character after
- the number.
-
- Returns a new PyObject representation of that number:
- PyInt, PyLong, or PyFloat.
- May return other types if parse_int or parse_float are set
- */
- char *str = PyString_AS_STRING(pystr);
- Py_ssize_t end_idx = PyString_GET_SIZE(pystr) - 1;
- Py_ssize_t idx = start;
- int is_float = 0;
- PyObject *rval;
- PyObject *numstr;
-
- /* read a sign if it's there, make sure it's not the end of the string */
- if (str[idx] == '-') {
- idx++;
- if (idx > end_idx) {
- PyErr_SetNone(PyExc_StopIteration);
- return NULL;
- }
- }
-
- /* read as many integer digits as we find as long as it doesn't start with 0 */
- if (str[idx] >= '1' && str[idx] <= '9') {
- idx++;
- while (idx <= end_idx && str[idx] >= '0' && str[idx] <= '9') idx++;
- }
- /* if it starts with 0 we only expect one integer digit */
- else if (str[idx] == '0') {
- idx++;
- }
- /* no integer digits, error */
- else {
- PyErr_SetNone(PyExc_StopIteration);
- return NULL;
- }
-
- /* if the next char is '.' followed by a digit then read all float digits */
- if (idx < end_idx && str[idx] == '.' && str[idx + 1] >= '0' && str[idx + 1] <= '9') {
- is_float = 1;
- idx += 2;
- while (idx <= end_idx && str[idx] >= '0' && str[idx] <= '9') idx++;
- }
-
- /* if the next char is 'e' or 'E' then maybe read the exponent (or backtrack) */
- if (idx < end_idx && (str[idx] == 'e' || str[idx] == 'E')) {
-
- /* save the index of the 'e' or 'E' just in case we need to backtrack */
- Py_ssize_t e_start = idx;
- idx++;
-
- /* read an exponent sign if present */
- if (idx < end_idx && (str[idx] == '-' || str[idx] == '+')) idx++;
-
- /* read all digits */
- while (idx <= end_idx && str[idx] >= '0' && str[idx] <= '9') idx++;
-
- /* if we got a digit, then parse as float. if not, backtrack */
- if (str[idx - 1] >= '0' && str[idx - 1] <= '9') {
- is_float = 1;
- }
- else {
- idx = e_start;
- }
- }
-
- /* copy the section we determined to be a number */
- numstr = PyString_FromStringAndSize(&str[start], idx - start);
- if (numstr == NULL)
- return NULL;
- if (is_float) {
- /* parse as a float using a fast path if available, otherwise call user defined method */
- if (s->parse_float != (PyObject *)&PyFloat_Type) {
- rval = PyObject_CallFunctionObjArgs(s->parse_float, numstr, NULL);
- }
- else {
- double d = PyOS_string_to_double(PyString_AS_STRING(numstr),
- NULL, NULL);
- if (d == -1.0 && PyErr_Occurred())
- return NULL;
- rval = PyFloat_FromDouble(d);
- }
- }
- else {
- /* parse as an int using a fast path if available, otherwise call user defined method */
- if (s->parse_int != (PyObject *)&PyInt_Type) {
- rval = PyObject_CallFunctionObjArgs(s->parse_int, numstr, NULL);
- }
- else {
- rval = PyInt_FromString(PyString_AS_STRING(numstr), NULL, 10);
- }
- }
- Py_DECREF(numstr);
- *next_idx_ptr = idx;
- return rval;
-}
-
-static PyObject *
_match_number_unicode(PyScannerObject *s, PyObject *pystr, Py_ssize_t start, Py_ssize_t *next_idx_ptr) {
/* Read a JSON number from PyUnicode pystr.
idx is the index of the first character of the number
@@ -1466,7 +905,7 @@ _match_number_unicode(PyScannerObject *s, PyObject *pystr, Py_ssize_t start, Py_
rval = PyObject_CallFunctionObjArgs(s->parse_float, numstr, NULL);
}
else {
- rval = PyFloat_FromString(numstr, NULL);
+ rval = PyFloat_FromString(numstr);
}
}
else {
@@ -1479,93 +918,6 @@ _match_number_unicode(PyScannerObject *s, PyObject *pystr, Py_ssize_t start, Py_
}
static PyObject *
-scan_once_str(PyScannerObject *s, PyObject *pystr, Py_ssize_t idx, Py_ssize_t *next_idx_ptr)
-{
- /* Read one JSON term (of any kind) from PyString pystr.
