Merge remote-tracking branch 'scottkmaxwell/master' into bigint_as_string-gh31

5 files changed, 85 insertions, 16 deletions

diff --git a/simplejson/__init__.py b/simplejson/__init__.py
index 3ee7893..88eb02d 100644
--- a/simplejson/__init__.py
+++ b/simplejson/__init__.py
@@ -138,12 +138,13 @@ _default_encoder = JSONEncoder(
     use_decimal=True,
     namedtuple_as_object=True,
     tuple_as_array=True,
+    bigint_as_string=False
 )
 
 def dump(obj, fp, skipkeys=False, ensure_ascii=True, check_circular=True,
         allow_nan=True, cls=None, indent=None, separators=None,
         encoding='utf-8', default=None, use_decimal=True,
-        namedtuple_as_object=True, tuple_as_array=True,
+        namedtuple_as_object=True, tuple_as_array=True, bigint_as_string=False,
         **kw):
     """Serialize ``obj`` as a JSON formatted stream to ``fp`` (a
     ``.write()``-supporting file-like object).
@@ -193,6 +194,10 @@ def dump(obj, fp, skipkeys=False, ensure_ascii=True, check_circular=True,
     If *tuple_as_array* is true (default: ``True``),
     :class:`tuple` (and subclasses) will be encoded as JSON arrays.
 
+    If bigint_as_string is true (not the default), ints 2**53 and higher
+    or lower than -2**53 will be encoded as strings. This is to avoid the
+    rounding that happens in Javascript otherwise.
+
     To use a custom ``JSONEncoder`` subclass (e.g. one that overrides the
     ``.default()`` method to serialize additional types), specify it with
     the ``cls`` kwarg.
@@ -203,7 +208,8 @@ def dump(obj, fp, skipkeys=False, ensure_ascii=True, check_circular=True,
         check_circular and allow_nan and
         cls is None and indent is None and separators is None and
         encoding == 'utf-8' and default is None and use_decimal
-        and namedtuple_as_object and tuple_as_array and not kw):
+        and namedtuple_as_object and tuple_as_array
+        and not bigint_as_string and not kw):
         iterable = _default_encoder.iterencode(obj)
     else:
         if cls is None:
@@ -214,6 +220,7 @@ def dump(obj, fp, skipkeys=False, ensure_ascii=True, check_circular=True,
             default=default, use_decimal=use_decimal,
             namedtuple_as_object=namedtuple_as_object,
             tuple_as_array=tuple_as_array,
+            bigint_as_string=bigint_as_string,
             **kw).iterencode(obj)
     # could accelerate with writelines in some versions of Python, at
     # a debuggability cost
@@ -225,7 +232,7 @@ def dumps(obj, skipkeys=False, ensure_ascii=True, check_circular=True,
         allow_nan=True, cls=None, indent=None, separators=None,
         encoding='utf-8', default=None, use_decimal=True,
         namedtuple_as_object=True,
-        tuple_as_array=True,
+        tuple_as_array=True, bigint_as_string=False,
         **kw):
     """Serialize ``obj`` to a JSON formatted ``str``.
 
@@ -272,6 +279,10 @@ def dumps(obj, skipkeys=False, ensure_ascii=True, check_circular=True,
     If *tuple_as_array* is true (default: ``True``),
     :class:`tuple` (and subclasses) will be encoded as JSON arrays.
 
+    If bigint_as_string is true (not the default), ints 2**53 and higher
+    or lower than -2**53 will be encoded as strings. This is to avoid the
+    rounding that happens in Javascript otherwise.
+
     To use a custom ``JSONEncoder`` subclass (e.g. one that overrides the
     ``.default()`` method to serialize additional types), specify it with
     the ``cls`` kwarg.
@@ -282,7 +293,8 @@ def dumps(obj, skipkeys=False, ensure_ascii=True, check_circular=True,
         check_circular and allow_nan and
         cls is None and indent is None and separators is None and
         encoding == 'utf-8' and default is None and use_decimal
-        and namedtuple_as_object and tuple_as_array and not kw):
+        and namedtuple_as_object and tuple_as_array
+        and not bigint_as_string and not kw):
         return _default_encoder.encode(obj)
     if cls is None:
         cls = JSONEncoder
@@ -293,6 +305,7 @@ def dumps(obj, skipkeys=False, ensure_ascii=True, check_circular=True,
         use_decimal=use_decimal,
         namedtuple_as_object=namedtuple_as_object,
         tuple_as_array=tuple_as_array,
+        bigint_as_string=bigint_as_string,
         **kw).encode(obj)
 
 
diff --git a/simplejson/_speedups.c b/simplejson/_speedups.c
index c25909e..4efccdc 100644
--- a/simplejson/_speedups.c
+++ b/simplejson/_speedups.c
@@ -89,6 +89,7 @@ typedef struct _PyEncoderObject {
     int use_decimal;
     int namedtuple_as_object;
     int tuple_as_array;
+    int bigint_as_string;
 } PyEncoderObject;
 
 static PyMemberDef encoder_members[] = {
@@ -2025,19 +2026,19 @@ static int
 encoder_init(PyObject *self, PyObject *args, PyObject *kwds)
 {
     /* initialize Encoder object */
-    static char *kwlist[] = {"markers", "default", "encoder", "indent", "key_separator", "item_separator", "sort_keys", "skipkeys", "allow_nan", "key_memo", "use_decimal", "namedtuple_as_object", "tuple_as_array", NULL};
+    static char *kwlist[] = {"markers", "default", "encoder", "indent", "key_separator", "item_separator", "sort_keys", "skipkeys", "allow_nan", "key_memo", "use_decimal", "namedtuple_as_object", "tuple_as_array", "bigint_as_string", NULL};
 
     PyEncoderObject *s;
     PyObject *markers, *defaultfn, *encoder, *indent, *key_separator;
-    PyObject *item_separator, *sort_keys, *skipkeys, *allow_nan, *key_memo, *use_decimal, *namedtuple_as_object, *tuple_as_array;
+    PyObject *item_separator, *sort_keys, *skipkeys, *allow_nan, *key_memo, *use_decimal, *namedtuple_as_object, *tuple_as_array, *bigint_as_string;
 
     assert(PyEncoder_Check(self));
     s = (PyEncoderObject *)self;
 
-    if (!PyArg_ParseTupleAndKeywords(args, kwds, "OOOOOOOOOOOOO:make_encoder", kwlist,
+    if (!PyArg_ParseTupleAndKeywords(args, kwds, "OOOOOOOOOOOOOO:make_encoder", kwlist,
         &markers, &defaultfn, &encoder, &indent, &key_separator, &item_separator,
         &sort_keys, &skipkeys, &allow_nan, &key_memo, &use_decimal,
-        &namedtuple_as_object, &tuple_as_array))
+        &namedtuple_as_object, &tuple_as_array, &bigint_as_string))
         return -1;
 
     s->markers = markers;
@@ -2054,6 +2055,7 @@ encoder_init(PyObject *self, PyObject *args, PyObject *kwds)
     s->use_decimal = PyObject_IsTrue(use_decimal);
     s->namedtuple_as_object = PyObject_IsTrue(namedtuple_as_object);
     s->tuple_as_array = PyObject_IsTrue(tuple_as_array);
+    s->bigint_as_string = PyObject_IsTrue(bigint_as_string);
 