- idx is the index of the first character of the term
- *next_idx_ptr is a return-by-reference index to the first character after
- the number.
-
- Returns a new PyObject representation of the term.
- */
- PyObject *res;
- char *str = PyString_AS_STRING(pystr);
- Py_ssize_t length = PyString_GET_SIZE(pystr);
- if (idx >= length) {
- PyErr_SetNone(PyExc_StopIteration);
- return NULL;
- }
- switch (str[idx]) {
- case '"':
- /* string */
- return scanstring_str(pystr, idx + 1,
- PyString_AS_STRING(s->encoding),
- PyObject_IsTrue(s->strict),
- next_idx_ptr);
- case '{':
- /* object */
- if (Py_EnterRecursiveCall(" while decoding a JSON object "
- "from a byte string"))
- return NULL;
- res = _parse_object_str(s, pystr, idx + 1, next_idx_ptr);
- Py_LeaveRecursiveCall();
- return res;
- case '[':
- /* array */
- if (Py_EnterRecursiveCall(" while decoding a JSON array "
- "from a byte string"))
- return NULL;
- res = _parse_array_str(s, pystr, idx + 1, next_idx_ptr);
- Py_LeaveRecursiveCall();
- return res;
- case 'n':
- /* null */
- if ((idx + 3 < length) && str[idx + 1] == 'u' && str[idx + 2] == 'l' && str[idx + 3] == 'l') {
- Py_INCREF(Py_None);
- *next_idx_ptr = idx + 4;
- return Py_None;
- }
- break;
- case 't':
- /* true */
- if ((idx + 3 < length) && str[idx + 1] == 'r' && str[idx + 2] == 'u' && str[idx + 3] == 'e') {
- Py_INCREF(Py_True);
- *next_idx_ptr = idx + 4;
- return Py_True;
- }
- break;
- case 'f':
- /* false */
- if ((idx + 4 < length) && str[idx + 1] == 'a' && str[idx + 2] == 'l' && str[idx + 3] == 's' && str[idx + 4] == 'e') {
- Py_INCREF(Py_False);
- *next_idx_ptr = idx + 5;
- return Py_False;
- }
- break;
- case 'N':
- /* NaN */
- if ((idx + 2 < length) && str[idx + 1] == 'a' && str[idx + 2] == 'N') {
- return _parse_constant(s, "NaN", idx, next_idx_ptr);
- }
- break;
- case 'I':
- /* Infinity */
- 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') {
- return _parse_constant(s, "Infinity", idx, next_idx_ptr);
- }
- break;
- case '-':
- /* -Infinity */
- 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') {
- return _parse_constant(s, "-Infinity", idx, next_idx_ptr);
- }
- break;
- }
- /* Didn't find a string, object, array, or named constant. Look for a number. */
- return _match_number_str(s, pystr, idx, next_idx_ptr);
-}
-
-static PyObject *
scan_once_unicode(PyScannerObject *s, PyObject *pystr, Py_ssize_t idx, Py_ssize_t *next_idx_ptr)
{
/* Read one JSON term (of any kind) from PyUnicode pystr.