     Py_INCREF(s->markers);
     Py_INCREF(s->defaultfn);
@@ -2189,8 +2191,21 @@ encoder_listencode_obj(PyEncoderObject *s, PyObject *rval, PyObject *obj, Py_ssi
         }
         else if (PyInt_Check(obj) || PyLong_Check(obj)) {
             PyObject *encoded = PyObject_Str(obj);
-            if (encoded != NULL)
+            if (encoded != NULL) {
+				if (s->bigint_as_string) {
+					int overflow;
+					PY_LONG_LONG value = PyLong_AsLongLongAndOverflow(obj, &overflow);
+					if (value < 0) {
+						value = ~value;
+					}
+					if (overflow || (value>>53)) {
+						PyObject* quoted = PyString_FromFormat("\"%s\"", PyString_AsString(encoded));
+						Py_DECREF(encoded);
+						encoded = quoted;
+					}
+				}
                 rv = _steal_list_append(rval, encoded);
+            }
         }
         else if (PyFloat_Check(obj)) {
             PyObject *encoded = encoder_encode_float(s, obj);
@@ -2395,6 +2410,18 @@ encoder_listencode_dict(PyEncoderObject *s, PyObject *rval, PyObject *dct, Py_ss
             kstr = PyObject_Str(key);
             if (kstr == NULL)
                 goto bail;
+			if (s->bigint_as_string) {
+				int overflow;
+				PY_LONG_LONG value = PyLong_AsLongLongAndOverflow(kstr, &overflow);
+				if (value < 0) {
+					value = ~value;
+				}
+				if (overflow || (value>>53)) {
+					PyObject* quoted = PyString_FromFormat("\"%s\"", PyString_AsString(kstr));
+					Py_DECREF(kstr);
+					kstr = quoted;
+				}
+			}
         }
         else if (skipkeys) {
             Py_DECREF(item);
diff --git a/simplejson/encoder.py b/simplejson/encoder.py
index 7f4f1cb..4c7601b 100644
--- a/simplejson/encoder.py
+++ b/simplejson/encoder.py
@@ -107,7 +107,7 @@ class JSONEncoder(object):
             check_circular=True, allow_nan=True, sort_keys=False,
             indent=None, separators=None, encoding='utf-8', default=None,
             use_decimal=True, namedtuple_as_object=True,
-            tuple_as_array=True):
+            tuple_as_array=True, bigint_as_string=False):
         """Constructor for JSONEncoder, with sensible defaults.
 
         If skipkeys is false, then it is a TypeError to attempt
@@ -160,6 +160,10 @@ class JSONEncoder(object):
 
         If tuple_as_array is true (the default), tuple (and subclasses) will
         be encoded as JSON arrays.
+
+        If bigint_as_string is true (not the default), ints 2**53 and higher
+        or lower than -2**53 will be encoded as strings. This is to avoid the
+        rounding that happens in Javascript otherwise.
         """
 
         self.skipkeys = skipkeys
@@ -170,6 +174,7 @@ class JSONEncoder(object):
         self.use_decimal = use_decimal
         self.namedtuple_as_object = namedtuple_as_object
         self.tuple_as_array = tuple_as_array
+        self.bigint_as_string = bigint_as_string
         if indent is not None and not isinstance(indent, basestring):
             indent = indent * ' '
         self.indent = indent
@@ -285,13 +290,13 @@ class JSONEncoder(object):
                 markers, self.default, _encoder, self.indent,
                 self.key_separator, self.item_separator, self.sort_keys,
                 self.skipkeys, self.allow_nan, key_memo, self.use_decimal,
-                self.namedtuple_as_object, self.tuple_as_array)
+                self.namedtuple_as_object, self.tuple_as_array, self.bigint_as_string)
         else:
             _iterencode = _make_iterencode(
                 markers, self.default, _encoder, self.indent, floatstr,
                 self.key_separator, self.item_separator, self.sort_keys,
                 self.skipkeys, _one_shot, self.use_decimal,
-                self.namedtuple_as_object, self.tuple_as_array)
+                self.namedtuple_as_object, self.tuple_as_array, self.bigint_as_string)
         try:
             return _iterencode(o, 0)
         finally:
@@ -327,7 +332,7 @@ class JSONEncoderForHTML(JSONEncoder):
 