@@ -1666,10 +1018,7 @@ scanner_call(PyObject *self, PyObject *args, PyObject *kwds)
if (!PyArg_ParseTupleAndKeywords(args, kwds, "OO&:scan_once", kwlist, &pystr, _convertPyInt_AsSsize_t, &idx))
return NULL;
- if (PyString_Check(pystr)) {
- rval = scan_once_str(s, pystr, idx, &next_idx);
- }
- else if (PyUnicode_Check(pystr)) {
+ if (PyUnicode_Check(pystr)) {
rval = scan_once_unicode(s, pystr, idx, &next_idx);
}
else {
@@ -1678,6 +1027,9 @@ scanner_call(PyObject *self, PyObject *args, PyObject *kwds)
Py_TYPE(pystr)->tp_name);
return NULL;
}
+ PyDict_Clear(s->memo);
+ if (rval == NULL)
+ return NULL;
return _build_rval_index_tuple(rval, next_idx);
}
@@ -1687,10 +1039,9 @@ scanner_new(PyTypeObject *type, PyObject *args, PyObject *kwds)
PyScannerObject *s;
s = (PyScannerObject *)type->tp_alloc(type, 0);
if (s != NULL) {
- s->encoding = NULL;
s->strict = NULL;
s->object_hook = NULL;
- s->pairs_hook = NULL;
+ s->object_pairs_hook = NULL;
s->parse_float = NULL;
s->parse_int = NULL;
s->parse_constant = NULL;
@@ -1712,28 +1063,11 @@ scanner_init(PyObject *self, PyObject *args, PyObject *kwds)
if (!PyArg_ParseTupleAndKeywords(args, kwds, "O:make_scanner", kwlist, &ctx))
return -1;
- /* PyString_AS_STRING is used on encoding */
- s->encoding = PyObject_GetAttrString(ctx, "encoding");
- if (s->encoding == NULL)
- goto bail;
- if (s->encoding == Py_None) {
- Py_DECREF(Py_None);
- s->encoding = PyString_InternFromString(DEFAULT_ENCODING);
- }
- else if (PyUnicode_Check(s->encoding)) {
- PyObject *tmp = PyUnicode_AsEncodedString(s->encoding, NULL, NULL);
- Py_DECREF(s->encoding);
- s->encoding = tmp;
- }
- if (s->encoding == NULL)
- goto bail;
- if (!PyString_Check(s->encoding)) {
- PyErr_Format(PyExc_TypeError,
- "encoding must be a string, not %.80s",
- Py_TYPE(s->encoding)->tp_name);
- goto bail;
+ if (s->memo == NULL) {
+ s->memo = PyDict_New();
+ if (s->memo == NULL)
+ goto bail;
}
-
/* All of these will fail "gracefully" so we don't need to verify them */
s->strict = PyObject_GetAttrString(ctx, "strict");
@@ -1742,8 +1076,8 @@ scanner_init(PyObject *self, PyObject *args, PyObject *kwds)
s->object_hook = PyObject_GetAttrString(ctx, "object_hook");
if (s->object_hook == NULL)
goto bail;
- s->pairs_hook = PyObject_GetAttrString(ctx, "object_pairs_hook");
- if (s->pairs_hook == NULL)
+ s->object_pairs_hook = PyObject_GetAttrString(ctx, "object_pairs_hook");
+ if (s->object_pairs_hook == NULL)
goto bail;
s->parse_float = PyObject_GetAttrString(ctx, "parse_float");
if (s->parse_float == NULL)
@@ -1758,10 +1092,9 @@ scanner_init(PyObject *self, PyObject *args, PyObject *kwds)
return 0;
bail:
- Py_CLEAR(s->encoding);
Py_CLEAR(s->strict);
Py_CLEAR(s->object_hook);
- Py_CLEAR(s->pairs_hook);
+ Py_CLEAR(s->object_pairs_hook);
Py_CLEAR(s->parse_float);
Py_CLEAR(s->parse_int);
Py_CLEAR(s->parse_constant);