 def _make_iterencode(markers, _default, _encoder, _indent, _floatstr,
         _key_separator, _item_separator, _sort_keys, _skipkeys, _one_shot,
-        _use_decimal, _namedtuple_as_object, _tuple_as_array,
+        _use_decimal, _namedtuple_as_object, _tuple_as_array, _bigint_as_string,
         ## HACK: hand-optimized bytecode; turn globals into locals
         False=False,
         True=True,
@@ -378,7 +383,7 @@ def _make_iterencode(markers, _default, _encoder, _indent, _floatstr,
             elif value is False:
                 yield buf + 'false'
             elif isinstance(value, (int, long)):
-                yield buf + str(value)
+                yield buf + str(value) if not _bigint_as_string or -(1<<53) <= value < (1<<53) else buf + '"' + str(value) + '"'
             elif isinstance(value, float):
                 yield buf + _floatstr(value)
             elif _use_decimal and isinstance(value, Decimal):
@@ -465,7 +470,7 @@ def _make_iterencode(markers, _default, _encoder, _indent, _floatstr,
             elif value is False:
                 yield 'false'
             elif isinstance(value, (int, long)):
-                yield str(value)
+                yield str(value) if not _bigint_as_string or -(1<<53) <= value < (1<<53) else '"' + str(value) + '"'
             elif isinstance(value, float):
                 yield _floatstr(value)
             elif _use_decimal and isinstance(value, Decimal):
@@ -503,7 +508,7 @@ def _make_iterencode(markers, _default, _encoder, _indent, _floatstr,
         elif o is False:
             yield 'false'
         elif isinstance(o, (int, long)):
-            yield str(o)
+            yield str(o) if not _bigint_as_string or -(1<<53) <= o < (1<<53) else '"' + str(o) + '"'
         elif isinstance(o, float):
             yield _floatstr(o)
         elif isinstance(o, list):
diff --git a/simplejson/tests/__init__.py b/simplejson/tests/__init__.py
index 8d2e443..12289b6 100644
--- a/simplejson/tests/__init__.py
+++ b/simplejson/tests/__init__.py
@@ -27,6 +27,7 @@ def additional_tests(suite=None):
 
 def all_tests_suite():
     suite = unittest.TestLoader().loadTestsFromNames([
+        'simplejson.tests.test_bigint_as_string',
         'simplejson.tests.test_check_circular',
         'simplejson.tests.test_decode',
         'simplejson.tests.test_default',
diff --git a/simplejson/tests/test_bigint_as_string.py b/simplejson/tests/test_bigint_as_string.py
new file mode 100644
index 0000000..cb81fab
--- /dev/null
+++ b/simplejson/tests/test_bigint_as_string.py
@@ -0,0 +1,23 @@
+from unittest import TestCase
+
+import simplejson as json
+
+class TestBigintAsString(TestCase):
+    values = [(200, 200), (2**53-1, 9007199254740991), (2**53, '"9007199254740992"'), (2**53+1, '"9007199254740993"'), (-100, -100), (-2**53, -9007199254740992), (-2**53-1, '"-9007199254740993"'), (-2**53+1, -9007199254740991)]
+
+    def test_ints(self):
+        for value_pair in self.values:
+            self.assertEquals('%s' % value_pair[0], json.dumps(value_pair[0]))
+            self.assertEquals('%s' % value_pair[1], json.dumps(value_pair[0], bigint_as_string=True))
+
+    def test_lists(self):
+        for value_pair in self.values:
+            l = [value_pair[0], value_pair[0]]
+            self.assertEquals('[%s, %s]' % (value_pair[0], value_pair[0]), json.dumps(l))
+            self.assertEquals('[%s, %s]' % (value_pair[1], value_pair[1]), json.dumps(l, bigint_as_string=True))
+
+    def test_dicts(self):
+        for value_pair in self.values:
+            d = {'value': value_pair[0]}
+            self.assertEquals('{"value": %s}' % value_pair[0], json.dumps(d))
+            self.assertEquals('{"value": %s}' % value_pair[1], json.dumps(d, bigint_as_string=True))