@@ -1772,8 +1105,7 @@ PyDoc_STRVAR(scanner_doc, "JSON scanner object");
static
PyTypeObject PyScannerType = {
- PyObject_HEAD_INIT(NULL)
- 0, /* tp_internal */
+ PyVarObject_HEAD_INIT(NULL, 0)
"_json.Scanner", /* tp_name */
sizeof(PyScannerObject), /* tp_basicsize */
0, /* tp_itemsize */
@@ -1903,7 +1235,7 @@ _encoded_const(PyObject *obj)
if (obj == Py_None) {
static PyObject *s_null = NULL;
if (s_null == NULL) {
- s_null = PyString_InternFromString("null");
+ s_null = PyUnicode_InternFromString("null");
}
Py_INCREF(s_null);
return s_null;
@@ -1911,7 +1243,7 @@ _encoded_const(PyObject *obj)
else if (obj == Py_True) {
static PyObject *s_true = NULL;
if (s_true == NULL) {
- s_true = PyString_InternFromString("true");
+ s_true = PyUnicode_InternFromString("true");
}
Py_INCREF(s_true);
return s_true;
@@ -1919,7 +1251,7 @@ _encoded_const(PyObject *obj)
else if (obj == Py_False) {
static PyObject *s_false = NULL;
if (s_false == NULL) {
- s_false = PyString_InternFromString("false");
+ s_false = PyUnicode_InternFromString("false");
}
Py_INCREF(s_false);
return s_false;
@@ -1941,13 +1273,13 @@ encoder_encode_float(PyEncoderObject *s, PyObject *obj)
return NULL;
}
if (i > 0) {
- return PyString_FromString("Infinity");
+ return PyUnicode_FromString("Infinity");
}
else if (i < 0) {
- return PyString_FromString("-Infinity");
+ return PyUnicode_FromString("-Infinity");
}
else {
- return PyString_FromString("NaN");
+ return PyUnicode_FromString("NaN");
}
}
/* Use a better float format here? */
@@ -1986,14 +1318,14 @@ encoder_listencode_obj(PyEncoderObject *s, PyObject *rval, PyObject *obj, Py_ssi
return -1;
return _steal_list_append(rval, cstr);
}
- else if (PyString_Check(obj) || PyUnicode_Check(obj))
+ else if (PyUnicode_Check(obj))
{
PyObject *encoded = encoder_encode_string(s, obj);
if (encoded == NULL)
return -1;
return _steal_list_append(rval, encoded);
}
- else if (PyInt_Check(obj) || PyLong_Check(obj)) {
+ else if (PyLong_Check(obj)) {
PyObject *encoded = PyObject_Str(obj);
if (encoded == NULL)
return -1;
@@ -2074,16 +1406,16 @@ encoder_listencode_dict(PyEncoderObject *s, PyObject *rval, PyObject *dct, Py_ss
static PyObject *empty_dict = NULL;
PyObject *kstr = NULL;
PyObject *ident = NULL;
- PyObject *key = NULL;
- PyObject *value = NULL;
PyObject *it = NULL;
+ PyObject *items;
+ PyObject *item = NULL;
int skipkeys;
Py_ssize_t idx;
if (open_dict == NULL || close_dict == NULL || empty_dict == NULL) {
- open_dict = PyString_InternFromString("{");
- close_dict = PyString_InternFromString("}");
- empty_dict = PyString_InternFromString("{}");
+ open_dict = PyUnicode_InternFromString("{");
+ close_dict = PyUnicode_InternFromString("}");
+ empty_dict = PyUnicode_InternFromString("{}");
if (open_dict == NULL || close_dict == NULL || empty_dict == NULL)
return -1;
}
@@ -2119,17 +1451,49 @@ encoder_listencode_dict(PyEncoderObject *s, PyObject *rval, PyObject *dct, Py_ss
*/
}
- /* TODO: C speedup not implemented for sort_keys */
-
- it = PyObject_GetIter(dct);
+ if (PyObject_IsTrue(s->sort_keys)) {
+ /* First sort the keys then replace them with (key, value) tuples. */
+ Py_ssize_t i, nitems;
+ items = PyMapping_Keys(dct);
+ if (items == NULL)
+ goto bail;
+ if (!PyList_Check(items)) {
+ PyErr_SetString(PyExc_ValueError, "keys must return list");
+ goto bail;
+ }
+ if (PyList_Sort(items) < 0)
+ goto bail;
+ nitems = PyList_GET_SIZE(items);
+ for (i = 0; i < nitems; i++) {
+ PyObject *key, *value;
+ key = PyList_GET_ITEM(items, i);
+ value = PyDict_GetItem(dct, key);
+ item = PyTuple_Pack(2, key, value);
+ if (item == NULL)
+ goto bail;
+ PyList_SET_ITEM(items, i, item);
+ Py_DECREF(key);
+ }
+ }
+ else {
+ items = PyMapping_Items(dct);
+ }
+ if (items == NULL)
+ goto bail;
+ it = PyObject_GetIter(items);
+ Py_DECREF(items);
if (it == NULL)
goto bail;
skipkeys = PyObject_IsTrue(s->skipkeys);
idx = 0;
- while ((key = PyIter_Next(it)) != NULL) {
- PyObject *encoded;
-
- if (PyString_Check(key) || PyUnicode_Check(key)) {
+ while ((item = PyIter_Next(it)) != NULL) {
+ PyObject *encoded, *key, *value;
+ if (!PyTuple_Check(item) || Py_SIZE(item) != 2) {
+ PyErr_SetString(PyExc_ValueError, "items must return 2-tuples");
+ goto bail;
+ }
+ key = PyTuple_GET_ITEM(item, 0);
+ if (PyUnicode_Check(key)) {
Py_INCREF(key);
kstr = key;
}
@@ -2138,18 +1502,20 @@ encoder_listencode_dict(PyEncoderObject *s, PyObject *rval, PyObject *dct, Py_ss
if (kstr == NULL)
goto bail;
}
- else if (PyInt_Check(key) || PyLong_Check(key)) {
- kstr = PyObject_Str(key);
+ else if (key == Py_True || key == Py_False || key == Py_None) {
+ /* This must come before the PyLong_Check because
+ True and False are also 1 and 0.*/
+ kstr = _encoded_const(key);
if (kstr == NULL)
goto bail;
}
- else if (key == Py_True || key == Py_False || key == Py_None) {
- kstr = _encoded_const(key);
+ else if (PyLong_Check(key)) {
+ kstr = PyObject_Str(key);
if (kstr == NULL)
goto bail;
}
else if (skipkeys) {
- Py_DECREF(key);
+ Py_DECREF(item);
continue;
}
else {
@@ -2163,10 +1529,6 @@ encoder_listencode_dict(PyEncoderObject *s, PyObject *rval, PyObject *dct, Py_ss
goto bail;
}
- value = PyObject_GetItem(dct, key);
- if (value == NULL)
- goto bail;
-
encoded = encoder_encode_string(s, kstr);
Py_CLEAR(kstr);
if (encoded == NULL)
@@ -2178,11 +1540,12 @@ encoder_listencode_dict(PyEncoderObject *s, PyObject *rval, PyObject *dct, Py_ss
Py_DECREF(encoded);
if (PyList_Append(rval, s->key_separator))
goto bail;
+
+ value = PyTuple_GET_ITEM(item, 1);
if (encoder_listencode_obj(s, rval, value, indent_level))
goto bail;
idx += 1;
- Py_CLEAR(value);
- Py_DECREF(key);
+ Py_DECREF(item);
}
if (PyErr_Occurred())
goto bail;
@@ -2195,9 +1558,8 @@ encoder_listencode_dict(PyEncoderObject *s, PyObject *rval, PyObject *dct, Py_ss
}
if (s->indent != Py_None) {
/* TODO: DOES NOT RUN */
+ indent_level -= 1;
/*
- indent_level -= 1;
-
yield '\n' + (' ' * (_indent * _current_indent_level))
*/
}
@@ -2207,8 +1569,7 @@ encoder_listencode_dict(PyEncoderObject *s, PyObject *rval, PyObject *dct, Py_ss
bail:
Py_XDECREF(it);
- Py_XDECREF(key);
- Py_XDECREF(value);
+ Py_XDECREF(item);
Py_XDECREF(kstr);
Py_XDECREF(ident);
return -1;
@@ -2227,9 +1588,9 @@ encoder_listencode_list(PyEncoderObject *s, PyObject *rval, PyObject *seq, Py_ss
Py_ssize_t i;
if (open_array == NULL || close_array == NULL || empty_array == NULL) {
- open_array = PyString_InternFromString("[");
- close_array = PyString_InternFromString("]");
- empty_array = PyString_InternFromString("[]");
+ open_array = PyUnicode_InternFromString("[");
+ close_array = PyUnicode_InternFromString("]");
+ empty_array = PyUnicode_InternFromString("[]");
if (open_array == NULL || close_array == NULL || empty_array == NULL)
return -1;
}
@@ -2285,9 +1646,8 @@ encoder_listencode_list(PyEncoderObject *s, PyObject *rval, PyObject *seq, Py_ss
}
if (s->indent != Py_None) {
/* TODO: DOES NOT RUN */
+ indent_level -= 1;
/*
- indent_level -= 1;
-
yield '\n' + (' ' * (_indent * _current_indent_level))
*/
}
@@ -2349,8 +1709,7 @@ PyDoc_STRVAR(encoder_doc, "_iterencode(obj, _current_indent_level) -> iterable")
static
PyTypeObject PyEncoderType = {
- PyObject_HEAD_INIT(NULL)
- 0, /* tp_internal */
+ PyVarObject_HEAD_INIT(NULL, 0)
"_json.Encoder", /* tp_name */
sizeof(PyEncoderObject), /* tp_basicsize */
0, /* tp_itemsize */
@@ -2406,19 +1765,42 @@ static PyMethodDef speedups_methods[] = {
PyDoc_STRVAR(module_doc,
"json speedups\n");
-void
-init_json(void)
+static struct PyModuleDef jsonmodule = {
+ PyModuleDef_HEAD_INIT,
+ "_json",
+ module_doc,
+ -1,
+ speedups_methods,
+ NULL,
+ NULL,
+ NULL,
+ NULL
+};
+
+PyObject*
+PyInit__json(void)
{
- PyObject *m;
+ PyObject *m = PyModule_Create(&jsonmodule);
+ if (!m)
+ return NULL;
PyScannerType.tp_new = PyType_GenericNew;
if (PyType_Ready(&PyScannerType) < 0)
- return;
+ goto fail;
PyEncoderType.tp_new = PyType_GenericNew;
if (PyType_Ready(&PyEncoderType) < 0)
- return;
- m = Py_InitModule3("_json", speedups_methods, module_doc);
+ goto fail;
Py_INCREF((PyObject*)&PyScannerType);
- PyModule_AddObject(m, "make_scanner", (PyObject*)&PyScannerType);
+ if (PyModule_AddObject(m, "make_scanner", (PyObject*)&PyScannerType) < 0) {
+ Py_DECREF((PyObject*)&PyScannerType);
+ goto fail;
+ }
Py_INCREF((PyObject*)&PyEncoderType);
- PyModule_AddObject(m, "make_encoder", (PyObject*)&PyEncoderType);
+ if (PyModule_AddObject(m, "make_encoder", (PyObject*)&PyEncoderType) < 0) {
+ Py_DECREF((PyObject*)&PyEncoderType);
+ goto fail;
+ }
+ return m;
+ fail:
+ Py_DECREF(m);
+ return NULL;
}
generated by cgit v0.12 at 2025-12-25 17:35:27 (GMT)